Progress in Biological Chirality

Foreword of the Editors Progress in Biological Chirality is based on papers presented at the 3^^^ Interdisciplinary Sym...

Author: Gyula Pályi (Editor) | Claudia Zucchi (Editor) | Luciano Caglioti (Editor)

23 downloads 2356 Views 25MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!

Report copyright / DMCA form

DOWNLOAD PDF

Foreword of the Editors Progress in Biological Chirality is based on papers presented at the 3^^^ Interdisciplinary Symposium on Biological Chirality (April 30-May 4, 2003, Modena, Italy), it, however, contains a number of important contributions from Authors who could not participate at this Symposium. We believe, that this meeting and the book are presenting some very important advances after the first (Serramazzoni, Italy, 1998) and the second (Szeged, Hungary, 2000) meetings of this series. The phenomenon of the biological chirality (that is: the high enantiomeric excess of chiral substances used by all living organisms) represents a very attractive challenge to scientists since at least 150 years [1]. Corresponding research was often of speculative character [2], yielding highly intellectual approaches, which, however, got experimental verification only very recently [3]. The intellectual efforts resulted several excellent reviews (e.g. [4]) and even books (e.g. [5]) on this topic. We believe that the problems of biological chirality could be resolved by an unusually broad interdisciplinary effort. This, in a word of rapidly increasing specialisation, requires exceptional dedication from the scientists who are dealing with these problems. This was characteristic for the Authors of a previous book [5b] and also for the Authors of the present volume. We feel particularly honoured by being chosen for the coordination of these efforts, what - we are firmly convinced - represent in fact (not only in the title) a progress in biological chirality.

^ Tjyula Palyi ^ University of Modena and Reggio Emilia

Claudia Zucchi University of Modena and Reggio Emilia

^uciano Cagwoti University X a Sapienza" Roma [1] L. Pasteur, CK Acad. Set 16 (1848) 535-538; ^ m Chim. Phys. 24 (1848) 442-459. [2] (a) W.H. Mills, Chem. Ind. 51 (1932) 750-759. (b) F.C. Frank, Biochim. Biophys. Acta 11 (1953) 459-463. (c) V. Avetisov and V. Goldanskii, Proc. Natl. Acad ScL USA 93 (1996) 11435-11442. [3] (a) A Szabo-Nagy andL. Keszthelyi, Proc. Natl. Acad Sci. USA 96 (1999) 4252-4255. (b) K. Soai, I. Sato, S. Komiya, M. Hayashi, Y. Matsueda, H. Imamura, T. Hayase, H. Morioka, H. Tabira, J. Yamamoto and Y.Kowata, Tetrahedron Asymmetry 14 (2003) 185-188. [4] (a) L. Keszthelyi, Quart. Rev. Biophys. 28 (1995) 473-507. (b) L. Mark6, Diss Savariensis 24 (1998) 1-64. (c) B.L. Feringa and R. A van Delden, Angew. Chem., Int. Ed. 38 (1999) 3418-3438. [5] D.B. Cline (Ed.) Physical Origin of Homochirality in Life, AIP Press, Woodberg (NY, USA), 1996. (b) G. Palyi, C. Zucchi and L. CagUoti CEjds.)Advances in BioChirality Elsevier, Amsterdam, 1999. (c)H. Brunner, Rechts Oder Links, Wiley-VCH, Weinheim, 1999. (d) C. Mc Manus, Right Hand, Left Hand, Weidenfeld & Nicholson, London, 2002.

Preface This book. Progress in Biological Chirality, is dealing with an exciting area of Natural Sciences. The topic is interesting for (at least) two reasons. One reason is, that the biological chirality (some call it "homochirality"), is of enormous importance for the highly selective and very finely tuned, concerted chemical reactions, which make up life. According to some scientists, this phenomenon is one of the essential requisites of the origin of life on Earth. As a consequence of these considerations, the understanding of the origin, evolution and interrelations while in act of biological chirality is of fundamental importance for fields ranging from theoretical biology to planning the structure of new pharmaceuticals. Another reason is of emotional character. Biological chirality today represents one of the great unresolved secrets of Natural Sciences. It is something like, as the gravity was in Galileo's and Newton's age: you see everywhere the result of a principle, but it is not yet sufficiently documented how it came to existence. As a result of efforts described in the present book, we know today much more about these challenges than only a few years earlier. This progress is the result of several interdisciplinary efforts from theoretical physics, chemistry and biology, through palaeontology, paleobotanies to preparative efforts at modelling some aspects of biological chirality by the synthesis and structural characterisation of specially designed molecules. A broad spectrum of such research efforts is presented in this book, which will give a useful information source to specialists and a remarkable piece of reading to all who are interested in problems how life is operating and how it originated.

Prof Ferdinando Taddei President National Academy of Sciences, Letters and Arts (Modena)

Prof Gian Carlo Pellacani Rector Universit^f Modena and^Reggio Emilia I •/

iAt^C(ul2»^^^^>«'«^-^

Prof Raymond Daudel President European Academy of Sciences, Arts and Humanities (Paris)

,/j^iy^^l Y

Board of the Referees Editors and Authors acknowledge thankfully the work of the following Referees: Vladik A. Avetisov N.N. Semenov Institute of Chemical Physics Russian Academy of Sciences Kossygina 4 117977 Moscow, Russian Federation [email protected] Krishnan Balasubramanian Department of Applied Science College of Engineering University of California, Davis Hertz Hall Bldg 661, PO Box 808, L-794, Livermore CA 94550 [email protected] Henri Brunner Institut fiir Anorganische Chemie Universitat Regensburg D-93040 Regensburg, Germany [email protected] Luciano Caglioti Dipartimento di Studi di Chimica e Tecnologia delle Sostanze Biologiche Attive Universita "La Sapienza" P.le A. Moro 5 00185 ROMA, Italy caglioti @axrma.uniromal .it Janos Csapo Institute of Chemistry Department of Biochemistry and Food Chemistry University of Kaposvar GubaS.u. 40 H-7400 Kaposvar, Hungary [email protected] Jerzy Dzik Institute for Paleobiology PAN University of Warszawa Twarda 51/55 00818 Warszawa, Poland [email protected]

Arrigo Fomi Department of Chemistry University of Modena and Reggio Emilia viaCampi 183 1-41100 Modena, Italy [email protected] Noriko Fujii Research Reactor Institute Kyoto University Noda, Kumatori, Sennan Osaka 590-0494, Japan [email protected] Romeu Cardoso Guimaraes Departamento de Biologia Geral Inst. Ciencias Biologicas Universidade Federal de Minas Gerais 31270-901 Belo Horizonte MG, Brazil [email protected] Lajos Keszthelyi Institute of Biophysics Biological Research Centre Hungarian Academy of Sciences Temesvari krt. 62 P.O. Box 521 H-6701 Szeged, Hungary [email protected] Dilip Kondepudi Department of Chemistry Wake Forest University Winston-Salem NC 27109 USA [email protected] Andras Liptak Carbohydrate Research Group Hungarian Academy of Sciences Egyetem ter 1 H-4010 Debrecen Hungary [email protected]

Progress in Biological Chirality Laszlo Marko Department of Organic Chemistry University of Veszprem Veszprem, H-8201, Hungary [email protected] Koichiro Matsuno Department of Bioengineering Nagaoka University of Technology Nagaoka 940-2188, Japan [email protected] Paul G. Mezey Canada Research Chair in Scientific Modelling and Simulation Department of Chemistry and Department of Physics and Physical Oceanography Memorial University of Newfoundland Saint John's, NF, CANADA A1B 3X7 [email protected], [email protected] Lynn Mihichuk Department of Chemistry University of Regina Regina, Saskatchewan, CANADA S4S 0A2 [email protected] Maria Minunni Dipartimento di Chimica Universita di Firenze Via del la Lastruccia 3 50019 SestoF.no(FI), Italy [email protected] Gyula Palyi Department of Chemistry University of Modena and Reggio Emilia viaCampi 183 1-41100 Modena, Italy [email protected] Vince Pozsgay National Institutes of Health 31 Center Dr. Rm 2A25, MSC 2423 Bethesda, MD 20892-2423 [email protected] Livia Simon Sarkadi Dept. of Biochemistry and Food Technology Budapest Univ. of Technology and Economics Muegyetem rkp, 3

H-1111 Budapest, Hungary sarkadi @ mail, bme.hu Gyorgy Steinbrecher International Working Group FUSION-B.F.R. Association EURATOM-MEC Department of Theoretical Physics, Physics Faculty, University of Craiova, Str.A.I.Cuza 13, Craiova-1 100, Romania [email protected] Eors Szathmary Department of Plant Taxonomy and Ecology Eotvos University Ludovika ter 2 H-1083 Budapest Hungary [email protected] Ferdinando Taddei Department of Chemistry University of Modena and Reggio Emilia ViaCampi 183 1-4II00 Modena, Italy Taddei .ferdinando @unimo .it Giovanni Torre Department of Chemistry University of Modena and Reggio Emilia ViaCampi 183 1-41100 Modena, Italy Torre.giovanni @unimo.it Riccardo Zanasi Dipartimento di Chimica Universita di Salerno Via Salvador Allende 1-84081 Baronissi (SA), Italia [email protected] Geoffrey Zubay Columbia University New York City 10027 USA [email protected] Claudia Zucchi Department of Chemistry University of Modena and Reggio Emilia ViaCampi 183 1-41100 Modena, Italy [email protected]

Progress in Biological Chirality G. Palyi, C. Zucchi and L. Caglioti (Editors) © 2004 Elsevier Ltd. All rights reserved.

Chapter 1 Origin of Biomacromolecular Homochirality: in Search of Evolutional Dynamics Vladik Avetisov Semenov Institute of Chemical Physics of the Russian Academy of Sciences, Kossygina 4, 119991 Moscow, Russia avetisov@chph. ras. ru

1.

Introduction Biomacromolecular homochirality is one of the important attributes of the molecular background of life. It is well known that the nucleotide sequences (A, T(U), G, C) in DNA and RNA polymeric chains, as well as the amino acid sequences in enzymes, are "nearly random". In this respect, they are "slightly edited random copolymers" [1-4]. At the same time, these biopolymers have a universal and remarkable trait, namely, DNA and RNA consist of D nucleotides, while enzymes involve solely L amino acids. From the viewpoint of chirality, the primary structures of DNA, RNA, and enzymes are highly specific. They are homochiral. There is no exception to this rule as far as key biomacromolecules are concerned. Biomacromolecular homochirality is assumed to be necessary for reproduction of genetic information. Indeed, macromolecular chains of DNA and RNA are known to be templates (matrix) on which complementary copies are assembled, and any homochiral A, T(U), G, C sequence can serve as a template suitable for assembling of complementary replica. In opposite, as it was shown [5], pairing between a chirally defect unit in replica and the corresponding unit in homochiral template is impossible unless chemical bonds along the chains are broken. The reproduction of "chiral mutations" with a significantly high probability is impossible due to complete loss of the template profile around of chiral defect [6-8]. During the DNA, RNA, and enzymes biosynthesis, enantiomeric configuration of units is under precise control. It is less than one "chiral defecf per a biomacromolecular chain. The control is provided by biochemical functions of high enantioselectivity: They are realized, for it's turn, by particular homochiral macromolecules, enzymes. This "exclusive circle" is crucial from the origin-of-homochirality problem, because selection of macromolecular sequences through template directed replication requires homochirality and, what are important, enantiospecific functions capable of maintaining the assemblage of homochiral sequences. Let us consider macromolecular chain containing N units of L or D configurations. The number of all possible chains differing in L, D sequences is equal to 2^ and grows exponentially with N. The value of M becomes commensurate with the particle number fluctuations in "laboratory scale" systems (-10'^^) at N^40. Therefore, whenever A^ does not

4

Progress in Biological Chirality

exceed a couple dozens, it is possible to choose such conditions for the chain assembling that all conceivable sequences including homochiral ones may be realized. The origin of homochiral macromolecules of such length has no specific physical problem. Given an appropriate condition for polymerization, the probability for any predefined sequence to appear, in particular, a homochiral one, does not vanish even when the choice of L or D configuration is purely random. However, there are strong statistical limitations for assembling of long L,D-sequences. Indeed, comparison of M with the particle number fluctuations in a "cosmological scale" system consists, for instance, of 10^" particles (the number of carbon atoms in the Solar system) revels that M becomes of the same order of magnitude at A^«130. This means that each realized sequence of L, D-units containing more than a hundred units is certain to be "unique" for the overwhelming majority of sequences cannot be realized in principle, simply because even the whole Universe is too small for that. At such level of complexity, which is related to the simplest biomacromolecules, a relative part of realizable sequences is vanishingly small regardless of physical or chemical conditions. There is no looking over all possible sequences during the evolution. Hence "the memory of random choice" under such statistical limitations simply means that, in principle, any primary pool of sequences must be suitable for replication, mutation, and selection. Therefore, some explanation should be invoked to account for that whether homochiral sequences could be selected and kept under the replication-mutation processes at prebiotic stage of evolution. It is for this reason, the problem of the origin of rather complex homochiral macromolecules is of special interest in the theory of evolution.

2.

Evolution as DifTusion First, we discuss the origin of biomacromolecular homochirality in a fi"amework of the widely known model of evolution suggested by M. Eigen and coworkers [9, 10]. This model has also been reviewed in a number of publications [11-19]. Let us consider sequences /, (/=1,2. .2^, N»\) of L-, D-units. All possible sequences of length N are usefiilly to be described by a metric sequences space, in particular, by A^dimensional hypercube. Each point in the sequences space represents a particular sequence. Each of A^' neighbors of a given sequence differs from the last by chiral configuration of only one N units, i. e. each neighbour differs in a "point mutation". A distance between two sequences, /, and //, is measured by a minimal number of successive unit steps (point mutations) allowing transitionfi-om/, to /;. It is called the Hamming distance. Let us introduce the equations describing replication-mutation processes. In the quasispecies evolution model [10], the equations have the form; dx -± = ( 4 e „ -B^ ~cp)x^^Y.^,Q,^x, , a - 1,2...2" (1) where x. =c,0)/^c'.(r) is the relative part of sequences // among all sequences, c^(r) is the concentration of/;. The rate parameters A^ and B^ define replication and degeneration of the sequence /,, respectively. The value {A^ - B^) is referred to fitness o. of the sequence //. Parameter (p means the velocity of an outflow across the system. The total concentration of sequences being constant, the value (p is equal to mean fitness over the system.

Origin of Biomacromolecular Homochirality: in Search of Evolutional Dynamics

5

^ = <^0) = ^ . 0 . x ^ . The relative probability to copy of sequence // xactlyandthe relative probability to make a "mutant copy" // under replication on the template h are defined by Qj^^ and Q.., respectively. Stochastic process, which determines evolutional dynamics, is defined by the dependence of relative probability to produce mutant copies, Qj^., on the length of mutation jumps, dj^^. In another words, the evolutional dynamics are defined by both the fitness values {o.} and the probabilities to jump on different distances in the sequences space. In particular, if the point mutations arise purely random, the probability to jump on distance
6

Progress in Biological Chirality

prebiotic evolution, there is the only way to resolve the origin-of-homochirality problem: The replication-mutation-selection processes among homochiral macromolecules must be supported by enantiospecific functions [20, 21], and we return to a sophistic: -"What was first, layer or egg?" Surely, we do not repudiate the traditional view on the origin of biological homochirality based on prebiotic stereochemistry. However, we deem it appropriate to comment on those approaches, which are directly related to the essence of the problem for the majority of researchers. It is believed, the there are three principle questions. - Did mirror symmetry breaking occur in the course of chemical or biological evolution on the Earth? - What was the possible cause of the symmetry breaking: effect of a chiral physical field (an asymmetrical factor) or a spontaneous breaking? - What was responsible for the "sign" of chirality, a causative factor or random one? Let us comment the most popular view on the phenomenon under the question referred in the literature in a variety of terms: "homochirality", "bio-chirality", "chiral purity", "mirror asymmetry of the bioorganic world". These terms reflect a historically formed opinion about the existence of two "molecular world", the symmetrical (racemic) world of nonliving nature and the fully asymmetrical (chirally pure) world associated with living nature. It needs to be emphasized, however, that it would be incorrect to make use of this distinction without due regard for the complexity of objects that form both molecular worlds. To proceed, there is the mirror symmetry breaking. Doubtless, biological systems are characterized by the prevalence of some enantiomers of chiral compounds over others. However, this fact in itself cannot be interpreted as mirror symmetry breaking. It only suggest that enantiospecific functions, similar to specific ones in biology at large, are subject of strong unification because changes in metabolism traits during evolution are rare events. The only fact, which may be interpreted as mirror symmetry breaking, is the absence of "mirrorantipode" life based on L nucleotides and D amino acids. Therefore, so fare as such a symmetry breaking is concerned, it should be born in mind that in the bioorganic world it is found only at the population level. It would be helpfiil to discover a universal mechanism that possibly could underlay chiral asymmetry in organic material and could be naturally realized either on the Earth or in the space. At the same time, the basic problem: how hypothetical features of chemical evolution could predetermine the origin of single branch of life remains beyond the scope of most publications. Attempts at the solution of the problem over more than a century have generated numerous ideas to explain formation of asymmetrical organic material which cover virtually the entire range of possible causes, from the asymmetrical effect of circularly polarized light or polarized gamma-radiation by neutron stars to asymmetrical crystallization due to "poor" mixing of large-scale volumes. Actually, there are two classes of factors able to break symmetry in chiral chemical systems: effects of chiral fields (local or global) and spontaneous symmetry breaking (for review see, for instance, [20, 22]). At present, there is no doubt that many natural factors can be responsible for asymmetrical formation of chiral organic compounds. How universal or effective one or another natural chiral factor may be is debatable. This issue is still popular. However, the question of whether asymmetrical formation of chiral organic material is possible under abiotic conditions has conclusively been

Origin of Biomacromolecular Homochirality: in Search of Evolutional Dynamics

7

given an affirmative answer. It is noteworthy, that substantial progress in our knowledge about asymmetrical processes in stereochemistry does not appear to be conductive to deep insight into the problem, largely because we do not understand the basic relationship between the asymmetry at the level of simple organic molecules, which may have arisen in the course of chemical evolution, and biomacromolecular homochirality. Therefore we assume that it is also important to search the answers fare from the traditional ways. The origin-of-homochirality problem may primarily be found an issue of the prebiotic dynamic type, rather than the asymmetrical reactions specificities. A major theoretical problem in this context is to describe evolution of macromolecules, perhaps like to RNA, and specific functions in increasing of their complexity. This task is not trivial since the number of potentially possible objects grows exponentially with increasing complexity and becomes physically infinite even for macromolecules of tens units. The last may result in specific evolutional dynamics, with error catastrophe being one possibility. A type of dynamics that drastically differ from the evolution of the Darwinian type will be considered below. The main interest will be focused on the characteristic types of the population degradation in the case of both short and long mutation jumps^ in the sequences space are realized.

3.

Outline of the Problem In order to define dynamics, it is necessary to give the fitness values for each sequence. Having assigned the fitness values, we introduce a fitness landscape on the multidimensional sequences space [12, 13, 23, 24]. The fitness landscape of evolutionary system is similar to the energy landscape of physical systems. The difference is that physical system moves to the minimum energy, and the evolutionary system tends to the maximum fitness. On this language, mutation jumps in the sequences space are constrained by the fitness landscape. What kind of fitness landscape is adequate to the pre-biological evolution? This is a rather vague question. For biological evolution, because selection of sequences is provided by specific functions, one can assume that a deterministic algorithm of the landscape introduction should be chosen. For prebiotic evolution, in opposite, one can proceed from an assumption that the fitness landscape is defined by a great number of random factors [23]. The landscapes of such kind have been studied in a broad variety of works devoted to the dynamics of complex systems in physics (see, for instant, [25-29]). The investigations show that, for the space with 2^ configurational states, the number of maxima on randomly generated landscape comprises about 2^. Therefore, the fitness landscapes, we are interested in, turn out to be extremely rugged with a tremendous number of "mountains" and "ravines". The most impressive thing is that the fitness maxima on the landscape are actually clustered in some sort of basins. On the coarsest scale, the landscape consists of large basins with the highest mountains. When decreasing the scale, one will obtain a more detailed map of these mountains basins. One will find that each of the largest basins consists of smaller basins with the lower mountains. If the scale will decrease once more, one will see that each of these smaller basins consists of yet smaller ones, etc. Such splitting of the large basins of sequences space into smaller and yet smaller ones is helpflil to present by a hierarchically branching tree (Fig. 1). Besides, hierarchical structure of relationships between the basins of sequences is similar to a hierarchy of relations between species established on the famous phylogenetic

8

Progress in Biological Chirality

Figure 1. Hierarchical structure of nested basins tree. Hierarchy is seemed to be a significant common trait of rugged multi-dimensional fitness landscapes. Figuratively speaking, where complexity and randomness have to exist jointly, a hierarchy emerges. Thus, one of the designing tasks under the question is the description of random motion constrained by hierarchical fitness landscape. We would like to note, that this task has intensively discussed during a relatively long period, starting from the 80-ies (see, for example, [30] and references in [31-33]). However, the useful theoretical tools for the study of such models have been developed only recently [31-33].

4.

Ultrametric Diffusion The existence of hierarchically nested basins of mountains allows one to define evolutional dynamics in terms of the basin-to-basin transitions (Fig. 2). It is assumed that transitions between the basins belonging to the same level of the basin hierarchy correspond to the "mutation jumps" of a certain scale realized in the sequences space. Within such basins a quasi-steady distribution of sequences is established during a relatively short (for given scale) time, while transitions between the basins determine long time behaviour of the evolutional system. The hierarchy of nested basins corresponds to the hierarchy of mutation jumps, by

Figure 2. Schematic presentation of the basin-to-basin dynamics

Origin of Biomacromolecular Homochirality. in Search of Evolutional Dynamics

... i • ^i ^i

9

i

^i , g , ^ ^ ^ g . i^

Ultrametric distant Figure 3. Ultrametricity of the basin-to-basin transitions

which the evolutional system walks in the sequences space. nested basins topology. Let us again consider 2^ sequences consist of L and D units. These sequences can be regarded as lattice sites; the set of 2^ sites constitutes a sequences space BN. One can define the basins-to-basin transitions on this lattice as follows (Fig. 3). Let us divide the sequences space BN into two mutually disjoint basins ^^.^(aj, a^ =1,2, each of which consists of 2^"^ sites. The probability of transitions between any two sites possessed by different basins B^_X^^) is equal to p^ by definition. Next, let us divide each basin ^^.^^i) into two smaller basins B^_Xa^a^\ a^ = 1,2, each having 2^'^ sites. The probability of transition between any two sites of different hdisim Bf^_^{a^a^) is equal to p^_^ if the basins Bj^_2{a^a^) lie in a common basin B^_Xa^), and it is equal to p^ if the basins 5;^_2(fli«2) ^^® found in different basins B^_^(a^). In other words, the probability of transition between sites of different basins BJ^_^ depends on the hierarchical level, N-\ or N, at which these basins merge into a common super basin. We proceed like this until we reach a bottom level iV with each basin B^{a^a^...a^) having only one site. As a resuh, the probability of transition between any two given sequences is specified by a single value p^ among p^>p^>...>p^>...>p^, according to the level Y=\,2...N on which these two sequences happen to be in a single basin ^^. The value y, therefore, can be chosen as a parameter for definition the distance between the sequences. The most interesting feature of in this construction consist in the hierarchy of nested basins of sequences, which is similar to the hierarchy of nested ultrametric spheres. An introduction to ultrametric spaces can be found, for instance, in [30]. Therefore, the random walk in uhrametric space is relevant to the basin-to-basin type dynamics. This idea was firstly arisen in physics of complex systems. Recent studies in this field [31-32] have shown that ultrametricity indeed can be directly used for description of random motion constrained by hierarchical landscapes. The approach is based on the/?-adic numbers natural for uhrametric spaces.

10

Progress in Biological Chirality

The general /?-adic equation describing a Markovian process of a random walk in ultrametric can be written as

where x e g^ is a/7-adic coordinate in ultrametric sequences space, i.e. a sequence, and t is the time. Equation (2) is the usual balance equation for the transitions on the sequences space. The only peculiarity is that/7-adic numbers describes the sequences space, i.e. the sequences space is ultrametric by definition. The function f{x,t) is the probability distribution over the sequences space. The function p{x\y) describes the rate constant for the mutation jumps transforming sequence y into sequence x. This function establishes a relation between the fitness landscapes and the mutations jumps. In particular, the condition/?(x | >') = p(>'| x) corresponds to the "degenerate" fitness landscapes. That is the case when summits of all mountains lie on the same level, but the depths of ravines, through which the transition ways pass, are different. An important example of the degenerated landscapes is a regular hierarchical landscape: The basins of sequences split into the same number of smaller basins, and the mutation jumps between different basins of the same hierarchical level have equal probabilities. A landscape of such type is described above. In this case the transition probability p{x \ y) depend only on ultrametric distant \x-y\p = p^ (for the problem under the question p=2). Hence the equation for random walk constrained by such fitness landscapes takes the form

'^=\p^^-y\^1JM-f{x,t)\iy

(3)

5.

Characteristic Laws of the Population Degradation It is interesting to understand how the system evolves under a hierarchy of the mutation jumps. Toward this end, we will study the degradation of population S{t) in a basin of localization of the inhial distribution • ^ ^ ^ [0 |x|^>l We will consider three essentially different fitness landscapes, namely, "logarithmic landscapes", "power landscapes", and "exponential landscapes". Logarithmic landscapes are characterized by the transition probabilities having a slow (logarithmic) incidence with respect to the mutation jump scale, p ( | ^ - > ' l ; , ) ~ M l + U - > ^ | ^ ) ] " , a > 0 . This type of landscapes may be associated with the conditions, under which both short and long mutation jumps are realized with commensurable probabilities. Contrary to logarithmic landscapes, exponential landscape are characterized by an exponential suppression of the long mutation jumps, >o(|;c->'|^)~exp[-a|x->'|^J. Finally, on can consider an "intermediate" kind of landscapes - power landscapes with power incidence of the transition probability with respect to the mutation jump length,

Origin of Biomacromolecular Homochirality: in Search of Evolutional Dynamics

11

The solution of the equation (3) for the above three hierarchical landscapes have been found in [33]. A sketch of these results is as follow. The long-time behaviour of population degradation constrained by logarithmic landscapes is described by the Kohlrausch-Williams-Watts law, S{t) ~ exp[- 0 / T)"" J 0 < a < 1 . The long-time behaviour of population degradation constrained by linear landscapes is described by the power decay law, S{t)- (t/r)"' 0 < a . The long-time behaviour of population degradation constrained by exponential landscapes is described by the logarithmic decay law, S(t) ~ a[ln(r/ r)]"^ 1 < a. These results certainly give us an understanding how evolution is governed by long mutation jumps. However, a major intrigue lies in the questions what kind of error catastrophe is characteristic for the evolutional dynamics constrained by hierarchical fitness landscape. The answer may give some reasons to search hierarchically constrained dynamics capable to selection of homochiral macromolecules at the prebiotic stages of evolution.

6.

References

O.B. Ptitsin and M.V. Volkenstain, /. Biomol. Struct. Dynamics 4 (1986) 137-156. J. Monod, Chance and Necessity: an Essay on the Natural Philosophy of Modern Biology, Vintage Books, New York, 1971. [3 C.-K. Peng, S.V. Buldyrev, A.L. Goldberg, S. Havlin, F. Sciortino, M. Simons andH.E. Stanley, Nature 356 (1992) 168-170. 14 V.S. Pande, A.Y. Grosberg and T. Tanaka, Proa Nat. Acad Sci. USA 91 (1994) 12972-12975. [5 V.I. Goldanskii, V.A. Avetisov and VV. Kuz'min, FEES Lett. 101 (1986) 181-183. [6 G.F. Joyce, G.M. Visser, C. A. A. van Boeckel J.H. van Boom, L.E. Orgel and J. van Westrenen, Nature 310(1984)602-604. [7; G.F. Joyce, A.W. Schwartz, S.L. Miller and L.E. Orgel, Proc. Nat. Acad Sci. USA 84 (1987) 4398-4402. V.A. Avetisov, V.I. Goldanskii and VV. Kuz'min, Physics Today 44 (1991) 33-41. [9: M. Eigen, Naurwissenschaften 58 (1971) 465-526. [10 M. Eigen, J. McCaskii and P.J. Schuster, J. Phys. Chem. 92 (1988) 6881-6891. 111 B. Derida and L. Peliti, Bull. Math. Biol. 53 (1991) 355-363. 112 P. Tarazona, Phys. Rev. A 45 (1992) 6038-6050. [13 S. Franz, L. Peliti and M. Sellitto, J. Phys. A.: Math. Gen. 26 (1993) L1195-L1199. S. Franz and L. Peliti, J. Phys. A.: Math. Gen. 30 (1997) 4481-4487. [U [15 S. Galluccio, Phys. Rev. E. 56 (1997) 4526-4539. [16 F. Slanina and M. Kotrla, Phys. Rev. Lett. 83 (1999) 5587-5589. [17 F. Slanina and M. Kotrla, Phys. Rev. E 62 (2000) 6170-6179. [18 L. Peliti, Europhys. Lett. 57 (2002) 745-748. [19 V.A. Avetisov, in: Fundamentals of Life (Eds. G. Palyi, C. Zucchi and L. Caglioti) Elsevier and Accademia Nazionale di Scienze, Lettere ed Arti (Modena), Paris, 2002, pp. 361-367. [20 V.A. Avetisov and V.I. Goldanskii, Physics-Uspekhi, 39 (1996) 819-835 [21 V.A. Avetisov and V.I. Goldanskii, Proc. Nat. Acad Sci. USA 93 (1996) 11435-11442. [22 V.I. Goldanskii and V.V Kuz'min, Physics-Uspekhi, 32 (1989) 1-49. [23 S. Kauffman, The Origin of Order, Oxford University Press, New York, 1993. [24 D. Sherrington, Physica D 107 (1997) 117-121. [25 M. Mezard, G. Parisi and M. Virasoro, Spin-Glass Theory and Beyond, World Scientific, Singapore, 1987. [26 H. Frauenfelder, Nature Struct. Biol. 2 (1995) 821-823. [27 D.T. Leeson and D.A. Wiersma, Nature Struct. Biol. 2 (1995) 849-852. [28 H. Frauenfelder and D.T. Leeson, Nature Struct. Biol. 5 (1998) 757-759. [29 D.J. Wales, M.A. Miller andT.R. Walsh, Nature 394 (1998) 758-760. [1 [2

12

Progress in Biological Chirality

[30] [31] [32] [33]

R. Raminal, G. Toulouse and M.A. Viiasoro, Rev. Mod. Phys. 58 (1986) 765-788. V.A. Avetisov, A.K. Bikulov and S.V. Kosyrev, J. Phys. A.:Math. Gen. 32 (1999) 8785-8791. V.A. Avetisov, A.K. Bikulov, S.V. Kosyrev, and V.A. Osipov, J. Phys. A.: Math. Gen. 35 (2002) 177-189. V.A. Avetisov, A.K. Bikulov and V.A. Osipov. J. Phys. A.: Math. Gen. 36 (2003) 4239-4246.

Progress in Biological Chirality G. Palyi, C. Zucchi and L. Caglioti (Editors) © 2004 Elsevier Ltd. All rights reserved.

Chapter 2 Carbon Monoxide Clusters in the Formation of D-Sugars, LAmino Adds, Ribonucleotides and Deoxyribonucleic Acids in Prebiotic Molecular Evolution on Earth Nigel Aylward, Neville Bofinger* School of Natural Resource Sciences, Queensland University of Technology, George St., Brisbane, Queensland 4000, Australia n. aylward@student. qui. edu. au

1.

Introduction A property of most biological molecules is molecular asymmetry or chirality (Greek for hand) [1], such that the molecules are not identical to their mirror images, termed enantiomorphs. Most sugars found in nature are D-sugars, which are related to Dglyceraldehyde [2], whilst most amino acids found in proteins belong to the L-stereochemical series related to L-glyceraldehyde. Since the discovery of the configuration of glucose [3], absolute asymmetric synthesis has turned out to be a major challenge [4]. Recent successful studies have used photochemistry with circularly polarised light [5], chiral selection based on the electroweak interaction (although disputed) [6], and the combination of a magnetic field and non-polarized Ught [7], amongst others [8]. This paper proposes a prebiotic route to asymmetric sugars and amino acids based on adsorption of carbon monoxide on a cyclic porphin template. This seems possible as porphyrins have been shown to be present as geological deposits on Earth [9], form chargetransfer complexes with nitrogen containing compounds [10], and exhibit magneto-optical rotatory dispersion [11]. The synthesis is based on spontaneous reactions in a presumed primeval atmosphere. Although geological records of conditions in the atmosphere and hydrosphere of the early Earth are almost completely lacking [12], an atmosphere composed of hydrogen cyanide, formaldehyde, ammonia, hydrogen, carbon monoxide and water should give a large range of organic compounds [13]. For this work the assumption is made that carbon monoxide, and water were present and photolysis of water could have lead to a reducing medium [14]. An examination of the total energy of these molecules and their transformation products; sugars and amino acids; suggests that the latter could have arisen from reactions in a primeval atmosphere that may have existed on Earth [12], and would be expected to occur somewhere in the Universe provided that a suitable reaction mechanism could be envisaged. For the spontaneous chemical formation of these biologically important molecules it is

14

Progress in Biological Chiiality

satisfying if all are formed from the same initial milieu of reactants by established chemical reactions that are kinetically feasible and preferably thermodynamically favoured. Here, one such set of reactions is proposed based on the availability of the primary reactants, thermodynamic, and kinetic considerations.

2.

A Theory of Asymmetric Induction As this work is applied theoretical chemistry a simplified theory is given here to justify the suggested asymmetric induction. The following considerations apply to an adduct of carbon monoxide with the magnesium porphin complex of approximately C4V symmetry as designated in Fig. 1. Porphin and its derivatives [15], have been treated by the free electron model [16, 17], as an 18-membered benzenoidal ring where the secular equation becomes identical in form to the LCAO secular equation for the problem [15]. Porphin and tetrahydroporphin have been treated under the point group D4h using LCAOs to obtain more accurate wavefunctions and molecular orbital energies [18]. These are also available from present day ah initio calculations. Using the free-electronmodel or particle on a ring [19], the energy levels are given by. En^

n - 0 , 1,-1,2,-2,.

where the quantization is evident, and the wavefunctions by,

yn((t))=-V^""

n = 0, 1,-1, 2,-2,

Magnetic Field

Electric Field

Rl

MgCOporphin (2) Figure 1. The orientation and charges of the MgCOporphin charge transfer adduct indicating the in-plane electric vector and the perpendicular magnetic vector of theradiationto cause an allowed A2 -> E transition

Carbon Monoxide Clusters in the Formation of D-Sugars, L-Amino Acids, ...

15

The quantum numbers of the highest filled level are + 4 or -4, and of the lowest unfilled level, + 5 or -5, hence the transition is fourfold degenerate [15]. The electric transition moment for this transition is given by the integral, < \|/* I m I \|/> dx and for light polarized in the x-direction, the integral is [15]: rp/47i <e-'^* le^* + e " * | e ' ^ ^ > ^ 0 where rp is the radius of the circle. A consideration of the four transition moments shows that the allowed bands correspond to changes of the quantum number of+1 or -1. The angular momentum in this model is quantized and perpendicular to the plane of the ring with a value, Mz = m hJ2n

m = 0, 1, -1, 2, -2,...

where m is the magnetic quantum number - the same as the quantum number for the zcomponent of angular momentum. Thus, the interaction of the orbital motion of an electron with a magnetic field is quantized, only the values, E = - m pm | H |

m = 0, 1,-1, 2,-2,...

being possible, where the atomic unit of magnetic moment is given by [20], Pm= -eh/4^mec Using time-dependent perturbation theory [20], the electric moment induced by the magnetic field of the radiation perpendicular to the porphin plane is given by an expression proportional to the product of the electric and magnetic transition moments, where, mko = <\i/k* | m | v|/o>

and,

jdzok = <¥o* | \iz\Wk>

and H is the applied field. The presence of the radiation magnetic field H [21], perturbs the ground and excited wavefunctions requiring that these are described using a summation of higher eigenstates. The presence of these is sufficient for there to be an oscillating electric transition moment perpendicular to the ring which is finite. Thus, the presence of the magnetic field pointing upwards (i.e. arising from the direction of the incident radiation). Fig. 1, together with the allowed in-plane A2 -> E electronic transition from the electric transition moment, should result in an oscillating in-phase electric field perpendicular to the plane, and an angular momentum of the ring electrons in accordance with Lenz's law [22]. The electric transition imparts sufficient energy to the charge transfer adduct [10], (calculated as 0.2479 h from HOMO to LUMO) for the activation energy to be overcome. Fig. 2, and the dissociation energy to be exceeded, such that the carbon monoxide molecule moves to the periphery of the ring system where a transition state consisting of an aziridine-ring is ormed. The

16

Progress in Biological Chiiality

Y DISTANCE(A)

X DISTANCE(A)

Figure 2. Total Energy (-1297 + X haitree) of the MgCO-porphin complex. Origin on the Mg"^ ion, x-axis through a pyrrole N atom, y-axis in-plane

carbon monoxide entity of the charge transfer complex is positively charged so that during half the cycle of the electric transition moment perpendicular to the ring this assists in increasing its height above the ring as shown in Fig. 3, and an increase in the energy of the system as shown in Fig 2.

R2 DISTANCE(A)

0.5

1 1.5 R1 DISTANCEfA)

Figure 3. Height of the caiixm monoxide entity above the ring as it traversesfromthe centre at coordinate (0,0,0) to near a pyrrole nitrogen atom. The coordinates of the caibon of the CO entity are at (2.3,0.6, 1.3). The ordinate A and abscissae are in miits of 10 '^ m (Angstrom)

Carbon Monoxide Clusters in the Formation of D-Sugars, L-Amino Acids,

17

-200

-400

R2 DISTANCE(A)

1

1

R1 DISTANCE(A)

Figure 4. The potential energy surface of the acetaldimine-carbon monoxide adduct on the surface of a porphin molecule. Magnesium ion at coordinate (0,0,0) with coordinated imine. Carbon monoxide entity at the transition state near a pyrrole nitrogen atom. The ordinate is the total energy (-1427.0 + X.10-3) (hartree)

This formation of a carbon monoxide adduct is postulated to be repeatable, with the subsequent formation of high energy reactive transition complexes. However, after the initial transition complex is formed the bonding in the porphin ring system is more fixed, with anticlockwise transition moments leading to reactions. On reduction these complexes should yield D-sugars. Moreover, the formation of charge-transfer complexes with imines predisposes them to form aziridine complexes hydrolysable to L-amino acids preferentially, as shown in Fig. 4.

3.

Methods of Calculation The computations tabulated in this paper used the SPARTAN (1994) and GAUSSIAN98(1998) packages, [23] using IBM RISC 6000 and SP2 computers. Equilibrium structures were calculated at the HF/6-31G* and MP2/6-31G* level. For the large porphins energies more accurate than MP2 were not available. For jobs run under GAUSSIAN98, geometry optimizations used the VTIGHT option. The basis set used in the MP2 geometry optimizations was the standard split-valence double zeta basis set, augmented with a d-type polarization function on the non-hydrogen atoms [24]. The lowest ab initio level of theory which provides an account of the dispersion forces and correlation correction to the dipole moment is the MP2 level (the second order MoUer-Plesset perturbation theory) [25]. This is a suitable level of theory for the systems considered in the present work. Since theoretical Hartree-Fock frequencies tend to exceed known experimental values, the calculated frequencies were scaled by a factor of 0.89, when used to derive zero point vibrational energies (ZPE) and thermochemical corrections to the total energies. This is

18

Progress in Biological Chirality

consistent with the recommendation of Pople et al. [26]. The zero-point energies were calculated at the HF level of accuracy. Calculations in this paper that use only HF energies without adding the zero-point energy are explicitly indicated with a subscript, e.g., A H(HF) The energy unit used for these calculations is the atomic unit for energy, the hartree. 1 hartree = 627.5095 kcal. mol"^ [23].

4.

Results

4.1 Total energies (hartrees) The total energies and zero point energies for the HF and MP2/6-31G* equilibrium geometries are given in Table 1. 4.2 The prebiotic synthesis ofD-sugars andL-amino acids The reactions proposed for the formation of D-glyceraldehyde and L-alanine involve the reactions of carbon monoxide either with itself (condensation) or with an imine as follows: Table 1. HF and MP2 /6-3IG* total energies (hartrees) for the respective equilibrium geometries Molecule (1) Mgporphin Mgporphin (triplet, 6-3IG**) (2) Mg(CO)porphin (3) MgporphinCO (4) Mgporphin(CO)3 (5) Mgporphin(CO-)3 (6) MgCOporphin(CO)2(CO-)2 (7) Mg(CO-)porphin(CO-), (8) Mg(C(OH)-)porphinCO(C(OH))2CH(OHr (9) Mg(C(OP=0-))-porphinCOC(OH)COC(OH)CH-0(10) Mg(C(OH)-)poiphinCOC=C(OH)CH(OH)^^ (11) Mg(C(OH)-)porphinCOUCHC(OH)CH(OH)^" (U-uridine) porphin Mg(CH3CH=NH)porphinCO Mg(CH3CH-NH)porphin \C0/ MgCOporphin(CO-CO)2 glyceraldehyde deoxyribose ribose CO H2 HO-P=0 0-P=0 OH" OH 2-methyl aziridine-3-one CH3-CH=NH uracil H2O

HF hartree

MP2 hartree

-1181.86208

-1185.12250 -1185.10392 -1298.13452 -1297.93784

-1294.60561 -1294.48601 -1519.80735 -1519.79776 -1745.17902 -1745.14346 -1747.89231 -2237.45751 -1672.19465 -2084.83363 -983.25693 -1427.59183 -1427.63029 -1745.17902 -341.64958 -494.58123 -569.43280 -112.73434 -1.12683 -491.05064 -490.49517 -75.32660 -75.38228 -245.77318 -133.07280 -412.47029 -76.01075

ZPE (HF) hartree

-1524.00686 -1749.95424 -1750.01626

-986.48501

0.31165

-1431.64414 -1749.97463 -342.56404 -495.93959 -570.97086 -113.02818 -1.14414

-246.48876 -133.49001 -413.63259 -76.19924

0.10256 0.17281 0.17816 0.00484 0.01033 0.01942 0.00711

0.08481 0.07392 0.02148

Carbon Monoxide Clusters in the Formation of D-Sugars, L-Amino Acids, 3C0

+

3H2 -

19

C3H6O3

glyceraldehyde AH(MP2/ZPE) = 0.00368 h

CH3-CH = NH + CO AH(MP2/ZPE) = 0.03481 h

CH3-CH - NH Rl \ / R2 CO

However, it is well known that these reactions do not proceed easily. A mechanism is needed whereby this could have occurred. The template proposed here is that of original porphin, and the mechanism is similar to present photosynthesis. However, the usual reliance on thermodynamic data or kinetic data will not justify the formation of the transition complex as photosynthesis as usually stated involves the input of electromagnetic energy without which it is less efficient. It is assumed here that the energy needed to form the complexes is provided by the radiation. 4.3 The formation of the magnesium-porphm'carbon monoxide complex In present day chlorophyll the porphin moiety often contains a magnesium ion. This structure has been taken as a basic template for the postulated sugar synthesis. It has been stated that there is not a distinction between the principal and subsidiary valences in the metallic complexes [18]. The reaction in Scheme 1 is proposed as the route by which carbon monoxide first complexes with magnesium chelated porphin.

Mg*porphin + CO

Mg-(CO)-porphin AH(HF) = -0.00919 h Scheme 1

(2)

If the Mg-porphin (triplet state) is in the triplet state, the enthalpy change is also favourable, as follows: A H(MP2/6-3iG**) ~ -0.10388 h

It is postulated that when the complex becomes excited from photolysis according to the equation, Mg-(CO)-porphin-

Mg-porphin-CO

20

Progress in Biological Chirality

the CO group is able to move through a transition state to the porphin ring, as shown. The energy of the product is almost identical to that of the excited reactant, but the excitation energy is more than the activation energy for the transformation.

Mg(CO)-porphin

(2)

Mg-porphinCO

(3)

AH(MP2) = 0.19668 h The potential energy surface for the magnesium-porphin-carbon monoxide comple is shown in Fig. 2. 4.4 The isomeric states of the magnesium-porphin complex The magnesium complex can be described classically by either of the two structures shown in Scheme 2.

Mg-porphin (la)

Mg-porphin (lb) Scheme 2

The actual structure may be resonant between them. However, when the CO group migrates to the ring it must pick a C=N bond from either structure (la) or (lb). The results of the following analysis are built on the assumptions of the model that the surface to react with the carbon monoxide is the same surface on which the radiation is incident. If the magnesium-porphin-carbon monoxide complex is considered approximately under the point group C4v, then the HOMO of symmetry A2 and LUMO of symmetry E, as are shown in Fig. 5. The A2 -> E transition between these states is allowed as previously illustrated [15]. The electric vector is expected to be in the plane of the porphin ring, with the magnetic vector perpendicular to the ring, as shown.

Carbon Monoxide Clusters in the Formation of D-Sugars, L-Amino Acids. ... (a)

^

21

(b)

Figure 5. (a) MgporphinCO HOMO and (b) MgporphinCO LUMO

This should result in a very small induced electric dipole moment perpendicular to the ring, whose direction should be in a direction to assist the carbon monoxide moiety to gain height above the ring whilst approaching the transition state, as shown in Fig. 3. With the magnetic vector perpendicular to the ring the electronic charge is expected to precess counterclockwise (when viewed from above) in accord with Lenz's law, to produce resonance form (la), whilst the induced electric dipole should also point upwards to assist the elevation of the carbon monoxide moiety, and result in the formation of a complex with form (la). Conversely, when the magnetic field of the radiation points downwards, the induced circulation of electron charge is expected to be clockwise, in accord with Lenz's law, but the induced electric dipole should point downwards thwarting the elevation of the carbon monoxide,moiety. The net result is that when reaction occurs it will be supposed to be with structure (la). If the magnesium ion subsequently takes other ligands then the CO group may preferentially react with those slightly counterclockwise to itself rather than those clockwise to itself Similarly, if clusters of CO groups are present on the ring, the transition moment is supposed to be more effective in linking them if moving counter-clockwise than clockwise i.e. it is supposed that the direction of the net electronic transition moment is important. The activation energy to bond the carbon monoxide entities is less than 80 kcal.mol" Wave functions for the LUMO suggest that it has a high TT* character as does the ground state of the magnesium-porphin-CO complex. This is here illustrated with the formation of a three-carbon bonded complex reducible to glyceraldehyde and glycerol.

Mg-porphin-(CO-)3 (5) radical AH(HF) = 0.00959 h

Mg-porphin-(CO)3 (4)

22

Progress in Biological Chirality

4.5 The formation of L-amino acids It is supposed that an imine residue such as acetaldimine coordinated to the magnesium ion can react with the high energy complexed CO group on the surface periphery of the porphin ring, but it can only approach for optimum reactivity in one orientation, as shown: Mg-porphin-CO + CH3CH=NH AH(HF) = -0.03303 h Mg.(CH3CH=NH)porphinCO AH(HF) = -0.03846 h

^ Mg-(CH3CH=NH)-porphin-CO

. Mg(CH3CH-NH)porphin

\ / CO

MgacetaldimineporphinCO complex The potential energy surface to form the acetaldimine - carbon monoxide complex is shown in Fig. 4. From this diagram the activation energy to form the 3-methyl aziridone is calculated as 0.08 h. A drawing of the transition state is given in Fig. 6.

Figure 6. Transition state of the MgacetaldimineporphinCO complex

Carbon Monoxide Clusters in the Formation of D-Sugars, L-Amino Acids,

23

4.6 The formation of carbohydrates from carbon monoxide and hydrogen The justification for the premises in this paper have a firm basis in the favourable energy changes involved in the conversion of carbon monoxide and hydrogen to sugars as exemplified by the following two equations involving the production of 2'-deoxyribose and ribose, respectively. 5 CO + 6H2

-^

C5H10O4 + H2O deoxyribose

The equation gives the enthalpy of reaction as, AH(MP2/zPE) = -0.03690 h This indicates that this is an exothermic reaction and may be spontaneous. Similarly, for the formation of ribose the equation is, 5 CO + 5H2

^ C5H10O5 ribose

The equation gives the enthalpy of reaction as, AH(MP2/zPE)= -0.01824 h Again, the enthalpy change suggests that the reaction could be spontaneous. However, the usual reliance on thermodynamic data or kinetic data will not justify the formation of the transition complex as photosynthesis as usually stated involves the input of electromagnetic energy without which it is less efficient. It is assumed here that the energy needed to form the complexes is provided by the radiation. 4.7 The formation of the pentoses The main interest is the formation of D-ribose and D-2'-deoxyribose from the bonding of 5 CO groups complexed to the porphin ring.

Mg(CO)porphin(CO)2(CO-)2 (6)

Mg-(CO-)-porphin-(CO-)4 (7) radical AH(MP2) = -0.06203 h

24

Progress in Biological Chirality

The pentacovalent complex is shown in Fig. 7. It is not unique as there are at least two other similar structures for the complex of entirely equivalent energy that might be expected to be reducible to D-arabinose and D-xylose. In the reduction of this complex it is assumed that the precessional movement of the electrons allows a proton to be bonded, with the increase of charge to +1, as shown.

+ 2H2 +

HT

Mg-(CO-)-porphin-(CO-)4 (7)

Mg(C(OH).)porphin-CO-(C(OH))2-CH(OH)'^ (8)

This reaction is energetically favoured. AH =-0.49519 h

Carbon Monoxide Clusters in the Formation of D-Sugars, L-Amino Acids, ... 4.8 The formation of ribose phosphate The supposition is that this just involves the addition of free radical electrons to the phosphorus atom shells of metaphosphite, whilst it is complexed to the Mg^^ ion, as shown. MgCOporphin(CO-CO)2 + HO-P=0 + H2

• Mg(COP02')porphin'(CO-(COH)2CHO-)

AH(HF) = -0.10102 h

*^

^

^

#

"-^^m' ^fe^

Mg-C(OPO-)-porphin-(CO-C(OH)-CO-C(OH)-CHO-) complex

W This may be compared to the following reaction involving charged species. Mg- COH)-porphin-(CO-(COH)2-CHOH)'^ -^ Mg-(COP02-)-porphin-(CO-(COH)2-CHO-) + 0-P=0' + H2 AH(HF) = -0.19686 h 4.9 The formation of a ribose ketene This reduced complex may lose hydroxyl ion to produce a ketene according to the energetically favoured reaction shown, HOC

+ H" Mg(C(OH)-)porphinCO(C(OH))2CH(OH)'^ (8)

+ H2O Mg(C(OH)-)porphinCOC=C(OH)CH(OH)^'^ (10)

AH =-0.31309 h

25

26

Progress in Biological Chirality

4.10 The formation of a D-2 '-deoxyhbose nucleoside This ketene is of high enough energy for the reaction with a nucleic acid base to be favourable, as shown,

+ U MgC(OH)-)porphinCOCC(OH)-CH(OH)^^ MgC(OH)-porphinCOUCHC(OH)CH(OH)'^ (10) (11) AH =-0.16868 h The structure of the nucleoside is shown in Fig. 8. Reduction of these complexes by the products of photolysis [14], should yield sugars, nucleosides or nucleotides. porphin-H20 + hv —• porphin^ H20-

+

H2O"

^ H- -f OH-

porphin^ + H2O —• porphin + H20^ H20^

•H^

+ OH

The reductions are always favourable as the molecules are of high energy resulting from the many photochemical excitations.

Figure 8. The structure of the nucleoside MgC(OH)-porphinCOUCHC(OH)CH(OH)^

Carbon Monoxide Clusters in the Formation of D-Sugars, L-Amino Acids, ...

27

eg. Mg(COH)porphin-(CO-(COH)2-CHOH)^ + BHi-^Mgporphin + HOCH2-(CH(OH))3-CHO + H ribose A H(HF) = -0.02209 h

5.

Discussion The exact sequence of reactions leading to ribose, deoxyribose, the nucleosides, and nucleotides, is open to conjecture, but it is postulated here to proceed via carbon monoxide oriented clusters excited by radiation. In the reduction and protonation of the Mg-porphin-(CO)5, it is expected that other acceptor molecules may be used apart from the proton. For example formaldehyde might lead to the hexoses. Similarly trimethyl sulphide cation may lead to the sugars rhamnose and fucose. The 1-CO group is similarly expected to be susceptible to attack by formaldehyde anion giving hexoses, whilst glyoxal anion might lead to heptoses. Finally, the ease of producing 2,3-unsaturated compounds leads to a strong indication of the very early formation of ascorbic acid in prebiotic molecular evolution.

6.

Conclusion From the postulates presented here it is clear that prebiotic paths to the biologically active sugars are conceivable. Further work at a higher level of accuracy may provide some more data to support this work and may change the enthalpy values tabulated in this paper.

7.

Acknowledgements Appreciation is expressed to Queensland University of Technology Centre for Instrumental and Developmental Chemistry for the equipment and facilities; to Mr. A. Lewis, S. Walsh, and Dr. J. Young and M. Barry, B. Savage, and A. Rasmussen of the Supercomputing Department, and Dr. A. Wiegand. N. Aylward is grateful for the award of a scholarship from QUT.

8.

References

[1] (a) G. Palyi, C. Zucchi and L. Caglioti, Eds., Fundamentals of Life, Elsevier and Accademia Nazionale di Scienze, Lettere ed Arti (Modena), Paris, 2002. (b) G. Palyi, C. Zucchi and L. Caglioti, Eds., Advances in BioChirality, Elsevier, Amsterdam, 1999. (c) I. Tinoco, K. Sauer and J.C. Wang, Physical Chemistry, Prentice-Hall Inc., Englewood Cliffs, N.J., 1978. (d) R. Janoschek, Ed., Chirality from Weak Bosons to the Alpha Helix, Springer-Verlag, Berlin, 1991. (e) E.H. Rodd, Ed., Chemistry of Carbon Compounds, Elsevier, Amsterdam, 1951. [2] A.L. Lehninger, Biochemistry, Worth Pubhshers Inc., N.Y., 1975. [3] C.S. Hudson,^^. Carb. Chem. 3 (1948) 1-22. [4] B.L. Feringa and R.A. van Delden, Angew. Chem., Int. Ed 38 (1999) 3418-3438. [5] N.P.M. Huck, W.F.Jager, B. de Lange and B.L. Feringa, Science 111> (1996) 1686-1688. [6] R. Berger and M. Quack, J. Chem. Phys. 112 (2001) 3148-3158. [7] G.L.J.A. Rikken and E. Raupach, Nature 405 (2000) 932-935. [8] B.L. Feringa, Science 292 (2001) 2021-2022.

28 [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22]

[23]

[24] [25] [26]

Progress in Biological Chirality R. S. Czemuszewicz, J. Porphyrins Phthalocvanines 4 (2000) 426-431. D.G. Whitten, I.G. Lopp and JP.D. Wildes, J.Am. Chem. Soc. 90 (1968) 7196-7200. P.J. Stephens, W. Suetaak and P.N. Schatz, J. Chem. Phys. 44 (1966) 4592-4602. (a) E. Nisbct and N. Sleep. Nature 409 (2001) 1083-1093. (b) G. Palyi and C. Zucchi, Orig. Life Evol. Biosphere 30 (2000) 172. F. Raulin, Orig. Life Evol. Biosphere 30 (2000) 116-117. F.K. Fong, Light Reaction Path of Photosynthesis, Springer-Verlag, Berlin, 1982. W.T. Simpson, J. Chem. Phys. 17 (1949) 1218-1221. (a) N.S. Bayliss, J. Chem. Phys. 16 (1948) 287-292. (b) J.R. Piatt, J. Chem. Phys. 17 (1949) 484-495. W.T. Simpson, J. Chem. Phys. 16 (1948) 1124-1136. H.C. Longuet-Higgins, C.W. Rector and JR. Piatt, J. Chem. Phys. 18 (1950) 1174-1181. P.W. Atkins, Molecular Quantum Mechanics, Clarendon Press, Oxford, 1970. W. Kauzmann, Quantum Chemistry, Academic Press Inc., N.Y.. 1957. J.T. Hougen, J. Chem. Phys. 32 (1960) 1122-1125. (a) D.C. Giancoli, Physics, Prentice Hall Int. Inc., Englewood Cliffs, New Jersey 1985. (b) R.L. Weber, K.V. Manning, M.W. White and G.A. Weygand, College Physics, McGraw-Hill Book Company, N.Y., 1959. (a) SPARTAN (1994) Ver. 3.1, Wavefunction Inc., 18401 Von Karman Avenue, Irvine, Cahfomia 92715, U.S.A. (b) Gaussian98 (1998) Users Reference. Gaussian Inc., Carnegie Office Park, Bldg.6., Pittsburgh, PA 15106. USA. P C . Hariharan and J.A. Poplc, Theoret. Chim. Acta 28 (1973) 213-222. C. MoUcr and M.S. Plesset, Phys Rev. 46 (1934) 618-622. J.A. Pople, H.B. Schlegel, R. Krishnan, D.J. De Frees, J.S. Binkley, M.J. Frisch, R.A. Whiteside, R.J. Hout and W.J. Hehre. Int. J. Quantum Chem. Symp. S 15 (1981) 269-278.

Progress in Biological Chirality G. Palyi, C. Zucchi and L. Caglioti (Editors) © 2004 Elsevier Ltd. All rights reserved.

Chapter 3 Molecular Clockworks as Potential Models for Biological Chirality Lajos Bencze,^'* Claudia Zucchi,^ Luciano Caglioti,'' and Gyula Palyi"*'* "^ Muller Laboratory, Department of Organic Chemistry, University of Veszprem, P.O. Box J58, Veszprem, H-8201, Hungary [email protected] Department of Chemistry, University ofModena and Reggio Emilia, Via Campi 183, Modena, 1-41100, Italy [email protected] ^ Department of Chemistry and Technology of Biologically Active Compounds, University "La Sapienza", P.le A. Moro 5,1-00185 Roma, Italy

1.

Introduction The origin of biological (homo)chirality is one of the last (?) classical great secrets of Natural Sciences [1]. Several elegant approaches have been postulated for the solution of this mystery [2]. These include the photochemical effect of circularly polarized sunlight in Nature [3], the influence of asymmetry of weak nuclear forces on molecular reactivity [4], autocatalytic amplification of mineral morphological (macroscopic) chirality [5] and other highly intellectual working hypotheses. We suspect that the interaction of coordinated ligands (prominently: chiral plus pro-chiral) may exert an, at least partial, influence on the origin of biological (or "pre-biological") enantioselection, on the maintenance of its level [6] or on the development of biological signal transduction systems [7]. Interestingly, the main focus of the research aimed at obtaining molecular level enantioselection is at the meeting point of two other great concerns of contemporaneous chemistry, namely (a) to achieve enantioselective syntheses by molecular catalysis [8] and (b) to miniaturize down to the molecular level [9]. Both are considered in the pioneering papers of Kelly et al. [10], reporting partially hindered molecular systems, where the reduction of rotational freedom produces brake/cogwheel like interactions. This principle, combined with application of photochemistry, is used also in the so-called molecular switches [11]. We have studied some intermediates of coordination catalysis from this standpoint. Here we summarize the results obtained with mononuclear complexes, alkylcobalt carbonyls, which are well known from C,C-bond making catalysis [12].

30

Progress in Biological Chiralin

RO

/^CH.

O

oc Co

CO

p^ -Ph Ph

Figure 1. Conformations of the of the ester group and of the triphenyl phosphine ligand (notation)

2.

Structural Studies X-Ray diffraction of crystalline phases of more than a dozen flexible ROC(0)CH2Co(CO)3PR'3 (1) type alkylcobalt carbonyl phosphine complexes showed [13, 14] that these complexes form enantiomeric conformations with achiral R and R'. Theseconformations are characterized by the opposed helical chirality of the PR'3 (mostly R' = phenyl) ligands [15] and each of these enantiomeric structures is accompanied by only one conformation of the ester fragment (Fig. 1). Essentially, the system exhibits 100% diastereomeric yield. This particular situation, which leads to a dramatic reduction in the number of the observed conformers, with respect to the number of statistically possible (energetically apparently equivalent) ones, prompted us to perform molecular mechanical and quantum chemical calculations on one of these compounds: [(ethoxycarbonyl)methyl]cobalt tricarbonyl triphenylphosphine, EtOC(0)CH2Co(CO)3PPh3. These studies enabled an attempt to be made at performing a preparative/structural evaluation of the theoretical considerations.

3.

Quantum Chemistry Equilibrium energy calculations [16, 17] showed that a//"statistically" possible, observed and non-observed stereoisomers are energetically equivalent in the ground state, within the limits of computational error. Thus, it should be concluded that the conformation preferences ("selection rules") observed for the crystalline phases are not (prevalently) of ''thermodynamic" origin. Consequently, the possible transformation routes ("racemization")

Molecular Clockworks as Potential Models for Biological Chirality

31

O—Et „ „ -^"^^r^

OC

ester fragment

^,vCO Co ^.^^

cobalt carbonyl fragment

/7

triphenyl phosphine ligand

Figure 2. Schematic representation of the ester, cobalt carbonyl and triphenyl phosphine fragments in the molecules

were analysed by quantum chemical methods [20]. These promised to afford greater insight into the forces operative in the intramolecular self-organization. Based on ground-state energy calculations, molecular graphical studies, rigid and relaxed rotor approach conformation analyses^ it is deduced that the mechanism of the inversion proceed with correlated motions of coaxial rotors, coupled bevel gear like rotors with different periodicity and the first example of a coupled conrotation ever observed in molecular systems. The most informative potential function is shown in Fig. 3. The results indicate that the rotation of the ester group (Fig. 2) to left or right proceeds with different energies because of steric crowding. As the ester group rotates, this induces a "concerted" rotation of the Co(CO)3 moiety in the equatorial plane of the complex. On the other hand, rotation of the Co(CO)3 moiety in the equatorial plane induces reorientation of the phenyl groups on the phosphine ligand through repulsive interactions with the ortho-H atoms of the Ph rings. Consequently, if the ester group spontaneously reorients from re position, then it will turn preferentially to left (towards the ester carbonyl group) causing clockwise (con)rotation of the Co(CO)3 moiety, which again generates M conformation of the PPhs ligand (Fig. 4). Thus, once rotated, the ester group will be in si position, accompanied by M-PPhs, while, starting from the ^/-ester, after rotation relY combination will develop, both being exactly what was observed experimentally in the crystalline phase. The steering effect of the ester group is also due to its rigidity, owing not only to the central sp^ carbon but also to the C-H O type H-bridge interaction [21], identifiable on the basis of the X-ray data, between one of the ethyl a-H-s and the ester carbonyl-0. It should be noted that this effect renders the two Et a-H atoms inequivalent, thus generating a centre of chirality on the ethyl-a-C atom. We analysed the correlation between the configuration of this carbon atom and the conformations developing in the crystalline phase, and found that the silM and re/P couplings always corresponded to the S a-C and R a-C configurations, respectively. ^ Racemization of the whole molecule involves the simultaneous rotation of 6 (formal) a-bonds, 16 dihedral angles and a total of 48 torsional interactions.

32

Progress in Biological Chirality 25

collision of the ester carbonyl group with the equatorial carbonyl ligands

collision of the ethoxy group with the equatorial carbonyl ligands

(0 0)

X

isomer _ re -50

175

-25

200

Dihedral Angle Co-C(H2)-Csp^=0 (©i), deg.

Figure 3. Conformational potential function of the ester group (torsion col) in EtOC(0)CH2Co(CO)3PPh3 The correspondence of the fragments of the EtOC(0)CH2Co(CO)3PPh3 molecule to the mechanism of parts of a classical (mechanical) clock [22] is shown in Fig. 4. These modelling studies revealed that the driving force of the high stereoselection is thermal energy, just as in the case of macromechanical devices, and the low-energy interconversion of enantiomers (when applicable) takes place through a clockwork mechanism, whereas the formation of the diastereomers corresponds to gear slippage, a malfunction also found in mechanical clocks. The ester fragment is always oriented quasi-parallel to the Co(CO)3 group (plane-to-plane angle 25°). Other possible conformations around the C{sp^)-Q{sp^) bond are not populated. Based on experimental data and density functional calculations it is concluded, that the

escapement

gear

balance

Figure 4. Mutual rotation of the fragments and analogy with mechanical clockwork in EtOC(0)CH2Co(CO)3PPh3

Molecular Clockworks as Potential Models for Biological Chirality

33

conformational chirality may be attributed mainly to the r|^ type coordination between the aester and the central cobalt atom [23]. This novel type of interaction (autosolvation) can open a new way for the consideration of esters as prochiral ligands invoking stereoselection amongst the incoming ligands and behaving as a source of chiral induction in catalysis, as well. Our X-ray studies [14, 24] on additional ROC(0)CH2Co(CO)3PR3 complexes showed different coupling of the si/re (ester) and MyP (phosphine) conformations, but always the same, or even larger reduction in the number of conformers. We attribute this to the influence of the R group on the potential curves. A greater reduction of the number of isomers was observed with chiral R or R' groups [25].

4.

Extension of the Clockwork-Principle We subjected the above structural/theoretical picture to a preparative/structural/spectroscopic analysis. First of all, it should be pointed out that the surprising early CD spectroscopic resuhs, obtained with chiral derivatives of complexes 1 [26], showing CD bands that could be attributed only to chiral perturbation of transitions involving the d electrons of the Co [23], are in excellent agreement with the novel structural and theoretical results [16, 23 d]. An interesting extension of the theoretical results obtained with an achiral complex 1 (R = Et) and structural observations with several additional achiral analogs was the introduction of centres of configurational chirality into the ester group. While at the achiral complexes the concerted development of chiral conformations resulted in the observation of only 50% of the statistically possible isomers, the introduction of, even racemic forms, of the 5Bu-group [25a] resulted in the development of only 25% of all possible isomers in the crystalline phase (Fig. 5). A solution spectroscopic (^H-NMR) study showed that this dramatic selection rule may be retained also in the liquid phase (Fig. 6). An additional possibility for a preparative/structural study can be deduced for other systems with C(5p^).CH2-Co-(CO)3-PPh3 sequences which should show comparable behaviour if our analysis is correct.

(a)

sQ\x (b)

(configuration)

S 1 R

S

12.5 % 1 si

Ester

25%

PPhs

50 %

R re

/////

S si

R

m

S

1R 1 re

Figure 5. Correlation of statistically possible isomers/conformers of ROC(0)CH2-Co(CO)3PPh3 complexes with those observed experimentally by X-ray diffraction. (a) R = achiral; (b) R = sBu; m observed combinations

34

Progress in Biological Chirality

Acetone - de

ppm

2.16

2.14

2.12

Figure 6. 'H-NMR spectrum of the a-CHj group of the 5BuOC(0)CH2Co(CO)3PPh3

A promising group of such compounds are the benzylcobalt carbonyl derivatives [27]. We have prepared and structurally characterized the following /r^iw5-RCH2Co(CO)3PPh3 complexes: R = para-tBuC^- [28] (Fig. 7), meta-ClCeiU' [29] as well as two orthosubstituted phenylacetylcobah tricarbonyl triphenylphosphine derivatives (Fig. 8) [30]. These benzyl complexes show in the crystalline phase quasi-parallel position of the aromatic (benzyl) ring and the Co(CO)3fragment,as well as two helical isomers (M and P, in 1:1 ratio) of the PPha group. The symmetric para-tBu derivative shows two (crystallographically) independent molecules in the P f phase and their enantiomers (generated by an inversion centre), that is, a picture indicating the existence of all four isomers. Most interestingly the w^/a-Cl-derivative, which is asymmetric, but the Cl-substituent is farther from the carbonyl ligands (thus generating not overly strong crowding) shows the presence of all four isomers too, but in a

C13'

C12'

C24

C19'

Figure 7. ORTEP drawing of the two independent molecules (one enantiomer each) in the ciystalline phase of p-tBuC6H4CH2Co(CO)3PPh3 [28]

Molecular Clockworks as Potential Models for Biological Chirality

35

vC16

€34

IC33

C20/

C21

Figure 8. ORTEP drawing of the o-MeC6H4CH2C(0)Co(CO)3PPh3 and o-PhC6H4CH2-C(0)Co(CO)3PPh3 molecules (one enantiomer each) [30]

ratio of 12:38 % for each PPhs conformation (which however are in 50:50 % ratio). This latter effect perfectly reflects the "slipping" malfunction of the escapement/gear connection in mechanical clockworks [22]. These observations are in harmony with the theoretical analysis, and suggest directions of future preparative/structural and theoretical work. The striking similarity between this molecular system and the mechanical clock presents new synthetic and theoretical directions which are now under investigation in our laboratories. One of the most important of these directions is the connection of clockworktype steric hindrance to chiral induction, a phenomenon which is of utmost importance from the viewpoints of theoretical biology [1], of synthetic organic chemistry [8, 31] and perhaps even of the origin(s) of life on Earth [32].

5.

Acknowledgement The authors acknowledge financial support for this research to the [Hungarian] OTKA program (Grant Number T-016326, T-035221), to the [Italian] Ministry of University and Research (MURST) and [Italian] National Research Council (CNR) as well as Prof R.D. Adams (Columbus, SC, USA) for his comments on the manuscript.

6.

References

[1] G. P^yi, C. Zucchi and L. Caglioti, in: Advances in BioChirality (Eds. G. P^lyi, C. Zucchi and L. Caglioti) Elsevier, Amsterdam, 1999, pp. 3-12. [2] (a) L. Keszthelyi, Quart. Rev. Biophys. 28 (1995) 473-507. (b) L. Markd, Diss. Savariensis 24 (1998) 1-64. (c) G. P^yi, K. Micskei, L. Bencze and C. Zucchi, Magyar Kern. Lapja 58 (2003) 218-223.

36

Progress in Biological Chiralit>

[3] A. Vitkin, Opt. Photonics News 111 (1996) 30-33. [4] (a) A. Szabo-Nagy and L. Keszthelyi, Proc. Natl. Acad. Sci. USA 96 (1999) 4252-4255 and refs. therein. (b) W.A. Bonner, Chirality 12 (2000) 114-126. [5] (a) K. Soai, T. Shibata, H. Morioka and K. Choji, Nature 378 (1995) 767-768. (b) K. Soai and T. Shibata, in: Advances in BioChirality (Eds. G. Palyi, C. Zucchi and L. Caglioti) Elsevier, Amsterdam, 1999, pp. 125-136. (c) K. Soai, S. Osanai, K. Kadowaki, S. Yonekubo, T. Shibata and I. Sato, /. Am. Chem. Soc. ill (1999) 11235-11237. (d)K. Soai, T. Shibata and I. Sato,^cc. Chem. Res. 33 (2000) 382-390. (e) K. Soai, in: Fundamentals of Life (Eds. G. Palyi, C. Zucchi and L. Caglioti) Elsevier and Accademia Nazionale di Scicnzc, Lettere ed Arti (Modena), Paris, 2002, pp. 427-435. [6J (a) G. Palyi, C. Zucchi, R. Boese, M. Szabo, R. Szilagyi and L. Bencze, 12"^ Internal Conf Origin Life (July 11-17, 1999, San Diego, CA, USA), Abstr. p. 42 (C.2.6). (b) G. Palyi, C. Zucchi and C. Hajdu, Atti Memorie, Accad Naz Sci Lett Arti (Modena) 316 [8/2J (2000) 389-415. [7] G. Varadi, M. Strobeck, S. Koch. L. Caghoti, C. Zucchi and G. Palyi, Critical Rev. Biochem. Mol. Biol. 34 (1999) 181-214. [8] (a) H. Brunner and W. Zetthneicr. Handbook of Enantioselective Catalysis, VCH, Weinheim, 1993, vol. 12. (b) J. Ojima, Ed., Catalytic Asymmetric Synthesis, VCH, New York. 1993. (c) C.C. Stinson, Chem. & Eng. News 76 (Sept. 21) (1998) 83-104; 77 (Oct. 11) (1999) 101-120; 78 (May 8) (2000) 59-70; 78 (July 10) (2000) 63-80; 78 (Oct. 23) (2000) 55-78; 79 (Oct. 1) (2001) 79-97; M. Jacoby, ibid. 80 (March 25) (2002) 43-46; A.M. Rouhi. ibid 80 (June 10) (2002) 43-50; 80 (June 10) (2002) 51-57; 81 (May 5) (2003) 45-55; 81 (May 5) (2003) 56-61. [9] (a) H.D. Gilbert Ed., Miniaturization, Reinhold, New York, 1961. (b) J. Rebek, Jr. Ace. Chem. Res. 17 (1984) 258-264. (c) K. Mislow, Chemtracts Org Chem. 2 (1989) 151-174. (d) J.-M. Lehn, Supramolecular Chemistry, VCH. Weinheim, 1995. (e) Special reports in: Nature 2000 (Aug. 20); Chem. & Eng. News 2000 (Oct. 16); Science 2000 (Nov. 24). [10] (a) T.R. Kelly, M.C. Bowyer, K. V. Bhaskar. D. Bebbington. A. Garcia, F. Lang, M.H. Kim and M.P. Jette, J. Am. Chem. Soc. 116 (1994) 3657-3658. (b) T.R. Kelly, 1. Tellitu and J.P. SesXclo, Angew. Chem., Int. Ed Engl. 36 (1997) 1866-1868. (c) A.P. Davis, ibid 37 (1998) 909-910. [11] B.L. Feringa, Ed., Molecular Switches, VCH-Wiley, Weinheim, 2001. [12] (a) V. Galamb and G. Palyi. Coord Chem. Rev. 59 (1984) 203-238. (b) I. Kovacs and F. Ungvary, ibid 161 (1997) 1-32. [13] G. Palyi, C. Zucchi, T. Bartik, T. Herbrich. C. Kriebel, R. Boese, A. Sorkau and G. Frater, Atti Accad. Sci. Bologna, Rend CL Sci. Fis. 281 [14/10] (1992/93) 159-167. [14] G. Palyi, K Alberts. T. Bartik, R. Boese, G. Frater, T. Herbrich. A. Herfurth, C. Kriebel, A. Sorkau, CM. Tschoemer and C. Zucchi, Organometallics 15 (1996) 3253-3255. [15] (a) P. Finocchiaro, D. Gust and K. Mislow, J. Am. Chem. Soc. 96 (1974) 3198-3205. (b) K. Mislow, D. Gust, P. Finocchiaro and R.J. Boettcher, Top Curr. Chem. 47 (1974) 1-28. (c) K. Mislow, Ace. Chem. Res. 9 (1976) 26-33. (d) H. Iwamura and K. Mislow, ibid 21 (1988) 175-182. (e) J. Polowin, S.C. Mackie and M.C. Baird, Organometallics 11 (1992) 3724-3730. (f) S.E. Gamer and A.G. Orpen, J. Chem. Soc, Dalton 7>fl«5. (1993)533-541. [16] (a) M.J. Szabo, L. Bencze, R.K. Szilagyi, G. Palyi and C. ZuQchi, Xlllth FECUEMConf Organomet. Chem. (Aug. 29 - Sept. 3, 1999, Lisboa, Vox^:)Abstr. (018). (b) L. Bencze, R.K. Szilagyi, M.J. Szabo, R. Boese, C. Zucchi and G. Palyi, in: Fundamentals of Life (Eds. G. Palyi, C. Zucchi and L. Caglioti) Elsevier and Accademia Nazionale di Scienze, Lettere ed Arti (Modena), Paris, 2002, 451-471. [17] A systematic comparison of available semiempirical Hamiltonians [18] yielded reasonable equilibrium geometries, with the PM3^'^ method, these were used for obtaining equihbrium heats of formation by the Spartan programme [19]. Heat of formation (kcalmol^). X-ray geometry re/P -1570.05, si/M -1569.62; optimized geometry, all combinations -1812.16 ^ -1812.21. Thus the crystal field effect is cca. +240 kcal-mol'. [18] (a) J.J.P. Stewart,/ Comput. Chem. 10 (1989) 209-220. (b) J.J.P. Stewart, ibid 10 (1989) 221-264. (c) J. Yu, W.J. Here, personal communication to R.K. Szilagyi. [19] SPARTAN 4.0 Wavefunction Inc, 18401 von Karman, Suite 370, Irvine, CA, 92715 (USA) [20] PM3TN* (see refs. 18, 19) [21] (a) J. Rebek. Jr. Ace. Chem. Res. 23 (1990) 399-404. (b) G.R. Desiraju, ibid. 24 (1991) 290-296. (c) Idem, Angew. Chem., Int. Ed Engl. 34 (1995) 2311-2327.

Molecular Clockworks as Potential Models for Biological Chirality

37

[22] (a) F.J. Britten, The Watch and Clock Marker's Handbook, Dictionary and Guide, Antique Collector's Club Ltd, Woodbridge, Suftolk (UK), 1976, 11*^ ed. (b) F.W. Britten, Horological Hints and Helps, Antique Collector's Club Ltd, Woodbridge, Suftolk (UK) 1996, 4^^ ed. [23] Autosolvation: (a) G. Palyi and G. Varadi, J. Organomet. Chem. 86 (1975) 119-125 (b) G. Palyi, Transition Met. Chem. 2 (1977) 273-275. (c) G. Palyi, M. Kovacs-Toplak and G. Yai^<^,AttiAccad Sci. Bologna, Rend CI. Sci. Fis. 281[14/10] (1992/93) 59-167. (d) M.J. Szabo, R.K. Szilagyi, andL. Bencze, Inorg. Chim. Acta 344 (2003) 158-168. [24] C. Zucchi, R. Boese, K. Alberts, T. Herbrich, G. Toth, L. Bencze and G. Palyi, Eur. J. Inorg Chem. (2001) 2297-2304. [25] (a) M.J. Szabo, R.K. Szilagyi, L. Bencze, R. Boese, C. Zucchi, L. Caglioti and G Palyi, Enantiomer 5 (2000) 549-559. (b) C. Zucchi, S. Tiddia, R. Boese, CM. Tschoemer, L. Bencze and G. Palyi, Chirality 13 (2001) 458-464. (c) C. Zucchi, D. Tunini, R. Boese, L. Bencze, R. Kurdi, L. Caglioti and G. Palyi, submitted (2004). [26] V. Galamb, G. Palyi and M. Kajtar, Inorg. Chim. Acta 53 (1981) LI 13-Ll 14. [27] V. Galamb, G. Palyi, F. Ungvaiy. L. Marko, R. Boese and G. Schmid, J. Am. Chem. Soc. 108 (1986) 33443351. [28] C. Zucchi, A. Comia, R. Boese, E. Kleinpeter, H. Alper and G. Palyi, /. Organomet. Chem. 586 (1999) 6169. [29] A. Comia, A.C. Fabretti, C. Zucchi and G. Palyi, to be published. [30] H. Alper, L. Bencze, R. Boese, L. CagUoti, R. Kurdi, G. Palyi, S. Tiddia, D. Tunini and C. Zucchi, J. Mol. Catal. A: Chemical 204-205 (2003) 227-233. [31] R.E. Gawley and J. Aube Principles of Asymmetric Synthesis, Pergamon-Elsevier, Oxford, 1996. [32] G. Gilat, in: Advances in BioChirality (Eds. G. Palyi, C. Zucchi and L. Caglioti) Elsevier, Amsterdam, 1999, pp. 47-68.

Progress in Biological Chirality G. Palyi, C. Zucchi and L. Caglioti (Editors) © 2004 Elsevier Ltd. All rights reserved.

Chapter 4 Diastereomers Do, What They Should Not Do Henri Brunner Institutfur Anorganische Chemie, Universitdt Regenshurg, 93040 Regensburg, Germany henri. brunner @chemie. uni-regensbnrg. de

1.

Enantiomer and Diastereomer Crystallization Enantiomers, related as image and mirror image, have the same energy content and as a consequence the same properties as far as scalars are concerned including solubility. With regard to vector properties they act sign-inverted. Crystallization from a solution containing enantiomers may lead to one of the following cases: conglomerate crystallization, racemate crystallization and crystaUization of a solid solution [1,2]. There is the well known discovery of Pasteur in 1848 [3]. He crystallized an aqueous solution of sodium ammonium tartrate and obtained mirror image single crystals which he could separate with a looking glass and a pair of tweezers. In his famous experiment he was lucky, because the temperature was below 27 °C. This was also true, when he had to repeat his experiment in the laboratory of Biot, before he was allowed to give his invited talk to the French Academy of Sciences. Today many conglomerate crystals are known. 65 space groups including some of the more frequent ones involve the crystallization of one enantiomer in a single crystal for symmetry reasons, which leads to the conglomerate situation. It is assumed that about 5-10 % of all the crystallizations of enantiomers give conglomerates, although the crystals usually are too small for manual separation (Figure 1, left side, conglomerates RfiAR and S,S,S,S). The vast majority of crystallizations from solutions containing enantiomers are racemate crystallizations (90-95 %). In the crystal lattice there are pairs of enantiomers, interrelated by symmetry planes, inversion centers etc. Sodium ammonium tartrate crystallizes from an aqueous solution as a racemate at temperatures above 27 °C. It then forms a monohydrate, whereas the conglomerate is the tetrahydrate (Figure 1, left side, racemates

R^SASASASAS). The case solid solution is a very rare type of crystallization of enantiomers restricted to a couple of examples only. The lattice contains the enantiomers randomly distributed (Figure 1, left side, solid solutions R,R,SARASA) Diastereomers, not related as image and mirror image, differ in energy content and consequently in all properties including solubility. The term diastereomers is mainly used in the resolution of enantiomers with the help of enantiomerically pure resolving agents. In this context diastereomers agree in the configuration of the resolving agent and differ in the

40

Progress in Biological Chiraliu

*M same energy content same solubilities,...

R,R,R,R and S,S,S,S

•

diffemrgM§^^^^^^, diffemnm&ti^^^^M

''•••ll*l;!iliii!!^

Racemates (90-95%) R,o,R,S,R,S,R,S,R,S Solid solutions (rare) R,R,o,R,R,R,S,R

Solid solutions (if similar) RS,RS,RS,SS,RS,RS

Figure 1. The different variants in the crystalhzation of enantiomers and diastereomers

configuration of the enantiomeric parts to be resolved. Although the overall molecular shapes of diastereomers are as different as those of different molecules, there will be parts which are similar (the resolving agent) and parts which are image-mirror image related (the racemate). From a solution containing two diastereomers normally they crystallize in separate lattices as two different compounds do, when they crystallize from the same solution. This, of course, is the basis for the resolution of racemic mixtures via conversion of the enantiomers to diastereomers. Occasionally, diastereomers are separated by chromatography, but predominantly it is fractional crystallization which is successfully used for the separation of diastereomers. Even today in the realm of enantioselective catalysis including the 2001 Nobel prizes to Knowles, Noyori and Sharpless, the vast majority of optical resolutions is done by fractional crystallization of diastereomers (Figure 1, right side "conglomerates" RS, RS, SR (and SS, SS, SS)). After separation, the diastereomers are re-converted to the enantiomers. Assume that the two diastereomers in an optical resolution crystallize in a 1:1 ratio in the same lattice (Figure 1, right side "racemates" RS,SS,RS,SSJIS,SS,RS). Then, repeated fractional crystallization does not lead to a separation. In this case, an enrichment of one enantiomer with respect to the other, e.g. by virtue of an asymmetric synthesis, does not show up in the crystalline phase, which as a rule is isolated and processed, but in the supernatant solution, which contains all the impurities and normally is discarded. This may be the reason why optical resolutions based on fractional crystallizations of diastereomers occasionally have been unsuccessflil.

Diastereomers Do, What They Should Not Do

41

Solid solutions with a statistical distribution of diastereomers tend to form the more the similar they are (Figure 1, right side, solid solutions RS,RS,RS,SS,RS,RS) [4, 5]. Probably, this phenomenon is the reason, why in many optical resolutions diastereomer separation by fractional crystallization is less straightforward than expected. The present article deals with the "racemate" crystallization of diastereomers, framed in Figure 1, in which the two diastereomers assemble in a ratio of exactly 1:1 in the same single crystal. As such a behavior is counter-productive for diastereomer separation it has impact on optical resolutions by fractional crystallization.

2.

The Diastereomers (Rm^c)- and (5RhyS'c)-[CpRh(LL*)CI] Compounds of the type [CpRh (LL*)C1] are synthesized in the reaction of [CpRhCl2]2 with a ligand LL*H in the presence of a base. Here, the ligand LL*H is (N-O)H prepared in a Schiffbase condensation from salicylaldehyde and (^-1-phenylethylamine, the most simple primary amine carrying a H, Me, Ph substituent combination at the chiral center [6]. In the reaction a new chiral center at the Rh atom is formed and two diastereomers {Rmi,Sc) and {S^,Sc) arise, which only differ in the metal configuration (Figure 2, upper part). The ^H-NMR spectrum of [CpRh(N-0)Cl] at room temperature showed only broad signals [6]. On cooling the sample to -60 °C in CD2CI2 solution, the spectrum became sharp and the signals of the two diastereomers could be observed separately. The ratio at 213 K was 77:23, an assignment to the diastereomers (Rm.Sc) and (Sm.Sc) not being possible. The coalescence

N"^-

V

^3^ fi %

,CH

H

^"3

^&>

Figure 2. The equilibrium between (REUi,SC)- and (SRh,SC)-[CpRh(N-0)Cl] (upper part). The diastereomers (RRh,SC)/(SRh,SC) crystallizing in the same single crystals as an inversion pair (lower part)

42

Progress in Biological Chirality

temperature Ji proved to be 283 K. The half-lives for the epimerization reaction (Rmi^Sc)[CpRh(N-0)Cl] — (^Rh,^c)-[CpRh(N-0)Cl] at 11.3 X in CD2CI2 were 31 ms for the forward reaction and 9.2 ms for the back reaction [6]. Again, an assignment to (^Rh,*S'c) and (Smi,Sc), respectively, could not be made. Slow difdision of petroleum ether into a toluene solution of [CpRh(N-0)Cl] at room temperature gave crystals suitable for X-ray structure analysis. Surprisingly, the two diastereomers {RRh,Sc)- and (*S'Rh,*S'c)-[CpRh(N-0)Cl] were present in a 1:1 ratio in the same single crystal. For the assignment of the metal configuration the priority sequence Cp > CI > O > N was used [7, 8]. Thus, from a rapidly interconverting mixture in solution the two diastereomers (R^,Sc) and (6'Rh,*Sc) crystallized in a 1:1 ratio in the same lattice. As expected, the bond lengths and bond angles of the two diastereomers (Rmx,Sc) and (*^Rh,-^c) are extremely similar [6]. In fact, the bond lengths from the Rh atom to its ligands deviate less than 1 % from each other. Figure 2, lower part, shows that the two diastereomers form an ''almost-racemate", connected by an "almost inversion center". In our analysis we place this inversion center exactly halfway between the two Rh atoms. The deviations of the four different substituents around the Rh atoms Cp(centroid), CI, O, N from inversion symmetry are marginal. Naturally, inconsistencies arise for the chiral nitrogen substituents which for both diastereomers have the same (S) configuration, excluding centrosymmetry. With 5.487 A the Rh-Rh distance in the "inversion pair" is extremely short. All the other Rh-Rh distances are much longer. Thus, in these "inversion pairs" two diastereomers assemble in the same way as two enantiomers do in a racemate related by an inversion center. In the middle the pairs of diastereomers {Rmi,Sc) and (^Rh,^c) are image/mirror image related, towards the outside, however, centrosymmetry is lost. The lattice is formed by translation of these inversion pairs.

3.

The Molecular Recognition Motif for the 1:1 Co-crystallization of the Diastereomers (/?RhA)- and (.S^RHA )-[(Cp)Rh(N-0)Cl] and Related Pairs of Diastereomers In the inversion pair (RRh,Sc)/(SKhySc) the two Rh atoms and the two Cp(centroids) define a "central plane" the dihedral angle Cpl-Rhl-Rh2-Cp2 being 174.9 ^ The angles Cpl-Rhl-Rh2 = 75.3 ° and Cp2-Rh2-Rhl = 78.5 ° show that the Rh-Cp centroids are slightly inclined towards Rhl-Rh2 forming a lying Z. This leaves the two Cp planes almost parallel to each other. Concerning the legs of the half-sandwich pianostools it is the Osai and CI substituents which point to the inside of the inversion pairs. This turns the N ligands, the chiral substituents of which disturb centrosymmetry, to the outside, the N atoms being close to the central plane. Molecular recognition occurs between two (Cp)Rh(Osai)Cl fragments with opposite metal configuration, which approach each other by inversion symmetry. With almost parallel Rh-Cp centroids in the central plane the fragments orient their Osai-Cl edges towards one another. There are 2x2 hydrogen bonds between the C-H groups of the Cp rings and the Osai and CI substituents of opposite molecules [distances (angles) C3-H-02 3.90 A (145.5 °), C29H-Ol 3.51 A (151.6 °), C2-H-.C12 3.65 A (153.1 °) and C28-H.-C11 3.51 A (136.9°)] giving rise to the motif of the inverted pianostools. The same molecular recognitionmotif is found in other half-sandwich complexes, for which 1:1 co-crystallization of two diastereomers in the same single crystal has been

Diastereomers Do, What They Should Not Do

43

Figure 3. Inversion pairs {RM,SC)' and (5'M,5'c)-[(Ar)MXY(NR*)] showing M-M distances of 5.6 - 5.8 A, the lying Z arrangement and hydrogen bonds C-H-X and C-H-Y

described in the literature [9, 10]. In these literature reports the phenomenon of 1:1 diastereomer co-crystallization has been stated, but not analyzed and understood. The inversion pairs of the compounds [(Cp)Ru(N-0)Cl] , N-0 = salicylaldimine anion derived from (5)-methyl valinate [11] (upper left of Figure 3), fS>l-phenylethylamine [12] (upper right), (i?)-l-hydroxybut-2-ylamine [13] (lower left), have distances and angles extremely close to iRmi,Sc)/(Smi,Sc) [6] (lower right). In the inverted pianostools of Figure 4 the diastereomers approach each other with their "racemic" sides. However, it should be kept in mind that it is not a "must" for such halfsandwich compounds to show 1:1 diastereomer co-crystallization. There are examples which crystallize as pure diastereomers (see discussion in refs. 9, 10). Thus, there is a delicate balance between the two situations 1:1 diastereomer co-crystallization and crystallization as a pure diastereomer. In solution the diastereomer ratios of such half-sandwich complexes vary depending on compound type and substitution pattern. In both crystallization aUernatives (diastereomer ratio 50:50 or 100:0) asymmetric transformations with respect to the metal configuration are involved [14-16], although they probably are not the reason for realizing one or the other.

4.

Analysis of the Molecular Recognition Motif in Compounds of the Type I(Ar)MXYZ], X and Y Being Electronegative Substituents The molecular recognition motif of the inverted pianostools is strictly obeyed in a dozen of compounds discussed in refs. 9 and 10, including the lying Z for the arrangement

44

Progress in Biological Chiralitv'

Ar(centroid)-M-M-Ar(centroid). Four of these pairs are shown in Figure 3 in a slight top view with respect to the 7i-bonded ligand. The M1-M2 distances are in the narrow range between 5.58 and 5.82 A. Similarly the X1-Y2 and Yl ••X2 distances are between 4.34 and 5.10 A, X being a halogen and Y the oxygen of the salicyl system. The distances Z1-Z2 range between 9.12 and 9.98 A indicating that the nitrogen substituents are far apart from each other pointing away from the center of the pairs. The C-H--X/Y hydrogen bonds within the pairs are all in the same range. The molecular recognition motif of Figures 2 and 3 explains the increased occurrence of 1:1 diastereomer co-crystallization of compounds [(Ar)MXY(NR*)] on the basis of a centrosymmetric arrangement of two mirror image (Ar)MXY fragments. A consequence should be that other compounds having the structural element (Ar)MXY with X and Y electronegative substituents should establish this molecular recognition motif as well, even if they have nothing to do with chirality. As the motif was increasingly found for (Cy)Ru compounds a search in the Cambridge File for [(Cy)RuLCl2] complexes was made, expecting pair formation as indicated in Figures 2 and 3 with the orientation of the ligands L to the outside of the pairs. There were 19 entries [10]. Ten nicely fitted the molecular recognition pattern. Four of them are shown in Figure 4 with ligands L = 4-cyanopyridine [17], nbutyl(diphenyl)phosphine [18], tribenzylphosphine [19] and tris(3-methylphenyl)phosphine [20]. Nine [(Cy)RuLCl2] systems did not fit this pattern. In these nine compounds the ligands L were similar to those which complied with the pattern. In these compounds there were no

Figure 4. The molecular recognition motif of the inverted pianostools for [(Cy)RuLCl2] compounds showing M-M distances of 5.6 - 5.8 A, the lying Z arrangement and hydrogen bonds C-H-X and C-H—Y

Diastereomers Do, What They Should Not Do

45

Figure 5. Racemate crystallization of [(Cy)Ru(N-0)Cl] derived from 8-hydroxyquinoline in the molecular recognition motif of the inverted pianostools

inversion pairs with Ru-Ru distances in the range 5.60-5.80 A. Thus, the situation corresponds to the diastereomers discussed above: The molecular recognition motif of the inverted pianostools is an attractive possibility, but it is not reinforced. There are always examples which could follow it, but fail to do so. The analysis given for the compounds [(Cy)RuLCl2] could easily be extended to related systems. A representative example is the compound [(Cy)Ru(N-0)Cl] [21] in which the unsymmetrical chelate ligand N-0 is the anion of 8-hydroxyquinoline (Figure 5). The chiral molecules form a racemate with a perfect inversion center between the two enantiomers. The Ru-Ru distance is 5.34 A and the hydrogen bonds C-H-0 3.279 A (angle 125.8 °) and C-H-Cl 3.674 A (angle 160.0 °) are fully conform with the molecular recognition pattern of the inverted pianostools. 5.

Relevance of the Inverted Pianostool Motif for Achiral Compounds, Enantiomers and Diastereomers As shown in Figure 4 for [(Cy)RuCl2L] derivatives, half-sandwich compounds of the type [(Ar)MLX2] tend to form the molecular recognition motif of the inverted pianostools in the solid state. Such a detail is overlooked as long as only a single molecule is considered as customary in today's publications. It is perceived only after the packing diagram has been taken into account. The left part of Figure 6 shows a representative example of an achiral [(Cy)RuLCl2] derivative with L = tribenzylphosphine. If enantiomers of the type [(Ar)MXYL] are crystallized, frequently the inverted pianostool motif is established resulting in a racemate, the usual situation in enantiomer crystallization as outlined in the introduction. As an example compound [(Cy)Ru(N-0)Cl] of Figure 5 is repeated in the middle of Figure 6.

46

Progress in Biological Chiralitv

^^^^^^S^i^^^^l^^,

interesting detail of packing diagram

Mmmtmmem

iliitf.

racennate crystallization

Figure 6. The molecular recognition motif of the inverted pianostools for achiral compounds (left), enantiomers (middle) and diastereomers (right) The most interesting case is the crystallization of diastereomers of the type [(Ar)MXY(NR*)]. Establishment of the inverted pianostool motif results in the interesting consequence that two diastereomers co-crystallize 1:1 in the same single crystal with the implications for the failure of isomer separation addressed in the introduction. Remarkably, the half-sandwich complexes showing 1:1 diastereomer co-crystallization not only comprise Rh compounds but also Ru, Os and Ir compounds, not only cyclopentadienyl

Z ^-••*' Figure 7. Generahzed representation of the inverted pianostool molecular recognition motif

Diastereomers Do, What They Should Not Do

47

derivatives but also benzene and cymene derivatives, not only chloro ligands but also iodo ligands and not only salicylaldiminates but also pyrrolylaldiminates. In addition, most of the compounds differ appreciably in the chiral nitrogen substituent [9, 10]. Thus, a generalized representation of the molecular recognition motif of the inverted pianostools with M - M distances of 5.6-5.8 A, the lying Z arrangement and hydrogen bonds between the C-H groups of the pianostool seat and electronegative substituents X and Y takes the form shown in Figure 7.

6.

References

[1] C.J. Welch, Chimlity 13 (2001) 425-427. [2] E.L. Eliel and S.H. Wilen, Stereochemistry of Organic Compounds, Wiley & Sons, New York, 1994. [3] G.B. Kaufiman, I. Bemal and H.-W. Schiitt, Enantiomer 4 (1999) 33-45. [4] B.S. Green, M. Lahav and D. Rabinovich, ylcc. Chem. Res. 12 (1979) 191-197. [5] L. Addadi, E. Gati, M. Lahav and L. Leiserowitz, Isr. J. Chem. 15 (1916-11) 116-123. [6] H. Brunner, A. Kolhiberger and M. Zabel, Polyhedron 22 (2003) 2639-2646. [7] C. Lecomte, Y. Dusausoy, J. Protas, J. Tirouflet and A. Dormond, J. Organomet. Chem. 73 (1974) 67-76. [8] H. Brunner, Enantiomer 2 {1991) 133-134. [9]H. Brunner, M. Weber and M. Zabel Angew. Chem. 115 (2003) 1903-1907: Angew. Chem. Int. Ed 42 (2003) 1859-1862. [10] H. Brunner, M. Weber and M. Zabel, Coord Chem. Rev. 242 (2003) 3-13. [11] H. Brunner, T. Zwack, M. Zabel, W. Beck and A. Bohm, Organometallics 11 (2003) 1741-1750. [12] S.K. Mandal and A.R. Chakravarty, J. Chem. Soc, Dalton Trans. (1992) 1627-1633. [13] H. Brunner, T. Neuhierl and B. Nuber, Eur. J. Inorg. Chem. (1998) 1877-1881. [14] H. BrumiQr,Adv. Organomet. Chem. 18 (1980) 151-206. [15] H. Brmmer,Angew. Chem. I l l (1999) 1149-1163;Angew. Chem. Int. Ed 38 (1999) 1194-1208. [16] H. Brunner, Eur. J. Inorg. Chem. (2001) 905-912. [17] D.K. Gupta, A.N. Sahay, D.S. Pandey, N.K. Jha, P. Sharma, G. Espinosa, A. Cabrera, M.C. Puerta and P. Valerga, J. Organomet. Chem. 568 (1998) 13-20. [18] G. Bruno, M. Panzalorto, F. Nicolo, C. G. Arena and P. Cardiano, Acta Crystallogr., Sect. C (Cr. Str. Comm.) 56 (2000) e429. [19] S. Serron, S.P. Nolan, Yu. A. Abramov, L. Brammer and J.L. Petersen, OrganometalUcs 17 (1998) 104110. [20] A. Hafner, A. Miihlebach and P.A. von der Schaaf, Angew. Chem. 109 (1997) 2213-2216; Angew. Chem. Int. Ed Engl 36 (1997) 1111-1114. [21] C. Gemel, R. John, C. Slugovc, K. Mereiter, R. Schmid and K. Kirchner, J. Chem. Soc, Dalton Trans. (2000) 2607-2612.

Progress in Biological Chirality G. Palyi, C. Zucchi and L. Caglioti (Editors) © 2004 Elsevier Ltd. Allrightsreserved.

Chapter 5 Molecular Origins of Life Homochirality as a Consequence of the Dynamic Co-Emergence and Co-Evolution of Peptides and Chemical Energetics Auguste Commeyras,'*'* Jacques Taillades,^ Helena Collet,^ Laurent Boiteau/ Odile Vandenabeele-Trambouze,^ Robert Pascal,^ Herve Cottet,^ Raphael Plasson,^ Jean-Philippe Biron,^ Eddy Souaid,^ Laurence Garrel/ Olivier Lagrille,^ Gregoire Danger,^ Jean-Christophe Rossi,^ Franck Selsis,^ Michel Dobrijevic" and Herve Martin"^ ^Organisation Moleculaire, Evolution & Materiaux Fluores (UMR5073-CNRS), Departement de Chimie- CC017, Universite de Montpellier 2, F-34095 Montpellier Cedex 5, France [email protected] ^Centro de astrobiologia - INTA (CSIC), Ctra de Ajalvir km 4,E- 28850 Torrejon de Ardoz, Madrid, Spain ^Laboratoire d'Astrodynamique, d'Astrophysique et d'Aeronomie de Bordeaux (L3AB, UMR5804~CNRS), 2 rue de VObservatoire, BP89, F'33270 Floirac, France ^Laboratoire Magmas et Volcans, Universite Blaise Pascal-CNRS-OPGC, 5 rue Kessler, F-63038 Clermont-Ferrand cedex 1, France

1.

Introduction There are some obvious arguments that are hard to circumvent, such as; no dynamics without energy, no evolution without dynamics, no life without evolution, which result in postulating the existence of dynamic co-evolutionary processes at the origin of Ufe. The idea of dynamic co-evolution at the time of the origin of life requires the addition of the energy component to the old debate about the priority of emergence between peptides and RNAs This debate is far from being resolved and continues to evolve, since it is assumed that the emergence of the RNA world is hard to imagine [1] without the aid of peptides and their catalytic effects [2]. Putting into perspective the conditions of the emergence of the first peptides and their subsequent catalytic effects may have constituted one of the first steps in this extraordinary mechanism of the pathway from inert to "alive". To participate in this debate, we have carefully analysed the formation of a-amino acids discovered in meteorites, and shown that they are derived from the following system {H2O, HCN, R^COR^ NH3, R^NH2, CO2, NaHCOs, B(0H)4", NO, O2}, involving thirty five different carbonyl compounds ammonia and two amines [3]. By studying the evolution of these systems, we deduced an experimental scenario which can explain the emergence of

50

Progress in Biological Chirality

exogenous and endogenous a-amino acids as well as the emergence of peptides on the primitive Earth. The core feature of this scenario is the "molecular dynamics" which can be shown by the following course of events.

Experimental The experimental conditions are described in the quoted references.

3.

Results and Discussion

3. J a-Amino acids and related compounds from a multi-component reaction model We start, in Figure 1, with some of the essential molecules of prebiotic chemistry (water, hydrogen cyanide, carbonyl compounds, ammonia or amines, carbon dioxide and borates). 3.1.1 Reversible reactions The first reaction between these molecules is the formation of the a-hydroxynitriles (I). This reaction is very fast. It is followed by the formation of amino alcohols (II). These amino alcohols lose the OH" anion to give iminium ions (III). These iminium ions react with the cyanide anion to give a-amino nitriles (IV). These a-amino nitriles then react with the initial carbonyl derivatives to give new a-amino alcohols (V). As before, these new amino alcohols (V) in turn lose OH" and, together with the cyanide anion, give a-amino dinitriles (VII). But, in competition with this reaction, the amino alcohols (V) also lose H^, leading to amino alcoholates (VIII) [4-10]. Carbon dioxide, when present in the medium, specifically reacts with the a-amino nitriles (IV) to give carbamic acids (IX) and carbamates (X) [11, 12].

B.-%„">

HCN + RiR2C=0 + H2O+ NH3 + CO2 + B(OH)4

^

'NH2

(II)

I-

B(OH)4

.(IV)

R2

(III)

i (IV)

CO, I

CN ORis

R / ' ^ C ^ r (XI) OH

F^

H

R2

(V)

riT^ H

(IX)

HO

R2

(VIII) BIN/"

R^^ CN HO

(VI) OH

^R2

H

<« ^^'^^ H O

I Figure 1. The initial ^stem, made up of 6 compounds [H2O, HCN, R^R^CO, NH3, CO2, B(0H)4 ] leads, at equilibrium, to all of the neutral or ionized molecules represented in this figure

Molecular Origins of Life: Homochirality as a Consequence of the Dynamic... 51 When boric acid is present in the medium it specifically reacts with the a-hydroxynitriles (I) to give boric esters (XI) and the corresponding borates (XII) [13, 14]. All these reactions are reversible. 3.1.2 Non reversible reactions In Figure 2, starting from these nitriles, we examine the different possibilities for escaping from this reversibility. The first two possibilities are reactions (A) and (B) seen in Figure 2. These reactions are the hydration of the nitrile frmctions by OH", leading competitively to a-hydroxy amides (XIIc) and a-amino amides (VIIIc). They are known, in prebiotic literature, as the Strecker reaction [15]. In addition to reactions (A) and (B), we have discovered three new reactions (C), (D) and (E) which enable this system to escape from equilibrium. The starting point for the first of these reactions (C) is the alcoholates (VIII). This leads to a-amino amides (VIIIc) via a reaction path that is catalyzed by the carbonyl derivatives [1618]. This reaction is chemically selective, and is part of the Strecker reaction. Its rate

RtsV^ONHj R^

CONH2

Rr^^NH2 "!•> (II). (iii>. (IV) (IX) R K /CO2H

(X)

R/^ ^OH Ris.^02H

L(C)

i4(D)

Ris ^ O N H a

(vmd) IH

B(OH)4

'p

(vnia)

(Xlla)

Ri

H2NCO HO2C

CONH2

H

"2 (Vinb2)

pN

"^^-j^C^^i "* (Vlllbl)

J

(Xb)

K r + R2

I-

R]R2CO (Xc)

R l \ ^C0NH2

I Ris/CO^

/Y^«2

RKX>ONH2

.CONH2

R2^^NH2

HO2C

(Xa)

^

' Ri R2

R2"^^NH2

(Vnic)

(F) •

r HNCO

R2>W"2 (Xd)

Figure 2. Five different and competitive irreversibie reactions (A), (B), (C), (D), (E) enable the reversible system to escape from reversibility-. To take into accomit the information from § 3.2, we have added the carbamoylation, reaction (F), to this figure

52

Progress in Biological Chirality

constant is very high compared with the previously mentioned reaction paths (A) and (B). When a small concentration of ammonia is used, the reaction (C) gives imino-diacids as proven by Laurence Garrel (personal communication). The discovery of these imino-diacids n Miller's type experiments [19], as in meteorites [20], confirms the relation between exogenous and endogenous a-amino acids, and gives the prebiotic status to the reaction (A) (B) and (C) [3]. The second reaction (D) starts from the carbamates (X). It leads to hydantoins (Xc) via the intermediary isocyanates (Xb). This reaction is also chemoselective [11, 21]. It is known as the Bucherer-Bergs reaction. Its rate constant is almost as high as that of the reaction (C). The third reaction (E) starts from the borates (XII), which are used as catalysts. This reaction is also chemoselective [13, 14]. It leads to the formation of a-hydroxy amides (XIIc). Its rate constant is lower than that of (C) and (D), but much higher than that of (A) and (B). All of the reactions (A) to (E) displace the species in equilibrium, preferentially giving aamino amides and hydantoins, and less effectively a-hydroxy amides. Thus the process is kinetically controlled, but only by the reactions (C), (D) and (E). Reactions (A) and (B) play a negligible role. The next step gives a-hydroxy acids (Xlld), a-amino acids (Vllld), and Ncarbamoylamino acids (Xd), all of which have the same rate. This means that the formation of these compounds is always kinetically controlled by the reactions (C), (D) and (E). The global process needs to be completed^ by the action of cyanate on a-amino acids (reaction F). This reaction is slow, so it does not interfere with the previous one, but it transforms quantitatively the free a-amino acids into N-carbamoyl amino acids [3, 22]. The consequence, for the primitive Earth, is that this whole set of reactions could have kinetically converged, not towards free a-amino acids, but rather towards N-carbamoyl amino acids. 3.2 Peptides from N-carbamoylamino acids through a primary pump At this point in the story we must ask why, N-carbamoyl amino acids were formed at all? The following step in this scenario is to go from N-carbamoyl amino acids to peptides. Our approach (Figure 3) is a molecular engine (also called a primary pump) fed with N-carbamoyl amino acids and frielled by a gaseous mixture of NO/O2 as the energy source [23, 24]. 3.2.1 Molecular engine: description The first step (1) is the formation of N-carbamoylamino acids in aqueous solution [22]. The second step (2) is the concentration of these N-carbamoyl amino acids. This could have taken place at low tide on the shores of the Hadean continent. In the dry phase, step (3), if NOx were present in the primitive atmosphere [24], it could nitrosate the N-carbamoylamino acids, and produce stoichiometric amounts of nitrous acid. These nitrosated intermediates are unstable and, through cyclization, lead to the Ncarboxyanhydrides of a-amino acids (NCAs). This reaction is quantitative at ambient temperature [25]. At the same time as producing NCAs, they form stoichiometric amounts of nitrogen and water. The nitrous acid stabilizes the NCAs for several hours. If the NCAs stay too long in an acid medium they are hydrolyzed into a-amino acids which are then recycled.

Molecular Origins of Life: Homochirality as a Consequence of the Dynamic... 53 H2N—C—C-OH I CH3 Non-racemic meteoritic a-methyl-aminoacids

ENERGY NO + 0 ,

^=-

NOx

N2 + H2O +CO2

l%<eeL<9.2% a-aminoacids of the reservoir (1+8 )Land(l-e ) D

Primitive ocean 5 < pH < 7

High rate of racemisation of a-aminoacids

R3

Slow rate of racemisation of peptides

O

R,

2

V

'^

\

/HNCO

® Pep-CO-N

NCO-

Peptides. Elongation Evolution Accumulation Homochirality

Figure 3. Molecular engine leading to the production of evolutionary sequential peptides, startingfromN-carbamoylamino acids andfromCNO', NO, O2, NaHCOs, H2O When, as represented in step (4), the rising tide changes the pH of these NCAs back to between 5 and 7 (pH of the primitive ocean [26]), two competitive reactions resuh. Some of the NCAs are hydrolysed through step (5a), and another part, form peptides in few minutes, through step (5b). In step (6), the peptides thus produced react with the cyanate, to give N-carbamoyl peptides. This reaction easily occurs at pH between 5 and 7. This reaction prevents subsequent elongation of the peptides, but this protection is removed in the next cycle [27]. In step (7), the N-carbamoyl peptides slowly hydrolyze. This hydrolysis leads to shorter peptides, down to a-amino acids themselves. These peptides and amino acids are in turn carbamoylated. The last step (8) is the racemization of the a-amino acids and peptides. This reaction is selective. Amino acids racemize faster than peptides. When it works, this primary pump is capable of doing beautiful things. 3.2.2 Elongation When the N-carbamoyl peptides of the cycle (n) arrive on the shore, they dry out and are immediately unprotected by the NOx. This reaction occurs under the same conditions as the formation of NCAs. At the next high tide, the unprotected peptides grow longer.

54

Progress in Biological Chirality

3.2.3 Evolution The evolution of peptides could have been brought about by their partial hydrolysis and lengthening. 3.2.4 Accumulation The formation rate of peptides is fast (their half lifetime takes only a few hours), whereas their hydrolysis is slow (with a half lifetime of about 400 years [28]). As a result, peptides could have accumulated. 3.2.5 Homochirality The rise of homochirality can be described as the consequence of the cooperative effects of reactions (5a), (5b), (7), (8) and by the dynamic of the system. Reaction (5a) has no selectivity, whereas (5b) is enantioselective. Note that the primitive Earth was seeded by non-racemic a-methyl-amino acids with enantiomeric-excess (ee) L between 1% and 9.2% [29, 30]. These a-methyl-amino acids can not racemize because of the stability of C-C bond. As a consequence, the a-amino acids of the reservoir were non-racemic with (1+s) L and (1-8) D. When these non-racemic amino acids react with a racemic NCA, peptides enriched by the L- enantiomer are formed through reaction (5b). Meanwhile the competitive hydrolysis by (5a) of the remaining NCA could give a-amino acids enriched by the D- enantiomer (this remains to be proved in next future). This reaction creates a dissymmetry between the two parts of the reservoir (peptides with an ee L and amino acids with ee D configuration), but globally the reservoir remains symmetric, to within values around the initial 8 due to a-methyl-amino acids that can not racemize. If now the amino acids, with an ee. D configuration, racemize faster than the peptides (reaction 8), the reservoir's dissymmetry increases beyond 8, and this increase is amplified during the next cycle. Now, with regard to the hydrolysis in step 7, this hydrolysis reaction can easily remove the hetero peptides (LD or DL), which hydrolyse faster than the homo peptides, while the difference in concentration between the remaining LL and DD groups due to the reaction (5b), is kinetically increased during the next cycles. This again remains to be proven even if such possibilities have been evoked [31]. (a) Importance of the dynamic of the system Figure 4 represents the different stationary states in these models as a function of the rate of peptide formation. At low rates when v < VB, the peptides formed are hydrolysed, leading to racemic stationary states. At higher rates when v > VB, the peptides accumulate and evolve towards homochirality and VB is the bifurcation point. If they function near the bifurcation point, such dynamic systems are unstable and they can go in unpredictable directions (L) or (D) [33]. But if an "external chiral perturbation" occurs, it favours the bifurcation in a predetermined direction.

Molecular Origins of Life: Homochirality as a Consequence of the Dynamic... 55

+1

Racemic state

/ k

Optically active states

Bifurcation/ point -1 Figure 4. Representation of stationary states as described by Iwamoto [32] and Kondepudi [33] dynamic models. Enantiomeric excess versus the rate of the molecular engine. The biftircation point corresponds to a precise rate of the reaction

(b) External perturbation In our model, the non-racemic a-methyl-amino acids could have been the external chiral perturbation. These a-methyl-amino acids could have either: initiated the reaction (5b) enantioselectively, or perturbated the racemization equilibrium (L) <=> (D) of the proteinic a-amino acids in the reservoir.

10'ee

start (f = 0) evaluation of dt

f

i (L)-AA hydrolysis

i

(L)-AA condensation I

(D)-AA hydrolysis

-^—

1

(D)-AA condensation I

increase t by dt

i

NCA hydrolysis

tbydt update concentrations

update concentrations

I y^

yes t > fmatx. ^ s _

Figure 5. Algorithm showing the functioning scheme of the primary pump. The evolution of the enantiomeric excess during the beginning of its functioning shows the break of symmetry and its amplification

56

Progress in Biological Chirality

Either scenario could have lead the amplification of dissymmetry towards a unique enantiomeric form (L). Measurements are required to test the validity of these hypothesis. At this point, it must be made clear that the argument concerning homochirality, is not yet based on knowing all the kinetic constants. Only a complete kinetic modelling based on real kinetic constants can confirm such a scenario. We believe it is a price that must be paid in order to study such dynamic systems. Some journal referees need to be convinced that this is in fact the case, (c) Kinetic modeling While waiting for this ideal kinetic model, we have constructed one (see the algorithm) using estimated (or real cf § 3.4) kinetic constants, (ku. > kLo, kracem a-amino acids > kracem peptides, khydroiyse LD et DL > khydroiyse LL and DD). During its fianctioning, this model leads to a break of symmetry (L) or (D) and to its amplification (see the graphic in Figure 5). i. 3 Environmental requirements Before proceeding we have got information showing that the primitive Earth may have satisfied all the necessary requirements to make this primary pump realistic [24]. These results gave us greater motivation to proceed in detailing the primary pump, looking for mechanisms and kinetic constants for every step. 3.4 Primary pump step by step Only some details of step (1), (3) and (5) have been obtained. 3.4. J Step (J): carbamoylation Although the rate law is complex [22], V = d[NCO]' / dt = [NCO]' x (ko + k4 x[C03]' + ks x [NH3]' + k6 x[AA]') k6, the kinetic constant for the carbamoylation of amino acids at 50°C and pH 6.5 was obtained. 1300 lO-^mol'^Ls-^forVal, 1700 lO-^/mof^Ls-^ for Gly, 3530 10-^/mor'Ls"^forThr. These kinetic constants are very low, but they look well adapted to the dynamic of the system. Only kinetic modelling will prove or disprove this. 3.4.2 Step (3): nitrosation The gaseous mixture NO/O2 react with N-carbamoyl amino acids even in solid phase [25, 34]. The picture (a) in Figure 6 shows crystals of N-carbamoyl valine. In presence of a gaseous mixture of NO/O2. Picture (b) shows the surface of the crystal at the beginning of the reaction. The bubbles are N2. The liquid is water and nitrous acid. Picture (c) is the Valine NCA. This reaction is complete at 20 °C in half an hour. The kinetic constants for the nitrosation of these N-carbamoyl amino acids were obtained : - by HNO2 in water (kHN02 = 0.1 s'^), the constant is low, - by N2O4 in dry phase (kN204 = 1 7 s'^) the constant is 17 times bigger. - but by N2O3 in solid phase (kN203 = l.OxlO^s'^) the kinetic constant is 1 billion times bigger than with nitrous acid [35, 36].

Molecular Origins of Life: Homochirality as a Consequence of the Dynamic... 57

Figure 6. (a/ crystals of N-carbamoylvaline), (b/ action of NO/Oson a), (c/ valine NCA)

Note that N2O4 and N2O3 are present in equilibrium in the gaseous mixture NO/O2. Figure 7 summarizes the reversible reactions between NO, O2, N2O4 and N2O3 in the gaseous phase. 3.4.3 Step (5): peptides synthesis Our work gives a prebiotic status to these NC As. Except for those of Bartlett [37, 38] that have to be reconsidered, no kinetic measurements of the polycondensation of NCA can be found in the literature, in aqueous solvent.

I 1/2NO I

jl 1/2

Colourless

1/202

O-O-NO

t

ArH^9S = -57.2 KJ.mof ^0*^298 = -65.93 KJ.mol'

1/2NO

l/2 0]>^o/o-NO

I (A)

(B)

N-0-]>Q

-—^

NQz

• ;

NC^*^

Brown ^ NNO '

^^n-'^

I ^=?==^ 0=N-]>Q

(C)

: i

t; « - ^ ^ ^ ° (D)

o NiOi

II Colourless

ArH^298--57.12KJ.mo-f ArG°298 = -4.73 KJ.mo-f

Bl

Colourless

ArH°298 = -40.56 KJ.mol /VG^98=1.86KJ.mof

Figure 7. Equilibrium between NO and O2 in the gaseous phase

58

Progress in Biological Chirality

(a) Oligocondensation ofNCA and enantioselectivity In aqueous solution NCAs give two competitive reactions with different kinetic laws: - Vhyd = khyd [HO][NCA] for hydrolysis kcondens [-NH2][NCA] for oligocondensation. For example [39] for valine NCA : - the kinetic constant for hydrolysis is khydrolysis = 4.40xl0'^s"\ - and for the formation of the dipeptides H-L-Val-L-Val-OH and H-D-Val-L-Val-OH kLL =1.65x10-^ s-^Lmor\ kDL= 1.14x10"^ s'^Lmol"^ The first conclusion is that in water the kinetic constants for the divaline formation are one hundred times bigger than those of hydrolysis. This is a very promising result for peptide synthesis in aqueous solution. The second conclusion, when considering the ratio KLL/KDL = 152, is that a large enantioselectivity is observed already for the dipeptide formation. This ratio should be higher for the synthesis of larger peptides due to a probable micro phase separation around octapeptides, (see the discussion on this point by Kricheldorf p 180185 [40]) and possible assisted polycondensations of NCAs by these microphase like membranes [41]. In the near future these measurements will be, for example, extended to the synthesis of HL-VAL-L-isoVal-OH and H-D-VAL-L-isoVAL-OH from L-isoVAL on NCA-VAL, L and D and related compounds, to see one of the possible roles of the non-racemic exogenous amino acids in the emergence of the homochirality (see § 3.2.5b). Now it should be noted that in water, the formation rate of peptides is a function of their pKa (only the non protonated amino group is active) and, paradoxically the PKa's of peptides are not very well known. (b) FastpKa measurement of peptides We have developed a fast method for measuring the pKa using capillary electrophoresis. In the figure 8 we see that the electrophoretic mobility of peptides depends on the pH.

Figure 8. Capillary electrophoresis of polyglycines. Electrophoretic mobilities are pH dependent

Molecular Origins of Life: Homochirality as a Consequence of the Dynamic... 59

25''C y = m1 +m2*exp(-m3*m0)-m4*m0 Error Value ml

7.9698

m2

14.82 2.2411 0.032703 0.0045027

m3 m4 1 Chisq

a •a

R^

0.041 3.5145 0.25004 0.0062972

0.99834

NA NA

JS 8-5

Figure 9. pKa of polyglycines from 1 to 9 residues

From these electrophoretic mobilities, we obtained the pKa's for various peptides such as the polyglycines (for details see PhD of Raphael Plasson 2003 University of Montpellier 2 to be published). The pKa of peptides decreases with their lengthening, (c) Lengthening From this information, we deduce that in aqueous solution, at equal concentrations, longer peptides (lower pKa) tend to lengthen faster than shorter peptides, by assuming that the kinetic lengthening constant is the same for all peptides. This also will be carefully controlled in the near future.

pH primitive ocean

Figure 10. In aqueous solution, at equal concentrations of peptides, the concentration of the non-protonated amino group of the longer peptides (lower pKa) is higher than the concentration of the non-protonated amino group of shorter peptides

60

Progress in Biological Chirality

L

ionic

W

*

VE VW I VVE VEV VEVE

current 8.00

10.00

12.00

14.00

16.00

18.00

20.00

22.00

24.00

Elution time/min 26.00

28.00

30.00

32.00

34.00

36.00

Figure 11. Molecular engine running with 2 amino acids Val (V) and Glu (E) (HPLC/MS,Q-TOF analysis)

(d) Chemoselectivity This physico-chemical approach needs to be completed by synthesis techniques because, in water, other parameters are involved in the lengthening. Figure 11 shows the results obtained by running the primary pump with two amino acids: V (valine) and E (glutamic acid). Of the many possible peptides, many but not all are formed. The distribution of peptides formed in these experiments shows an effective chemoselectivity (Figure 12) during the lengthening process. Starting with 2 amino acids

4 /4 Dipeptides formed

4/8 Tripeptides formed

4/16 Tetrapeptides formed

5/32 pentapeptides formed

VVVV EVVV

..»•

VEVV

VEVW W E W

EEVV VVVE EWE VEVE

- ^

VEWE WEVE

EEVE VVEV EVEV VEEV EEEV WEE EVEE VEEE EEEE

Figure 12, Chemoselectivity during the formation of peptides from 2 amino acids as starting materials

Molecular Origins of Life: Homochirality as a Consequence of the Dynamic... 61

(e) Selectivity due to hydrophobia With four amino acids (V, E, G, A) as the starting materials, MALDI-TOF analysis shows that the length of the peptides increases with the cycle number, we go from 800 Da for cycle 1, to 1000 Da for cycle 2, to 1300 Da for cycle 7... etc. But insoluble peptides are also formed, and are therefore missing in the soluble fraction. The insoluble fraction (separated after 7 cycles) is made up of peptides whose molecular weight is distributed around 800 Daltons, whereas the molecular weight of the soluble compounds is higher than 2000. So there is a selectivity process due to hydrophobia that is complementary to the others. The stimulating aspect of this point is that the existence of this stress (due to hydrophobia) eliminated the insoluble peptides. The only peptides capable of surviving (that is to say of evolving) were the soluble ones, in other words those, with good sequences, that were best adapted to their environment. Note that the inappropriately adapted species, the insoluble ones, are completely recycled after hydrolysis. We can therefore hypothesize that in aqueous solution, the selectivities described above could have considerably decreased the number of peptides formed and thus making realistic that the prebiotic synthesis of peptides occurred without the genetic code.

Cycle #2

'*'*i^,^ 1380

1920

2460

3000

Cycle #7 (soluble fraction)

Cycle #7 (insoluble fraction)

Figure 13. The molecular engine was run with 4 amino acids (V, E, G, A). MALDI-TOF analysis for the peptides formed is given for the cycle 1, 2, and 7 (soluble fraction and insoluble fraction)

62

Progress in Biological Chirality

3.5 Chemical energy Energy itself must have evolved. In the course of this research, we moved from radiation energy to NOx anhydrides. These NOx compounds then evolved towards a second group of anhydrides, by which we mean the NCAs. Why not try to find an affiliation between NCAs and ATP or equivalent? Energetically speaking this affiliation is correct, and nobody knows the origin of these fundamental molecules which must have been synthesized in a continuous flow. 3.6 Terrestrial energy In order to work, the primary pump requires an alternation of dry and aqueous periods for instance on tidal beaches. In addition to chemical energy, a terrestrial source of energy was also required in order to allow "primary pump efficiency"; this energy was not directly involved in the chemical reactions, but initially it created favourable conditions for their development. 3.7 Need of extension of kinetic modelling The global system already presented constitutes a primary pump, fed with a-amino acids, fuelled by a gaseous mixture NO/O2 which recycles the essentials of its material. With about fifty carbonyl compounds this molecular engine is capable of synthesizing all the a-amino acids of the primitive Earth. From these the homochiral peptides could have been built, using only twenty of them. To understand and appreciate the evolution of such systems, kinetic modelling is needed.

4.

Conclusion To summarize, we have presented a scenario for the dynamic co-evolution of peptides and energy on the primitive Earth. In this scenario homochirality and sequential peptides appear as a consequence of this dynamic co-evolution. This scenario is not necessarily the right one, but it is chemically original and has the virtue that it can be experimentally tested. During the course of these experiments, we hope to open up a gateway to the emergence of the catalytic activities of peptides. We have already begun to approach this aspect of the work [42].

5. Acknowledgements This work was supported by the European Community program COST D27 Prebiotic Chemistry and Early Evolution; I'Universite de Montpellier 2, le Ministere de VEducation Nationale et le Centre National de la Recherche Scientifique (Departement de Chimie et Institut National des Sciences de I 'Univers).

6.

References

[1] R.F. Gesteland, T.R. Cech and J.F. Athins, Eds., The RNA word second edition. Cold Spring Harbor Laboratory Press, 1999, p. 709. [2] C. De Duve, Clues from present-day biology: the thioester world, in: The molecular origin oflifQ, (Ed. A. Brack) Cambridge University Press, 1998, p. 219-236.

Molecular Origins of Life: Homochirality as a Consequence of the Dynamic... 63 [3] A. Commeyras. J. Taillades. H. Collet, L. Boiteau. R. Pascal O. Vandenabeele. A. Rousset. L. Garrel. J.C Rossi, H. Cottet, J.P. Biron, O. Lagrille, R. Plasson, E. Souaid, F. Selsis, M. Dobrijevic and H. Martin, Approche dynamique de la synthese des peptides et de leurs precurseurs sur la Terre primitive, in: Les traces du vivant. (Eds. M. Gargaud, D. Despois and J.P. Parisot) Presses Universitaires de Bordeaux, 2003, p. 115-163. [4] J. Taillades and A. Commeyras, Systemes de Strecker et Apparentes: I-Etude de la decomposition en solution aqueuse des a-alcoyl-aminonitriles tertiaires. Mecanisme d'elimination du groupement nitrite. Tetrahedron 30 (1974) 127-132. [51 J. Taillades and A. Commeyras, Systemes de Strecker et Apparentes: II-Mecanisme de formation en solution aqueuse des a-alcoylaminoisobutyronitrile a partir d'acetone, d'acide cyanhydrique et d'ammoniac, methyl ou dimethylamine. Tetrahedron 30 (1974) 2493-2501. [6] J. Taillades and A. Commeyras, Systemes de Strecker et Apparentes: Ill-Etude en solution aqueuse de la stabilite et des conditions de synthese des a-aminonitriles tertiaires. Importance des protons portes par le groupement amine. Tetrahedron 30 (1974) 3407-3414. [7] M. Bejaud, L. Mion and A. Commeyras, Systemes de Strecker et Apparentes. VII- Etude stereochimique des a-aminodinitriles, produits secondaires de la synthese selon Strecker des a-aminonitriles. Tetrahedron Lett. 34 (1975) 2985-2986. [8] M. B6jaud, L. Mion, J. Taillades and A. Commeyras, Systemes de Strecker et Apparentes. IV-Etude comparative de la reactivite des a-aminonitriles secondaires et tertiaires en solution aqueuse entre pH 10 et 14. Hydrolyse des a-aminonitriles secondaires et son importance dans la formation prebiotique des acides amines naturels. Tetrahedron 31 (1975) 403-410. [9] M. Bejaud, L. Mion and A. Conuneyras, Systemes de Strecker et Apparentes. VI- Stabilite des aminonitriles en fonction du pH. Bull Soc. Chim. France 1-2 (1976) 233-236. [10] M. Bejaud, L. Mion and A. Commeyras, Systemes de Strecker et Apparentes. VIII- Etude thermodynamique des systemes acetaldehyde-acide cyanhydrique-monomethylamine ou ammoniac en solution aqueuse. Stabihte en fonction du pH des cyanhydrines, a-aminonitriles et a-aminodinitriles formes. Bull. Soc. Chim. France 9-10 (1976) 1425-1430. [11] A. Rousset, M. Lasperas, J. Taillades and A. Commeyras, Systemes de Strecker et Apparentes. XIFormation et stabihte de I'a-carboxyaminonitrile intermediaire essentiel dans la synthese des hydantoines selon Bucherer-Bergs. Tetrahedron 36 (1980) 2649-2661. [12] A. Rousset, M. Lasperas, J. Taillades and A. Commeyras, Systemes de Strecker et Apparentes. XVComportement d'a-alcoylaminonitriles en presence de CS2 et de CO2. Bull Soc. Chim. France 5-6, II (1984)209-216. [13] J. Jammot, R. Pascal and A. Commeyras, Hydration of cyanohydrins in weakly alkaline solutions of boric acid salts. Tetrahedron Lett. 30 (1989) 563-564. [14] J. Jammot, R. Pascal and A. Commeyras, The influence of borate buffers on the hydration rate of cyanohydrins: evidence for an intramolecular mechanism. J. Chem. Soc, Perkin Trans. LI (1990) 157-162. [15] A. Strecker, y^ww. Chemie 75 (1850) 25-29. [16] R. Pascal, J. Taillades and A. Commeyras, Systemes de Strecker et Apparentes. IX- L'acetone, catalyseur d'hydratation des a-aminonitriles tertiaires en solution aqueuse basique. Bull Soc Chim. France 3-4 (1978) 177-184. [17] R. Pascal, J. Taillades and A. Commeyras, Systemes de Strecker et Apparentes. X- Decomposition et hydratation en milieu aqueux basique des a-aminonitriles secondaires. Processus dliydratation autocatalytique et catalyse par I'acetone. Tetrahedron 34 (1978) 2275-2281. [18] R. Pascal, J. Taillades and A. Commeyras, Systemes de Strecker et Apparentes. XII- Catalyse par les aldehydes de Thydratation intramoleculaire des a-aminonitriles. Tetrahedron 36 (1980) 2999-3008. [19] S.L. Miller, Production of some organic compounds under possible primitive Earth conditions. J. Am. Chem. Soc. 11 (1955) 2351-2361. [20] S. Pizzarello and G.W. Cooper, Molecular and chiral analyses of some protein amino acid derivatives in the Murchison and Murray meteorite. Meteoritics Planetary Science, 36 (2001) 897-909. [21] J. Taillades, J. Brugidou, R. Pascal, R. Sola, L. Mion and A. Commeyras, Nouvelles voies de synthese d'acides a-ammcs. Actualite Chimique (1986) 13-20. [22] J. Taillades, L. Boiteau, I. Beuzelin, O. Lagrille, J. Biron, W. Vayaboury, O. Vandenabeele-Trambouze, and A. Commeyras, A pH-dependent cyanate reactivity model: application to preparative Ncarbamoylation of amino acids. Perkin Trans. 2 (2001) 1247-1253.

64

Progress in Biological Chirality

[23] A. Commeyras, J. Taillades. H. Collet. L. Mion. L. Boiteau, O. Trambouze-Vandenabeele, H. Cottet J.P. Biron, F. Schue, O. Giani, O. Lagrille, R. Plasson, W. Vayaboury, H. Martin, F. Selsis, M. Dobrijevic and M. Geffard. La Terre matrice de la vie: emergence avant-gardiste des peptides sur les plages de I'Hadeen., in: L'environnement de la Terre primitive, (Eds. M. Gargaud D. Despois and J.P. Parisot) Presses Universitaires de Bordeaux, 2001, p. 361-380. [24J A. Commeyras, H. Collet, L. Boiteau, J. Taillades, O. Vandenabeele-Trambouze, H. Cottet, J.-P. Biron, R. Plasson, L. Mion, O. Lagrille, H. Martin, F. Selsis and M. Dobrijevic, Prebiotic synthesis of sequential peptides on the hadean beach by a molecular engine working with nitrogen oxides as energy sources. Polymer Internat. 51 (2002)661-665. [25] H. Collet, C. Bied, L. Mion, J. Taillades and A. Commeyras, A new simple and quantitative synthesis of a-amino acid-N-carboxyanhydrides. Tetrahedron Lett. 37 (1996) 9043-9046. [26] S.J. Mojzsis, R. Krishnamurthy and G. Arrhenius, Before RNA and after: geophysical and geochemical constraints on molecular evolution, in: The RNA word second edition (Eds. R.F. Gesteland, T.R. Cech and J.F. Athins) Cold Spring Harbord, 1999 p. 1-47. [27] H. Collet, L. Boiteau, J. Taillades and A. Commeyras, Solid phase decarbamoylation of N-carbamoyl peptides and monoalkylureas using gaseous NOx: a new simple deprotection reaction with minimum waste. Tetrahedron Lett. 40 (1999) 3355-3358. [28] A. Radzicka and R. Wolfenden, Rates of uncatalysed peptide bond hydrolysis in neutral solution and the transition state affinities of proteases. J. Am. Chem. Soc. 118 (1996) 6105-6109. [29] JR. Cronin and S. Pizzarello. Enantiomeric excesses in meteoritic amino acids. Science 275 (1997) 951955. [30] S. Pizzarello and J.R. Cronin, Non-racemic amino acids in the Murray and Murchison meteorites. Geochim. Cosmochim. Acta, 64 (2000) 329-338. [31] N.E. Blair and W. A. Bonner, A model for the enantiomeric enrichment of polypeptides on the primitive Earth. Origin Life Evol. Biosphere 11 (1981) 331-335. [32] K. Iwamoto and M. Seno, On a chemical system related to absolutely asymmetric synthesis. J. Chem. Phys. 76(1982)2347-2351. [33] D.K. Kondepudi and G.W. Nelson. Weak neutral currents and the origin of biomolecular chirahty. Nature 314(1985)438-441. [34] J. Taillades, H. Collet, L. Garrel. 1. Beuzelin, L. Boiteau, H. Choukroun and A. Commeyras, N-Carbamoyl amino acid solid-gas nitrosation by NO/NOx: A new route to oligopeptides via a-aminoacid Ncarboxyanhydride. Prebiotic imphcations. J. Mol. Evol. 48 (1999) 638-645. [35] O. Lagrille, Nitrosation de N-carbamoylamino acides solides par le melange gazeux NO/O2. Synthese de N-carboxyanhydrides (Anhydrides de Leuchs), in: Departement de chimie. Universite de Montpellier 2, Montpelher, Fr. 2001, p. 211. [36] O. Lagrille, J. Taillades, L. Boiteau and A. Commeyras, N-carbamoyl derivatives and their nitrosation by gazeous NOx-A new promissing tool in stepwise peptide synthesis. Eur. J. Org. Chem. (2002) 1026-1032. [37] P.D. Bartlett and D.C. Dittmer, A kinetic study of the Leuchs anhydrides in aqueous solution. II.,. J. Am. Chem. Soc. 79 (1957) 2159-2161. [38] P.D. Bartlett and R.H. Jones, A kinetic study of the Leuchs anhydrides in aqueous solution. I. /. Am. Chem. Soc. 79 (1957) 2153-2158. [39] R. Plasson, J.P. Biron, H. Cottet A. Commeyras and J. Taillades, Kinetic study of a-amino acids Ncarboxyanhydrides polymerisation in aqueous solution using capillary electrophoresis. J. Chrom. A., 952/1-2(2002)239-248. [40] H.R. Kricheldorf, a-Amino acid-N-carboxy-anhydrides and related heterocycles. Synthesis, properties, peptide Synthesis, polymerisation. Heterocycles. Springer-Verlag, Berhn, 1987, 213 pages. [41] PL. Luisi, P. Walde, M. Blocher and D. Liu, Research on the origin of life: membrane-assisted polycondensations of amino acids and peptides. Chimia, 54 (2000) 52-53. [42] R. Pascal, Catalysis by Induced Intramolecularity: What can be learned by mimicking enzymes with caibonyl compounds that covalently bind substrates?". Eur. J. Org. Chem. (2003) 1813-1824.

Progress in Biological Chirality G. Palyi, C. Zucchi and L. Caglioti (Editors) © 2004 Elsevier Ltd. All rights reserved.

Chapter 6 Use of Amino Acids and Amino Acid Racemization for Age Determination in Archaeometry J. Csapo,^'* M. Collins,^ Z. Csapo-Kiss,^ E. Varga-Visi/ G. Pohn" and J. Csapo' Jr. ""University ofKaposvdr, Faculty ofAnimal Science, H'740l Kaposvdr. P.O.Box 16, Guba S. u, 40, Hungary csapo@mail atk. u-kaposvar. hu ^ Fossil Fuels and Environmental Geochemistry, NRG, Drummond Building, University of Newcastle-upon-Tyne, Newcastle-upon-Tyne, NEl 7RU, United Kingdom ^Janus Pannonius University Faculty of Natural Sciences, Natural Geography Department, H-7624 Pecs, Ifjusdgu, 6, Hungary

1.

Introduction

/. / Age determination from fossil bones The subject of this paper is a survey on the possibilities of the determination of the age of archaeological samples (bones and teeth) containing proteins based on transformations of the amino acids therein. The first possibility is to estimate chronological age based upon the extent of racemization of the L-amino acids which are the constituents of native proteins. Amino acid contents in fossil shell, bone and tooth samples from were first reported by Abelson in 1954 [1]. In 1967, Hare and Abelson [2] reported that D-amino acids in fossils resulted from conversion of L-amino acids of protein. It was found that the older the fossil the higher the D/L ratio and, after a certain age, amino acids occurred in racemic form. The ratio of D-allo-isoleucine and L-isoleucine content in a fossilised shell sample was found to be 0.32 and the fossil was estimated to be 70 000 years old [3]. This is considered as the first application of amino acid racemization (AAR) - or more correctly epimerization (isoleucine has two chiral centres) - in geochronology. Subsequently, racemization of amino acids was used for age determination of various materials containing protein [4, 5, 6, 7, 8, 9, 10]. Isoleucine was given special attention because it could be reliably and consistently separated from the diastereoisomeric D-alloisoleucine by either ion exchange or reversed phase chromatography. Some errors of age determination based on AAR were reported [5]. Temperature, pH, soil composition and various contaminants would also impact upon the extent of racemization (and hence age estimation) in fossil bone samples. Indeed Marshall [11] in reporting the troubled history of racemization dating of bone highlighted the mismatch between early AAR estimates of Palaeoindina bones [7] and subsequent radiocarbon dating of the same amino

66

Progress in Biological Chirality

acids. Marshall [11] reports Milford Wolpoff s opinion that many people regarded AAR as "some kind of joke". Since various changes in temperature during the past the reaction temperature of racemization can only be estimated and not accurately determined. Therefore in this, as in earlier studies, DL ratios of independently dated samples have been used to 'calibrate' the rate. These data are then compared with data obtained from the analysis of amino acids in samples of unknown age. To make the comparison more accurate, the antecedents of samples of known age when analysed were the same as or similar to those of unknown age. All of the samples analysed in this study originated from the similar Hungarian environment. The samples were excavated from a depth of approximately 1.5-2.0 metres (the depth is very important because of the temperature; the yearly average temperature at this depth is ll^'C. The pH of the soil at this depth in most cases was very near to neutral and never exceeded 7.0 (the pH ranged between 5.5 and 6.8). Therefore, approx. 250 fossil bone samples previously analysed by the radiocarbon method were collected from various Hungarian museums, and their D- and L-amino acid contents were determined so far. The analyses were accomplished in the Radiocarbon Laboratory of the Nuclear Research Institute of Debrecen from collagen isolated from bone according to the generally accepted methods of radiocarbon laboratories. The DL ratio was plotted against time to produce a calibration curve. This curve can be used to estimate the age of samples of unknown age after their D- and L-amino acid contents have been determined. The DL ratio for 2-3 amino acids was determined for each sample and the mean value of ages estimated from calibration curves was used to derive the final age estimate. 1.2 Individual age determination from teeth The possible most accurate estimation of age at the time of death is a significant part of the study of historical populations. In everyday practice age means the time span between the birth and death of any individual human being. It is traditionally measured in years, months, weeks, etc. In other words it is calendar or chronological age. The usual age estimation methods of historical anthropology observe and analyse the changes occurring on the skeleton and on the teeth with the progress of age. Growth, maturation and aging are all processes that leave characteristic traces on human bones and teeth, therefore the estimation of the so called skeletal and dental ages can be estimated. Skeletal and dental ages come from the biological age of any given individual. If and when the person's biological age is close to his or her calendar age our estimate for his or her skeletal (or dental) age can fall very close to his or her calendar age [12]. This way we are able to produce an indirect estimation of calendar age via biological age determination. When estimating age the anthropologist analyses ontogenesis, the biological aging of the human organism. These processes are determined by a multitude of external (environmental) and internal (genetic) factors. Chronological age runs at a steady pace, but the passing of biological age can be and really is very varied between individuals and even within one human frame too, among its constituent parts. This is the reason for which we cannot achieve a truly accurate calibration of our age measuring methods compared to the passing of chronological time: this sort of calibration could be nothing else but an individual one. As biological and calendar ages often don't concur, chronological age in paleoanthropology may be estimated only within certain intervals (5 years at least). There can be no other realistic aim for us but to score within these limits with the greatest plausibility possible. Some

Use of Amino Acids and Amino Acid Racemization for Age Determination in Archaeometr>^ constituents of the human body, such as teeth, are easy to examine even centuries after the death of the individual, as their changes are not significantly affected by the environment. Their analysis could significantly improve the accuracy of age estimation. Helfman and Bada [13] were the first to declare that the aspartic acid racemization process within teeth could be utilised to estimate the age of living animals and of humans. Then they estabhshed the reaction rate coefficient of aspartic acid for human teeth at 8.29x10""* year"\ but they measured it at T-STxlO""* year"^ a year later [14]. Bada and Brown [15] produced a calibrating diagram by plotting In (H-D/L)/(l-D/L) against time. They found a satisfactory level of conformity between the actual age data and the data estimated on the basis of amino acid racemization. Gillard et al [16] analysed the D-Asp content of molars and premolars. They failed to establish significant differences between the D-Asp contents of the same teeth when taking samples fi-om several parts of their crowns and roots. Ohtani and Yamamoto [17] found significant differences between estimated and actual ages when comparing the D-Asp contents of dental enamel and dentin substance. They detected that racemization (kAsp: 5.75x10"^ year'^) is much faster in dentin than in enamel (kAsp: 4.47x10"'* year'^). They concluded that age could be more accurately estimated on the basis of dentin (±3 years) than on the basis of enamel (±2-11 years). They also calculated post-mortal reaction speed coefficients for an average temperature of 15 "^C (dentin: kAsp. 9.70x10"^ year'^ and enamel: 1.330x10"^ year-^). Ritz et al. [18] analysed the dentin substance of third molars' roots and they arrived at the conclusion that racemization of the root's dentin was somewhat different to that of the crown's dentin. They produced a special calibration diagram for those cases where age has to be determined but the crown substance was damaged or impaired. They established that the degree of Asp racemization was multiplied in acid solvent proteins compared to non-acid solvent ones. Ohtani [19] examined Asp racemization on central and lateral incisors, on first and second molars and he also treated the averages of these. He found a close correlation of actual ages and D/L Asp ratios. Ohtani concluded that racemization within deciduous teeth was a good indicator for the estimation of individual age, but racemization of permanent was far less useful for the same purpose. Ritz et al. [20] have reported the application of this method to biopsy samples of dentine. Thanks to the strictly regulated nature of the sampling process ages estimated on the basis of Asp racemization presented a very close conformity to actual ages. Ritz et al. established a margin of error not larger than ±year for 46.4% of the cases analysed and the error of age estimation never went beyond the ±5 years limit. Improvements in chromatographic methods have increased the number of chiral amino acids which can be separated in a single analysis. Nevertheless the ready separation, rapid racemization and high concentration of aspartic acid in teeth, means that it is still the most widely studied amino acid. We have previously estimated the racemization half-lives of different amino acids [21, 22]. It was concluded that D-enantiomers of amino acids with faster racemization than that of Asp promise to be just as good indicators of individual age as Asp, and amino acids falling of slower racemization rate could also provide useful information on age. We present in this paper two case studies: one regarding the age determination of fossil bone samples and the other concerning the estimation of individual age at the time of death. The development of the calibration curves used as well as the correlation of our results (based

67

68

Progress in Biological Chirality

on amino acid racemization) with standard paleoanthropological methods will be also described in this report. 2.

Materials and Methods

2. / Amino acid determination method The determination of the DL amino acid ratio in hydrolyzed proteins requires chromatographic methods which are suitable for the simultaneous separation and highly sensitive detection of several amino acids and their enantiomers. In order to be able to measure the changes of the amino acid composition and/or the enantiomeric ratio of the individual amino acids in very small archaeological samples the use of the highly selective and sensitive high-performance liquid chromatographic method (HPLC) is inevitable. Precolumn derivatization of a-amino acids using o-phthaldialdehyde (OPA) and various thiols as the reagents leads to fluorescent formed derivatives, while another reagent, 9-fluorenylmethyl chloroformate (FMOC-Cl) transforms a-amino and imino acids to fluorometrically active derivatives. Both types of derivatives have good chromatographic properties enabling several amino acids to be separated and measured within one chromatographic run at very low concentration levels. If the aim is the enantiomeric separation of amino acids, chiral derivatization agent should be used which transforms the amino acid enantiomers to pairs of diastereomeric derivatives separable on achiral HPLC columns. The chiral 1-(9-fluorenyl)-ethyl chloroformate (FLEC) reagent [23], and also the OPA/thiol reagent are suitable for this purpose if a chiral thiol is used in the latter case [24]. Both types of reagents were successfully used in our study. If the aim is only the separation of the isoleucine and D-allo-isoleucine, no precolumn derivatization is necessary since they are diastereoisomers, and the separation and quantitation can be made by amino acid analyser without chiral selectors. The prerequisite of a successful chromatographic analysis is the complete hydrolysis of the protein content of archaeological samples. All hydrolyses were carried out using 6M hydrochloric acid. If the aim is the determination of the enantiomeric ratio, it is very important to keep the extent of racemization at the lowest possible level. The effect of the reaction times and temperatures on the extent of the hydrolysis and racemization has been carefully investigated [25]. 2.2 Hydrolysis of proteins with reduced racemization The following materials were used for testing the racemization during hydrolysis: bovine ribonuclease, lysozyme, cytochrom C, fossil bone sample, and individual free amino acids as follows: L-aspartic acid, L-glutamic acid,, L-threonine„ L-alanine, L-valine, L-phenylalanine, L-histidine and L-tryptophan. The classical Moore and Stein [26] method using 6M hydrochloric acid medium for 24 h at 110 ""C was compared with hydrolysis in the same medium but at higher temperature (160-180 "^C) with shorter reaction times (15-60 min) using closed Pyrex tubes as reaction vessels. The use of microwave oven was also attempted. The method of Einarson et al. [23] was used to monitor the racemization (HPLC column: Kromasil C8 5 |am, 250 x 5.6 I.D.; eluent: gradient system A = 40% methanol in 9.5mM phosphate buffer at pH = 7.05, B = acetonitrile; pre-column derivatization reagent: ophthalaldehyde/2,3,4,6-tetra-0-acetyl-1 -thio-(3-D-glucopyranoside (OPA/TATG)).

Use of Amino Acids and Amino Acid Racemization for Age Determination in Archaeometry 3.

69

Results and Discussion

3. L Hydrolysis studies With the use of higher hydrolysis temperature e. g. 160 ""C, 170 ""C or 180 T , much shorter reaction time is adequate to achieve complete hydrolysis related to classic hydrolysis conditions when 24 hour heating is required for the same purpose at 110 ''C. Even splitting of the most stable bonds adjacent to Val, He and Leu is completely accomplished at 160 ''C for 60 min, at 170 ^C for 45 min and at 180 T for 30 min. The rate of the racemization in the course of the hydrochloride acid catalyzed protein hydrolysis decreases in the following order: aspartic acid, glutamic acid, threonine, phenylalanine, alanine, valine, histidine, as shown in Table 1 on the examples of lysozime, cytochrom and bone sample. Tryptophan was almost completely decomposed under the examined conditions [25]. The rate of the racemization of peptide bound amino acids in 6M hydrochloric acid between 110 and 170 ""C is 4-7 times higher than that of the free amino acids. Microwave-promoted hydrolysis is advantageous if racemization occuring due to this sort of treatment has no importance related to the subject of the study. If minimizing racemization is a prerequisite of the success of the research, this method is definitely disadvantageous. As seen in Table 1, the degree of racemization in case of high temperature - short reaction time methods is lower by 20-55% than at 110 "C for 24 h. Two optimal conditions were selected 160 ''C for 60 min or at 170 ""C for 45 min, and these were selected for the age estimation studies. Table 1. D- amino acid content of lysozyme (A), cytochrome (B) and fossil bone (C) hydrolysed by 6 M hydrochloric acid at different temperatures for different times Amino acid Asp Glu Thr Ala Val Phe His

A

110 "C for 24 h B

C

A

160 *C for 60 min. B

C

6.62 4.58 3.62 2.99 2.11 3.31 1.83

7.0i 4.61 3.74 3.21 2.24 3.42 1.89

7.89 5.93 4.38 4.02 2.53 3.64 2.38

3.27 2.79 2.29 1.69 1.69 3.19 1.64

3.42 2.84 2.31 1.65 1.84 3.37 1.67

4.15 3.61 3.14 2.13 2.33 3.57 2.01

A

170 T for 45 min. B

C

A

180 T for 30 min. B

C

3.29 2.81 2.11 1.72 1.71 2.89 1.52

3.57 2.89 2.23 1.77 1.82 3.11 1.60

4.42 3.74 3.04 2.11 2.27 3.60 1.99

3.84 3.51 2.87 2.81 2.54 2.97 1.79

3.99 3.63 3.14 2.89 2.57 2.83 1.93

4.67 3.92 3.42 3.04 2.82 3.20 2.11

Amino acid Asp Glu Thr Ala Val Phe His

The values refer to the percentage of racemization expressed as the ratio [D/(D+L)]xlOO. Each value is the mean of three determinations

70

Progress in Biological Chirality

3.2. Age determinations offossil bones 3.2.1. Epimerization of isoleucine. The bone samples were washed in running distilled water, dried in a vacuum drying oven and ground to produce powder material as fine as flour. Apolar contaminants were removed with petroleum ether in a Soxhlet extractor. The free amino acids were extracted by O.IM HCl solution for 16 hours. The nitrogen content of the residue was determined by Kjel-Foss nitrogen analyser. Sample size (200-2.000 mg residual material containing approx. 10-20 mg protein) was dependent on nitrogen content. Samples were weighed and hydrolysed with 6M HCl at 170 ""C for 30 min. The sample was then lyophylized, the residue was dissolved in water, and the precipitated silicate compounds were separated from the liquid containing free amino acids by centrifrigation. The pH of the solution was raised to pH=9 to precipitate metal hydroxides which were filtered off The hydrolysed solution was neutralised and evaporated to dryness by lyophylization. An aliquot of hydrolysed material was dissolved in a citrate buffer solution of pH=2.2, and isoleucine and D-allo-isoleucine were determined by LKB 4101 type amino acid analyser as described by Csapo et al. [27]. 3.2.2 Racemized amino acids determined with HPLC. The other D- and L-amino acids were separated by the method of Einarsson et al. [23]. The data obtained from the analyses on 24 fossil bone samples of known age from various Hungarian museums are summarised in Table 2 [10, 21]. Linear relationship was found between the D/L ratio and the age of the samples determined by the radiocarbon method. Seven amino acids (His= histidine, Phe= phenylalanine, Asx= aspartic acid / asparagine, Glx= glutamic acid / glutamJne, Ala= alanine, Ile= isoleucine, Val= valine) were investigated. These may be considered as being the most suitable for age determination because some of them show very fast racemization (His, Phe, Asp), while others show medium (Glu and Ala) and very slow racemization (He, Val); analytical data for other amino acids analysed are not presented in Table 2. None of the ratios lower than 0.1 or higher than 0.7 presented in Table 2 because, in these cases, the accuracy of age determination was doubtful. (The two acidic amino acids, aspartic acid and glutamic acid are exceptions, which mount up approx. 30% of the protein content of the bone and the tooth, so due to this the low rate of racemization resuhs in a well measurable D-enantiomers. The half life of amino acid racemization, i.e. the time required for the ratio of D- to L-amino acid to reach 0.333, was calculated by interpolation or extrapolation from the data of Table 2 and is presented in Table 3. The temperature of the samples during burying was approximately 1IX, while the pH of the soil in the surroundings of the bone was between 5.5 and 6.8. From the data of Table 2 and Table 3 it is evident, that D-His, D-Phe, D-Asp, D-Glu and D-Ala contents can be used for the age determination of samples which are 2-12.000, 320.000, 5-35.000, 8-70.000 and 10-80.000 years old, respectively. Age of samples older than 30.000 and 50.000 years can be determined on the basis of He and Val content, respectively. Data presented in Table 2 were corrected using the D-amino acid content of a fresh pig bone to correct for induced racemization. Concentrations of the D-form for all of the amino acids were negligible. However, all analyses were corrected for the small concentrations present in fresh pig bone. Studying the calibration curves, it can be concluded that in the case of D/L ratio lower than 0.1, the D-amino acid content is too low (with the exception of the aspartic acid and glutamic acid) and age determination is uncertain. It is obvious that the calibration curves can be used

Use of Amino Acids and Amino Acid Racemization for Age Determination in Archaeometry

71

Table 2. D/L ratios for various amino acids concerning ages of fossil samples determined by the radiocartwn method Age of samples determined by the ^'^C corrected method (year) 2200 2800 3110 3240 4630 5460 6850 11200 12400 15600 18600 20200 22600 25400 28600 30400 32500 36900 44600 46800 54300 62200 65000 72400

His

0.138 0.162 0.181 0.199 0.253 0.312 0.419 0.618 0.682

-

Phe

The D/L ratios for various amino acids Ala Glu Asp

_ 0.101 0.109 0.128 0.179 0.225 0.252 0.442 0.473 0.561 0.654 0.689

-

_ 0.109 0.128 0.171 0.271 0.289 0.378 0.432 0.491 0.543 0.580 0.621 0.643 0.702

-

_ 0.091 0.126 0.143 0.178 0.209 0.233 0.256 0.275 0.311 0.250 0.355 0.395 0.481 0.500 0.527 0.606 0.340

-

_ 0.112 0.131 0.158 0.192 0.209 0.228 0.246 0.289 0.321 0.343 0.381 0.465 0.483 0.510 0.586 0.613 0.652

lie

Val

-

-

0.099 0.118 0.134 0.142 0.169 0.188 0.199 0.221

Table 3. Half lives of racemization and epimerization of various amino acids found in Hungarian fossil bone samples* Amino acids Histidine Phenylalanine Tyrosine Aspartic acid Serine Threonine Glutamic acid Alanine Isoleucine Leucine Valine

Half life (year) 5500 8500 8600 13500 16500 17000 28500 32000 110000 140000 180000

The half life of amino acid racemization, i.e. the time required for the ratio of D- to L-amino acid to reach 0.333, was calculated from the data of Fig. 2. The temperature of the samples during burying was approximately 11 °C, while the pH of the soil in the surroundings of the bone was between 5.5 and 6.8.

0.100 0.115 0.119 0.136

72

Progress in Biological Chirality

for age determination most satisfactorily in the linear range (between 0.1 and 0.5 where Damino acids are present in well detectable amounts). The optimum D/L ratio for each sample can be found by analysing the amino acids best suited for age determination, e.g., for fossil bone samples of 11.200 years of age the D/L ratios for His, Phe, Asp and Ala are 0.682, 0.473, 0.271 and 0.112, respectively. In this case the D/L ratios of Phe and Asp are recommended for determining the age of samples; however, the D/L ratios of His and Ala can be used to confirm the estimate based on the ratios of Phe and Asp. The applicability of calibration curves is presented finally. As an example, one bone sample with unknown age and originated from the end of the Neolithic Age or from the beginning of the Copper Age according to the archaeologists was analysed for L- and Damino acids and the following results were obtained (the unknown sample was excavated from a depth of 1.85 metres and the pH of the soil was 6.50): L-His: 0.0697 mg, D-His: 0.0289 mg, D/Lnis= 0.428. Age calculated from calibration curve: 7100 years; SEM = 337. L-Phe: 0.0543 mg, D-Phe: 0.0138 mg, D/Lphe=0.254. Age calculated from calibration curve: 6950 years; SEM = 191. L-Asp: 0.1346 mg, D-Asp: 0.0245 mg, D/LASP=0.182. Age calculated from calibration curve: 6900 years; SEM = 465. The estimated age of the sample is the mean value of the above estimates, or 6980 years. This mean value has a standard error of 202 years and the 95% confidence interval would be 6554 to 7406 years. Of course the described method includes the analytical error of age estimation by the ^"^C method, but the effects on AAR of temperature, pH and the composition of soil have been eliminated. The D/L ratio for 2 to 3 amino acids should be determined for each sample, and the mean value of estimated ages based on calibration curves is considered the best estimate of age of the fossil sample. We have utilised this method very successfully for dating fossil bone samples from Hungary. The difference between the data from the calibration curve and those from ^"^C dating was generally negligible. From 1994 we have determined the age of several human bones and different species of animals (cattle, horse, red deer) to the greatest satisfaction of our archaeologist colleagues. However, our calibration curves should not be used in other environments, due to different conditions (temperature, pH). However, based on these results, other calibration curves can be formulated for each environment on the basis of the methods described here. The great advantage of this method compared to the other methods is that very low sample size is required. 2-10 mg sample size containing O.l-l.O mg protein is sufficient for age determination based on AAR. The radiocarbon age determination method can be used only for samples not older than 80.000 years. By means of AAR (especially by the means of He, Leu and Val) the age of the samples can be estimated up to 300-500 thousand years. 3.2.3. Age estimation of teeth based on the D-aspartic and D-glutamic acid contents. Our research project was carried out in collaboration with the Institute of Odontology and Earth Sciences Centre of Gothenburg University. We analysed two recent tooth sample series to establish the so called calibrating diagrams. In 1998-99 we gathered 22 teeth from the dental surgery of Pannon Agricultural University's Faculty of Animal Sciences in Kaposvar and we measured the D- and L-aspartic acid contents of them. When planning the sample we attempted to include individuals with in the largest possible age envelope (17-62 years), and

Use of Amino Acids and Amino Acid Racemization for Age Determination in Archaeometry

73

we also tried to select enough individuals from each age group to have a comprehensively representative sample. Table 4. contains the data produced by analysing the Kaposvar dental sample of 1998. We calculated ln(l+D/L)(l-D/L) correlations both for aspartic and glutamic acids besides D/L aspartic and D/L glutamic acid ratios. D/L ratios as well as the ln(l+D/L)(l-D/L) ftmction were presented as a function of age. We calculated correlations of known ages and the D/L ratios of the two amino acids by linear regression. We found a very close positive relation between D/L ratio and age in case of aspartic acid contents. The value of r was 0,91 for the D/L ratio as well as for the calculated function. When analysing glutamic acid we concluded that the values of r fell between 0,98-0,99 for the D/L ratio as well for the calculated function. Our examination of this dental sample of 22 teeth also led us to the conclusion that D-aspartic acid is a useful indicator for the estimation of individual age if treated to the analytical methods (protein hydrolysis, derivative production, separation and identification of D- and Lenantiomers) we applied. We also drew the conclusion that D-glutamic acid content is also suitable for accurate age estimation beside D-aspartic acid, though D-glutamic acid is present in teeth in a smaller concentration because of its different racemization half-period. That is the reason why it is more difficult to measure and its scoring is a more demanding job for researchers. Our first conclusions were based on a numerically small sample but we supported them by analysing 102 dental samples in 1999. At the same time we opened up our field of research from comparatively young age groups towards older ones. Our 1999 examinations produced r=0,93 positive correlations between D/L ratios and individual ages both in case of D-aspartic

Table 4. D-amino acid content of teeth of different age Age (year)

17 20 21 22 22 24 24 25 27 28 31 32 35 40 42 43 43 44 46 53 53 62

lii(l+D/L)/(l-D/L)

D/L ratio

Asp

Glu

Asp

Glu

0.034 0.035 0.036 0.037 0.038 0.039 0.038 0.041 0.042 0.043 0.044 0.044 0.047 0.050 0.052 0.053 0.053 0.053 0.055 0.059 0.060 0.065

0.017 0.017 0.019 0.019 0.020 0.021 0.019 0.021 0.021 0.021 0.022 0.022 0.024 0.026 0.026 0.027 0.028 0.027 0.028 0.031 0.030 0.033

0.068 0.070 0.072 0.074 0.076 0.078 0.076 0.082 0.084 0.086 0.088 0.088 0.094 0.100 0.104 0.106 0.106 0.106 0.110 0.118 0.120 0.130

0.034 0.034 0.038 0.038 0.040 0.042 0.038 0.042 0.042 0.042 0.044 0.044 0.048 0.052 0.052 0.054 0.056 0.054 0.056 0.062 0.060 0.066

74

Progress in Biological Chirality

and D-glutamic acids. The relation of an individual's age and the D-aspartic acid content of his or her tooth was presented in Figure 1, the correlation of age and D-glutamic acid was presented in Figure 2. These two correlations are eminently suitable to estimate the age of any individual in the age envelope of 40-86 years on the basis of the D-aspartic and D-glutamic content of his or her tooth. The results of our 1999 research work confirmed those arrived at in 1998, so we may state the existence of an extremely close link between any individual's age and the D-aspartic acid content of his or her tooth, and the D/L aspartic acid ratio on the basis of the analysis of a numerically large sample of teeth. Another one of our assumptions was also proven correct: it was not just D-aspartic acid content but also D-glutamic acid content that could be used to estimate the age of an individual if a sufficiently sensitive method of analysis was provided to measure the small concentration of D-glutamic acid present. In the third phase of our work we tried to apply the calibration diagrams produced by our amino acid racemization method on tooth samples originating in historical times. The age of individuals from Avar cemetery were estimated by the above mentioned anthropological methods in advance. The average of their age estimated by "traditional" anthropological methods, D/L aspartic and D/L glutamic ratios as well the age of these bone samples calculated on the basis of D/L aspartic acid and D/L glutamic acid was presented in Figure 3. In Figure 3 we also presented these ages by linear regression and this way it also presented the correlation of ages estimated by anthropological and by the two D-amino acid contents. It was evident from this figure that this correlation of ages estimated by the traditional and the new D-aspartic acid content methods was extremely close. When comparing the results of anthropological age estimation to those of the method based on amino acid racemization the value of "r" exceeded 0.9 with both amino acids.

50

60

70

80

Age of life (year)

Figure 1. Linear regression between the age of life of the individuals and the D/L aspartic acid ratio of their teeth

40

50

60

70

80

Age of life (year)

Figure 2. Linear regression between the age of life of the individuals and the D/L glutamic acid ratio of their teeth

Use of Amino Acids and Amino Acid Racemization for Age Determination in Archaeometry

75

E cp Q

^5

30

20

30

40

50

60

70

80

Age of life (year) estimated by anthropological methods

Figure 3. Linear regression between the age of life estimated by anthropological methods and D-amino acid content We investigated whether are there any differences in the D-aspartic acid and D-glutamic acid content of teeth from the same skull. For the sake of increasing the accuracy of the method we determined the concentration of the D-aspartic acid and D-glutamic acid and the D/L ratio on both two amino-acids of the incisors, eye teeth, premolar teeth and molar teeth taken from the skull of a female's skeleton estimated 40-45 years old by anthropological method. From our researches it seems that the incisors contain more glutamic acid than the others, and the aspartic acid content of the first incisor and the first molar tooth is also higher than the others (Figure 4). Examining the D-amino acids and the D/L ratio we ascertained that the D-aspartic acid and the D-glutamic acid content and the D/L ratio of the two incisors are the highest among all the teeth, followed by the eye teeth, the premolar teeth and the molar teeth (Figure 5). We received the least amount of D-amino acid content and D/L ratio at the third molar tooth. From our experiments it seems that the later growing teeth contain less Damino acids than the ones coming out earlier, which shall be taken into consideration at the age determination based on D-amino acid concentration of teeth. At the comparative experiments it is practical to use teeth from the same age, because the differences among the teeth of the individual can be bigger than the difference between individuals, which makes age determination uncertain. The significance of the anthropological utilisation of the amino acid racemization based age estimation method can be summarised as followings: A single tooth or even a fragment with some enamel is sufficient for amino acid based analysis. It could be an especially useful characteristic when treating poorly preserved remains. The amino acid racemization based age estimation method is built on an exact foundation of natural sciences. Its application was clear-cut, and therefore free of intra- and interpersonal errors. Amino acid racemization

76

Progress in Biological Chiralitv 1,4

Figure 4. L-aspartic acid and L-glutamic acid content of different teeth samplesfromthe same skull (Ii, I2 = incisor: C = canine; Pi, P2 = premolar; Mi, M2, M3 = molar)

0,06-

S 0,05o o 3

0,04-

I

•g

"o 05

o 0,03E 05

Q 0,02-

2 0,010,00-

I I D-aspartic acid I M I D/L Asp ratio I I D-glutamic acid • I D/L Glu ratio

mMz-+I1

in I2

Figure 5. Composition of the different teeth samplesfromthe same skull (Ii, l2= incisor C = canine; Pi, P2= premolar, Mi, M2, M3 = molar)

provides help where it is the most urgently needed: in the estimation of age within the most difficult adult group. The results produced by the amino acid racemization based method presented a fine correlation to those of other "classic" anthropological methods. When used in combination they could confirm each-other's results. For theoretical considerations this method was developed into a different entity compared to all routine age determination methods of anthropology. In contrast to all the other methods, it did not take into consideration the genetically programmed evolution of the organism, nor its responses to the environment or its physiological adaptation. It measured the structural

Use of Amino Acids and Amino Acid Racemization for Age Determination in Archaeometiy

77

alterations of amino acids, which are processes independent of circumstances of life and genetical facilities. In fact, amino acid racemization measured the passing of chronological time and not that of biological time, and therefore it gave a completely new meaning to the word age in historical anthropology. The advantage of the method elaborated by us compared to the previous ones can be summarized as follows: - we do not only determine the age of the individual by the aspartic acid, but by the D/L ratio of the aspartic acid and the glutamic acid, and the average of the two estimations are considered as the real age; - if different teeth are available from the different individuals then we can correct the age of the individual at the time of death by analysing the teeth of the same skull and the accuracy of the analysis can be improved if different teeth are available. 4.

Conclusion After reviewing previous attempts to use the extent of amino acid racemization for the determination of the age of archaeological samples containing proteins the authors describe their own approach. After an optimised protein hydrolysis with low racemization the D- and L-amino acid content in fossil bone samples of known age (radiocarbon method) was determined by HPLC after precolumn derivatization. Based on the obtained half-lives of racemization and plotting the D/L ratio as a function of time for various amino acids, calibration curves were obtained which can be used for the age determination of fossil bone samples in the range of 2000-500 000 years. Another method is presented for the determination of age at death based upon the racemization of aspartic acid in teeth. In addition to aspartic acid, D-glutamic acid was also found to be suitable for the estimation of age. Calibration curves based on these investigations were used for the age estimation of more than 200 skeletons of unknown age from the different historical periods. The correlation coefficient between our results and those obtained using standard paleoanthropological methods was over 0.9. From our experiments it seems that the later forming teeth racemized more slowly than those which erupt earlier, and this must be taken into consideration in age determination. Comparative experiments suggest that differences in extent of racemization of the same individual can be larger than the difference between individuals.

5.

References

[1] P.H. Abelson, Amino acids in fossils. Carnegie Inst. Washington Yearb. 53 (1954) 97-108. [2] P.E Hare and P.H. Abelson, Racemization of amino acids in fossil shells. Carnegie Inst. Washington Yearb. 66(1967)526-536. [3] P.E. Hare and R.M. Mitterer, Laboratory simulations of amino acid diagenezis in fossils. Carnegie Inst. Washington Yearb. 67 (1968) 205-212. [4] J.F. Wehmiller and P.E. Hare, Racemization of amino acids in marine sediments. Science 173 (1971) 907914. [5] K.M, Williams and G.G. Smith, A critical evaluation of the application of amino acid racemization to geochronology and geothermometiy. Orig. Life 8 (1977) 91-144. [6] G.H. Miller and P.E. Hare, Amino acid geochronology: Integrity of the carbonate matrix and potential of moUuscan fossils. In: Biogeochemistry of amino acids (Eds. P.E. Hare, T.C. Hoering and K. Jr. King) J. Wiley and Sons, NY, 1980, pp. 415-425.

78 [7 [8 [9 [10 [11 [12 [13 [14 [15 [16 [17; [18 [19 [20 [21 [22 [23

[24

[25

[26

[27

Progress in Biological Chirality J.L. Bada, Aspartic acid racemization ages of California Paleoindian skeletons. Am. Antiquity 50 (1985) 645-647. J.L. Bada, Amino acid racemization reactions andtiieirgeochemical implications. Ann. Rev. Earth Planet Sci. 13(1985)241. J. Csapo, I. Pap and L. Kolto, Archaeological age determination of fossil bone samples based on amino acid racenuzation and epimerization. Anthropologia Hungarica 1 (1988) 67-86. J. Csapo, Z. Csapo-Kiss, L. Kolto and I. Papp, Age determination of fossil bone samples based on the ratio of amino acid racemization. Archaeometry '90, Birkhauser Verlag, Basel 1990, pp. 627-635. E. Marshall, Racemization dating: Great expectations. Science 248 (1990) 799. D.H. Ubelaker, Human Skeletal Remains, Excavation, Analysis, Interpretation. Taraxacum, Washington, 1989, pp. 63-95. P.M. Helfman and J.L. Bada, Aspartic acidracemizationin tooth enamel from living hmnans. Prog. Natl. Acad. Sci. USA 72(1975)2891-2894. P.M. Helfman and J.L. Bada, Aspartic acidracemizationin dentine as a measure of ageing. Nature 262 (1976)279-281. J.L. Bada and S.E. Brown, Amino acidracemizationin living mammals: biochronological apphcations. Trends in Biochemical Sciences September 3-5, 1980. R.D. Gillard, A M Pollard, P. A. Sutton and D.K. Whittaker: An improved method for age at deatii determination from the measurement of D-aspartic acid in dental collagen. Archaeometry 32 (1990) 61-70. S. Ohtani and K. Yamamoto, Estimation of age from a tooth by means ofracemizationof an amino acid, especially aspartic acid - Comparison of enamel and dentin. J. Forensic Sciences 37 (1992) 1061-1067. S. Ritz, H.W. Schiitz and C. Peper, Postmortem estimation of age at death based on aspartic acid racemization in dentin. Its applicabihty for root dentin. Int. J. Legal Medicine 105 (1993) 289-293. S. Ohtani, Age estimation by aspartic acid racemization in dentin of deciduous teeth. Forensic Science International 68 (1994) 77-82. S. Ritz, R. Stock, H.W. Schutz and H.J. Kaatsch: Age estimation in biopsy specimens of dentin. Int. J. Legal Medicine 108(1995) 135-139. J. C s ^ , S. Nemethy, S. Folestad A. Tivesten, T.G. Martimand Z. Csapo-Kiss, Age determination based on amino acid racemization. A new possibility. Amino Acids 1 (1994) 317-325. J. Csapo, Z. Csapo-Kiss and J. Jr. Csapo, Use of the amino acids and amino acid racemization for age determination in archaeometry. Trends in Analytical Chemistry 17(1998) 140-148. S. Einarsson, S. Folestad and B. Josefsson, Separation of amino acid enantiomers using precolumn derivatization with o-phthalaldehyde and 2,3,4,6,-tetra-O-acetyl-l-thio-p-D-glucopyranoside. J. Liquid Chrom. 10(1987) 1589-1596. S. Einarsson, B. Josefsson, P. Moller and D. Sanchez: Separation of amino acid enantiomers and chiral amines using precolumn derivatization with (+)-l-(9-fluorenyl)ethyl chloroformate and reversed-phase liquid chromatography, ylwfl/. Chem. 59(1987) 1191-1198. J. Csapo, Z. Csapo-Kiss, L. Wagner, T. Talos, T.G., Martin, S. Nemetiiy, S.Folestad and A. Tivesten, Hydrolysis of proteins performed at high temperatures and for short times with reduced racemization, in order to determine the enantiomers of D- and L-amino acids. Analytica ChimicaActa 339 (1997) 99-107. S. Moore and W.H. Stein: Chromatographic determination of amino acids by the use of automatic recording equippment. In: Methods in Enzymology (Eds. S.P. Colowick and N.O. Kaplan), 6 (1963) 819831. J. Csapo, I. Toth-Posfai and Z. Csapo-Kiss, Optimization of hydrolysis at determination of amino acid content in food and feed products. Acta Alimentaria 15 (1986) 3-21.

Progress in Biological Chirality G. Palyi, C. Zucchi and L. Caglioti (Editors) © 2004 Elsevier Ltd. All rights reserved.

Chapter 7 Enantiomeric Enrichment in Nonracemic Conglomerates. A Possible Component of the Solution to the Problem of the Origin of Biochirality Stanley I. Goldberg Department of Chemistry, University of New Orleans, New Orleans, Louisiana, 70148, USA sgoldber@uno. edu

1.

Introduction There is a growing conviction, based on theoretical and experimental studies, reviewed by Bonner [1], that polyribonucleic acids had to have been initially homochiral in order to have been able to carry out the essential tasks of replication and translation. With even the slightest deviation from homochirality the two helical strands of polyribonucleic acids cannot achieve complete pairing of complimentary bases required for these processes. Homochirality, therefore, is not only an indispensable condition of contemporary life, but it also appears to have been a requirement for the initial emergence of life. Any solution to the biochirality problem, therefore, must provide either enantiopure prebiotic chiral biomolecules or, as suggested by Wald [2], racemic prebiotic homochiral biopolymers. Schemes involving partially resolved material which proceed on to living systems before they achieve the homochiral condition of contemporary life are unacceptable. The present report provides a possible component to a solution in accord with the enantiopure prebiotic stereochemical constraint. It combines the observational evidence for the presence of nonracemic samples of several amino acids in 4.5 billion year old Murchison meteorite, which fell in Australia in 1969 [3, 4, 5], and in the 1949 Murray Kentucky carbonaceous meteorite [6] with experimentally determined properties of partially resolved conglomerates [7] to show how global accumulations of enantiopure biologically relevant material could have formed on prebiotic Earth.

2.

Extraterrestrial Nonracemic Chiral Influence The discovery of nonracemic material in the Murchison and Murray meteorites is tantamount to the discovery of the earlier action of a nonracemic chiral influence that was extraterrestrial and gave rise to selective enantiomer depletions in prebiotic racemic material prior to its dehvery to Earth's surface. Whether that chiral influence was the circularly polarized portion of the synchrotron radiation (ultraviolet) associated with neutron stars, as postulated by [8], or the infrared (and presumed ultraviolet) circular polarization recently

80

Progress in Biological Chirality

observed in the Orion OMC-1 star formation by [9] (or some other source) will have to await ftiture clarification. The important fact is that large amounts of nonracemic chiral material was available on prebiotic Earth. Although the meteorites' more abundant enantiomer in every case was found to an L-amino acid, the levels of enantiomeric purity were low, below the 100% level required to meet the homochirality prerequisite discussed above. What was needed therefore was a readily available natural prebiotic process to transform that partially resolved material into global accumulations of enantiopure biomaterial.

3.

High Amplifications of Enantiomeric Purity by Crystallizations from Solutions of Nonracemic Conglomerates Although it is true that in the absence of a nonracemic chiral influence enantiomers are identical, this is not the case when an enantiomer on the molecular level is compared to an aggregated state of enantiomers within the same system. In such circumstances, the enantiomer in excess may be discriminated from those in the aggregated or associated state. This was experimentally demonstrated in the ^H-nuclear magnetic resonance spectra of highly structured chiral monolayers [10] and in partially resolved samples of strongly hydrogenbonded compounds by [11]. Molecular aggregation reaches a maximum, of course, in the crystalline state, and most enantiomers exhibit properties different from those of the crystalline racemic modifications with which they exist in equilibrium. Of these, the most important is solubility. In partially resolved conglomerates, it is the enantiomer in excess that is always first to crystallize [7] and the crystals may be obtained in near enantiomeric purity. Thus in 1956 Amiard [12] allowed aqueous solutions of nonracemic threonine to cool from 80"^ to 20''C and obtained crystals enriched in the more abundant enantiomer in amounts which, on average, were about two-thirds greater than the original excess. A single recrystallization gave enantiopure threonine. Watanabe and Noyori [13] found when they gradually cooled an aqueous solution of glutamic acid hydrochloride (13% enriched with the L enantiomer) they rapidly obtained almost enantiopure crystals, (98.7% L enantiomer). Nohira et al [14] used 20 mg samples of enantiopure crystals to seed solutions containing 15.5 g of racemic chiral amine salts to obtain, after only one recrystallization of initially deposited precipitate, between 1.6 and 3.4 g of enantiomerically pure material. Also to be noted is the rather wide spread occurrence of conglomerates. By examining 1308 compounds in Beilstein, [15] estimated that between 5 and 10% of all chiral neutral organic compounds crystallize as conglomerates. These facts when combined with current models of comets and the comments added below constitute an element of the solution to the biochirality problem. 4.

Mechanisms for Formation and Isolation of Deposits with High Enantiomeric Purity from Nonracemic Aqueous Solutions of Conglomerates Nondestructive delivery [16, 17, 18] of organic compounds, including chiral material along with huge amounts of water, principally by comets, to the Earth's surface during its first billion years is considered [16, 17, 19] to have been the major source of the planet's biomass. Some of these aqueous solutions, containing nonracemic chiral substances among the dissolved material, would flow into surface features where they would exist as extensive shallow lakes (Figure l.A), eventually depositing their solutes or part of their solutes (crystallization or fractional crystallization. Figurel.B), owing to evaporative concentration

Enantiomeric Enrichment in Nonracemic Conglomerates 81

6

D

Figure 1. Possible mechanisms for formation and isolation of highly enantiomerically enriched precipitatesfromnon-racemic aqueous solutions of conglomerates on early Earth. A - Initial aqueous solution fills an extensive hallow depression; B - concentration of solution owing to evaporation brings about deposition of material highly enriched in the originally more abundant enantiomer; C tectonic slippage or collision of a small impactor removes part of the rim of depression, allowing the supernatant to flow away; or, D - an earthquake like disturbance shifts solid material into the depression, absorbing the supernatant and forming an isolated stratum of enantiomerically enriched material

and/or temperature lowering. Earthquakes, avalanches, or an impacting meteors could breech the banks of the lakes, allowing the supematents to drain away leaving isolated deposited material (Figure l.C), or shift solid material into the depression, absorbing the supernatant and forming a stratum (Figure 1. D). In some cases, these isolated deposits will include, in a highly enriched state, the more abundant enantiomer of partially resolved conglomerates. Thus, during the long era when there were continuous, nondestructive deliveries of nonracemic chiral material to the surface of early Earth, one could expect such processes to be repeated many times in many different locations, resulting in substantial global accumulations of enantiopure biomaterial and setting the stage for the emergence of homochiral life.

82

Progress in Biological Chirality

References [1] W. A. Bonner, in: D-Amino Acids in Sequences of Secreted Peptides of Multicellular Organisms, (Ed. P. JoUes) Birkhauser Verlag, Basel, Switzerland 1998, pp. 159-188. [2] G. W2\d,Ann. N. Y. Acad Sci. 69 (1957) 352-368. [3] M.H. Engel and B. Nagy, Nature 296 (1982) 837-840. [4] M.H. Engel and S.A. Macko, Nature 389 (1997) 265-268. [5] J.R. Cronin and S. Pizzarello, Science 275 (1997) 951-954. [6] J.R. Cronin and S. Pizzarello, in: Abstracts of the 9th ISSOL Meeting, San Diego, California (USA) 11-16 July, 1999, p. 41. [7] J. Jacques, A. Collet and S.H. Wilen, Enantiomers, Racemates, and Resolutions, John Wiley & Sons, New York, 1981, pp. 181-182. [8] E. Rubenstein, W.A. Bonner, H.P. Noyes and G.S. Brown, Nature 306 (1983) 118. [9] J. Bailey, A. Chrysostomou, J.H. Hough, T.M. Gledhill, A. McCall, S. Claik, F. Menard and M. Tamma, Science 281 (1998) 672-674. [10] M.V. Stewart and EM. Amett in: Topics in Stereochemistry (Eds. N.L. Allinger, E.L. Eliel and S.H. Wilen) John Wiley & Sons. New York, 13 (1982) 195-262. [11] B.S. Jursic and S.I. Goldberg, J. Org Chem. 57 (1992) 7172-7174. [12] G. Amiard, Bull. Soc. Chim. France (1956) 447. [13] T. Watanabe and G. Noyori, Kogyo Zasshi 70 (1967) 2174-7; Chem. Ahstr. 69 (1968) 874431. [14] H. Nohira, M. Kai, M. Nohira. J. Nishikawa. T. Hoshiko, and K. Saigo, Chemistry Letters (1981) 951-952. [15] J. Jacques, M. Leclercq and M.-J. Brienne, Tetrahedron 37 (1981) 1727-1733. [16] C.F. Chyba, Nature 343 (1990) 129-133. [17] C. F. Chyba, P.J. Thomas, L. Brookshaw and C Sagan, Science 249 (1990) 366-373. [18] J.M. Greenberg, A. Kouchi, W. Niessen. H. Irth, J. van Paradijs, M. de Groot and W. Hermsen, J. Biol. Fhys. 20 {1994) 61-10. [19] J.M. Greenberg, in: Physical Origin of Homochirality in Life, (Ed. D.B. Cline) American Institute of Physics. Woodbury, New York. 1996, pp. 185-210.

Progress in Biological Chirality G. Palyi, C. Zucchi and L. Caglioti (Editors) © 2004 Elsevier Ltd. All rights reserved.

Chapter 8 Genetic Code: a Self-Referential and Functional Model Romeu Cardoso Guimaraes'* and Carlos Henrique Costa Moreira^ '^Dept Biologia Geral, Inst. Ciencias Biologicas, Universidade Federal de Minas Gerais, 3J270-901 BeloHorizonteMG, Brazil romeucg@icb. ufmg. br ^Dept. Matemdtica, Inst Ciencias Exatas, Universidade Federal de Minas Gerais, 31123-970 Belo Horizonte MG, Brazil

1.

Introduction Fundamental constituents of living beings are the polymers proteins and nucleic acids. Their sequences are mutually correspondent through the genetic code, a defining pillar of life's homogeneity [1]. The code is the set of correspondence rules between the amino acids and nucleotide base triplets. In cells, sequences of segments of DNA are transcribed into messenger RNA, which are translated into proteins. In this context, the code is a mechanism for translation. It can be said that the symbol or signal (significant) resides in the nucleic acid, the encoded (signifier) 'meaning' being the correspondent amino acids and proteins. In the context of an isolated system of protein synthesis, it is not necessary to consider DNA. Translation involves only RNA, and mRNA, can be considered genes [2]. This text presents a fijll model for construction of the code but remains synthetic in various more technical sections whose analytical treatment is forwarded to other presentations. The model indicates that the system was originally only coding for protein synthesis, not decoding pre-existing mRNAs. The presentation is organized in two topics. It starts describing the matrix of correspondences and the three basic constituents of the system. It is indicated that the clustering of similar attributions in the matrix can be well described by its division into sectors and quadrants. Some rules prevailing in the development of the system are individually reviewed and others introduced: amino acid types and biosynthesis routes, and their relationships with aRS classes; the locations of aRS classes and of the specific punctuation system in the sectors and quadrants. The topic is closed with the presentation of a scheme for the temporal order in the occupation of boxes containing more than one attribution, derived fi'om a rule of precedence of purines over pyrimidines in the development of attributions and in the evolution of variant sets, and of a mechanism involving the participation of tRNA pairs in protein synthesis. It is then described the systemic rationale followed in devising the temporal succession of attributions entering the code. Our analysis of the hydropathy correlation could point out the first attributions, which were also shown to belong in the main protein-stabilizing and RNA-binding set of amino acids, thereby able to

84

Progress in Biological Chirality

build a stabilized nucleoprotein working system. The development of a nonspecific punctuation system could be proposed, based on the distribution, along strings, of proteinstabilizing then destabilizing amino acids, followed by the introduction of the specific punctuation system, derived from the tRNA pairing mechanism. Amino acid sets characteristic of the protein secondary structures and of the DNA-binding motifs were adequately placed in the temporal scheme. The model was able to propose that genes were originated during the process of development of the coding system and defined in simultaneity with the definition of the products of the code, at the moments of their mutual coupling.

2.

Ingredients

2.1 Correspondence between triplets and amino acids Correspondences in the standard code are established by a system of 20 catalysts, the aminoacyl-tRNA synthetases (synthetases, aRS), one for each of the amino acid substrates. They are proteins, so that the notion of self-reference should already be introduced [3]. At some point in the evolution of the system, products became catalysts, participating in the process of their own formation; the system became cyclic and auto-catalytic. Other substrates are, according to the kind of organisms or organelles, 22-45 tRNAs (transfer RNA, transporters or adaptors). Each reaction synthesizes an aminoacyl-tRNA (ac-tRNA). The problem of deciphering the formation of the system is then a '3-body problem', at the heart of the process of building the cellular nucleoprotein system [4]. Nevertheless, much of present research is still devoted to the hypothesis that the code was formed by direct interactions between RNA and amino acids, as a '2-body problem' [5-7]. Such approaches may be relevant to a part of the process but certainly not to its totality. They usually privilege thermodynamic aspects and consider important the concentration of amino acids abundant in pre-biotic syntheses, partially overlapping the precursor or of independent biosynthetic origin, in the 3' C row (GNC triplets; Val, Ala, Gly, Asp, Glu) [6, 8]. The first four in this list were highlighted [9] on the basis of their wide variety of properties, which would be adequate for formation of all protein conformations together. Ninio [10], among others, valued the homogeneity of simple boxes in the core (Pro, Arg, Ala, Gly), but the two first ones are biosynthetically derived. A 'hidden' pattern of mRNA triplets, which would help ribosomes to maintain the regularity of reading in the triplet frame [11], was updated to GCU [12], indicating the origin of the coding system by Ala codons, composing the first 'words' of mRNA. Our systemic '3-body' approach accepts important participation of thermodynamic stability of triplets, but does not privilege this entirely. Thermodynamic data [13] referring to tRNA dimers but supposedly applicable to codon:anticodon minihelices, indicated that stability of the paired anticodon loops is much greater than what would be expected from helices of common RNA triplets. They are equivalent to helices of heptamers and relatively independent from the base composition of the triplets included. Most frequently, this text will refer to anticodons, their set being the genetic anticode. The standard anticode is considered the largest, with 45 anticodons, and containing only standard bases. The triplets generated by combinations of the 4 bases are 64 (4^) codons. The order of codons along the mRNA sequences dictates that of amino acids in proteins through pairing of anticodons with codons. Three codons do not have their correspondent tRNAs. They may be called non-sense, due to

Genetic Code: a Self-Referential and Functional Model

85

not encoding an amino acid, but determine the end of synthesis of a protein and are called terminators or stop (X). Codes containing more than one signal or symbol (codons, anticodons) for a meaning (amino acid, punctuation) are called degenerate. These can be read without ambiguity in one direction only; in the present case, from codons or anticodons to amino acids. The reading is made by synthetases, recognizing in the class of tRNA corresponding to one amino acid the attributes adequate to the class and accepting the ambiguity or degeneracy. A greater portion of the development of complexity, with reduced degeneracy, resides in the 5' position. Three of the 8 tetracodonic boxes (Ser, Leu, Arg) belong in the group of hexacodonic attributions. Tricodonics are He, with 5' (R, U) ambiguity, and X, with complex constitution: 3' A plus central and 5' Y. The other 4 split boxes are central U, 5' R or Y. Met and Trp are monocodonic, utilizing 5' C. Both tRNAs of Met, for elongation, and of fMet, for initiation (this with a slipped pDiN CAU; Section 3.14), are read by the same synthetase. In Table 1, each box shows the 4 bases in the 5' position of the triplets. These are necessary for identification of attributions in the 8 complex boxes, containing more than one attribution. Simple boxes contain only one attribution. In the standard anticode, only the bases in bold show up. Base A in this position is rare [14], here indicating only the occurrence of its complement (U) in codons. In X, 5' in parenthesis, only the codons occur, anticodons having been deleted. The pDiN approach is not a revival of the '2 bases out of 3' proposal [8, 15] for formation of the early code. It just describes main components of the triplets. 2.2 Matrix ofpDiN The order of bases follows their hydropathy (Figure 1), from hydrophobic to hydrophilic, A-G-C-U [6, 16]. This is different from the one sedimented in the traditional bibliography, AG-U-C. Nonetheless, the order through the aqueous interactions is the physiologic one. This is also indicated by various other symmetries depicted by the hydropathic matrix. An order in the distribution of simple and complex boxes can be described by the pDiN constitution (Figure IB). The 8 simple boxes are in the core (G and C in the pDiN) and in the non-axial boxes with central R, the 8 complex in the tips (A and U in the pDiN) and in the non-axial Table 1. Genetic anticode matrix. Read through columns the central base, through rows the 3' base, of the principal dinucleotides (pDiN; central -^3') of anticodons, which define the boxes of the matrix A G C U A G C U A G C U A G C U

AA

AG

AC

AU

Phe Phe Leu Leu Leu Leu Leu Leu Val Val Val Val He He Met, fMet He

A G C U A G C U A G C U A G C U

GA

GG

GC

GU

Ser Ser Ser Ser Pro Pro Pro Pro Ala Ala Ala Ala Thr Thr Thr Thr

A G C (U) A G C U A G C U A G C U

CA

CG

CC

CU

Cys Cys Trp X Arg Arg Arg Arg Gly Gly Gly Gly Ser Ser Arg Arg

A G (C) (U) A G C U A G C U A G C U

UA

UG

UC

uu

Tyr Tyr X X His His Ghi Gin Asp Asp Glu Glu Asn Asn Lys Lys

86

Progress in Biological Chirality

Ho RR

AA

Mx YR

AA AG

GA GG

CA CG

UA UG

AC AU

GC GU

cc cu

UC

RY

uu YY Ho

AU

"-.^ ^

GA

~" CA

AG 1 AC

GG

CG

GC

X

UA "^«« **

UG 1

^C

CC

,^^ "^ GU _ . . CU

(A)

" UU

(B)

Figure 1. Symmetries in the pDiN matrix and the precedence of central purines. (A) The 16 pDiN boxes are separated in homogeneous (Ho) and mixed (Mx) sectors, each of these divided into quadrants, respectively, RR : YY, and RY : YR. Axial and non-axial boxes are distinguished: Ho axis (bases repeated) bold, Mx axis (bases self-complementary) underlined. (B) The core of the axes or of the matrix is constituted by G and C bases, the tips by A and U. Non-axial boxes, italics, build a circle around the core and internal to the tips. Simple (tetracodonic) boxes in bold (core and non-axial with central R); complex boxes are the tips and non-axial with central Y

boxes with central Y. The core versus tips contrast may be related t o thermodynamic properties [6, 10] but the central purine properties involved should be related to their reactivity or specificity, e. g. bulkiness and hydrophobicity, in the interactions with the synthetase active centers.

2.3 Amino acid types and biosynthesis routes An important advance in understanding the process o f structuring the code w a s introduced by W o n g [17, 18], indicating that amino acids belonging to a biosynthesis route were regionalized, localized close to each other and regularly along the matrix. The indication originated from the proposal o f determination by biosynthesis routes is that the code was

Class 2 Group

6

5

4

3

pDiN 2

1 Ser Thr Pro

His Gly Ala

GA GU GG UG CC GC

Asp UC Asn UU

b Lys Phe

Class 1 1

AG AC AU AU CA CG

3

4

5

Group

Leu Val lie Met Cys Arg

Glu UG

UU

AA

2

Ghi

b (Lys)

CA UA

Trp Tyr

c

Figure 2. Synthetase classes and groups. The schematic, adapted from [26], is an attempted consensus maximizing the balanced sizes of the groups. The rare LysRSl is in parenthesis. Attributions in the CU box and Leu^^, involved with the hexacodonic attributions, are not shown

Genetic Code: a Self-Referential and Functional Model (A) ^ Asp(Ho2) ^ ^

(B)

Asn (Ho2) Lys(Ho2^'>) ThrrM.9. Thr (Mx2) ^ <

^^^ ^^^^

p^^^^^,

^ Pro(Ho2) < ^

- . Leu (Hoi) ^ ^Val(Mxl) [possibly Ala (Mx2)]

Met(Mxl) [or Asp -^ Met]

Gin (Mxl) Glu(Hol) ^ > Arg(Mxl) [orAsp^Arg]

Ser (H02)

87

Cys (Mxl) Trp (Mxl)

Phe (Ho2)

— • Tyr (Mxl)

Ser (Ho2)

— • Gly (Ho2)

Glu(Hol) or Gin (Mxl)

— ^ His (Mx2)

(C)

s

c,w

P

R — E D

I,M

T

N,K

Figure 3. Amino acid biosynthesis routes. Sectors, Ho and Mx; classes numbered; in parenthesis, the rare LysRS class 1. (A) Precursors and their derivations, in famiUes. (B) Amino acids of independent origin and nonconsensual derivationsfromothers belonging in the code. All propositions, except the derivation of His, are consistent with our model and summarized in (C) formed gradually, in co-evolution with the successive abundance of available amino acids, thereafter becoming incorporated into the synthetase system. Attributions of precursor amino acids were fixed and, when their derivations became abundant, these were attributed to tRNAs similar to the previous one. So, the code was established inside a metabolic system already in a complexification process and availability of amino acids was a limiting factor. It became open to discussion if such clustering, driven by biosynthesis relatedness, was more significant than those due to synthetase properties. The latter should be more akin to amino acid constitution, reflecting either similarity in chemical reactivity or steric properties. Some authors [19, 20] admit that the latter seems to be more coherent with groups inside the synthetase classes (Figure 2). Other aspects of the regionalization will be discussed in Section 2.5. Amino acid properties are multitude and may lead to different classifications [7^ 21], usually attempting quantification of attributes, ranking and averaging of different qualities, with questionable consensuses. Our choice was for some qualitative attributes thought most relevant for formation of the code, biosynthesis relatedness being one of them. Biosynthesis routes have received many revisions [22-24]. Only the consensual routes are adopted here (Figure 3). Amino acids of independent origin or with non-consensual derivations are 7, so

88

Progress in Biological Chiralit\' Table 2. Distribution of synthetase classes and punctuation signs. *, deviants from de aRS class / pDiN type rule; •**, the hexacodonic aRSl expansions and the rare LysRSl; the Ho sector bold Tyr 1 AA 2^' 1 UA CA Phe GA Cys Ser 2 X 1 Leu Trp X 2* His 1* Arg UG AG 1 Leu GG 2 CG Pro Gin 1 2 AC 1 Val Asp GC 2* 2 Ala Gly 1* Glu 1 Asn 2 AU He Ser 2 GU Thr 2* 2(1*) Met 1 Lys l^-> Arg fMet

cc cu

uc uu

that models obeying the biosynthesis derivation rule are constrained only by the relationships in the families of 3 precursors and 10 derived. Class transfers from precursors to derived are: class 2 -> 1 (n=5) and class 1 -> 2 (Glu -> Pro). 2.4 Aminoacyl-tRNA synthetases The 20 synthetases of the standard code are divided in 2 classes or families (Figure 2), with internal homogeneity, formed by duplications of an original enzyme. Each class contains 10 components, distributed in the matrix as in Table 2. LysRS is atypical, class 1 in some organisms. These duplications, followed by variations enabling adequation of the new enzyme to new substrates, should have been limiting factors in configurating the system. Class 1 aminoacylates the 2' hydroxyl of the terminal adenosine of tRNA, while aRS2 does so on the 3' hydroxyl, PheRS being also atypical, aminoacylating in the aRSl fashion [25]. The presentation in the form of dendrograms reflects degrees of similarity of sequences of the proteins. The most similar to each other are in level 1, the most different from all others in levels 5 or 6. The most common interpretation is phylogenetic, saying that the proximity to the root would indicate longer time of divergence from an ancestor to the class or group, and distance from the base more recent branches. Such interpretation should not be taken at face value. Criticisms may be valid, due to the algorithms generating the trees being dependent on not firmly based assumptions. The phylogenetic trees would depend on divergence rates being regular, while these are generally not well known and depend on choices of adequate informative sites', leaving aside some special types of mutations. The maximized numerical symmetry of groups (Figure 2) is only approximate, being better qualified as balance or equilibration. In the standard code, the classes are composed of 10 enzymes each. Variations are of 11 class 1 to 9 class 2, in organisms where LysRSl replaces LysRS2 [27], but may be wider when AsnRS or GlnRS are dispensed with, such amino acids being obtained by amidation of the respective Asp-tRNA and Glu-tRNA. A synthetase is known with double attribution, for Pro and Cys [28]. Inside the groups of each class, variations may be more complex due to formation of functional subgroups, more valued by other authors. In [29], Arg-, Glu- and GlnRS are grouped together due to their tRNA-binding before that of amino acids; GlyRS is the simplest and PheRS the most complex, this being atypical in the acylation mode. In a consensual phylogeny [30], it is proposed to group GlyRS with PheRS, and AlaRS with HisRS, the most basal in the class. SerRS2 and AlaRS2, together with LeuRSl are also special due to not binding the anticodon stem-loops of tRNA [31-33]. It is clear from the distribution in the sectors, that aRS2 shows preference for the Ho and aRSl for the Mx sector.

Genetic Code: a Self-Referential and Functional Model

89

Synthetases not obeying this mle sum 25-30% in each class, and are called deviants. Hexacodonic expansions of aRSl are not counted as deviants and LysRS is class 1 in a few organisms, not being counted as standard. Other regularities are apparent, most evident the preference of aRS2 for central G and of aRSl for central A. It will be shown below that the code organization followed the aRS preferences for sectors, other regularities being consequent and secondary. A compelling evidence for antiquity of aRS2, relative to aRSl [34-35], stems from the observation that the former corresponds to simpler and smaller amino acids, the latter to bulkier and complex ones (Table 3). Class 2 corresponds also to a greater number of amino acids precursor to biosynthesis routes, protein N-end stabilizers, characteristic of simpler protein conformations and nucleic acid binders, and to more degenerate attributions (Table 4). The ensemble of these data indicated the temporal order, starting the construction of the code by the Ho sector and aRS2. Class 2 predominates over class 1 (7/1 amino acids) below a threshold of around 70 A, the reverse (1/9) above the threshold. The 2 others, the largest of aRS2, are activated by the atypical aRS systems. It is indicated that LysRS2 and PheRS2 active center pockets were not fully adequate to receive such bulky amino acids and with extreme hydropathy so that, in some organisms, LysRS 1 replaced the usual class 2 and PheRS2 adopted exceptional conformation, resulting in aminoacylation on the hydroxyl typical of aRSl. 2.5 Regionalization and the diagonal route The traditional explanation for the processes leading to the regionalized distribution of attributions in the matrix (clustering of similar amino acids and codons), has been that of Table 3. Amino acid side chain volume and amino acid protein residue hydropathy. Volumesfrom[36], hydropathy from [37]. The rare LysRS 1 in parenthesis, acylation site not known. The exceptional PheRS2, acylating the T hydroxyl of the terminal adenosine of the tRNA, in the aRSl mode, is marked aRS2

Hydropathy

A

Gly Ala Ser Pro Asp

0.423 0.258 0.508 0.677 0.962

3 31 32 32.5 54

Asn Thr

0.809 0.438

56 61

His

0.573

aRSl

Hydropathy

A

Cys

0.111

55

Glu Val Ghi

0.935 0.069 0.841

83 84 85

Met He Leu (Lys) Arg

0.059 0.029 0.066

105 111 111

0.982

124

Tyr Tip

0.361 0,325

136 170

96

Lys

1.000

119

Phe^

0.000

132

90

Progress in Biological Chirality

selection against errors derived from the non-clustered distributions (supposed to have occurred), which is the 'minimization of errors' hypothesis, reviewed in [23, 38]. Determinist contributions from the co-evolution of the routes of amino acid biosynthesis and of tRNA derivation [17], and from aRS specificities, tend to reduce the weight of the minimization hypothesis. The iterative selection process might have combined some properties distributed mainly along columns with others concentrated along rows, thereby reaching some demarcation of homogeneous islands, which could resemble the sectors and quadrants. Properties distinctively distributed along columns are, e. g., hydrophobic amino acids concentrated along the central A and hydrophilic along the central U, and small amino acids along the central G while, through rows, main regularities are the concentration of small amino acids and precursors to biosynthesis families along the 3' C, and of the tendency of biosynthesis families to spread along rows. Our approach through the diagonals [39-40], is more easily conductive to the regionalization, now in quadrants, offering a simpler mechanism, less costly in the number of iterations. The space available for distribution of variations was reduced from the beginning, thereby making easier the work of selection. The diagonal path is like following 'the midle way', of the hypothenusa of triangles instead of the cathets. Table 4 depicts a synthesis of the properties of amino acids belonging in sectors and quadrants of the matrix. Each of the properties will be detailed in the correspondent sections. The distinction of sectors stems from the formation of palindromic tRNA pairs, which remain restricted to the realm of each sector (Section 2.11). The complementarity of pDiN, in pairs, is correlated to some divergent or 'opposite' properties of the corresponding amino acids, such as the hydropathy (Section 3.2). Since synthetase classes were organized coherently with the sectors and these are divided into complementary quadrants, each class developed divergent properties, following the complementarity. It is expected, due to the 25-30% overlap between sectors in each class, that the averages of amino acid properties in the classes should be less divergent than the averages of sectors (Table 4). It can be said that the code organization resulted from close interaction between tRNA and Table 4. Regionalization in quadrants. Data on quadrants, sectors and classes are shown as averages (with decimals) or as numbers of occiurences. Highestratiosshown Degeneracy (Table 1) RR (4.0) / any of the central Y quadrants, YY or YR (2.3) = 1.7 Amino acids in aRS classes (Table 2) YY (5 aRS2 /1 aRSl) / YR the central Y quadrants (1 aRS2 / 5 aRSl) Amino acid hydrophilicity (Table 3) YY (0.773) / RY (0.171) = 4.52, the 3' Y quadrants Amino acid side chain volume (Table 3) YR (111.0) / YY (57.8) = 1.92. the central Y quadrants Destabilization of protein N-ends (Table 9) YR (6.8) / RY (2.4) = 2.83. the quadrants of the Mx sector Amino acids characteristic of protein conformations (Table 12) Aperiodic (coils and turns): Ho or aRS2 (5) / Mx or aRSl (0) Helices: YR (3) / RR (1), the 3' R quadrants; also Mx or aRSl (5) / Ho or aRS2 (3) Strands: Mx (6) / Ho (1); also aRSl (5) > aRS2 (2) Nucleic acid-binding amino acids (Table 11) RNA-binding: RR (1.55) / YR (0), the 3' R quadrants; also Ho or aRS2 > Mx or aRSl DNA-binding: RY (0.6) / RR (0), the central R quadrants; also Mx > Ho, but aRS2 > aRSl Binding both equally: Mx or aRSl (5) / Ho or aRS2 (0)

Genetic Code: a Self-Referential and Functional Model

91

synthetases, both strong determinants of the regionalized distribution of amino acids through the quadrants. It is a simplification to state that the code was constructed only by the synthetases since their distribution in the matrix followed tRNA complementarity. The distribution of degeneracy reflects the participation of these two principal components. Quadrants were the divisions most frequently producing higher contrasts (8 of the 11 characters). The rule of preference of aRS classes 1 and 2 for, respectively, the sectors of Mx and Ho pDiN, was generally obeyed, with the exception at the amino acids characteristic of DNA-binding motifs, indicating that this character was associated to the attributions deviant from the rule. 2.6 Variant codes The standard code was once thought universal but the finding of evolutionary variants showed that it is only canonic. Variants did not change significantly the standard (Table 5). They contributed, otherwise, to a general rule saying that the sites where they incided upon were evolutionarily labile and also considered of late introduction. The attribution they lead to is considered the original; the variant would be returning to the primary one. Events 'forcing' the changes were, mainly, the development of genomes with disequilibrated base with high G+C contents as in Micrococcus luteus, or high A+U as in organelles or some organisms such as mycoplasmas [14]. Variations found in eukaryotes occurred mainly in unicellulars (ciliates, yeasts, algae) and incided upon the same sites found, more frequently, in bacteria. Our evaluation showed that the most labile sites are the boxes involved with the punctuation system, indicating its late introduction and the involvement of various protein factors, forming a complex system, with more sites of variation. Another labile site is the CU box, especially the Arg dicodonic attribution. Among all sites, most labile are the 5' Y attributions. 2.7 First settlers of complex boxes The rule of purine precedence, firmed for the distribution of simple and complex boxes, can be extended to the definition of the first occupiers of complex boxes, the ones presently attributed to 5' R triplets. It is indicated the first occupier developed tetracodonic degeneracy, retained the 5' R and conceded 5' Y to new occupiers. This is not intended to mean that the structure with simple boxes-only ever existed. The fifth box entering the code (UC), right after the Gly, Pro and Ser group (Section 3.10), already developed complexity (Asp, Glu). Evidences supporting this conclusion stem from various considerations. (1) In the 3 boxes Table 5. Variant codes. Compiled from [41]. Some changes were not completely clarified as to the destination of the change, only knowing that the cited one was lost, and are not detailed 5'Y(13U,4C,5Y) Specific punctuation (11) CA(4) X->W, C UA(4) X ^ Y , Q , L , A AU(3) I(5'U)<—^M

Type of base affected 5'R(2) Type of box affected CU(7) R -> S, G, X S->G

Others (6) j^AG _^ J

3

S^^^X R^° : 2 losses

92

Progress in Biological Chirality

containing the specific punctuation signs, all aRSl, it is indicated that introduction of the punctuation system was a late event. There are variant codes where Ile^— returns to Met, ^^ to Trp and X'^^^ to Tyr. So, it is accepted that Met is the second occupier of the AU box and fMet the third, Trp the second occupier of the CA box and X the third, and X the second of the UA box. (2) In the other 5 complex boxes, 5' R are always aRS2, considered older than aRSl, the former having conceded 5' Y to the latter. (2a) At formation of the 3 hexacodonic attributions, two class 1 synthetases, LeuRS (primarily NAG) and ArgRS (primarily NCG) expanded their specificities, so that Phe was reduced from NAA to RAA and Ser NGA, NCU to NGA, RCU. So, the dicodonic attributions of Leu and Arg are secondary, their expansion involving 3' ambiguity: Leu 3' R, Arg 3' G or U. Such events should have been late, due to adaptive ajustment to amino acid usage in proteins. (2b) The other 3 complex boxes are central U. It is indicated that the precursor AspRS2 conceded 5' Y to the other precursor GluRSl. The other events followed the biosynthesis derivations: from Glu to GlnRSl, directed to the HisRS2 box, this being of independent origin, and from Asp to AsnRS2 then LysRS, the latter class 2 or 1. For Asn and Gin, there are also the evidences that they can be obtained, in some organisms, through amidation of, respectively. Asp- and Glu-tRNA [4243], the specific synthetases having been developed later, and that GlnRS was developed in some lineages, later being transfected to other. The temporal order of occupation of boxes is summarized: (1) primary settlers are the 20 amino acids and the original attributions of synthetases. It is indicated that Gly preceded Pro in the GG box. The first occupiers of simple boxes retained G, C and U, those of complex boxes retained 5' G or, in He, G and U; their sum is of 15 synthetases (9 class 2 / 6 class 1), PheRS2 being atypical (aRSl-like acylation); (2) additional occupiers of complex boxes correspond to 5' Y. They correspond to the atypical LysRS (class 2 standard, sometimes class 1), 4 class 1 and the secondary attributions: two expansions of aRSl plus the punctuation signs, these also occurring in aRSl boxes. 2.8 RNA replication The activity of RNA replication should precede installation of the coding system. So, it should not be a strong limiting factor, as should be considered the systems of amino acid biosynthesis, requiring generation of new enzymes in metabolic routes, or of generation of new synthetases, by duplication of earlier ones and adaptive variation. The ribosomal peptidyl-transferase activity (ppT; Section 2.12) also precedes installation of the coding system. It is non-specific, accepting any ac-tRNA. Its presence guarantees anabolic direction to the aRS activities. These would, by themselves, be bidirectional and tend to equilibration, subjected to influences of substrate and product proportions. Primitive synthetase activities, before their assumption by peptides, could have been amino acid-binding RNAs [44] or ribozymes, according to prevalent hypotheses and to what has been established for ppT [4546]. Ribosomes would work as ribozymes, their substrates being ac-tRNA [47], and as actRNA sinks, where amino acids are covalently linked in peptides and tRNA returned to the cytosol pools. The ppT activity resides in the RNA of the large ribosomal subunit (LrRNA). The RNA of the small ribosomal subunit (SrRNA) has a 'pDiN' AA (positions 1492 and 1493) which, together with a G (position 530), interact with the minihelix formed by the paired codon and anticodon in the ribosome interior [48-49]. 2.9 tRNA derivation The proposal of tRNA derivation with incorporation of the complementary strands into the

Genetic Code: a Self-Referential and Functional Model

93

code was investigated [50-51] and demonstrated plausible, but they could not offer a structure for the code. We reanalyzed their published data, organizing them according to the palindromic hypothesis for the code structure. It could be indicated that some tRNAs of the Ho, but not of the Mx sector, presented still detectable and significant conservation of complementary similarity (not shown). These data could be additional evidence for our model, adopting a temporal order where tRNA of the Ho sector preceded, in abundance and consequent recruitment into the system, those of the Mx sector. 2.10 Protein synthesis without translation Our model is similar to various others in admitting gradual and successive incorporation of attributions, each one with an amino acid and a restricted tRNA class (5' N-pDiN). Otherwise, an important difference is introduced, of not requiring an mRNA to be translated. In models assuming that the coding system arose to fiilfill the function of mRNA translation [38], at least one of two problems are added: if the mRNA was heterogeneous in base composition, the first attributions should be widely ambiguous or, if specific to certain codons, it is required that those not accepted by the specificities should work as terminators. The possibility that the mRNA was homogeneous in base composition would avoid such problems but is not considered plausible [52]. Our proposal avoids such problems abandoning the hypothesis that the original system was of translation, becoming only one of protein synthesis. The possibility of natural occurrence of an mRNA to be translated is not considered reasonable while there is no way of accomplishing the task. So, the rising system was entirely self-referential at the beginning, constituted only by the pre-existing components; ribosomes capable of receiving the ac-tRNA substrates and presenting the ppT activity; primitive synthetases, also possibly ribozymic; and their substrates tRNAs and amino acids. Such system may seem too complicated to have been original but there are possibilities of simplification, considering indications that tRNA and their derivations could have been at the foundation of all component RNAs. A minimal size for such system might be estimated to be comprized of a few primitive tRNAs. The tRNAs have been proposed as 'primordial genes' [53-54]. They are very similar to the 5S rRNA [55-56]. Primitive tRNAs could have been ribozymes, derived from mini-stem-loops or hairpins [57]. The homologies between tRNA and rRNA [58-60] are spaced with a harmonic distance of 9 bases [61-62], indicating the size of a minimal primordial repeat. After three cycles of duplicative elongation they reach the size of present tRNA: 9 - 1 8 - 3 6 - 7 2 bases, and further cycles that of rRNAs, as derivations from poly-tRNA-like structures. It has also been proposed that poly-tRNA could work as codon chains or primitive mRNA [59]. This proposal has similarities with that of [63], taking tRNA clusters present in actual genomes as relics of ancient coding systems. 2J1 tRNA fishing in pairs Formation and thermal stability of tRNA dimers paired through anticodon loops has been studied as a model for codon-anticodon pairing [13, 64]. Dimer formation has been tested on in vitro protein synthesis systems, with bacterial ribosome preparations and poly-U mRNA [65-66]. Dimers formed in the incubation mixture were inhibitory to translation and their formation was interpreted as possibly regulatory in vivo. When complementary tRNAs are sequestered as dimers, only the type in excess would be available for incorporation into ribosomes. Cells could utilize this mechanism for regulating amounts of protein types produced. Such studies have not been extended to proposals for generation of the code.

94

Progress in Biological Chirality

Doubts still remain about the in vivo occurrence of dimers and, in case they do occur, about the types of dimers relevant to this aspect of physiology. For instance, if it were required a standard pair of bases in the central position, possibilities are reduced to two columns, but 4 perfect pairs can still be formed with the anticodons of each box (Table 6). Which would be the physiologic and meaningful for formation of the code or translation regulation? Pairs of pDiN in each sector are, apparently, complementary in parallel, but represent pairs of antiparallel complementary palindromic triplets, where the external bases are the same. Even considering our restriction that, for formation of the code, meaningfiil were the ones formed by palindromic anticodons, such wide possibilities of pairings, together with our proposal of development of complete degeneracy in the 5' position for the first settler of a split box, where the pairs might have been formed by the first occupiers in the tetracodonic stage or by the second one (Table 7), pose options to be solved through other means. It is thought meaningful that in the four cases presenting this problem, first occupiers of the boxes are always aRS2 and the second occupiers are aRSl in three cases, LysRS being atypical. It is proposed that dimers formed with both anticodons palindromic promoted configurations which were specifically adequate for aminoacylation, ribozymic and enzymic. Synthetase active centers, in their coupling to tRNA, or ribosomal active centers, in their coupling to codon-anticodon pairs, were specific for the type of surfaces or clefts between bases, presented by the palindromic triplet pairs (Table 6, bold). These are composed of triplets with: in the Ho sector, both surfaces planar, homogeneously R and y; in the Mx sector, one with a hill (R) and the other with a valley (y) fitting each other in the central position; in all cases. Table 6. Steric constraints in the formation of palindromic pairs and estimated thermodynamic stability. (A) The bulky purines are presented in capitals, the smaller pyrimidines in lower case. Each column shows the possible perfect pairs formed with the 4 members of a box and members of each of the boxes in the column with the complementary central base. Hexacodonic expansions in parenthesis. (B) Thermodynamic stability of the pahndromic pairs (bold in A). Calculations according to [67], including penalties only for terminal A:U pairs. Terminal A:U or G:U pairs are considered equivalent [68]. These estimates may not be directly applicable to hehces formed with actual anticodons [13]

RRR yyy

(A)

yRR Ryy

(B) ^^

^^'^^ AFT

Phe AAA uuu Lys Phe GAA cuu Glu

Ser AGA ucu (Arg) Ser GGA ecu Gly

Leu AAG uuc Lys Leu GAG cue Glu

Pro AGG ucc (Arg) Pro GGG eee Gly

Val AAc uuG Asn Val GAc cuG Asp

Ala AGc ucG Ser Ala GGc ccG Gly

lie AAu uuA Asn He GAu cuA

Thr AGu ucA Ser Thr GGu ccA

Asp

Gly

(Leu) cAA Guu Gin

Ser cGA Gcu Arg

Leu cAG Guc Ghi

Pro cGG Gcc Arg

Val cAe GuG His

Ala eGe GeG Arg

Met cAu GuA His

Thr cGu GcA Arg

(Leu) uAA Auu X

Ser uGA Acu X

Leu uAG Auc X

Pro uGG Ace Trp

Val uAc AuG Tyr

Ala uGc AcG Cys

He uAu AuA Tyr

Thr uGu AcA Cys

-0.96 -6.20

-3.53 -15.48

-4.43 -22.92

-6.52 -26.78

-4.35 -21.84

-5.78 -25.52

-1.53 -9.63

-3.45 -14.40

RRy yyR

yRy RyR

Genetic Code: a Self-Referential and Functional Model

95

Table 7. Attributions in the pairs of palindromic triplets. In four pairs, choices are possible between the present attribution of the palindromic triplet and the proposed tetracodonicfirstoccupier (bold) of the box Ho sector Mx sector AAAPhe UUU Asn. Lys GAG Leu CUCAsp, Glu

AGASer UCUSer,Arg GGGPro CCCGly

ACA Cys UGUThr GCGArg CGCAla

AUATyr UAU He GUGHis, Gin CAC Val

with a symmetry center and the same bases at the extremities of the triplets. The forbidden structures are between sectors, with two-base plateaus or valleys ending in steep edges. The model of tRNA fishing specific for palindromic triplets was demonstrated for the location of the termination system, determined by the initiation system (Section 3.14). Specific fishing may not incide upon the dimers with highest thermodynamic stability but should have been determined by properties specific to the palindromic configuration or of other participants of the process. It should still be tested for utilization in translational regulation, through codon usage in present organisms, looking for regularities in the types of pairs with high contrasts. Triplet pairs were found between members of boxes containing the palindromic triplets in anticodon sets from all kingdoms. Table 8 presents data on the most reduced sets, from eubacteria and their derived. In Eukaryotes and Archaea, they are more abundant and do not present special novelties (not shown). Most necessary for the search on the simplified anticodon sets, wobbling rules developed for codon-anticodon pairings [41] were adapted for anticodon pairs (not shown): the 3' bases of codons and anticodons were considered equivalent, accepting various types of 5' bases in anticodons. None of the 5' G of the YY quadrant are able to form palindromic pairs with the 3' R of the RR quadrant, except the unusual Gli"^— , of fungal mitochondria. It is significant the introduction of modifications of restricted specificity in He and Met (L and ^C), guaranteeing formation of pairs with Tyr, only not needed in chloroplasts. 2A2 Successive addition of tRNA in pairs With the ingredients detailed above, it is possible to propose a model for the process of formation of the code, with successive incorporation of complementary tRNA (Figure 4). When an ac-tRNA is formed, it meets its pair and forms a dimer. In this situation, the tRNA anticodon works as codon to the ac-tRNA. Complementary tRNA are abundant and promote continuous formation of dimers but these are associated dynamically and more weakly, through hydrogen bridges and base stacking forces, than the amino acids in peptides, covalently linked. So, the ppT activity competes with advantage over the aRS and the tRNA dimers. In the situation of protein synthesis without mRNA, the ribosomal sites acceptor of the incoming tRNA and of the codon-anticodon helix (decoding site) are empty and the dimer can get in. So, dimers are not inhibitory but integral participants of protein synthesis. In mRNA translation, tRNA: ac-tRNA dimers are not capable of entering the ribosome and become inhibitory. At an intermediate stage in formation of the system the ribosomal decoding site acquired greater complexity and became apt to receive codon chains (such as poly-tRNA or the first exogenous mRNA) for translation. It is possible that tRNA or derivatives of them became parts of the decoding site, and dimers inhibitory. Liberation from inhibition could be obtained by aminoacylation of the sequestering tRNA. When both

96

Progress in Biological Chirality

Table 8. Triplet pairs between members of boxes containing the palindromic triplets in simplified anticodes. Modified bases or nucleosides are indicated by * or their symbols: I, inosine; Q, queuosine; L, lysil-C; ^^C, formil-C. Anticode sets from [14]. Eubacteria: Escherichia coli (Ec), Micrococcus luteus (Ml), mycoplasmas (My); chloroplasts (CI); mitochondria of vertebrates (V) and fungi (F) Phe Phe

Asn Lys

Leu Lvs

Ho sector

Ser Ser

GGA

GGA UCU* Ec

UGA UCU C1,F

UGA UCU* My

*UGA UCU* Ec

GAG CUU* Ec GGG

UAG

UAG CUU* My,F

ecu*

ecu

UAG CUU V UGG CCA F UCG CGU F,V

ucu CI

Asp GAG Glu

Pro Gly Arg Ala His Val Gin Val Cys Mx sector

GAA UUU* Ec, My, V

Ser Arg

Leu Leu

GAA UUU C1,V UAA UUU Cl.V

Thr Trp Thr Tyr He Tyr Met

cue Ml

cue Ml

UGG GGG CCC Ec,Ml My, CI, F, V Ec ICG ICG ICG CGC CGU CGU* Ml My Ec GUG GUG QUG CAC CAU CAU* My, CI, F, V Ml Ec UUG CAU C1.V GCA GCA UGU UGU* Ml,My,CLF,V Ec UCA UGU V QUA GUA GUA UAL UAL UAU My Ec CI GUA UA^C V

UCG CGA F

members of dimers are ac-tRNA, they are equally substrates for ppT. Aminoacylation developed specificity for the palindromic pairs, which formed the code, other types of pairs possibly remaining regulatory. The cycle is initiated by the synthetase activity, uniting a tRNA and an amino acid in an ac-tRNA. This can enter the ribosome directly or form a dimer with a complementary tRNA. The ac-tRNA entering the ribosome is united to a previous one, through the ppT activity, forming proteins and liberating tRNA; these are also replicated. Amino acids derive from biosynthesis and from protein degradation. In the absence of mRNA, dimers can enter the ribosome and the complementary tRNA works as a codon for the

Genetic Code: a Self-Referential and Functional Model

97

mRNA dimer

ITT

<

A

ac-tRNA

aRS amino acid

tRNA

protein

A

A

Figure 4. Protein synthesis: tRNA dimers and mRNA translation ac-tRNA. In mRNA translation, inhibition by dimers is overcome by aminoacylation of the complementary tRNA.

3.

The Functional Code

3.1 Entry order A critical step in studies of formation of the code is that of finding a guiding line, like a tip of Ariadne's thread, which could conduct to the entrance/exit from the labyrinth. The matrix is like a 'magic box', small and full. Nucleic acid complementarity suggests various possibilities of internal complementarity. Starting to decipher the construction of the code is finding the one representing physiology. 3.2 Hydropathy order One of the main indicators of order in the code structure is the hydropathy correlation (Figure 5). Hydrophobic amino acids are concentrated in the central A column, the most hydrophobic base, hydrophilic ones in the complementary column, with central U, the most hydrophilic base [69]. Utilization of pDiN, instead of bases [16] improved the understanding of the correlation but that work was developed at a time when determinations of hydropathy of protein amino acid residues were just beginning [70]. The few non-correlated attributions were also of no help for indicating some structure underlying the correlation. Our studies were based on a scale of hydropathy of protein amino acid residues (GMR), averaged and normalized from published individual ones [37, 71] and could indicate a rationale. The outliers formed palindromic pairs of anticodons: Ser ""^^^ Ser and Pro ^'"^^ Gly [72], from now on called the GPS group. The GMR scale differed, on the average, only 9% from the scale normalized from hydropathies of amino acid molecules in solution (GMM) [37, 71] but introduced a few major corrections: protein residues of Pro and Trp depicted increased hidrophilicity, those of Ala and Gly increased hydrophobicity. Highest correlations were

98

Progress in Biological Chirality 1.0

Q O

<

O

z < o

o Q >-

'X,

HYDROPATHY OF DINUCLEOTIDES

Figure 5, Hydropathy correlation. aRS 1, aRS2, outliers circled, * deviantsfromthe aRS class / pDiN rule, (*) hexacodonic expansions. Linear regressions for aRSl (I, M, V, C, W, Y, Q, R) non-deviants and for aRS2 (F, A, T. H, K D, K) non-outiiers. See also [37]

depicted by aRSl attributions to Mx pDiN and by the aRS2 attributions, excluding the outliers (the GPS group). Other aRSl attributions (deviants and hexacodonic expansions) remained inside the global correlation area, of all attributions minus the outliers. 3.3 RNA and RNP worlds Since the significant correlation was obtained with properties of protein amino acid residues, it is indicated that it was established by the protein synthetases, through the amino acid residues in their active centers. It is also possible that amino acid molecules, as substrates inside the enzyme active centers, could already depict residue-like hydropathy behavior, not anymore like those of free molecules in solution. As corollary, it is indicated that the noncorrelated attributions were realized by different mechanisms, non-proteic, be ribozymic or by peptides of special constitution, incapable of producing the correlation. Assuming that the

Genetic Code: a Self-Referential and Functional Model

99

first peptides elaborated in the sequencial process, were constituted only by the GPS group, it could be said, through a kind of 'circular' reasoning, that they were not adequate for producing the correlation, e. g., due to their homogeneous hydroapatheticity. Were the first synthetase activities ribozymic, it is admissible that possible correlations dictated by the RNA world could follow rules different from those of the protein world. The hydropathy correlation refers to amino acids and anticodons but the sites involved are, in present tRNA molecules, distant from one another. It could be proposed that, at some early time in the development of the interactions among aRS, tRNA and amino acids, the anticodon and receptor sites were either physically closer to each other, later becoming separated, or could be placed temporarily close together [6]. The correlation is understood as a requirement for the approximation of the substrates. For productive approximation, both should have coherent hydropathies. When one is hydrophilic, carrying some water molecules with it, the reaction with a hydrophobic one would be hampered. This proposition is only reminiscent of the stereochemical proposals for the origin of the code, of the '2-body' approaches. Alternatively, but with some more complication, the correlation could be mediated by other tRNA identity sites [73-76], which should be related to anticodon hydropathic properties. 3.4 Systemic criteria of order Other sources of information were prospected for materials and processes relevant to understanding the code. Their composition showed the code as a compact picture (a 'hologram') of various cell processes. This search was directed by the hypothesis that the first functions of proteins should have been of RNA stabilization, but encompassing three main attributes: (a) Metabolic stability. The known RNA instability and the stability of some protein types indicate that stabilization of the nucleoprotein system is due to proteins. So, a system should be built becoming stabilized due to properties of the products, (b) Autocatalysis. The scheme of products participating in the system of their own formation becomes one of auto-catalytic, self-referential and self-feeding cycles, (c) Cognitive binding. The following scenario indicates the necessity for a cognitive rationale. Primitive RNAs would compose a wide 'space of variation' of types and functions, some of these participating in the process of protein synthesis. In the same way, proteins synthesized composed another large spectrum. Some were more stable, others capable of RNA binding, or of participation in protein synthesis. Only when these three properties were associated in one special protein class, the system started becoming stabilized. The cognitive aspect was then added, when this class of proteins was adjusted for binding to the same special class of RNA involved with Table 9. The N-end rule of protein stabilization. Data from [78], adapted in grades [79], ftom strong stabilizers (grade 1) to strong destabilizers (grade 9). Tlie Ho sector bold Cys 3 ^ ^ Phe 9 Ser 2 Trp 9 \ X Leu 9 Pro 1 Arg 8 ^^ ^ Vail

Ala^

"" ^ Met 1 fMet

Thr 2

Gly 1

^^ ^ Asn 5 , „ ^y" *

100 Progress in Biological Chirality their synthesis. 3.5 Protein stability and punctuation systems The kind of amino acid at the N-end of proteins is an important component of systems determining their half-lives. Amino acids were classified according to their properties of stabilizing the N-ends and, consequently, the whole proteins, configurating the N-end rule (Table 9). We incorporated these data to the study of the code, starting from the obvious coincidence of fMet with the property of strongly stabilizing the N-ends, and of the stop signs with neighbor amino acids (Tyr, Trp) strongly destabilizing the N-ends. The tRNA of another strong stabilizer (Ala) is also part of the bacterial system for resuming translation by ribosomes stalled on defective mRNAs [77]. The known punctuation system, which we call specific, due to incorporating complex modifications in the tRNA sets and introducing new protein factors, is localized in the Mx sector. Its quadrants are strongly contrasting as to the protein N-end stabilization property (Table 4), in special coincidence with the tips of the Mx axis (AU, UA). We consider such properties primary and idiosyncratic to amino acids, proteolytic systems being developed in accordance with the amino acid properties. The 7 strong stabilizers (grades 1-2) are characteristic of simple boxes, only Met residing in a complex box and becoming involved with specific initiation. Conversely, the 5 strong destabilizers (grades 8-9) are characteristic of complex boxes; two of them (Arg, Leu) became involved with the aRSl hexacodonic expansions. 3.6 Polar distribution of amino acids in proteins In the same way as we considered the N-end rule, derived from original amino acid properties which determined specificities of proteolytic systems, so it was done with the preferential amino acid localization in the heads and tails of proteins. The hypothesis was investigated that proteins, generally and from the beginning, depicted a polar organization, with N-end stabilizing amino acids concentrated in the N-ends, destabilizers pushed to the tails. Such organization should have left imprints in the code organization. In fact, data from [78, 80-81] showed evidence of the polar organization (Table 10). Again, no mention had been made on the relevance of such data for understanding the code organization. It was just said that the two types of data, on polar organization of protein ends and N-end stabilization were mutually compatible. In the Mx sector, involved with specific punctuation, the preferential occurrence in tails of Tyr and Cys, of the boxes with stop codons, were added, and the preferential occurrence of Thr was defined right after Met. The preferred occurrence of Val at the heads and of Arg at tails kept coherence with the rule. 3.7 Non-specific punctuation It became clear, in addition, the participation of the Ho sector, more precisely the Ho axis, in another punctuation system, which we call non-specific, involving only the polar distribution of stabilizing and destabilizing amino acids. It should have preceded the specific system, when the code contained only the Ho sector. The strongest stabilizers of the GPS group (Gly, Pro), at the core of the Ho axis, were preferred in the heads of proteins, and the strong destabilizers (Lys, Phe), at the tips of the axis and activated by the atypical catalysts, were preferred at the tails of proteins. It is indicated that a code containing only the Ho sector was already able to produce proteins with structural and ftinctional organization. The polar distribution of amino acids according to the

Genetic Code: a Self-Referential and Functional Model 101 Table 10. Polar organization of protein sequences. Datafrom[80-81], adapted with elimination of ambiguities. Positions numbered whenever available: from the head or from the tail inwards. The Ho sector in bold Preferred in the head -^ Val Metl

Pro — Thr2

— Gly —

Phel,2 _ ~

Preferred in the tail — Cys2 — ArgL2 "~

~

Tyr — Asn Lysl,2

property of N-end stabilization could have been followed also by concentration of other functions, besides stability, in the heads of sequences. 3.8 RNA binding hy proteins These data were obtained from publications containing sequences of protein domains or motifs specifically involved in binding of various types of RNA [82] or DNA [83]. Sequences were analysed by a standard procedure: sites where 2 amino acids accounted for at least 67% of occurrences were considered conserved; sites internal to these, in the domains or motifs, not reaching this score were considered variable and became the controls. In this way, amino acids more abundant in the conserved than in the variable sites could be identified (Table 11). The 9 RNA-binding domains or motifs were: translation release factor 1 (eRFl; complete sequence) [84]; C5 protein of ribonuclease P (complete sequence); RNP or RRM domain; Kappa H domain (KH) [85]; double stranded RNA-binding domains (PS) [86-87], one of them Pro-rich (DSP) [88]; the only motif highlighted as 'unusual' (Pro-rich) in ribosomal LJLI protein; the characteristic motifs of the aRSl and aRS2 active centers. Of the 9 DNA-binding domains or motifs, 8 are compiled: helix-turn-helix (HIH), zinc fingers (Znf), beta sheets (Beta), helix-loop-helix (HLH) and their basic stretches (HLHb\ leucine zippers (LeuZ) and their basic stretches (LeuZb), homeoboxes (Horn). The hormone receptors are still being examined. The basic stretches are considered apart, due to behaving possibly differently from the other amino acids, e. g., interacting more closely with the backbone than with the nucleotide bases. 3.9 RNP and DNP worlds The most characteristic RNA-binding amino acids were Gly and Pro, and the whole GPS group is present among these. The 4 primary stabilizers of protein N ends belong in the RNAbinders group but the other amino acids characteristic of aperiodic structures (Asp, Asn) are not participants of nucleic acid-binding motifs. Conversely, specific DNA-binding amino acids belong in the set of deviants from the rule of preference of synthetase classes for pDiN sectors. This was the only instance of a constant association between an amino acid property and the character of belonging in the group of deviants. Among the 5 specific DNA-binders, there are the .3 aRS2* (Ala, Thr, His), GluRSl*, sharing the palindromic pair with the mostly RNA-binder LeuRSl*, and CysRSl, sharing the palindromic pair with ThrRS2*. Such association of deviants with the DNA-binding property indicates that introduction of DNA in the genetic system was determinant of the deviant attributions and that the code was completed already in the DNP (deoxyribonucleoprotein) realm. The data also indicate that RNA-binding is a main attribute of amino acids belonging in the Ho sector of the code. It is

102

Progress in Biological Chirality

suggested that this association was the result of specificity of these amino acids for RNA stretches rich in homogeneous clusters. Being this set also rich in protein stabilizing amino acids, the binding would produce stabilized ribonucleoproteins. These RNAs would become enriched among the ensemble of types and the ones most available for recruitment into the code. This would be an additional evidence for precedence of the Ho sector over the Mx one. Detailed investigation of this proposed specificity will presumably be difficult on the present RNP complexes, but could be conducted by modeUng. It should be tested, for instance, the suggestion that poli-Gly would bind more strongly to and become organized by poly-G, polyC or their double helices, than to poly-A or poly-U. 3.10 Stages information of the code Data available allow the proposition of a model for generation of the system in 4 schematic Table 11. Nucleic acid-binding amino acids. N-end stabilization grades from Table 9; characteristic of protein conformations from Table 12; aperiodic, coils and turns. Basic stretches (italics) not counted for the RNADNA ratios RNA

Pro Leu Phe Lys Val Ser Met

RFL RNP. KH, DS, Lll,aRSLaRS2 RFLDSRLlLaRS2 RF1,RNP, KH,L11,C5 RNP, DSP, aRS2 DS. aRSK C5 RNP, KH, DS, C5 aRSl aRSl

Arg

aRS2, C5

He Tyr Gin Trp

KH,aRSl RNP, DSP DS C5

Glu

aRS2

Ala

DSP

His Thr Cys

aRSl

Gly

Asn Asp

aRS Class

Stabilization

Conformation

Ho

2

1

aperiodic

^11 5/3 3/2 3/2 4/3 1/0 1/0

Ho Ho Ho Ho Mx Ho Mx

2 1* 1' 2(1-)

1 2 1

1 9 9 8 1 2 1

aperiodic helix strand helix strand aperiodic heUx 6 Ho / 2 Mx

111

Mx

1

8

hehx

111 111 1/1 1/1

Mx Mx Mx Mx

1 1 1 1

6 9 5 9

1/2

Ho

1*

4

strand strand helix strand 5Mx hehx

1/2

Mx

2*

2

hehx

1/2 0/1 0/1

Mx Mx Mx

2* 2* 1

7 2 3

hehx strand strand lHo/4Mx

0/0 0/0

Ho Ho

2 2

5 4

aperiodic aperiodic 2 Ho

DNA

RNA / , Sector DNA

HTRZnf

111

Beta, HLH Beta, HLH, LeuZ Hom, Znf Horn, Znf + LeuZb Beta, HTH, LeuZ

HLH, Hom + HLHb, LeuZb HLH, HTH HLH, Znf HTH HTH Beta, LeuZ + HLHb Beta, HTH + HLHb, LeuZb HLH, Znf HTH Znf + HLHb

Genetic Code: a Self-Referential and Functional Model

103

stages (Figure 6), starting with attributions where members of the two pairs of boxes are exclusively aRS2 and ending with those where members in the pair are exclusively aRSl. In the more complex intermediate stages, boxes in a pair are occupied by both synthetase classes [89]. Stage 1. The GPS group of amino acids initiate the system. They are homogeneously simple and hydroapathetic, strong stabilizers of protein heads and RNA-binders. It is possible that peptides synthesized at this stage were not sequentially ordered. If they could work as synthetases, their constitution with only those amino acids would not be capable of establishing the hydropathy correlation. Alternatively, the synthetases responsible for these attributions were originally ribozymes, later replaced by the protein synthetases and conserving the correspondences. Anticodons are in the Ho sector, at the core and its continuation along the hemi-columns with central G and C; the whole set ends up rich in G+C [90] and forms 2 paUndromic pairs [72]. Stage 2, The Ho sector is completed and a numerous set of amino acids with varied properties is added. The new attributions occupy the external hemi-columns, again in the direction of G+C rich triplets to the tips of the matrix. It starts with AspRS2, adopting an anticodon probably derived from one of Pro^^ , with a central base mutation, G -> U instead of G:C. This develops complete degeneracy and concedes YUC to GluRSl. A duplication in this class of synthetases adopts the tRNA of the palindromic pair and Leu. Entry of aRSl in this early stage may be related to the functions of tRNA stabilization (Section 3.9) and with primordial DNA entry in the genetic system, to which Glu and Leu are good binders. Next entries are of amino acids derived from Asp and Glu. Asn and Gin may be formed by amidation of Asp- and Glu-tRNA, respectively. The first was soon replaced by AsnRS2 but the second may have remained longer. AsnRS2 developed complete degeneracy and conceded YUU to Lys, the next derivation from Asp. In consequence of the amino acid being bulky and of extreme hydrophilicity, this attribution became atypic: it was adopted by an aRSl in some organisms but the adoption by an aRS2 is considered the standard one. The palindromic pair to Lys^^ was fished and adopted by PheRS2, and these formed the atypical pair of catalysts. Difficulties similar to the one with Lys, the bulkiness and extreme hydrophobicity of Phe, lead PheRS2 to adopt a special conformation, so that aminoacylation became realized upon the 2' position, as is regularly done by aRSl. Strict obedience to amino acid biosynthesis routes leads to the proposition that Stage 1 contained only Gly and Ser. The precursor (Glu) to Pro enters only in Stage 2. Consequently, still greater internal homogeneity is obtained for Stage 1, indicating that both Gly and Ser were initially octacodonic, respectively GG:CC and GA:CU. The amino acid repertoire of Stage 2 allows formation of proteins with complex properties. Polar organization of sequences was already established, with Gly, Pro, and Ser at the heads, Lys and Phe at the tails of peptides, indicating the formation of linearly ordered codon chains (genetic information). Peptide stability and their ability for RNA binding were established in the previous stage, not being limiting from Stage 2 onwards. Stage 3, It is possible that duplications generating the Ala-, His- and ThrRS2 attributions happened earlier, but there is no indication that their palindromic pairs had received attributions until this stage. They are placed here in accordance with their homogeneity of being aRS2* and related to the DNA-binding property of the amino acids. A temporal order in

104

Progress in Biological Chirality

the stage was reached considering thermodynamic stability of the triplets and the succession in biosynthesis derivation of the amino acids. The pair of core boxes is placed at the beginning, with Ala and Arg. The former is considered, by many authors, one of the earliest to enter the code but Arg is considered late, possibly having been preceded by another amino acid. Proposals on the first occupier of the Arg box are many [8]: another basic one, ornithine, somewhat more hydrophobic than Arg due to having an amino-propyl instead of the guanidino-propyl; beta-Ala, nor-Val, nor-Leu, and the acidic alfa-amino-isobutyrate and guanidoacetate. Our data could authorize another suggestion, that Ala was originally octacodonic (GC:CG), in the same way as was done with the other pair of core boxes (Gly:Gly, instead of Gly:Pro). In this way, the list of possible predecessors of Arg becomes simplified and a full ancestral homogeneity in the core of the matrix is obtained. Next comes the pair with Val:His, Gin. The two first amino acids are considered of independent biosynthesis derivation, the last in this category to enter the code. HisRS2 developed full degeneracy and conceded YUG to GlnRSl. This replaced the Gin derivation by amidation of Glu-tRNA. It was developed in some lineages, later transfected to others [43]. The last in the stage is the pair of triplets with the lowest thermodynamic stability, and Thr:Cys, Trp. From this stage onwards, enter the sulflirated amino acids, most of the aromatics, all derived from biosynthesis families and no more hydrophilic ones. Stage 4, The tips of the Mx axis are filled, later getting involved with specific punctuation. Amino acid derivation is through rows but that of tRNAs through the diagonal. The hexacodonic expansions of aRSl are developed. A main 'anomaly' of Arg is of having more codons than the actual usage of the amino acid in present proteins [14]. It may be suggested, on the basis of Ser having been originally octacodonic, that the force leading to the expansion of Arg was ex vacuo, resulting from the primary reduction of Ser^^. In fact, there are variant codes where Arg^^ is dispensed with. For the Leu expansion, mechanisms may have combined reduction of Phe usage and increase in that of Leu. The original Phe and Ser degeneracies were found excessive by the types of proteins effectively used in cells, therefore being reduced and conceding space to other amino acids. This rationale should be tested through examination of trends in codon usage in the different kingdoms. 3.11 Both synthetase classes coupled to a tRNA Another instance of symmetry among components of the code is well known [29]: the active centers of synthetases of different classes bind to the acceptor stem-loop of tRNA in specularly symmetric fashion: class 1 approaches directly the minor groove of the double helix, class 2 the major groove. This explains why class 1 enzymes practice aminoacylation upon the 2' hydroxy 1 of the adenosine of the -CCA end and class 2 on the 3' hydroxyl. It has been suggested [91] that the symmetry could be derived from the two classes deriving fi"om complementary strands of the same genetic stretches, but an advance was the demonstration that enzymes of the two classes could be accommodated simultaneously upon one same tRNA acceptor stem [19-20, 26]. Another mechanism of'union of opposites' was introduced in the process of building the code. Our mechanism of complementary tRNA fishing, carrying amino acids of opposite hydropathies, is amplified, a tRNA dimer being also able to fish and carry two couples of synthetases, each tRNA with the symmetric classes. A model for the formation of the system will then acquire the shape of a multidimensional network, attempting to represent clearly the apparent entanglement of the components.

Genetic Code: a Self-Referential and Functional Model 105 Ho Secto r

Mx Sector

2

1

3

4

2

1

3

3

X Sector

3

3

1

2

4

3

1

2

Stage

tRNA

aRS2

[1]

•CC GG» •GA

Gly Gly? Pro Ser Ser (N)

•UC

Asp (N)

Ho GA* [2]

•UU AA»

[3] Mx

{Asn Asn (N) (R) Lys (Y) atypical Phe (N) atypical

•GC GC» •UG

Ala Ala? His (N)

CA« •GU

Thr

AC«

aRSl

punctuation

— (R) Arg (Y) expansion [4] — (R) Glu (Y) Leu Gin}

— (R) Leu (Y) expansion [4] Arg — (R) Gin (Y) Val Cys(N)(R) Trp(Y)

•AU

He (N) (R,U) Met (C AU)

AU«

Tyr (N)

[4]

— (R) — (C) X (U) — {R,U) — (C) fMet (CAU) — (R) X (Y)

Figure 6. Stages in formation of the code. [ ], stages; • or (), 5' bases; { }, obtained by amidation

3.12 The structural model The rules governing the concomitant binding of the synthetase couples to the acceptor stem of a tRNA, without steric clashes, were detailed [19-20, 26]: there are some compatibilities among components of the 'cognate' groups of the synthetase classes (2a-la, 2b-lb and 2c-lb or Ic), other ones being forbidden. Considering that a primordial function of synthetases would have been of stabilization and protection of tRNA, the concomitant binding of one of each class to the same tRNA would reach maximal effect. This may be called an RNA

106 Progress in Biological Chirality chaperone function, borrowing the term commonly applied to interactions among proteins [11]. They advanced also the proposition that, after the stabilization binding, the 'pre'synthetase peptides developed the catalytic function, each one in the couple following its own differentiating path. If the primordial synthetase function was accomplished by the tRNAs themselves, it was 'locally' transferred to the 'chaperones'. This 'structural' model, based on the structural symmetry of synthetases, was found not compatible with our 'functional' model, based on protein ftinctions of intrinsic stability and composition related to the RNAbinding ability of amino acids. The 'structural' and 'functional' models could also be classified, respectively, in the categories of'monofactorial' and 'systemic' [92]. In plastic systems such as the biochemical, strict obedience or wide applicability of rigid rules should not be expected. Nevertheless, application of rules should be explored to exhaustion with the methodologic purpose of clarifying its scope and limits. In the exploration, acceptance of many 'exceptions to the rules' or ad hoc arguments should also be avoided. Following these lines, the functional model adopts rigidly the succession from Ho to Mx pDiN and in palindromic pairs. The structural model highlights various symmetries. Concomitant binding of the two enzyme classes to a tRNA leads to the proposition that the classes developed 'in concert' with one another, coetaneously. Our model ended up with the opposite suggestion, that VA of the couples of synthetases associated with tRNA pairs were of the same class: GlyRS2/ProRS2, Ser"^^ RS2/ Ser^^ RS2, GluRSl/LeuRSl, LysRS2^^^ /PheRS2^', GlnRSl/ValRSl, IleRSl/TyrRSl. The proposition of the structural model could be adopted for, at most, the three tRNA pairs harboring synthetases of different classes: the AlaRS2a/ArgRSla and ThrRS2a/CysRSla, and possibly also HisRS2aA^alRSla, which form compatible dockings on tRNAs, but not ThrRS2a/TrpRS 1 c. In the seven cases of 5'Y concessions, possible compatibility with the structural model was found in two (AspRS2b/GluRSlb, SerRS2a/ArgRSla). The structural model values strongly the PheRS/TyrRS docking: the atypic conformation of PheRS made it structurally non-specific, accommodating any of the lb or Ic but at the same time, it became the only pair possible for TyrRS. We added an explanation for both atypic PheRS and LysRS, derived from having to accommodate amino acids too bulky for the class 2. In adopting the model of 'concerted' evolution of the different classes, it is proposed that the five progenitors of the groups which developed multiplicity should present ambiguity, posteriorly developing specificity. Such rationale accepts the proposition that the code arose as a system for translation of pre-existing mRNA and adopts a frequent principle in evolutionary studies, of passing from generalist to specialist entities [93], which is not strictly necessary. It is equally possible to generate complexity by successive addition of 'specialist' components to the systems, each one with its proper degree of indeterminacy or variability. A clear conflict between the models refers to the temporal development of the code. The structural model, adopting coetaneity and 'concerted' addition of members of different classes but compatible groups, rejects the proposals of precedence of class 2 over class 1 and says that this does not offer an explanation, leaving to fortuitous coincidences, the symmetry of numbers of elements in the classes and groups (Ribas de Pouplana, Personal Communication, January 02, 2002). We propose that the numerical balance could be derived from the protecting and stabilizing function of the couplings, not necessarily involving the catalytic portion of the structural model. Up to what we could gather from proponents of the structural model (Ribas de Pouplana, Personal Communication, January 15, 2003) no sequence of entry of attributions, with the necessary physiologic correlations, was advanced. Strict obedience to

Genetic Code: a Self-Referential and Functional Model 107 the stmctural model would exclude the possibility of transfers of synthetases between incompatible groups of the classes and between enzymes of the same class. Compatible groups would have to be completed inside themselves, in a way similar to the self-contained sectors that the functional model proposes. Our result is at least partially expected due to the succession model adopted, from Ho to Mx and from aRS2 to aRSl. Generation of such intraclass coherence in the palindromic pairs could have been obtained through the well-known mechanism of the DNP world [94]: genes with intense activity (in the DNP world, transcription) frequently fix duplications which can evolve 'in concert', now meaning intraclass coherence. Incorporation of a synthetase to the code would facilitate the next one being of the same class. The first stages of the code would correspond to high rate of aRS2 duphcations, followed by those of aRSl. In spite of this, some details remain significant, besides the class homogeneity in Stages 1 and 4, further contributing to the distinction of sectors. In the Ho sector: (a) both class 1*, GluRS and LeuRS, belong to a palindromic pair, as well as (b) both atypical attributions, LysRS class 2 or 1 and PheRS2 but aRSl-like, make another palindromic pair. In the Mx sector, there is possible support to the model of tRNA stabilization by coupled synthetase classes: all palindromic pairs in Stage 3 are occupied by synthetases of groups 2a and la, capable of simultaneous docking upon a tRNA. Early entry of aRSl, in Stage 2, adopting Ho pDiN, may have benefited the system due to improved stabilization of tRNA, besides possibly being induced by entry of DNA, to which both Leu and Glu are good binders. The rare LysRS 1 was fixed when, in some organisms, a duplication leading to LysRS2 was not available or it was not adequate to receive such bulky amino acid. Incorporation of tRNA with Mx pDiN to the system is related to DNA-binding amino acids which were still attributed by the three aRS2 deviants. Entry of DNA may have introduced complex sequence mRNA, requiring the Mx pDiN. It can be indicated that aRSl duplications, initiated in Stage 2, were retarded relative to those of aRS2 which became deviants (Ala-, His- and ThrRS). This 'historical' rationale, involving partial asynchrony (retardation) of components, has been presented previously [37]. Higher class coherence could be hypothesized for the primitive steps of Stage 3: (a) the predecessor of Arg could have been AlaRS2, this becoming octacodonic, as indicated above; (b) the location of HisRS2 and CysRSl, or that of the tRNAs in the corresponding pairs, could have been exchanged. It is thought that the proposition in (a) could have chances to be found in variant codes, reflecting a possible situation in a progenote or cenancestor, where Lys would be the only basic amino acid; demonstration of the possibility in (b) will require further investigation. 3.13 Protein conformations Our model indicates (Table 12) evolution of peptides synthesized, along the course of evolution of the code, being composed initially by the aperiodic (coils and turns) portions of proteins, in agreement with an earher proposition [95]. A path of increased complexity was identified: helices were soon added, formed by interactions among neighboring amino acids in a string. Sheets were the last addition, resulting from interactions among distant segments of strings. The model is not in agreement with the early introduction of strands [96] and late of helices [97]. 3.14 Specific punctuation The approach based on pDiN suggested an apparent relationship of inversion between the AU and UA boxes, containing the initiation and one of the stop signs. The investigation

108

Progress in Biological Chirality Table 12. Protein conformations in the stages of formation of the code. In parenthesis: Pro may have entered in Stage 1; Gin may have been obtained by amidation of Glu-tRNA, in Stage 2. Distribution of amino acids in conformations from [98] Amino acid

Stage

Coils and turns

Gly Ser Asp Asn Glu Pro Leu Lys Phe Ala Arg His Gin Val Thr Cys Trp He Met

1

X X X X

2

Helices

Strands

X (1)

X X X X

3

X X X X

(2)

X X X X X

4 X

X

T>T

discarded the hypothesis of inversion but provided an explanation for the relationship among all three boxes containing the specific punctuation signs [79]. 3.14.1 Initiation The fMet pDiN is different from that of Met. The latter, utilized in elongation, maintains the standard configuration (central base - ^ 3 ' , CAU) but the fMet pDiN is slipped, becoming 5' "> central (CAU): the same occurs with the respective codons. Initiation codons are various while the anticodon is unique, configurating a new kind of 'degeneracy'. Alignment of the initiation codons demonstrates the slippage (Figure 7). There are various consequences of the slippage. Translation of mRNA is placed in phase, according to the precise modules of

Initiation

Elongation

Codons

5'

N[UG1

Anticodons

3'

"U [A C\

Codons

5'

NfUG]

G U

N

Anticodons

3'

^U [A C1

C A

N^ N_N

Amino acids

A C

N

^ w

fMet

-

Thr

N N

N

N_J^

N"'

~ N N

N N^

fMet ~ Val ~ Amino acids Figure 7. Initiation pDiN slippage. The second amino acids in the chains shown are either Thr, the preferred in this position (Section 3.7), or Val, the other strong stabihzer of the RY quadrant (Section 3.5), and preferred in the N-ends of proteins (Section 3.6). Standard pDiN underlined, initiation pDiN in brackets. W, wobble position

Genetic Code: a Self-Referential and Functional Model 109

triplets, in a thermodynamically strong way. The inverted' position of the fMet pDiN, recalling the inverted phosphodiester bond in the eukaryotic mRNA cap structure, eliminates the wobble position between the first two triplets. The two pDiN become contiguous, forming a tetranucleotide with all standard base pairs, two of them G:C. In the case of fMet-Thr, the two pDiN form a direct repeat structure, with inverted positions in the two strands; in the case of fMet-Val, the structure is of inverted repeats. Before the introduction of this process, determination of the reading frame would have to rely only upon the correct positioning of tRNA in the ribosomal decoding sites. Initiation is described, instead of the mere entry of fMet, as the bimolecular reaction between the two first ac-tRNAs, with second order kinetics, forming the first peptide bond.

5' 3'

A. Conflicts between the main] initiation codon and anticodons of the YR quadrant Initiation codon A [U G] G A [U G] A [U G] A [U G] A Anticodons of the YR quadrant A C A A c A Cys A C A A C A P P p P p A C G A c G A C G A c G w P P P p A c C Trp A c C A C c A c C P p P P p P A c X A c U A c u A c U u P p w P p P w Tyr

A A

X

A A

Arg

u u

A

u p

G

A

u p

C P U w

A

u

A

u u p

A

G

c

G

c

G

G

c

G

c

C P U w

G

c

A

G

G

G G

c u u

G

c u p

G

G

u p

u

C

u

P U w

A P G w C

p

c

G Gin

u

A

G

G His

u

G

u

A P G w C U A P G w C

p

G

u

U

^Figure 8. Part I

A

u

P w A u P w A U P w A U P w G w G w G w G w G w G w G w G w

A G C U P

C P

A

c p c

G C

p

c p

U P

u

A

w U G w U C w U U w P

A

u

P A P A P A P

w

G w G w G w G w G w G w G w G w

u

w U w U w

A G C P U w

C P

A

c p c

G

p

C P U w

u

A

p

c w

u

G

w

C P u U w w

u w

110

Progress in Biological Chiiality B. Pairings between the initiation anticodon and the stop codons Initiation anticodon [C A] U [C A] U [C A] U Stop codons A G U A G U A G U w P P A A U A A U A A U P P G A u G A u G A U

5'

C. Pairings between the main initiation codon and the stop codons Initiation codon A [U G] A [U G] A [U Stop codons A A G U A G U w w P A A U A A U A P P w G G A U G A U

G] G

U

A

U

A

u

D. Pairings between the initiation anticodon and anticodons of the YR quadrant Initiation anticodon _5;_ [C A] U [C A] U [C A] U Anticodons of the YR quadrant A C N Cys, Trp, X A C N A C N Tyr,X

A

U

N

A

U N

A

U N

N

G

C

N

G

C

N

R p, w Y

G

U

R

G

U

R

P Arg His

G P G P G

c u

P

p

p

G U Y Y /,P Figure 8. Part II. Fishing of anticodons correspondent to stop codons by the initation system. Standard (p) and G:U (w) base pairs; elongation pDiN underlined, initiation (slipped) pDiN in square brackets. The conflictive triplets considered decisive and meaningful (bold) involve a standard base pair ^ ) with the central base of the initiation codon. The second and third columns consider the different dinucleotides of the triplets. In (A), the 5' bases of the Thr (A) and Val (G) codons are added, respectively, to the third and fourth columns Gin

u

G

U

3.14.2 Termination Specific termination is necessary for precise configuration of protein tails. We could demonstrate that the stop signs were generated and localized through fishing by the initiation anticodon, with the slipped pDiN. The mechanism was coherent and superimposed upon the properties of the amino acids inhabiting the termination boxes, of protein N-end destabilization (Section 3.5). The search for relationships or possible interactions between the pDiN of initiation and termination boxes required tests with pairings not only between triplets but also between

Genetic Code: a Self-Referential and Functional Model

111

dinucleotides, and considering the two possibilities of dinucleotide slippage (Figure 8). Pairs meaningful for the physiology of punctuation should necessarily be coherent for the 3 boxes of the system. A single conflictive situation was observed, coherent for both termination boxes and the initiation system: the slipped pDiN of the initiation codon was found to be able to pair with the pDiN of tRNA correspondent to the boxes containing the stop signs, these in the standard configuration, thereby competing with the initiation anticodon (Figure 8.A). The conflict was defined by pairings of doublets of the initiation and termination partners, allowing for G:U base pairs, and requiring formation of a standard base pair in the central position, but could be extended to triplets, incorporating the 5' base of the second codon. The conflict was eliminated and solved with deletion of the tRNA correspondent to stop codons. Other mechanisms intervened in determination of the precise constitution of the punctuation triplets, such as in the choice of 5' C, monospecific in the wobble rules, for Met, fMet and Trp. The specific punctuation system may have extended the polymers. The initiation fMet may have been added to the N-end, instead of being only a modification of Met pre-existing in this position. Stop codons may have been added to the mRNA, instead of their tRNA having been only substituted by the Release Factor 1. In fact, repetitions of stop codons are frequent. 5.15 Genetic and systemic information A 'systemic concept of the gene' and the concept that 'the system defines the gene' were expressed earlier [99]. Nonetheless, even a decade after the original proposal [100], another study [101] was not able to provide complete clearness and precision to the proposal. Now it becomes possible to advance a systemic definition of primordial genes. The scenario constructed in the present work (Section 3.4) is that both RNAs and peptides participating in the primitive protein synthesis system formed wide 'spaces' of variation, with extensive heterogeneity of types. Among the RNAs, some participated more directly in the process; also among peptides, some were m.ore stable, others better RNA-binders, others even participating in the process of their own synthesis. The components becoming relevant for formation of a self-feeding system were mutually identified as genes and products through a cognitive process [102]. A sub-class of peptides combining, simultaneously, the properties of stability, RNA binding and participation in the protein synthesis process was able to bind, cognitively, to the same RNA that participated in their synthesis. At the cognitive meeting, the proteins, 'surprised', 'said' to 'their' R N A - you are my genes! -, and these, also 'surprised', 'said' to them-youaremy life! (Figure 9). In this way, the system became stabilized and selffeeding, auto-catalytic. In the simultaneity of the event, RNA became genes and the peptides recognizing them became products. Definition of genes and products occurs in the process of their mutual recognition, partners defining each other at the same time. The 'chicken or egg?' problem is dissolved. During the construction of the code, genes arose in consequence of the binding promoted by proteins. The event may be located at the time of Stage 2 of the code development, where there is already the possibility of linear ordering in the peptide sequences, dictated by the stability properties, and also of the arisal of genetic 'string memories'. Three stages are delineated in mRNA structuration. The triplet reading-frame is defined by the SrRNA decoding site, together with the anticodon modules. The 'hidden' auxiliary RNY triplet pattern will be examined in a separate study. Finally, the stage of ordering the types oftripletsinthe sequences, initially defined by the properties of protein

112 Progress in Biological Chirality

Figure 9. Systemic definition of gene. When the cognitive meeting occurrecL and RNA, 'surprised', recognized the 'others' to 'themselves' N-end stabilization. Information is defined as the set of interactive properties of elements [103]. In the interactions, information is exchanged or shared. The system, the organized collectivity, gives sense and reality to the information in the elements. While these are dispersed, not organized, their interactive properties may fill the whole lists of possibilities, where all potentials are variably and continuosly shaped and reshaped, unconstrained. When the linear order of protein sequences, which participated in the system, became fixed in genes, thereby guaranteeing a repetitive memory (source), it can be said that a part of the information located in various parts of the system became sedimented in the strings [1]. Other genes, which were added on top of the basic translation process, building a wider system, should be all coherent and participant in refeeding and stabilizing the basic modules. 3.16 Chemistry, history, self-organization and selection Three terms in this title are borrowed from [23], to which we add self-organization. They are discussed together, in order to organize various themes in the text and to introduce the two latter ones. The path described in development of the code obeys some strict rules of chemistry. Firstly, that of tRNA fishing in palindromic pairs, following from the Ho pDiN sector to the Mx sector. Then the rules of precedence of aRS2 over aRS 1, of triplets with higher over those with lower thermodynamic stability, and of central and 5' purines over pyrimidines. Any concessions to other rationalities would have to incorporate many ad hoc arguments. The order resulted adequate in relation to functional criteria. With guaranteed stability of peptides and their binding to RNA, in Stage 1, such properties do not constrain the next stages. Amino acids entering in that order produced satisfactory coherence with the development of protein conformations, fi"om aperiodic structures to helices then strands. Historic events became evident already in Stage 2. Data on Proline biosynthesis were incorporated, introducing a sugestion for improvement of Stage 1. Synthetase specificities were also successfully accommodated in the two pairs of boxes of Stage 2. Selection is an obligatory component of all evolutionary processes. In fact, it was introduced since the beginnings of studies on the organization of the matrix [69]. A main argument in the Darwinian proposition is the observation of adaptafions, in the sense of adequacy between structures and function. It becomes easy but almost trivialized to propose that the mechanism underlying adequacies was the selection of variants, among which the more adaptive, the successful ones in producing function, remained. The observation that the code is organized is usually described by the regionalization, which would result in minimization of mutational damage, or optimization [21, 104-106]. Mutafions, occurring at replication, transcription or

Genetic Code: a Self-Referential and Functional Model

113

translation would, most frequently, in the regionalized matrix, lead to substitution of an amino acid by another similar one. Many statistical tests have been elaborated showing that the standard code is situated among 'the best' in sets composed of various alternative codes artificially generated. Otherwise, such reasoning remains empty if the components of the process were not explicated. As strange as it may seem to selectionists, clarification of the molecular processes involved introduces chemical determinisms, which resuh in reduction of the space available for selection. For instance, the very mechanisms of tRNA fishing and recruitment of synthetases of coherent classes can be better classified as self-organized, as well as those of thermodynamic stability of the triplet pairs. 3,17 Selection focus So much obvious as the operation of selection, it is important to point out the steps where it should be claimed with emphasis. It acts upon the functionality of the system, its 'phenotype'. Most fimdamental should be the types of amino acids composing the code and the proteins derived from it. It can be said that the standard code became partially 'frozen' at some stage, but without loss of evolvability. This is evidenced by the addition of some important monocodonic attributions, best known being those of seleno-cysteine and pyrrolysine [107], and by the evolution of variant codes. Some amino acid substitutions can be proposed in the course of evolution of the standard code, as we did in the case of Gly^^ by Pro and, following earlier propositions, indicating a simpler alternative (Ala) for the predecessor to Arg. The main modus operandi of selection should have been through differential codon and amino acid usage, but its study is of high complexity. At a most basic level, is the process of organization of the linear sequence of codon and amino acid types, which define the specificity of biologic functions. This is called genetic information, a fraction of the information which was distributed in the system and became stored in genes [103]. Our model of sequential addition of amino acids to the system indicates that a polar organization of sequences, with concentration of protein-stabilizing and RNA-binding amino acids in the amino-terminal portions was among the first focus of selection. Further development should have been upon protein conformations. After the aperiodic structures, coinciding with the attributions referred to in the first stage, helices then strands were added. The last are the most complex, with larger and more hydrophobic amino acids, to inhabit the inner microenvironments of proteins. Differential codon usage should also be advocated to explain the allocation of (a) monocodonic attributions with 5' C, the least wobbler of standard bases, exactly to amino acids which became directly involved with initiation (Met) or conceding the neighboring triplet (Trp) to one of the stop signs; (b) concessions of 5' Y to the hexacodonic expansions, Phe to Leu and Ser^^ to Arg. The procedures of sequential addition of attributions and of disposing off of an mRNA to be translated make the process eminently constructive, of assemblage and addition of parts. The constructive process is like the bottomup building of pyramids [4]; the construction starting with the presence of an mRNA from the beginning, to be translated, would be like a top-down building of pyramids. A complex space of codons to be filled [108] was not real from the beginning but only virtual, a universe of possibilities. Our model becomes, then, one of high probability. As corollary, it can be indicated that calculations of probabilities, applied to the organization of the matrix, are not relevant as 'tests' of the hypothesis.

114

Progress in Biological Chirality

4.

Acknowledgments Support from FAPEMIG and CNPq; discussions and help from Andre Ribeiro de Oliveira Cavalcanti, Luiz Jose Delaye Arredondo, Marcello Barbieri and Savio Torres de Farias.

5.

References

[1] R.C. Guimaraes, Life. In: Universes of Knowledge (Eds. R.P. Martins and H. Mari) Fac. Letras UFMG, Belo Horizonte. Brasil 2002. pp. 149-200 (in Portuguese). [2] R.C. Guimaraes, The gene as an RNA molecule. Ciencia e Cultura Supl. 38 (1986) 939-940 (in Portuguese). [3] R.C. Guimaraes. The self-referential genetic code. 13^^ Intemat. Conf. Origin Life, ISSOL, Oaxaca Mexico, 2002, Abstract p. 104. [4] R.C. Guimaraes. Linguistics of biomolecules and the protein-first hypothesis for the origins of cells. J. Biol. Phys. 20 (1994) 193-199. [5] R. Ferreira and K. R. Coutinho, Simulation studies of self-replicating oligoribotides, with a proposal for the transition to a peptide-assisted stage. /. Theor. Biol. 164 (1993) 291-305. [6] J. Lehmann, Physico-chemical constraints connected with the coding properties of the genetic system. J. Theor. Biol. 202 (2000) 129-144. [7] E.N. Trifonov, Consensus temporal order of amino acids and evolution of the triplet code. Gene 261 (2000)139-151. [8] MA. Jimenez-Montano, Protein evolution drives the evolution of the genetic code and vice-versa. BioSystems 54 (1999) 47-64. [9] K. Ikehara, Y. Omori, R. Aral and A. Hirose. A novel theory on the origin of the genetic code: a GNCSNS hypothesis. J. Mol. Evol. 54 (2002) 530-538. [10] J. Ninio, Approches Moleculaires de I'Evolution. Masson, Paris, 1979 (in French). [11] A.M. Poole. D.C. Jeffares and D. Penny. The path from the RNA worid. J. Mol. Evol. 46 (1998) 1-17. (12] E.N. Trifonov and T. Bettecken, Sequence fossils, triplet expansion, and reconstruction of earliest codons. G^we 205 (1997) 1-6. [13] H. Grosjean and C. Houssier, Codon recognition: evaluation of the effects of modified bases in the anticodon loop of tRNA using the temperature-jump relaxation method. In: Chromatography and Modification of Nucleosides (Eds. C. W. Gehrke and K.C.T. Kuo) Elsevier. Amsterdam, 1990, pp. A255A295. [14] S. Osawa, Evolution of the Genetic Code, Oxford Univ. Press, Oxford, UK, 1995. [15] U. Lagerkvist Two out of three: an alternative method of codon reading. Proc. Natl. Acad. Sci. USA 75 (1978) 1759-1762. [16] J.C. Lacey Jr and D. W. Mullins Jr, Experimental studies related to the origin of the genetic code and the process of protein synthesis: a review. Origin Life Evol. Biosphere 13 (1983) 3-42. [17] J.T.F. Wong, A co-evolution theory of the genetic code. Proc. Natl. Acad Sci. USA 72 (1975) 1909-1912. [18] J.T.F. Wong, Evolution of the geneUc code. Microbiol. Sci. 5 (1988) 174-181. [19] L. Ribas de Pouplana and P. Schimmel, Two classes of tRNA synthetases suggested by sterically compatible dockings on tRNA acceptor stem. Cell 104 (2001) 191-193. [20] L. Ribas de Pouplana and P. Schimmel, Aminoacyl-tRNA synthetases: potential markers of genetic code development. Trends Biochem. Sci. 26 (2001) 591-596. [21] A.R.O. Cavalcanti, E.S. Leite, B. Barros Neto and R. Ferreira, On the classes of aminoacyl-tRNA synthetases, amino acids and the genetic code, in press. [22] B.K. Davis, Evolution of the genetic code. Prog. Biophys. Mol. Biol. 72 (1999) 157-243. [23] R.D. Knight, S.J. Freeland and L.F. Landweber, Selection, history and chemistry: the three faces of the genetic code. Trends Biochem. Sci. 24 (1999) 241-247. [24] T.A. Ronneberg, L.F. Landweber and S.J. Freeland, Testing a biosynthetic theory of the genetic code: fact or artifact? Proc. Natl. Acad Sci. USA 97 (2000) 13690-13695. [25] D. Moras. Structural and functional relationships between aminoacyl-tRNA synthetases. Trends Biochem. Sci. 17 (1992) 159-164. [26] P. Schimmel and L. Ribas de Pouplana, Formation of two classes of tRNA synthetases in relation to editing functions and genetic code. Cold Spring Harbor Symposium on Quantitative Biology, in press.

Genetic Code: a Self-Referential and Functional Model

115

[27] M. Ibba, J.L. Bono, PA. Rosa and D. Soil, Archaeal-type lysyl-tRNA synthetase in the Lyme disease spirochete Borrelia burgdorferi Proc. Natl. Acad. Sci. USA 94 (1997) 14383-14388. [28] C. Stathopoulos, T. Li, R. Longman, U.C. Vothnecht, H.D. Becker, M. Ibba and D. Soil, One polypeptide with two aminoacyl-tRNA synthetase activities. Science 287 (2000) 479-482. [29] J.G. Amez and D. Moras, Structural and fimctional considerations of the aminoacylation reaction. Trends Biochem. Sci. 22 (1997) 211-216. [30] G.M. Nagel and R.F. Doolittle, Phylogenetic analysis of the aminoacyl-tRNA synthetases. J. Mol Evol. 40(1995)487-498. [31] C. W. Carter Jr, Cognition, mechanism, and evolutionary relationships in aminoacyl-tRNA synthetases. Ann. Rev. Biochem. 62 (1993) 715-748. [32] J. Cavarelli and D. Moras, Recognition of tRNAs by aminoacyl-tRNA synthetases. FASEBJ. 7 (1993) 79-86. [33] S. Cusack, Aminoacyl-tRNA synthetases. Curr. Opinion Struct. Biol. 1 (1997) 881-889. [34] JR. Brown and W.F. Doolittle, Root of the universal tree of life based on ancient aminoacyl-tRNA synthetase gene duplications. Proc. Natl. Acad. Sci. USA 92 (1995) 2441-2445. [35] G.M. Nagel and R.F. Doolittle, Evolution and relatedness in two aminoacyl-tRNA synthetase families. Proc. Natl. Acad Sci. USA 88 (1991) 8121-8125. [36] R. Grantham, Amino acid difference formula to help explain protein evolution. Science 185 (1974) 862864. [37] R.C. Guimaraes and C.H.C. Moreira, Genetic code structure and evolution - aminoacyl-tRNA synthetases and principal dinucleotides. In: Fundamentals of Life (Eds. G. Palyi, C. Zucchi and L. Caglioti) Elsevier and Accademia Nazionale di Scienze, Lettere ed Arti (Modena), Paris, 2002, pp. 249-276. [38] M. Di Giulio, On the origin of the genetic code, J. Theor. Biol 187 (1997) 573-581. [39] R.C. Guimaraes, Non coding sequences original, genes derived: dinucleotide and hydropathic specificities of the aminoacyl-tRNA synthetase classes. Internal Symp. Soc. Mol Evol, Guanacaste, Costa Rica (1997) 6. [40] R.C. Guimaraes, Genetic code: hydropathic, dinucleotide type and aminoacyl-tRNA synthetase class organization,. In: Exobiology - Matter, Energy and Information in the Origin and Evolution of Life in the Universe (Eds. J. Chela-Flores and F. Rauhn) Kluwer, Dordrecht, 1998, pp. 157-160. [41] R.D. ICnigJit, S. J. Freeland and L.F. Landweber, Rewiring the keyboard: evolvability of the genetic code. Nature Rev. Genel 2 (2001) 49-58. [42] A.W. Cumow, M. Ibba and D. SGU, Transfer RNA-dependent asparagine formation. Nature 382 (1996) 589-590. [43] K.C. Rogers and D. Soil, Divergence of glutamate and glutamine aminoacylation pathways: providing the evolutionary rationale for mischarging. J. Mol Evol 40 (1995) 476-481. [44] M. Yams, Amino acids as RNA ligands: a direct-RNA-template theory for the code's origin. J. Mol Evol. 47(1998)109-117. [45] G.W. Muth, L.O. Donnelly and S.A. Strobel, A single adenosine with a neutral pKa in the ribosomal peptidyl-transferase center. Science 289 (2000) 947-950. [46] J. A. Piccirilli, T.S. McConnel, A.J. Zaug, H.F. NoUer and T.R. Cech, Aminoacyl esterase activity of the tetrahymena ribozyme. Science 256 (1992) 1420-1424. [47] P. Nissen, J. Hansen, N. Ban, P.B. Moore and T.A. Steitz, The structural basis of ribosome activity in peptide bond synthesis. Science 289 (2000) 920-930. [48] A.P. Carter, W.M. demons Jr, D.E. Brodersen, R.J. Morgan-Warren, T. Hartsch, B.T. Wimberly and V. Ramakrishnan, Crystal structure of an initiation factor bound to the 30S ribosomal subunit. Science 291 (2001)498-501. [49] J.M. Ogle, D.E. Brodersen, W.M. demons Jr, M.J. Tarry, A.P. Carter and V. Ramakrishnan, Recognition of cognate tRNA by the 30S ribosomal subunit. Science 292 (2001) 897-902. [50] S.N. Rodin, S. Ohno and A. Rodin, Transfer RNAs with complementary anticodons: could they reflect early evolution of discriminative genetic code adaptors? Proc. Natl Acad. Sci. USA 90 (1993) 4723-4727. [51] S.N. Rodin, S. Ohno and A. Rodin, A On concerted origin of transfer RNAs with complementary anticodons. Origin Life Evol Biosphere 23 (1993) 393-418. [52] R.C. Guimaraes and V. A. Erdmann, Evolution of adenine clustering in 5S ribosomal RNA. Endocytobiol Cell Res. 9(1992)13-45. [53] M. Eigen, The origin of genetic information: viruses as models. Gene 135 (1993) 37-47. [54] M. Eigen and R. Winkler-Oswatitsch, Transfer-RNA, an early gene? Natunviss. 68 (1981) 282-292.

116

Progress in Biological Chirality

[55] R.C. Guimaraes and R.H.R. Ferreira, Evolution of nucleic acid complexity: dinucleotides in the genetic code and 5S rRNA, 72"' Intemat. Conf. Origin Life, ISSOL, San Diego, 1999, Abstract p. 86. [56] MP. Staves and J.C. Lacey Jr, On the probability of a common origin for tRNA and 5S ribosomal RNA. Z Naturforsch. C44 (1989) 296-306. [57] M. Di Giulio, Was it an ancient gene codifying for a hairpin RNA that, by means of direct duplication, gave rise to the primitive tRNA molecule? J. Theor. Biol. Ill (1995) 95-101. [58] DP. Bloch, B. McAithur, R. Widdowson, D. Spector, R.C. Guimaraes and J. Smith, tRNA-rRNA sequence homologies: evidence for a common evolutionary origin? J. A4ol. Evol 19 (1983) 420-428. [59] DP. Bloch, B. McArthur, R. Widdowson, D. Spector, R.C. Guimaraes and J. Smith, tRNA-rRNA sequence homologies: a model for the generation of a common ancestral molecule and prospects for its reconstruction. Origin Life Evol. Biosphere 14 (1984) 571-578. [60] DP. Bloch, B. McArthur, R.C. Guimaraes, J. Smith and MP. Staves, tRNA-rRNA sequence matches from inter- and intraspecies comparisons suggest common origins for the two RNAs. Braz. J. Med. Biol. Res. 22(1989)931-944. [61 ] AD. Nazarea, D.P. Bloch and A.C. Semrau, Detection of a fundamental modular format common to transfer and ribosomal RNAs: second order spectral analysis. Proc. Natl. Acad. Sci. USA 82 (1985) 53375341. [62] Editorial. Basic modular format in tRNA's and rRNA's. Science 229 (1985) 1254. [63] K. Ohnishi. S. Hokari, H. Shutou. M. Ohshima, N. Furuichi and M. Goda, Origin of most primitive mRNAs and genetic codes via interactions between primitive tRNA ribozymes. Genome Informatics 13 (2002)71-81. [64] H. Grosjean, C. Houssier and R. Cedergren, Anticodon-anticodon interactions and tRNA sequence comparison: approaches to codon recognition. In: Structure and Dynamics ofRNA (Eds. P.H. Knippenberg and C.W. Hilbers) Plenum, New York, 1986, pp. 161-174. [65] T. Yamane. D.L. Miller and J.J. Hopfield, JJ 1981 Interaction of elongation factor Tu with the aminoacyltRNA dimer Phe-tRNA:Glu-tRNA. Biochemistry 20 (1981) 449-452. [66] D.L. Miller, T. Yamane and J.J. Hopfield. Effect of tRNA dimer formation on polyphenylalanine biosynthesis. Biochemistry 10 (1981) 5457-5461. [67] T. Xia. J. SantaLucia, M.E. Burkard, R. Kierzek, S.J. Schroeder, X. Jiao, C. Cox and D.H. Turner, Thermodynamic parameters for an expanded nearest-neighbor model for formation of RNA duplexes with Watson-Crick base pairs. Biochemistry 37 (1998) 14719-14735. [68] S.M. Freier, R. Kierzek, J. A. Jaeger. N. Sugimoto, M.H. Caruthers, T. Neilson and D.H. Turner, Improved free-energy parameters for predictions of RNA duplex stability. Proc. Natl. Acad. Sci. USA 83 (1986)9373-9377. [69] C.R. Woese, D.H. Dugre. W.C. Saxinger and AS. Dugre, The molecular basis for the genetic code. Natl. Acad Sci. USA 55 (1%6) 966-974. [70] J. Kvte and R.F. Doolittle. A simple method for displaying the hydropathic character of a protein. J. Mol. Biol. 157 (1982) 105-132. [71] R.C. Guimaraes and C.H.C. Moreira, Average dinucleotide and amino acid residue hydropathy scales fit the genetic code correlafion. Internal Conf Origin Life, ISSOL, San Diego, 12 (1999) 87. [72] R.C. Guimaraes. Anticomplementary order in the genetic coding system. If^ Internal Conf Origin Life, ISSOL, Orleans, 1996, Abstract p. 100. [73] M.E. Saks and J.R. Sampson. Evolution of tRNA recognition systems and tRNA gene sequences, J. Mol. £'vo/. 40(1995)509-518. [74] P. Schinunel, An operational RNA code for amino acids and variations in critical nucleotide sequences in evolution. J. Mol Evol 40 (1995) 531-536. [75] S.N. Rodin, A. Rodin and S. Ohno, The presence of codon-anticodon pairs in the acceptor stems of tRNAs. Proc. Natl Acad Sd. USA 93 (1996) 4537-4542. [76] S.N. Rodin and S. Ohno. Four primordial modes of tRNA-synthetase recognition, determined by the (G,C) operational code. Proc. Natl Acad Sci. USA 94 (1997) 5183-5188. [77] AW. Karzai, ED. Roche and R.T. Sauer, The SsrA-SmpB system for protein tagging, directed degradation and ribosome rescue. Nature Strucl Biol 1 (2000) 449-455. [78] A. Varshavsky, The N-end rule: functions, mysteries, uses. Proc. Natl Acad. Sci. USA 93 (1996) 1214212149. [79] R.C. GuimarSes, Two punctuation systems in the genetic code. In: First Steps in the Origin of Life in the Universe (Eds. J. Chela-Flores, T. Owen and F. Raulin) Kluwer, Dordrecht, 2001, pp. 91-94.

Genetic Code: a Self-Referential and Functional Model

117

[80] I.N. Berezovsky, G.T. Kilosanidze, V.G. Tumanyan and L. Kisselev, COOH-terminal decamers in proteins are non-random. FEES Lett. 404 (1997) 140-142. [81] I.N. Berezovsky, G.T. Kilosanidze, V.G. Tumanyan and L. Kisselev, Amino acid composition of protein termini are biased in different maimers. Protein Engin. 12 (1999) 23-30. [82] K. Nagai and I.W. Mattaj, RNA-Protein Interactions, IRL/Oxford University Press, Oxford, UK, 1996. [83] C O . Pabo and R.T. Sauer, Transcription factors: structural families and principles of DNA recognition. Ann. Rev. Biochem. 61 (1992) 1053-1095. [84] H. Song, P. Mugnier, A.K. Das, H.M. Webb,DR. Evans,M.F. Tuite,B.A. HemmingsandD. Barford, The crystal structure of human eukaryotic release factor eRFl - mechanism of stop codon recognition and peptidyl-tRNA hydrolysis. Cell 100 (2000) 311-321. [85] H. A. Lewis, K. Musurunu, K.B. Jensen, C. Edo, H. Chen, R.B. Darnell and S.K. Burley, Sequencespecific RNA binding by a Nova KH domain: implications for paraneoplastic disease and the fragile X syndrome. Cell 100 (2000) 323-332. [86] A. Ramos, S. Grunert, J. Adams, DR. Micklem,MR. Proctor, S. Freund, M. Byeroft, D, StJohnstonand G. Varani, RNA recognition by a Staufen double-stranded RNA binding domain. EMBOJ. 19 (2000) 997-1009. [87] J.M. Ryter and S.C. Schultz, Molecular basis of double-stranded RNA-protein interactions: structure of a double-stranded RNA binding domain complex with double-stranded RNA. EMBO J. 17 (1998) 75057513. [88] D.R. Micklem, J. Adams, S. Griinert and D. StJohnston, Distinct roles of two conserved Staufen domains in 'oskar" mRNA localization and translation. EMBO J. 19 (2000) 1366-1377. [89] R.C. Guimaraes, C.H.C. Moreira and R.C. Melo, The functional genetic code. Origin Life Evol. Biosphere 32 (2002) 464-465. [90] R. Ferreira and A.R.O. Cavalcanti, Vestiges of early molecular processes leading to the genetic code. Origin Life Evol. Biosphere 11 (1997) 397-403. [91] S.N, Rodin and S. Ohno, Two types of aminoacyl-tRNA synthetases could be originally encoded by complementary strands of the same nucleic acid. Origin Life Evol. Biosphere IS (1995) 565-589. [92] R.C. Guimaraes, A systemic manifesto. Intemat. Soc. Study Origin Life Newsletter 22 (1995) 12. [93] H. Takagi, K. Kaneko and T. Yomo, Evolution of genetic codes through isologous diversification of cellular states. Artificial Life 6 (2000) 283-305. [94] W.H. Li, Molecular Evolution, Sinauer, Sunderland MA, USA, 1997. [95] J. Jurka and T.F. Smith, p-tum-driven early evolution: the genetic code and biosynthetic pathways. /. Mol Evol. 25 (1987) 15-19. [96] M. Di Giulio, The p-sheets of proteins, the biosynthetic relationships between amino acids, and the origin of the genetic code. Origin Life Evol. Biosphere 26 (1996) 589-609. [97] M.L. Chiusano, F. Alvarez-Vahn, M. Di Giulio, G. D'Onofrio, G. Ammirato, G. Colonna and G. Bemardi, Second codon positions of genes and the secondary strucmres of proteins. Relationships and implications for the origin of the genetic code. Gene 261 (2000) 63-69. [98] T.E. Creighton, Proteins: Structures and Molecular Properties. Freeman, New York, 1993. [99] M.I.M.C. Pardini and R.C. Guimaraes, A systemic concept of the gene. Gen. Mol. Biol. 15 (1992) 713721. [100] M.I.M.C. Pardini, yf Systemic-Functional Concept of the Gene, MSc Thesis, Instituto de Biociencias, UNESP, Botucatu, SP, 1989 (in Portuguese). [101] R.C. Guimaraes and C.H.C. Moreira, The systemic concept of th$ gene - a decade after. In: Self Organization - Interdisciplinary Studies 2 (Eds. I.M.L. D'Ottaviano and M.E.Q. Gonzalez) Cole^So CLE 30, Ed. UNICAMP, Campinas, Sao Paulo, Brazil, 2000, pp. 249-280 (in Portuguese). [102] R.C. Guimaraes, An evolutionary definition of life - from metabolism to the genetic code. In: Fundamentals of Life (Eds. G. Palyi, C. Zucchi and L. CagUoti) Elsevier and Accademia Nazionale di Scienze, Lettere ed Arti (Modena), Paris, 2002, pp. 95-108. [103] R.C. Guimaraes, The concept of information in biology, In: Meeting the Cognitive Sciences 2 (Eds. M.E.Q. Gonzalez and M.C. Broens) EdUNESP, Mariha SP, 1998, pp. 69-93 (in Portuguese). [ 104] M. Di Giulio, The extension reached by the minimization of the polarity distances during the evolution of the genetic code. J. Mol. EvoL 29 (1989) 288-293. [105] M. Di Giulio, M. Rosaria-Capobianco and M. Medugno, On the optimization of the physicochemical distances between amino acids in the evolution of the genetic code. /. Theor. Biol. 168 (1994) 43-51.

118

Progress in Biological Chirality

[106] A.R.O. Cavalcanti, B. Barros Neto and R. Ferreira, On the classes of aminoacyl-tRNA synthetases and the error minimization in the genetic code. J. Theor. Biol 204 (2000) 15-20. [107] J.F. Atkins and R. Gesteland, The 22nd amino acid. Science 296 (2002) 1409-1410. [108] J.E.M. Homos and Y.M.M. Homos, Algebraic model for the evolution of the genetic code. Phys. Rev. Z,^//. 71(1993)4401-4404.

Progress in Biological Chirality G. Palyi, C. Zucchi and L. Caglioti (Editors) © 2004 Elsevier Ltd. All rights reserved.

Chapter 9 Specific Symmetry of Living Systems Victor A. Gusev Sobolev Institute ofMathematics, Siberia Division Russian Academy of Science, 4 Koptyuga str., Novosibirsk 630090, Russia vgus@math. nsc. ru

1.

Introduction At the microscopic scales the morphological grace of living objects is transformed into formless heap. On the other hand, nonliving crystals on micro and macro scales have graceful geometric shapes. There are many myths created by people about effects of minerals on the fate and health of the owners. Are living systems deprived of this magic feature? Our aim is to find where and what kind of symmetry presents into living systems. And for this purpose, i.e. to discover specific living symmetry, we will examine life from the point of view non-biology but mathematics and physics. 2.

About Fibonacci Numbers, Golden Section and Five-Symmetry It is common knowledge [1] that most tree's twigs are arranged in a spiral and the same applies to cone's scales of conifers and others plants. The ratios object's and spiral's turns are: 2/1, 3/1, 5/2, 8/3, 13/5, 21/8, 34/13, 55/21... (Ser. 1). All these are numbers of Fibonacci series: 0, 1, 1, 2, 3, 5, 8, 13, 21, 34, 55, ..., mn, (mn+mn-i), ... (Ser. 2). As it is well known, m , \ + \l5 . lim—^=^:= 1.6180339..., where q = On the other hand, q is the coefficient of self-similarity in the golden section defined by the equation of the solution (f = q -^ 1. The members of Ser. 1 can be described by — ^ = —ii__«d_ « (^r^ = ^ +1. We call attention to Fig. 1. If we divide the big pentagon as related as L// = ^ + 1. We call attention to the fact that the angle between two nearest corners of the big and small pentagons (Fig. 1) is equal to 27i/10. After 10 moves the initial pentagon makes a full rotation. The same number of nucleotides and molecules of deoxyriboses baring to one turn of DNA helix. Recall that deoxyribose contains pentoses. Nonliving natural crystals have axial symmetry by second, third, fourth and sixth, but not fifth order. Living systems have in addition fifth ordinal symmetry. Why and what purpose?

120

Progress in Biological Chirality

Figure 1. Schematic illustration of pentagonal symmetry, Fft)onacci numbers and golden section

3.

The Symmetries of the Genetic Code To look for specific symmetry in the genetic code we must use methods of discrete mathematics - arithmetic, algebra and matrices*. The basic idea, which gave impetus to research of genetic code by group-theoretic method was formulated by Rumer [2]". He discovered that purine-pyrimidine inversion C<->A, G^U(T) (Rumer transformation) transforms two octet groups of standard genetic code list to each other. The essence of grouptheoretical method is to search for symmetry and regularities in the structure of the genetic code after its matrix mapping [3-12]. Notice that search of symmetries and integer characteristics are not defined by aesthetic prejudices of investigators. As well known, presence of any symmetry of structure or process indicates presence of selection rules or the rules of exclusion [13]. In other words, the presence of symmetry indicates presence of conservation law, but its physical sense is not always clear. Papers [4, 5] introduced vector notation for nucleotides as column vector.

and congruent row vector:

v^

=(CGUA)*

The exterior multiplication of these vectors gives the doublet-matrix:

^CC CG CU CA ^ GC GG GU GA UC UG UU UA I ^AC AG AU AA j

(1)

In the next step the authors [4, 5] define the nucleotide's "power" d as dc=4, da=3, dA=2, du=l, and doublet-nucleotide power as dij=di+dj, where /,y—»• A, U, G, C. The matrix (1) can be represented in digital form after the definition:

7654 6543 V5432y

(2)

The main properties of the matrix: 'the symmetry of this matrix is so high that it is singular, ' "Now, concerning the interpretation(s) of this new kind of small numbers coincidences, we have proposed some personal thought coming from a theoretical physicist which could be, perhaps, one of the theoretical physicists undergoing the great migration towards Biology, Erwin Schrodinger announced a long time ago " [9]. " Danckwerts and Neubert [3] rediscovered this purine-pyrimidine symmetry of genetic code in 1975. "" This definition of vectors clearly views the analogy with matrix notation of quantum mechanics of Dirac.

Specific Symmetry of Living Systems

121

i.e. its determinant is equal to zero detD^O, the rank and defect are equal to two. This follows from exterior multiplication (1). It is remarkable that spur trD=20. This value is equal to the sum of numbers in the side-diagonal. Notice that in the side-diagonal and in parallel lines are ^'power-equaT' doublets. The inversion of matrix around the side-diagonal, amounts to a purine-pyrimidine inversion, no exchanges matrix ofproperty'' [4]. We do not give here all symmetries and digital regularities of genetic code are obtained in papers [4, 5], only cite the author's conclusion: ''Thus, algebraic method provides insight into root of genetic code. Symmetry into genetic code is manifested in selection of triplets for relevant amino acids. The model is constructed on the base of differ power of nucleotides C, G, U, A unique determines of relevant their amino acids'\ The author's conclusion is great argument that genetic code is no frozen accident. Really, it difficult to image a "nature plan" providing genetic code of symmetry [12]. ''The concept of nucleotides' power-determination introduced here allows imaging the genetic texts as sequence number from 1 to 4 {really in quaternary number system)"". [4], In refs. [8-11] the author analyzes the structure of genetic code by group-theoretical method. He analyzes genetic code as chemical structure of nucleotides by this method. One of the goals is finding connection between chemical contents of nucleotides and structure of genetic code. We do not analyze all the interesting resuhs about genetic code symmetry represented by Negadi, citing only one. The total sum of the matrix pattern elements transformed with each other after purine-pyrimidine transformation is characterized by numbers 44, 128, 84 and others. The author notices that the same "magic" numbers (44, 128, 84) are presented in modem physical theory of super-gravitation. May it be coincidence or chance? But in general terms the physical magic numbers of quantum mechanics, for example, magic numbers of nucleons in the nucleus and numbers of electrons in molecules, correspond to stable states. We can assume that the coincidence of biological and physical magic numbers is nature's prompt as this genetic code must be originated in our Universe but not in an other. The obvious conclusion springs to mind: The genetic code is not a product of chemical and prebiotic evolutions'"" but of the chemical structure of 4 nucleotides, 20 canonical amino acids and mapping the triplets to amino acids are determined by primary particles and chemical elements of their own origin in the observable Universe. This point of view of the genetic code might be formulated as a fundamental law of nature. If so, the structure of genetic code obtains the same status of noninterpretation as the physical laws of conservation of energy, linear momentum and angular momentum. The three last laws are implication of time's symmetry and space's homogeneity and isotropy, according to Noether's theorem [15]. The time, space, energy, linear momentum and angular momentum are degree values. On the other hand, all of symmetries and arithmetic and algebraic curiosities discovered in the genetic code are pure abstract numbers. We can propose that to understand the sense of Life and its attribute the genetic code and how they originated, it is necessary to find the symmetries connecting the genetic code with dimensionless properties of the Universe.

4.

About Chirality As it is well known, natural non-biological molecules are forming crystal structures as a

' From an other concept, but to the same conclusion reached by M. Eigen [13].

122

Progress in Biological Chirality

racemic mixture. All stmctural elements are involved in the process of transforming the genetic information presented by chiral molecules D sugars and L amino acids. In such a manner, in living systems the principal equivalence of right and left is broken. In other words, in living systems one of the critically important laws of nature is broken. All the biochemical processes are time-unidirectional - "from the cradle to grave, but no inverse" by figurative expression. In other words, time flies only one way in living systems. If so, the enzymes' geometric structure must be chiral, right or left, based on the fundamental physical CPTtheorem [16]. Enzymes direct all of biochemical processes. Hence it follows that DNA and RNA must be chiral too. The choice of right or left chirality in this case is predetermined by the enzymes' chirality. In order that enzymes make their key biochemical fianctions for muhiplication of living systems - replication, transcription and translation - they must be combined with DNA and RNA and move along the polymers. A necessary condition can be realized if enzymes and DNA/RNA polymers have the same chirality. By analogy: the nut (peptide) of right thread might be screwed on the right but not on the left screw (DNA/RNA). Thus, chirality of peptides and DNA/RNA and its monomer amino acids and nucleotides is determined by time-unidirectional biochemical processes. The requirement of time-unidirectionality i.e. of cause-and-effect self-coordination in biochemical processes in the living systems produces the need for structural elements, informational molecules (DNA, RNA and peptides) must consist of atoms, which allow the easy formation of chiral polymers [16]. It follows that the choice of carbon as a base of life on the Earth is determined too.

5.

Conclusion Apparently, the life phenomenon and its origin can not be adequately realized merely by biological categories. In other words, fundamental physical laws in observed Universe define not only the structure of micro and macro world and evolution of a nonliving matter, but also control the living matter by the determination of genetic code uniformity and monochirality of informational polymers. Only with the presence of living systems {Homo sapiens as a final step of biological evolution) can the Universe be observed - it is a single and principle criteria for its physical reality. Physical laws determine the origin of Life in our Universe. The genetic code's curiosities described above do not appear as a chance from this point of view. On the contrary, they force scientists to look for deep connection between laws of living and nonliving nature. '7« fact, argument is that: would be understood what is the Universe without before understanding what is the Life? " [17]. To summarize: 1. The biological objects have two global specific types symmetry - structural and functional. 2. The structural symmetry subdivided into micro and macro symmetries. 3. The macro symmetry represented by Fibonacci numbers, golden section and fivesymmetry. 4. The micro symmetry represented by monochirality of informational molecules and genetic code's symmetry. 5. The functional symmetry determined by cause-and-effect connections and timeunidirectional biochemical process.

Specific Symmetry of Living Systems

123

6. The time-unidirectional and monochirality symmetries of DNA/RNA and peptides are linked together and determined by global space-time symmetry of our Universe.

6.

Epilogue Our analyses of specific biological symmetries would be incomplete if the author does not express his own point of view on these problems. On the basis of discussion of the life phenomena and its attributes, the genetic code may be considered as a fundamental law in the same way as laws of conservation of energy, linear momentum and angular momentum. Only under this condition we can have a little hope to repeat the natural experiment, i.e. to construct in vitro something similar to a cell of microorganism. Otherwise the chance of success is near zero.

7,

Acknowledgement

The Siberia Division of RAS fimd of interdisciplinary research, project N142, 2003, Russia is acknowledged for financial support.

8. [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17]

References V.A. Kizel, Physical Causes Dissymetry of Living Systems, Science, Moscow, 1985 (in Russian). Yu.B. Rumer, Dokl. Akad Nauk SSSR 21 (1966) 1393-1394 (in Russian). H.J. Danckwerts and D. Neubert, J. Mot. Evol. 5 (1975) 327-332. D. Duplij and S. DupUj, Biophys. Bull. Kharkov Univ. 488 (2000) 60 (in Russian). D. Duplij and S. Duplij, Biophys. Bull. Kharkov Univ. 497 (2000) 1 (in Russian). M.A. Jimenez-Montano, C.R. de la Mora-Basanez and T. Poschel, BioSystems 39 (1996) 117-125. M.A. Jimenez-Montano, BioSystems 54 (1999) 47-64. T. Negadi, PREPRINTLPTO/negadi/NoDS101/J4/No01/2002. T. Negadi 2 / * Meeting of the International Colloquium on Group Theoretical Methods in Physics, Paris, July 15-20, 2002. T. Negadi, PREPRINTLPTO/negadi/NoD3}01/14/No02/2003. T. Negadi, Int J. Quant Chem. 94 (2003) 65-74. V.I. Shcherbak, BioSystems 70 (2003) 187-209. J.P. Elliott and P.G. Dawber, Eds., Symmetry in Physics, Macmillan Press Ltd, London, 1979. A. Babloyantz, Ed., Molecules, Dynamics, and Life, John Wiley & Sons, New York, Chichester, Bribaa Toronto, Singapore, 1986. L. Landau and E. Lifshitz, Eds., Course of Theoretical Physics, vol. 3, Science, Moscow, 1990 (in Russian). V. Gusev, Fundamentals of Life (Eds. G. Palyi, C. Zucchi and L. Caglioti) Elsevier and Accademia Nazionale di Scienze, Lettere ed Arti (Modena), Paris, 2002, pp. 545-552. A.D. Linde, Ed. Elementary Particle Physics and Inflationary Cosmology, Science, Moscow, 1990 (in Russian).

Progress in Biological Chirality G. Palyi, C. Zucchi and L. Caglioti (Editors) © 2004 Elsevier Ltd. All rights reserved.

Chapter 10 Origin of Biomolecules - Origin of Homochirality C. Hajdu* and L. Kesztheiyi** "^Department of Inorganic and Analytical Chemistry, University of Debrecen, Debrecen 40 JO, Hungary hajducs@tigris. kite, hu ** Institute of Biophysics, Biological Research Centre, Szeged H 6701, Hungary kl@nucleus. szbk u-szeged. hu

1.

Introduction The problem, how life originated on Earth and probably elsewhere in the Universe, has been and is still a fundamental question [1]. Answers are numerous starting from religious scriptures to attempts based on scientific information. Two trends dominated the thinking about origin of life until the 19^^ Century. The Platonic trend, dead matter was made living by some spiritual principle and the Democritian trend, spontaneous generation of life from dead matter. The Democritian trend in its simple terms, however, had to be discarded after the famous experiments of Spallanzani and Pasteur that demonstrated that life from non-living entities did not emerge. As we will see later presently a more sophisticated combination of the two trends might be the answer. In 1848 Pasteur discovered a phenomenon intimately connected with the living organisms what he named "dissimilarity" and we call it after 175 years homochirality of biomolecules. Homochirality of biomolecules means, in simple terms, that only L-amino acids and D-sugars occur in living systems with negligible exceptions though L- and D-forms (enantiomers) of these molecules are synthesized in equal number if one starts from achiral precursors. Pasteur's works opened up the way to investigate how did the complex, information-rich molecules that govern the complicated processes of life, arise from the inorganic matter. Major steps were made in 1920-ies by Oparin and Haldane developing the concept of "primordial soup" that contains large quantity of preformed building blocks of the molecules playing role in life. They assumed that the building blocks have been synthesized from inorganic matter directed by the laws of physics and chemistry. Oparin stated that the origin of life is an integral part of the evolution of Universe [2]. As science developed in the 20* Century his note became more and more authentic. It is now well accepted that to understand the origin of life the facts about the Universe, the laws that direct its evolution, the events occurring in it, all have to be included in the search for the origin of life. Conservative estimates date the appearance of the first living organisms on Earth to 3.8-3.5 billion years ago [3], This event was surely preceded by the accumulation of their building

126

Progress in Biological Chiiality

blocks, the biomolecules (amino acids, carbohydrates, lipids and heterocyclic compounds [4]) in substantial quantity, as conceived by Oparin and Haldane. It is generally accepted that homochirality of chiral biomolecules (i. e. L-amino acids and D-sugars) is and has been a uniqueness of life. Theoretical arguments suggest that the homochirality preceded life,, i. e. it emerged during the accumulation of the biomolecules, though observational proofs do not exist. In this paper we try to follow how the Universe developed from Big Bang, when immense energies were acting, to reach such stage of evolution when low energy processes could synthesize molecules. The environments for syntheses are the interstellar space or our habitat, the planet Earth or, possibly, other planets. Thus, first we present a timeline of the evolution of Universe, Solar System and Earth then we survey the recent ideas about the origin of biomolecules and homochirality. A large number of experiments demonstrate that the biomolecules could have been synthesized on Earth. As the space technology advanced more and more research and imitation experiments were devoted to show that they might have arrived from outside sources. Connecting the origin of homochirality to the origin of biomolecules the following possibilities arise: 1) biomolecules and homochirality were established on Earth, 2) biomolecules came from space, homochirality evolved on Earth, or 3) both came from space.

2.

Timeline of the Evolution of Universe, Solar System and Earth Physicists, cosmologists formulated a rather real theory on the time course of the Universe from Big Bang (about 12-15 billion years ago) to present sketched in Fig. 1 [5]. The time region and the assigned events are well founded with theories and observations regarding the elemental composition of the known matter. Uncertainties exist for the inflation period (earlier than lO'^s from Big Bang) and also from the calculations that the hitherto identified matter is not enough to understand the gravitation effects in the Universe. The missing matter (dark matter) and energy (dark energy) amount to about 95 % of what is necessary to keep the Universe in its present expanding state [6]. Great effort is devoted to search of this dark matter. We do not know whether the findings will help to understand life. In Fig. 1 an arrow shows the time of the origin of our Solar System 5 billion years ago. It 10-15x10 year \

modern galaxies

/~

--SxiO'yr 10%ear

\

early galaxies \

/

star formation \

/

first atoms \ \

synth. of nuclei

/ /

3'yr 3 min

/

\ protons, / \ neutrons/

10"'s lO^'s

\ Big Bang J L

/ inflation Planck age

10-^^s

Figure 1. Timeline of the evolution of the Universe from Big Bang to present Arrow signs the formation of the Solar System at 5 billion years ago

Origin of Biomolecules - Origin of Homochirality

127

1 -5

-4.6

t

1

1

1 Origin of i Solar System Earth

-4

-3.8

-3|5

1

1

•!

1

1

I'

<:>

0

xlO'yr

1

1

Solid Earth • First organic I First cells molecules 1 -^chemical <—k-> biological evolution 1 Meteorites Space molecules

Figure 2. Timeline of the evolution of Solar System, Earth and life happened in the period when the Universe already contained modem galaxies like ours. Fig. 2 follows the origin of Earth, 4.6 billion years ago, its evolution and the timeline of the origin of life and its evolution to present. The organic molecules appeared within about 200 million years after the solidification of the Earth crest (4 billion years ago) and accumulated for additional 300 million years when living cells already existed [3]. This period, the chemical evolution, was followed with the biological evolution that is still going on. Zero year, the present time is characterized with new developments, discovering biomolecules in meteorites and also in space. We restrict our review to chemical evolution: try to summarize the information on the origin of biomolecules and the homochirality.

3. Origin of Biomolecules 3. J Origin on Earth Surveying the research of the origin of biomolecules the following strategy may be recognized: 1) selecting possible environment; 2) finding energy source, similar to that existing in Nature; 3) choosing a procedure how to imitate the synthesis of compounds; 4) performing reliable experiments. In the first successful experiments it was assumed that H2, H2O, CO2, N2 and CH4 were present in the Earth atmosphere as starting materials for synthesis of the first organic molecules [7]. A very important feature of this environment was that in contrast to the present it did not contain oxygen. For energy source and procedure electric discharge through the mixture of the above gases was chosen. The electric discharge imitated the lightning. Surprisingly, the product contained different amino acids. Following this pioneering study solar light [8], X-ray [9], P-[10] and y-radiation [11] and heat [12] were applied. All these energy sources were surely present on Earth 4 billion years ago. As a result of these experiments formation of organic molecules, even with complicated structure, amino acids, carbohydrates, lipids and heterocyclic compounds [4], was detected.

128

Progress in Biological Chirality

In the late 1970-ies more and more information accumulated that the Earth atmosphere was not reducing but oxidizing containing mainly CO, CO2, N2, H2O gases [13]. In such environment the above processes could not occur. Kobayashi's group irradiated such gas mixture by protons and other high-energy particles imitating cosmic radiation that surely bombarded the Earth. They found amino acids and other biomolecules in the products [14]. The late 1970-ies revealed other important possible environment on Earth: the deep see hydrothermal vents [15]. Wachtershauser elaborated a detailed theory how pyrite, NiS, FeS compounds, abundant at such places could catalyze the production of biologically important molecules in presence of heat energy. Experiments confirmed the ideas: even dimers of amino acids were synthesized under circumstances imitating the deep see hydrothermal vents [16]. These and other experiments [17] emphasize the importance of this terrestrial environment for syntheses in presence of high heat source and catalysts. 3.2 Biomolecules from space Oro in a seminal paper from 1961 pointed out that biomolecules could have arrived from space [18]. Since that time more and more research is devoted to collect information not only about biomolecules but even living organisms on planet Mars and moons Europa and Titan not mentioning the program of Search for Extraterrestrial Intelligence (SETI). The Solar System could have passed through interstellar clouds, the early Earth was bombarded by comets and meteorites. As we now know (see Fig. 2) all these objects contain biologically important molecules [19-21]. Infrared spectroscopy found about hundred molecules in these clouds. Similar molecules have been produced in laboratories imitating the possible cosmic circumstances [22]. Molecules in comets were in situ studied at the comet Halley [23]. Amino acids were found in meteorites [21]. Recent investigations try to avoid criticism that terrestrial impurities plaid role in measurement by observing the C - C ratio (expressed as 5^^C = [(^^C/^^C)organic / (^^C/^^C)i„organ,c - 1] X 10^) 5^^C is in the average -30 from the earliest studied kerogens (in the Issua finding) through the present living systems on Earth while that for amino acids found in Murchison meteorite is « + 30 [23, 24]. Many authors accept that enough evidence has been accumulated that the first organic molecules on the Earth were of cosmic origin from the interstellar grains, comets and meteorites [25, 26]. According to the published estimations the quantity of these molecules were sufficient to start higher order processes like polymerization, compartmentalization and building up information [27]. At the end of this paragraph we present estimation about the evolution of the complexity of molecules. In Fig. 3 the complexity is characterised by the mass of the largest molecules (for example DNA).

4. Origin of Homochirality 4.1 Origin on Earth In general, ideas about the origin of homochirality may be divided into two groups: it happened via chance process or it was caused by some asymmetric physical force. In establishing homochirality the first step is to have a small asymmetry between the quantity of enantiomers the so-called enantiomeric excess (e.e.) which gets amplified by other processes. The discussions whether homochirality originated by chance or it was the result of a certain

Origin of Biomoiecules - Origin of Homochirality

129

10 T

6

8

10

12

14

16

Time (billion years)

Figure 3. Estimated timeline of the evolution of molecules. M means the mass of molecules in Daltons asymmetric physical agent will continue for long time probably until other, extraterrestrial living systems will be contacted. If living organisms based on D-amino acids and L-sugars will be detected then chance origin becomes almost certain. Without this development, efforts to find an effective physical force will continuously be pursued. The theoretical background of the chance process was thoroughly discussed in recent works [28, 29]. These studies clearly demonstrated that homochirality could evolve in a developing system where chiral molecules were formed from achiral sources, grown autocatalitically and competed against each other for the material resources. Though for this bifurcation theory, convincing experimental demonstration with biomoiecules, has not yet been reported. The recent work of the Soai group may be considered the closest approach to bifurcation [30]. The outcome of this process, however, is decided by the actual state of fluctuation at the bifurcation point, i. e. by chance. The requirements of the bifurcation processes probably exclude their effectiveness in space. Considering the asymmetric physical forces we have to distinguish two groups, global forces acting mainly on Earth and universal forces, that produce asymmetry everywhere in the Universe. These forces might cause a small e. e. that could be amplified. The first group involve asymmetric adsorption; circularly polarized light; electric, magnetic and gravitational fields: light and magnetic field, electric and magnetic fields, gravitation and rotation; Gilat's idea (magnetic moments around amino acids) [31]. Presently only one asymmetric universal force is known: the parity violating weak interaction that may act in two ways. The beta radiations from radioactive decays are polarized. Their spin points backward for P" particles and forward for P^ particles. Their interaction was expected different for L- and D-molecules. After the first investigators of this possibility it is named Vester-Ulbricht process. Until now no reliable experiment was reported. The other way of parity violating weak interaction is via the parity violating weak nuclear current that contributes, though in extremely small measure, to all interactions in molecules (for details see ref [33] and references therein). One of the authors of this paper (L. K.) recently reviewed the different works on the above topics [31-33] therefore we will deal only with the latest results. A quite intriguing result was reported by Shinitzky et al. [34] that D-tyrosine crystallizes

130

Progress in Biological Chirality Table 1, Distribution of Left and Right Quartz Crystals Number of crystals

Left

Right

Refs.

16807 27053

8607(50.5%) 13481 (49.8%)

8320(49.5%) 13572 (50.2 %)

[41] [42]

from saturated solutions much faster than L-tyrosine. Independent confirmations of these results would be highly desirable. Another very interesting and also surprising study was reported by Ribo et al. [35]. In water solution of achiral diprotonated porphyrins chiral homoassociates are formed. Repeating the experiments many times showed that the signs are statistical. When the solutions are stirred the chiral signs of the resulting homoassociates depend about 85% on the rotation direction. These experiments are somehow similar to those performed by Kondepudi's group crystallizing NaC104 and other molecules under stirring [36, 37], In these studies, however, the distribution of chiral signs was statistical. The preference in chiral selection by the direction of stirring the solution should be reinforced by independent investigators. If correct then one could imagine that vortices appearing in water could act as stirrers. Such vortices could also form around hydrothermal vents. One of the recurrent ideas for inducing homochirality is the adsorption on asymmetric surfaces like quartz crystals found everywhere on Earth. L-amino acids are adsorbed more on L-quartz crystals and vice-versa [38]. The problem, however, is the distribution of L- and D-quartz crystals on Earth. This problem was surveyed by Klabunovski and Thiemann [39]. Table 1 shows that the deviation from 50 - 50 % is small and even opposite in the two studies. Thus, this new compilation indicates that asymmetric adsorption on quartz crystals is not a real source of homochirality though Soai's group demonstrated that small excess of one crystal form can induce large asymmetry in autocatalytic systems [40]. In the area of Vester-Ulbricht process no new report appeared in the reviewed period which probably means a diminishing interest. There were, however, important new developments in the field of parity-violating neutral current. In theory, new calculation of the parity-violating energy difference between enantiomers (PVED) [43, 44] ended up with values of an order of magnitude larger than the values obtained earlier [45]. New relativistic calculation confirmed the dependence of PVED on the 6th power of theatomic number Z in the asymmetry centre [46]. The theory also pointed out that the effect of parity-violation could be even larger in chemical reaction [47]. Experimental confirmation of these calculations is important not only from the point of view of biological homochirality but also from basic knowledge about parity-violating effects. Because of the strong Z dependence it is advantageous to use those molecules that have high Z atoms in the asymmetry centre. We may divide the experimental approaches into two groups: microscopic and macroscopic measurements. In the first group the energy difference is measured with some very high-resolution spectroscopic method. Lasers [43], nuclear magnetic resonance (NMR [48] and IVlossbauer-effect [49]) were discussed as possible experimental approaches. A Mossbauer experiment on L- and D-tris(l,2-ethanediamine) iridium (III) complex provided only an upper limit of 4x10'^ eV being 3.6x10^ times larger than PVED [49]. Table 2 collects data on possible Mossbauer nuclides. PVED values are given according to the new calculations and are related to the line centre determined as 10% of the line width. It is seen that the best candidate for a successful study is the nuclide 73Ta^^\

Origin of Biomolecules - Origin of Homochirality

131

Table 2. Mossbauer-nuclides for Possible Measurement of Parity-Violating Energy Difference (PVEP). (Central value means 10% of line width) Nuclide

Linewidth

y-energy inkeV

73Ta'^' 32Ge^^ .3Np^' iilr'"' soZn"^' 26Fe"

6.23 13.3 59.6 73 93.3 14.4

PVED/ central value

Putative PVED ineV

rineV

6.2x10'^ 2.2x10-^ 2.5x10' 4.2x10-^ 4.2x10"^ 1.5x10'^

8.1x10-^^ 6x10-^' 3.5x10-" l.lxlO" 3.9x10"^' 1.5x10-^'

1.3x10-^^ 2.7x10-^^ 1.4x10-^ 2.6x10"^ 9.3x10" 9.6x10-^

It should be noted, however, that calculations how PVED appears in the isomer shift of a Mossbauer line are not known. In a recently reported Mossbauer measurements on chiral iron complexes a difference of (1.9±1.0)xl0-10 eV was found [50]. The difference does not seem statistically significant as seen in Fig. 4 where we delineated the published data with error. Nevertheless, that difference may be considered an upper limit for PVED. Using Table 2 this upper limit is « 104xPVED. In this measurement, however, the difference was 2%, 5 times less than the conservative estimate of 10 % of the line width in Table 2. Sophisticated infrared spectroscopy on vibrational modes of CHFClBr enantiomers has shown an upper limit for the difference of 13 Hz while theory predicted 2 mHz [51]. Another high-resolution spectroscopic method (NMR) was also considered as a possible way to measure PVED directly [48]. The authors concluded that in cases of high Z value spin one half nuclei such as Pt, Tl, Xe in the asymmetry centre of the enantiomers the difference could be measured with ultra-high-resolution technique. Suitable microscopic or, with other word, direct method to measure PVED are only in fijture plans therefore macroscopic or indirect methods may be used at least to demonstrate the theoretically forecasted PVED. Macroscopic methods, like polymerization or

•

' 11

(0 0.32-

E E

J

\_

1

1

T

I

—r

1

^

(

1

CO

ci) 0.30E o

•

• D

(0

L D •

()

'

1

1

1

2

3

1

4

No of meas.

Figure 4. Delineation of isomer shift data and their errorfromref. [48]: B for L-, < for D-form of Fe complex

132

Progress in Biological Chirality

crystallization of enantiomers amplify the small e. e. [52, 53] and simulate the mode how could PVED generate asymmetry during chemical evolution. In our approach we deah with crystallization of racemic mixture of L- and D forms of sodium ammonium tartrate, tris(l,2ethanediamine) cobak(III) and tris(l,2-ethanediamine) iridium(III) complexes [54]. Performing a large number of crystallizations the distribution measured by circular dichroism was shifted from zero value outside error for Co and Ir complexes while the distribution for tartrate crystallization did not show shift larger than error (Fig. 5). This study appears to have provided an experimental demonstration of the parity-violating weak interaction in molecules. According the main critics of such macroscopic experiments it is questionable whether the preferential crystallization is due to the parity-violating weak interaction or to some unknown chiral impurity [43]. For the actual experiment, in which many controls closed out the influence of chiral impurities, this reasoning is not acceptable even more because the influence of impurities surely does not increase with atomic number as was found in experiment (Fig. 5) and expectedfromtheory. In spite of this result we consider PVED an open question as a determining factor for the origin of homochirality. 4. 2 Homochiralityfromspace As the information on possible space source of biomolecules evolved suggestion for homochirality from space also appeared. Bonner and coworkers assumed that the circularly polarized light of neutron stars could induce e.e. on molecules synthesized at the interstellar grains and then transport it to Earth [55-58]. The space origin of homochirality gained fiirther support from two observations. The meteorites contained more L- than D- amino acids [23, 24] and circularly polarized infrared radiation was discovered in the Orion Nebula [59]. Infrared radiation has low energy to decompose molecules therefore it was inferred that the Orion Nebula contains also circularly polarized u. v. light. The e. e. in meteorites was assigned to the effect of these circularly polarized light sources. This hypothesis was criticized in ref [33] emphasizing its negligible probability. Therefore it seems to be usefiil to

r

1

'

1 — I — 1 — 1 — 1 — t — 1 — 1 —

1—'—1—'—1—;

Ir

100

^^ "b X

10

T

. "

r

1

•^

•1

1 Co

X **=

^

•

I

x:

(/) CD >

•(0

;j 0.1 r

1

'

•i

C

0

]

1-

0.01

T—

0

1 — ' — 10

1—1—1—1—1—1—1—1—1 20 30 40 50 60

'

1 70

'

\ 80

atomic number

Figure 5. Relative shifts of the distributions of ciystallisation depending on the atomic number of atom in the asymmetry centre (data takenfromref [54])

Origin of Biomolecules - Origin of Homochirality

13 3

call attention to alternative explanations. Salam pointed out that quantum mechanical cooperative and condensation phenomena might give rise to second order phase transition below a critical temperature Tc [60, 61] (transformation of D-amino acids to L-amino acids that have lower energy due to parityviolating energy differences). Tc is probably very low. The experimental verification of this phenomenon has not proved successful under terrestrial conditions [62, 63]. The negative results of the experiments performed in laboratories at very low temperatures for short time do not mean, however, that such transitions may not occur in the Cosmos during very long time and at low temperatures. This possibility, offering an understanding of e. e. of amino acids in meteorites was already considered by Figureau et al. [62]. Thus, presently the best candidate for inducing e. e. in meteorites is the Salam-process based on parity-violating weak interaction.

5.

Conclusion In the introduction questions were raised about the location (Earth or space) of origin of biomolecules and homochirality and also about philosophical {Platonic or Democritiah) trends. There are no indisputable arguments to make decision in these questions. We favor the 13

terrestrial origin. In the case of biomolecules the value of 8 C in meteorites could advice us. Accepting the value as a characteristics of the extraterrestrial organic matter it is difficult to 13

understand the immediate change in the value of 5 C from +30 to -30 (note that the present living organisms accomplish this change from 0 to -30 [64, 65]). It is easier to return to the terrestrial origin of biomolecules. In the case of homochirality we already considered the low probability of space origin. We think that the whole paper demonstrates the more complicated way of the Democritian trend, living from not living matter. Here only the first events were considered, even these are complicated and not yet clearly known. The most important events, like the origin of molecules with information, their repHcation and the formation of first cell, were not touched. The ambition of Science is to understand the origin of life based on the known Universe (matter and laws). But: not all matter is known (only about 5%), not all laws are known (? %). The final problem is the origin of matter and its laws. We have to assume, as a modern version of Plutonian trendy a Creator or God who established the harmony of matter and its laws. 6.

References

[1] G. Palyi, C. Zucchi and L. Caglioti, Eds., Fundamentals of Life, Elsevier and Accademia Nazionale di Scienze, Lettere ed Arti (Modena), Paris, 2002. [2] A.J. Oparin, The Origin ofPrebiological Systems (Ed. S.W. Fox) Academic Press, New York, 1965, pp. 91-96. [3] (a) J.W. Schopf, A.B. Kudryavtsev, D.G. Agresti, T.J. Wdowiak and A.D. Czaja, Nature 416 (2002) 7376. (b) M.D. Brasier, O.R. Green, A.P. Jephcoat, A.K. Kleppe, M.J. Van Kranendonk, J.F. Lindsay, A. Steele and N.V. Grassineau, Nature 416 (2002) 76-81. [4] J. Brooks and G. Shaw, Origin and Development of Living Systems, Acad. Press, London-New York, 1973. [5] M. Rees, Sci. Am., Separate Issue on the Frontiers of Space (2003) 82-85. [6] D.B. Cline, Sci. Am. 288/3 (2003) 28-35. [7] S. L. Miller, Science 117 (1953) 528-529.

13 4 [8 [9 [10 [11 [12 [13 [14 [15 [16 [17 [18 [19 [20 [21 [22 [23 [24 [25 [26 [27 [28 [29 [30 [31 [32 [33 [34 [35 [36 137 [38 [39 [40 [41

[42 [43 [44 [45 [46 [47 [48 [49 [50 [51 [52

Progress in Biological Chirality K. Bahadur. Nature 173 (1954) 1141-1143. K. Dose and B. Rajewsky, Biochim. Biophys. Acta 25 (1957) 225-226. T. Hasselstrom, M.C. Heniy and B. Murr, Science 125 (1957) 350-351. R. Paschke, R.W.H. Chang and D. Young, Science 125 (1957) 881-883. S.J. Fox and K. Harada J. Am. Chem. Soc. 82 (1960) 3745-3751. J.M. Kasting, Orig. Life. Evol. Biosphere 20 (1990) 199-231. K. Kobayashi, M. Tsuchija, T. Oshima and H. Yanagawa, Orig. Life. Evol. Biosphere 20 (1990) 99-109. J.B. Corliss, J. Diamond, L.I. Gordon. J.M. Edmond, R.P. van Hertzen, R.D. Ballard, K.K. Green, D. Williams, A. Bainbridge, K. Crane and T.H. van Andel, Science 203 (1979) 1073-1083. C. Huber and G. Wachtershauser, Science 216 (1997) 245-247. H. Oasawara, A. Yoshida, E. Imai, H. Honda, K. Halori and K. Malsuno, Orig. Life. Evol. Biosphere 30 (2000)519-526. J. Oro, Nature 190 (1961) 389-390. A.G.G.M. Tielens and S.B. Charley, Orig. Life. Evol. Biosphere 27 (1997) 23-51. P. Ehrenfreund and S. B. Charmley, Annu. Rev. Astronom. Astr. 38 (2000) 427-460. M.H. Engel and B. Nagy, Nature 296 (1982) 837-840. J.M. Greenberg, Sci. Am. 283/6 (2000) 70-75. JR. Cronin and S. Pizzarello, Science 275 (1997) 951-955. M.H. Engel and S. A. Macko, Nature 389 (1997) 265-268. J.M. Greenberg, Sci. Am. 250/6 (1984) 124-130. C.F. Chyba, P.J. Thomas, L. Brookshaw and C. Sagan, Science 249 (1990) 366-373. C.F. Chyba and C. Sagan, Nature 355 (1992) 125-132. V.A. Avetisov and V.I. Goldanskii, Physics- Uspekhi 39/8 (1996) 819-831. V.A. Avetisov and V.I. Goldanskii, Proc. Natl. Acad Sci. USA 94 (1996) 11435-11442. K. Soai, I. Sato, T. Shibata, S. Koniya, M. Hayashi, Y. Matsueda, H. Inamura, T. Hayase, H. Morioka, H. Tabira, J. Yamamoto and Y. Kowata, Tetrahedron Asymmetry 14 (2003) 185-188. L. Keszthelyi, Quart. Rev. Biophys. 28 (1995) 437-504. L. Keszthelyi, in: Fundamentals of Life (Eds. G. Palyi, C. Zucchi and L. Caglioti) Elsevier and Accademia Nazionale di Scienze, Lettere ed Arti (Modena), Paris, 2002, pp. 379-387. L. Keszthelyi. Orig. Life Evol. Biosphere 31 (2001) 249-256. M. Shinitzky, F. Nudelman, Y. Barda, R. Haimovitz, Effie Chen and D. W. Deamer, Orig. Life Evol. Biosphere 32 (2002) 285-297. J.M. Ribo, J. Crusat, F. Sagues, J. Claret and R. Rubires, Science 292 (2001) 2063-2066. D.K. Kondepudi, R. J Kaufman and N. Singh. Science 250 (1990) 975-976. K. Asakura, A. Ikumo, K. Kurihara, S. Osanai and D.K. Kondepudi, J. Phys. Chem. 104 (2000) 26892694. W.A. Bonner, Orig Life Evol. Biosphere 21 (1991) 59-111. E. Klabunovski aod W. Thiemann, Orig. Life Evol. Biosphere 30 (2000) 431-434. K. Soai, S. Osanai, K. Kadowaki, S. Yonekubo, T. Shibata, I. Sato, J. Am. Chem. Soc 121 (1999) 11235. C. Palache, G. B. Berman and C. Frondel, Relative frequences of left andrightquartz, in C. Frondel ed. The System of Mineralogy of J. D. Dana and E. S. Dana, Yale Univ. 1837, 1892. 7th ed. V. 3 Silica minerals. J. Wiley and Sons Inc. New York, London, 1962, p. 17. C. Frundel, American Mineralogist, 63 (1978) 17-27. (a) M. QaacKAngew. Chem. Int. Ed 41 (2002) 4618-4630. (b) D.A. Singleton, L.K. Vo, J. Am. Chem. Soc. 124 (2002) 10010-10011. F. Faglioni and P. Lazzeretti, Phys. Rev. E65 (2002) 11904. A. MacDermott, Orig. Life Evol. Biosphere 25 (1995) 191-199. J.K. Laerdahl and P. Schwerdtfeger, Phys. Rev A 60 (1999) 4439-4453. A. Soncini, A. Ligabue and P. Laz^xretti, Phys. Rev. E 62 (2000) 8395-8399. J.B. Robert and AI. Barra, Chirality 13 (2001) 699-702. L. Keszthelyi, J. Biol. Phys. 20 (1994) 241-245. A S . Lahamer, S.M. Mahurin, R.N. Compton. D. House, J.K. Laerdahl. M. Lein and P. Schwerdtfeger, Phys. Rev. Lett. 85 (2000) 4470-4473. A. Daussy, T. Marrel, A. Amy-Klain, C.T. Nguyen, C.J. Borde and C. Chardonnet, Phys. Rev. Lett. 83 (1999) 1554. Y. Yamagata, J. Theor. Biol. 11 (1966)495-498.

Origin of Bioraolecules - Origin of Homochirality

13 5

[53] Y. Yamagata, H. Sakihama and K. Nakano, Orig. Life 10 (1980) 349-355. [54] (a) A. Szabo-Nagy and L. Keszthelyi, Proc. Natl. Acad. Sci. USA 96 (1999) 4252^255. (b) A. SzaboNagy and L. Keszthelyi, in: Advances in BioChirality (Eds. G. Palyi, C. Zucchi and L. Caglioti) Elsevier, Amsterdam, 1999, 367-376. [55] E. Rubinsteia W A. Bonner, H.P. Noyes and G.S. Brown, Nature 306 (1983) 118-119. [56] W.A. Bonner and E. Rubinstein, BioSystems, 20 (1987) 99-111. [57] J.M. Greenberg, A. Kouchi, W. Niessen, H. Irth, J. van Paradijs, M. de Groot and W. Hermsea J- Biol. Phys. 20(1994)61. [58] W.A Bonner, J.M. Greenberg and E. Rubinstein, Orig. Life Evol Biosphere 29 (1999) 215-219. [59] J. Bailey, A. Chrysosyomou, J.H. Hough, T.M Gledhill, A. McCall, S. Clark, F. Menard and M. Tamura, S'c/e/ice 281 (1998) 672. [60] A. Salam,y. Mol. Evol. 33 (1991) 105-113. [61] A. Salam, in: Chemical Evolution: Origin of Life (Eds. C. Ponnamperuma and J. A. Chela-Flores) Deepak Publishing, Hampton, Virginia, 1993, pp. 101-117. [62] A. Figureau, E. Duval and A. Boukenter, in: Chemical Evolution: Origin of Life (Eds. C. Ponnamperuma, and J. A. Chela-Flores) Deepak Publishing, Hampton, Virginia, 1993, pp. 157-164. [63] R. Navarro-Gonzales, R.K. Khaima, and C. Ponnamperuma, in: Chemical Evolution: Origin of Life (Eds. C. Ponnamperuma, and J.A. Chela-Flores) Deepak Publishing, Hampton, Virginia, 1993, pp. 135-155. [64] M. Schidlowski, in: Fundamentals of Life (Eds. G. Palyi, C. Zucchi and L. Caghoti) Elsevier and Accademia Nazionale di Scienze, Lettere ed Arti (Modena), Paris 2002, pp. 307-329. [65] S.J. Mojzsis, G. Arrhenius, K.D. McKeegan, T.M. Harrison, A.P. Nutman and C.R.L. Friend, Nature 384 (1996)55-59.

Progress in Biological Chirality G. Palyi, C. Zucchi and L. Caglioti (Editors) © 2004 Elsevier Ltd. All rights reserved.

Chapter 11 Chiral Crystal Faces of Common Rock-Forming Minerals Robert M. Hazen Geophysical Laboratory and NASA Astrobiology Institute, Carnegie Institution of Washington, 5251 Broad Branch Road NW, Washington, DC 20015-1305, USA r. hazen@gl. ciw. edu

1.

Introduction Chiral crystalline surfaces provide effective environments for chiral molecular discrimination in both natural and industrial contexts [1]. Such surfaces have been cited for almost 70 years in reference to their possible role in the origins of biochemical homochirality [2-7]. In the past decade, fiirthermore, chiral crystal surfaces have received attention for their potential applications in the chiral selection and purification of pharmaceuticals and other molecular products [8-12]. Many recent studies have focused on the behavior of chiral surfaces of cubic close-packed (CCP) metals, including copper, silver, gold and platinum [13-23]. Single crystals of these metals, which can be modified by cutting, polishing and annealing faces with high Miller indices, display surfaces with chiral "kink" sites, even though the three-dimensional CCP structure is intrinsically achiral. Theoretical studies of these metal surfaces have demonstrated the potential for significant differences in adsorption energies of D- versus Lmolecules [14, 21-23], while experiments provide indirect evidence for chiral selectivity [13, 15-19]. Considerably less attention has been focused on the wide variety of chiral oxide and silicate mineral surfaces, which are ubiquitous in Earth's crust. Such surfaces provide the most abundant and accessible local chiral geochemical environments, and thus represent logical sites for the prebiotic chiral selection and organization of essential biomolecules. This chapter summarizes the geological occurrence, physical properties, crystal morphology and surface structures of some of the most common of these natural surfaces, including crystal faces of quartz (Si02), alkali feldspar [(Na,K)AlSi308], clinopyroxene [(Ca,Mg,Fe)Si03], and calcite (CaC03). One or more of these minerals is present in most common rocks in Earth's crust, as well as on the Moon, Mars and other terrestrial bodies, so chiral crystal environments are correspondingly ubiquitous [24, 25].

2.

Chiral Environments on Mineral Surfaces: General Considerations Many natural crystals are "euhedral" - bounded by a set of planar faces. These natural

138

Progress in Biological Chirality

Figure 1. Crystals commonly display three types of chiral surface featm-es, illustrated here in idealized drawings, (a) A periodic two-dimensional chiral arrangement of atoms in a plane; these atoms may be coplanar or they may occur at slightly different heights, (b) A terrace step that is chiral along a step edge (red line) (c) A kink site that provides a chiral center (X)

crystal growth surfaces, or "terminations," may be represented as the intersection of a plane with a three-dimensional periodic atomic structure. Such surfaces are usually defined in terms of a set of three integers, known as Miller indices, which relate the orientation of the terminal plane to integral intercepts of the three crystallographic axes [26, 27]. For a given unit cell, every possible planar termination has a unique corresponding set of Miller indices. A chiral crystal surface is defined as any terminal arrangement of atoms that cannot be superimposed on its reflection in a mirror perpendicular to the surface. Such crystal surfaces display three common types of chiral environments. Some atomic surfaces are chiral because the periodic two-dimensional structure of the exposed surface lacks mirror symmetry (Figure la). These surface atoms may be coplanar or they may display significant topography. In either case such a surface, if chiral, is not superimposable on its reflection in a perpendicular mirror. Many crystal surfaces possess perpendicular mirror symmetry and thus are inherently achiral. Nevertheless, such faces often feature steps in the atomic structure that intersect the mirror symmetry operator at other than right angles (Figure lb). Under these circumstances, local environments immediately along the step edge are chiral, even though most of the crystal face is achiral. A third type of chiral environment, a local "chiral center," may occur on any crystal face. Chiral centers commonly arise on surfaces in which both planar regions and steps possess mirror symmetry, as in the case of face-centered cubic metals. In these cases the steps may be "kinked" to provide a chiral center at the kink site (Figure Ic) [13]. Two distinct types of symmetry conditions lead to chiral crystal surfaces. A few minerals are inherently chiral because their crystallographic space group lacks any of the so-called "improper" symmetry operators, including mirrors, glide planes, an inversion center or a rotoinversion operator [26, 27]. Thus, in minerals such as quartz (space group P3i21 or P3221) every surface is chiral and there exist so-called left- and right-handed structural variants, which are not superimposable and thus related to each other by mirror symmetry [28, 29]. Most minerals possess space group symmetries that incorporate at least one mirror symmetry operator, and thus the mineral is intrinsically achiral. Nevertheless, as noted above, a crystal termination will be chiral if no perpendicular mirror symmetry operator intersects

Chiral Crystal Faces of Common Rock-Forming Minerals

13 9

that termination. This condition is met by one or more common crystal growth surfaces of many common rock-forming minerals. These faces, which have received little attention in terms of their chiral properties, provide the primary focus of this chapter. In addition to chiral planes, most crystal surfaces possess etch pits, growth steps, twin boundaries, dislocations or other nonperiodic features that provide numerous local chiral centers on an otherwise achiral surface environment. These ubiquitous local chiral features may have been important in fostering chiral molecular processes, but they are not in the scope of this review. Before examining the characteristics of specific chiral mineral surfaces, it is important to emphasize that all of these natural chiral surface environments occur in both left- and righthanded variants in approximately equal proportions. No evidence exists for an enantiomeric excess of any chiral mineral feature [30, 31]. Nevertheless, the widespread occurrence of local chiral environments provided the prebiotic Earth with innumerable sites for experiments in chiral selection and organization - experiments that may have led, through a process of chiral amplification [32-35], to a fortuitous, self-replicating homochiral entity. These minerals.

Figure 2. Common crystal forms of quartz include the hexagonal prism m (100), the dominant rhombohedron r (101) and the secondary rhombohedron z (Oil). Left- and right-handed quartz (a and b, respectively) may be distinguished by two additional forms, denoted ^ (111) and x (511). Most crystals, such as the 3.2-cm diameter specimen from Montgomery County, Arkansas (c), display only the /w, r, and z faces. Less common specimens, such as the 3.5-cm diameter right-handed crystal from Betroka, Madagascar (d), develop the additional forms

140

Progress in Biological Chirality

furthermore, represent an untapped library of chiral surfaces for possible industrial applications. The following section examines four common groups of rock-forming minerals that routinely display chiral crystal growth faces.

3.

Common Chiral Crystal Faces of Minerals

3. J Quartz Quartz (SiOi, trigonal space group P3i21 or P3221, a = 4.91 A, c = 5.41 A) is the predominant colorless mineral in most beach sand and is a principal component of many igneous, sedimentary and metamorphic rocks. Quartz is the only common rock-forming mineral that occurs in both right and left-handed variants. This structural distinction arises fi-om the silicatefi*ameworkthat incorporates either right- or left-handed helices of comerlinked Si04 tetrahedra [28, 36]. Three common crystal faces, illustrated in Figure 2, provide important chiral surfaces for study [37]: the ubiquitous (100) prism faces (denoted m in Figure 2), the dominant (101) rhombohedral termination (r), and the (Oil) rhombohedral termination (z), which is typically less well developed than (101). Note, however, that these three crystal forms are generally insufficient to distinguish right- fi-om left-handed specimens. This distinction can be made,

Figure 3. The (100), (101) and (Oil) surface structures of quartz (Si02), viewed from above (a, c, and e, respectively) and tilted 3° from horizontal (b, d, and f, respectively). Oxygen and silicon atoms are shown in red and blue, respectively. Positions of terminal oxygen atoms are indicatai by yellow Xs. In each drawing the caxis projection is vertical and each drawing presents an area 15 x 15 A

Chiral Crystal Faces of Common Rock-Forming Minerals 141 however, if the (111) and (511) faces (s and x, respectively) are present (Figures 2a and 2b, respectively). The surface structures of the three common quartz forms (/w, r, and z), while all chiral, are markedly different from each other, as illustrated in Figure 3. Above the point of zero charge of quartz (pH « 2.5), the quartz surface charge is negative [38-40]. In addition, silicon atoms typically remain tetrahedrally coordinated, so oxygen atoms (perhaps bonded to H, depending on pH) are expected to define the crystal terminations [41, 42]. Given this assumption, the surface structures of the (100), (101), and (011) faces are well constrained. The (100) prism face has zigzag bands of oxygen atoms separated by channels approximately 1.5 A wide and 2.0 A deep (Figure 3a). Note, therefore (Figure 3b), that the "surface" oxygen atoms are not coplanar. This feature is of critical importance in modeling surface interaction of quartz and other minerals. By contrast, the (101) face can be modeled with a more planar surface with a distribution of oxygen atoms that is much closer to an achiral array (Figure 3c and d). The (011) face presents yet a different character, with a denser chiral array of surface oxygen atoms (Figure 3e and f). These three faces also differ in the coordination of terminal oxygen atoms. On the (101) face all oxygen atoms are coordinated to a single silicon atom, whereas all oxygen atoms on the (Oil) face are coordinated to two silicon atoms. The (100) face, by contrast, features both one- and two-coordinated oxygen atoms. These marked differences in surface distribution of oxygen atoms explain, for example, the dramatically different adsorption characteristics of hematite (Fe203) on (101) versus (Oil) rhombohedral faces of some natural quartz crystals (Figure 4). These differences also point to the necessity of studying any surface interactions, such as selective adsorption of organic molecules, on individual faces rather than on powdered material. Given the striking differences in surface structures, the adsorption behavior of a molecule on one surface can bear little relationship to adsorption on any other face.

...^^:; •

^

^

'

;

fi§^

' " ;{P-- ^

Figure 4. Hematite (Fe203) preferentially deposits on (101) faces of quartz, while (011) faces remain largely micoated (-^1-mm diameter crystals from Paterson, New Jersey). This phenomenon results from significant differences in the surface structures of these two rhombohedral faces (see Figure 3c and d versus 3e and f)

142

Progress in Biological Chirality

J-'X

m

Figure 5. Common caystal faces of feldspar include the chiral form m (110), which is often well developed in orthoclase (a) and albite (b). The 7- x 7-cm specimen of alkah feldspar (c) from Ethiopia displays these faces

3.2 Alkali Feldspar Feldspars, including the alkali feldspar series (Na,K)AlSi308 and the plagioclase feldspar series (NaSi,CaAl)AlSi208, are among the most common rock-forming minerals in Earth's crust [24, 25]. These framework aluminosilicates are major constituents of most igneous rocks and they provide the principal repositories of alkali and alkaline earth cations. Feldspars form significantfractionsof many sedimentary and metamorphic rocks, as well. A variety of alkali feldspars, including both orthoclase (Figure 5a: KAlSiaOg, monoclinic space group C2/m, a = 8.56 A, b = 13.0 A, c = 7.19 A, p = 89.1°) and albite (Figure 5b: NaAlSiaOg, triclinic space group CI, a = 8.14 A, ^ = 12.8 A, c = 7.16 A, a = 94.3^ P = 116.5", y = 87.7"), commonly have well-developed (110) faces. This form occurs in enantiomeric pairs in many natural crystals (Figure 5c). The surface structures of feldspar are less well constrained than those of quartz because of uncertainties in the terminal oxygen configurations near alkali cations. It is likely, for example, that oxygen coordination of alkali cations near the crystal surface in an aqueous environment will vary as a fimction of pH. Uncertainty also arises from the occurrence of different ordered distributions of silicon and aluminum atoms in tetrahedral coordination, as well as the facile exchange of alkali and alkaline earth cations between the crystal surface and aqueous solution [43]. Given these uncertainties, one possible configuration of oxygen atoms at the (110) chiral surface is illustrated in Figure 6. In this example of an orthoclase surface structure with, potassium atoms retaining their full 7-coordination, oxygen atoms are arrayed in rows approximately parallel to [001], as illustrated in Figure 6a. This surface displays significantly nonplanar topography as a consequence of the oxygen atoms selected (Figure 6b). A different choice of terminal oxygen atoms (for example removing the highest rows of atoms in Figure 6b) would significantly increase the surface exposure of positively-charged alkali cations. 3.3 Clinopyroxene Clinopyroxenes, the most common of all ferromagnesian rock-forming minerals, incorporate a diverse group of species with the general formula (Ca,Mg,Fe)Si03 [36]. Pyroxenes are major components in many igneous and metamorphic rocks in both the Earth's crust and upper mantle. They occur commonly in both orthorhombic and monoclinic varieties, but it is the latter that most commonly offer chiral crystal growth faces. The most

Chiral Crystal Faces of Common Rock-Forming Minerals

143

Figure 6. One possible (110) chiral surface structure of orthoclase, which is a member of the alkaU feldspar group. Silicon, oxygen and potassium atoms are shown in blue, red and turquoise, respectively. Terminal oxygen atoms are marked with yellow Xs. The [001] axis is vertical and the area is 15 x 15 A. Note that terminal oxygen atoms are chosen in this model so that potassimn is fully coordinated, which effectively shields potassium atomsfromthe surface

common clinopyroxene structure, as typified by the mineral diopside (CaMgSi206, monoclinic space group C2/c, a = 9.75 A, b = 8.90 X c = 5.25 A, P = 105.6"), features chains of comer-linked silicate tetrahedral that are crossed-linked by divalent Mg and Ca cations in 6- and 8-coordination, respectively. The most common chiral clinopyroxene face is the ubiquitous (110) perfect cleavage plane, which is designated m (Figure 7a). This face also occurs on crystals, occasionally in

Figure 7. Clinopyroxene [(Ca,Mg,Fe)Si03] displays several chiral faces (a and b), including the common (110) cleavage plane (designated w), and occasionally the (111), (221) and (2 21) forms (designated w, o, and v, respectively), (c) The 1.3-cm diameter crystal of diopside (CaMgSi206)fromXinjiang, Uygur Province, China, displays both the (110) and the (111) chiral forms

144

Progress in Biological Chirality

combination with the ( H I ) , (221) and ( 2 21) chiral faces [37]. In addition, four (llO)-type faces often combine with pairs of (100) and (010) faces to form an 8-sided crystal prism (Figure 7c). Such elongated crystals, which parallel the silicate chain, represent a distinctive morphology of clinopyroxenes. The fact that the (110) surface is also a perfect cleavage surface in clinopyroxene raises the possibility of obtaining large, freshly exposed chiral surfaces from cleaved samples for studies of chiral molecular interactions. Ambiguity arises when attempting to model the (110) surface structure of clinopyroxene. As in quartz and feldspar, the silicon atoms are assumed to remain tetrahedrally coordinated. The coordination of divalent cations, however, is less certain and will likely vary depending on the environment of the crystal. Figure 8 illustrates three different possible terminal atomic arrangements for the (110) surface of diopside. In the first configuration (Figure 8a and b) calcium atoms near the surface are coordinated to seven rather than eight oxygen atoms, thus exposing both positively-charged calcium and negatively-charged oxygen atoms at the surface. Alternatively, magnesium may be partially coordinated near the surface in at least two possible configurations (Figure 8c through f). If Mg is four-coordinated near the surface, then a quasi-linear pattern of approximately planar surface atoms results (Figure 8c and d). If magnesium is five-coordinated near the surface, then a more complex surface structure results, with both positively-charged magnesium atoms and oxygen atoms at three different

Figure 8. Three possible terminations for the (110) surface of diopside (CaMgSi206). Ca, Mg, Si and O are turquoise, green, blue, and red, respectively. Each 15 x 15 A drawing has the [001]-axis projection vertical. X and + indicate O atoms and cations near the surface. Small Xs in (e) are O atoms that are significantly below other surface atoms, but may participate in surface binding

Chiral Crystal Faces of Common Rock-Forming Minerals

145

heights relative to the surface (Figure 8e and f). The adsorption characteristics of (110), consequently, will depend critically on the as yet unknown cation coordination at the surface. Calcite (CaC03; rhombohedral space group R3c), the principal mineral of limestone and marble, is of special interest in studies of chiral selection by mineral surfaces. Calcite was one of the most abundant marine minerals on the early Earth and calcite crystal surfaces would 3.4 Calcite have been widely present in prebiotic environments [44, 45]. Calcite is also one of the most common biominerals; it is strongly bonded to proteins in the shells of many invertebrates [46,47]. The potential for calcite to interact with chiral molecular species has been underscored by studies of surface growth topology, which may be strongly affected by the presence of L versus D amino acids [48]. The literature on calcite is confused by the common use of four different axial systems, each of which results in a different set of Miller indices for any given plane [36]. Two of these sets of axes are based on inconvenient rhombohedral unit cells (in which one axial length and one interaxial angle are specified). Most authors prefer the simpler hexagonal setting (in which two orthogonal axial lengths, a and c, are specified) and that convention is used in this chapter. However, additional confusion arises from the existence of two different axial conventions for the hexagonal unit cell. One set of axes, based on the classic morphology of the calcite cleavage rhomb, results in the so-called "cleavage rhomb unit cell" or "morphological unit cell" (a = 10 A; c = 8.5 A in the hexagonal setting). This cell is invariably used to describe twinning, cleavage, and crystal forms [36,37]. In this setting, the Miller indices for the common cleavage rhomb are (101). Alternatively, the so-called "structural unit-cell" (a = 5 A; c = 17 A in the hexagonal setting) is the minimal unit cell determined by x-ray methods. In this case the axial orientations are identical to the morphological cell, but the a axis is halved and the c axis is doubled. Thus, for example, Miller indices for the cleavage face (101) in the hexagonal morphological setting become (104) in the hexagonal structural setting. When working with calcite surfaces, therefore, it is critical to specify both the unit cell and the Miller indices in order to avoid ambiguity. The most common calcite crystal form is the scalenohedron, in which adjacent faces have mirror-related surface structures (Figure 9). This form, with Miller indices (211) in the hexagonal morphological cell or (214) in the structural cell, is of special interest because of its ability to adsorb D and L amino acids selectively [7]. Modeling the (211) scalenohedral surface is complicated by the nature of the calcite structure, which has a halite or NaCl-type face-centered cubic arrangement of alternating Ca cations and rigid CO3 anions. A few surfaces, such as the perfect rhombohedral cleavage [(101) or (104) in the morphological or structural settings, respectively], present a uniform surface structure of coplanar Ca and CO3 (Figure 10). This surface incorporates a gHde plane operator and is thus achiral. Most calcite crystal surfaces, however, intersect coplanar arrays of Ca and CO3 groups so that the surface must incorporate steps and kinks, in much the same way as high-index planes of face-centered cubic metals are stepped and kinked [13]. Thus, the common calcite scalenohedral faces [(211) or (214) in the hexagonal morphological or structural settings, respectively] display a complex chiral surface topology that is not easily, or unambiguously, modeled. Figure 11 displays a possible surface configuration, based on the assumption that all surface oxygen atoms are associated with CO3 groups. This requirement leads to prominent

146

Progress in Biological Chirality

Figure 9. Calcite, CaCOs,frequentiyoccurs with (a) the chiial scalenohedral form [designated v; (211) or (214) in the hexagonal morphological or structural setting, respectively], as well as (b) the rhombohedral form [designated r; (101) or (104), respectively], which is also the common cleavage plane, (c) A doubly-terminated crystal from Elmwood Mine, Tennessee, displays a well-formed scalenohedron

2-A high surface steps (Figure lib, arrow). These steps are parallel to [018] in the hexagonal structural setting (or [012 ] in the hexagonal morphological setting), and are spaced approximately 12 A apart. This topology, with its linear array of chiral binding sites, may provide a natural template for the synthesis of homochiral polypeptides [7].

Figure 10. (a) The calcite rhombohedral cleavage [(101) or (104) in the hexagonal morphological or structural settings, respectively] presents a surface in which Ca cations andrigidCO3 anions alternate. The surface has glide plane symmetry (vertical yellow lines) and so is achiral. (b) The cleavage surface topology is revealed in a view that is tilted 6°fromthe horizontal. Ca, C and O atoms are turquoise, blue and red, respectively. Each drawing is approximately 15 x 15 A, and the c-axis projections are vertical

Chiral Crystal Faces of Common Rock-Forming Minerals

147

Figure 11. (a) The structm-e of the scalenohedral face of calcite [(211) or (214) in the hexagonal morphological or structmBl settings, respectively] features a chiral arrangement of positive (+) and negative (X) charge centers near the crystal termination. Ca, C and O atoms are turquoise, blue and red, respectively. In this 20 x 20 A view the ( 0 1 8 ) axis in the hexagonal structural setting [equivalent to the (012) axis in the hexagonal morphological setting] is vertical ~ an orientation that provides a useful image of the surface structure, (b) A view of this surface tilted 3° from horizontal (projected almost down the [01 8 ] axis) reveals the irregular surface topology, including 2-A-deep steps (yellow arrow) that result from the oblique intersection of layers of Ca and rigid CO3 groups with the surface (yellow line)

3.5 Other common chiralfaces of rock-forming minerals In addition to quartz, feldspar, pyroxene and calcite, numerous other minerals display chiral surfaces. Most of these species are rare or their chiral forms are seldom expressed. However, two other particularly common minerals, amphibole and gypsum, deserve mention with regard to their common chiral crystal faces. The amphibole minerals include a varied suite of hydrous chain silicates that are often found in igneous and metamorphic rocks [36]. This compositionally diverse group commonly conforms to the formula (Na,K,Ca)2(Mg,Fe,Al)5Si8022(OH)2. Amphiboles are structurally related to pyroxenes and, like pyroxenes, occur in both orthorhombic and monoclinic forms. The latter clinoamphibole group frequently displays chiral crystal faces and cleavage surfaces. These amphiboles, such as tremolite, Ca2Mg5Si8022(OH)2, and actinolite, Ca2(Mg,Fe)5Si8022(OH)2 (both monoclinic space group C2/m; a = 9 . 8 A , b = 1 8 . l A , c = 5.3 A, P= 104.7*'), routinely develop the chiral (110) and (Oil) forms, designated m and r, respectively, in Figure 12, as well as less common (120) form [37]. In addition, the (110) plane is a perfect cleavage in all clinoamphibole species and so offers the potential for exposing fresh chiral surfaces for study. However, as with clinopyroxenes, the detailed surface structure of clinoamphiboles with be strongly dependent on coordination of monoand divalent cations at the surface. Further characterization of these faces thus represents a promising research opportunity. Gypsum (CaS04 2H20: monoclinic space group C2/c; a = 5.7 A, b = 15.2 A, c = 6.3 A, p=

148

Progress in Biological Chirality

Figure 12. (a) The (110) and (011) forms of clinoamphibole, designated m and r, respectively, are relatively common chiral crystal surfaces, (b) The chiral (120) form (designated e) also occurs, though infrequentiy. (c) The 2-cm diameter crystal of actinolite [nominally Ca2(Mg,Fe)5Si8022(OH)2]fromMpwa-Mpwa, Tanzania, displays both the m and r forms

113.8*'), the most abundant natural sulfate, is a common marine evaporite mineral that readily forms euhedral crystals with chiral (110) and (111) faces, as illustrated in Figure 13 [36,37]. Thick gypsum deposits are found around the globe in many regressive sedimentary sequences. Crystal growth is extremely rapid under appropriate evaporative conditions; natural euhedral crystals may achieve lengths in excess of 10 cm in several days (R.Lavinski, personal communications). Gypsum has proven to be of special interest in studies of the interactions of chiral surfaces with chiral molecules. Growth of (110) and (111) faces, in particular, are dramatically influenced by the presence of chiral solute molecules [49]. Thus, for example, D and L alanine have been shown to thwart the growth of enantiomeric (110) faces, producing highly

ar^> b

Figure 13. (a) Gypsum, CaS04 2H2O, conunonly developstiiechiral (110) and (111) forms, which are designated m and / respectively, (b) The (1 11) form (designated n) is also seen occasionally, (c) A euhedral gypsum crystalfromGui Lin, Guanxi Province, China, 7.5 x 2.5 cm

Chiral Crystal Faces of Common Rock-Forming Minerals

149

distorted crystals. This phenomenon has been invoked to explain the occurrence of uniformly asymmetric gypsum crystals from a Miocene evaporite deposit in Poland - an environment presumably dominated by L amino acids [50]. However, in spite of these intriguing morphological curiosities, the chiral surfaces of gypsum remain difficult to study because of the mineral's high degree of solubility in water.

4.

Conclusions This brief overview points to several important conclusions regarding chiral crystalline surfaces. 1. Chiral crystalline faces are ubiquitous in nature. Quartz, feldspar, pyroxene, calcite, amphibole and gypsum provide a weahh of enantiomeric atomic surfaces in virtually every common crustal rock on Earth and other terrestrial bodies. Furthermore, any irregular mineral fracture surface will provide an additional variety of local chiral environments. In addition, hundreds of other candidate crystal growth faces also occur in nature. Most of those surfaces occur either on relatively rare minerals [i.e., the (111) and (221) faces of topaz, Al2Si04(OH)2, which occasionally forms crystals > 1 m in length] or are crystal forms that are rare in well-developed specimens [i.e., the unusual (111) form of the common mineral olivine, (Mg,Fe)2Si04]. A few of these less common surfaces will be present in most geochemical settings. Each of these surfaces has a specific atomic structure that represents a possible location for the selection, concentration and assembly of chiral organic species from the indiscriminately racemic prebiotic molecular soup into the homochiral macromolecules of life. 2. Most mineral surfaces are not chiral: With the exceptions of quartz (for which all crystal faces are inherently chiral) and calcite (for which the predominant scalenohedral face is chiral), most crystal grov^h surfaces on most minerals are achiral. Care must be taken, therefore, when studying minerals for their ability to induce chiral molecular separations. In this regard, special note should be made of layer hydroxides and layer silicates, including micas, chlorites and clays, which have been invoked in prebiotic processes of molecular selection and organization [51-54]. All layer silicates develop primarily the achiral (001) basal surfaces and therefore cannot impose a chiral environment. 3. Different forms of a crystal typically display very different surface structures: The (100), (101) and (Oil) forms of quartz are dramatically different. Each form has a different chiral surface distribution of atoms and a different atomic topography. Chiral interactions of molecules, therefore, are expected to differ for these different surfaces. In this regard it is significant that several previous experiments have employed powdered minerals (notably left- or right-handed quartz) in the hopes of inducing a chiral selective effect [2-5,55]. While such experiments may yield fortuitous enantiomeric excesses in the product suite, this use of powdered material greatly reduces the hope of discerning a structural mechanism for the observed chiral effect. The use of well-documented chiral crystalline surfaces is therefore much to be preferred. 4. Some surfaces are ''more chiral" than others: The distribution of surface charges on some chiral faces, such as the (101) form of quartz, closely approximates an achiral

150

Progress in Biological Chirality

configuration. Other faces, such as those of the calcite scalenohedron, deviate significantly from their enantiomer. These differences point to the possible utility of a "chirality index" that measures the misfit of a chiral surface with its enantiomer [56]. 5. In some instances the surface structure is ambiguous: The surface structures of feldspar and clinopyroxene, for example, depend on the coordination numbers of monovalent and divalent cations at or near the surface - structural details that will depend strongly on the surface environment. Small changes in cation coordination can result in significant changes of the chiral surface charge distribution. 6. So-called "flat" crystal faces may be stepped: The surface of the calcite scalenohedron typifies an atomic configuration in which coplanar structural elements intersect a surface obliquely. This situation results in a stepped surface that may provide a linear array of chiral centers. Such an array may facilitate the condensation of homochiral polymers. These distinctive attributes of chiral mineral surfaces point to significant opportunities for future studies at the dynamic interface between crystals and their environments.

5.

Acknowledgements I am grateful to Aravind Asthagiri, Robert Downs, Gozen Ertem, Mary Ewell, Andrew Gellman, James Kubicki, David Sholl and Henry Teng for useful discussions and constructive reviews of the manuscript. All crystal structure drawings were made with the program XtalDraw [57], courtesy of Robert Downs. Photographs of mineral specimens were generously supplied by Dr. Robert Lavinsky, President of Arkenstone in Garland, Texas.

6. [1] [2] [3] [4] |5] [6] [7] [8] [9] [10] Ill] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24]

References R.M. Hazen and D.S. Sholl, Nature Materials 2 (2003) 367-374. R Tsuchida. M. Kobayashi and A. Nakamura, J. Chem. Soc. Japan 56 (1935) 1339-1341. G. Karagounis and G. Coumoulos, Nature 142 (1938) 162-163. A. Amariglio, H. Amarigho and X. Duval Helv. Chim. Acta 51 (1968) 2110-2111. W.A. Bonner, P.R. Kavasmaneck, F.S. Martin and J.J. Flores, Ohg. Life 6 (1975) 367-376. N. Lahav, Biogenesis: Tfieories of Life's Origin, Oxford University Press, NY, 1999. R.M. Hazen, T.R. Filley and G.A. Goodfreind, Proc. Natl. Acad. Sci. USA 98 (2001) 5487-5490. B. Kahr, S. Lovell and J.A. Subramony, Chirality 10 (1998) 66-71. B. Kahr and R.W. Gumey, Chem. Rev. 101 (2001) 893-951. S.C. Stinson, Chem. Eng. News 79 (May 14, 2001) 45-46. M. Jacoby, Chem. Eng News 80 (March 25, 2002) 43-46. AM. Rouhi, Chem. Eng. News 80 (June 10. 2002) 43-44. C.F. McFadden, P.S. Cremer and A.J. Gelhnan, Langmuir 12 (1996) 2483-2487. D.S. ShoU. Langmuir 14 (1998) 862-867. A. Ahmadi, G. Attard, J. Feliu and A. Rodes, Langmuir 15 (1999) 2420-2424. G.A. Attard, J. Phys. Chem. B 105 (2001) 3158-3167. D.S. Sholl, A. Asthagiri and T.D. Power,./. Phys. Chem. B 105 (2001) 4771-4782. J.D. Horvath and A.J. Gellman, J. Am. Chem Soc. 123 (2001) 7953-7954. J.D. Horvath and A.J. Gellman, J. Am. Chem. Soc. 124 (2002) 2384-2392. A. Kuhnle, T.R. Linderoth. B. Hammer and F. Besenbacher. Nature 415 (2002) 891-893. Z. Sljivaneanin, K.V. Gothelf and B. Hammer. J. Am. Chem. Soc. 124 (2002) 14789-14794. T.D. Power and D.S. SholL Top. Catal. 18 (2002) 201-208. T.D. Power, A. Asthagiri and D.S. ShoU, Langmuir 18 (2002) 3737-3748. F.J. Turner and J. Verhoogen, Igneous andMetamorphic Petrology, Mc-Graw-Hill, NY, 1960.

Chiral Crystal Faces of Common Rock-Forming Minerals

151

[25] F.J. Pettijohn, Sedimentary Rocks, Harper & Row, NY, 1957. [26] F.D. Bloss, Crystallography and Crystal Chemistry, Holt, Reinhart and Winston, NY, 1971. [27] M.B. Boisen, Jr. and G.V. Gibbs, Mathematical Crystallography, Mineralogical Society of America, Washington, 1985. [28] L. Bragg, G.F. Claringbull and W.H. Taylor, Crystal Structures of Minerals, Cornell University Press, Ithaca, NY, 1965. [29] J.V. Smith, Geometrical and Structural Crystallography, John Wiley, NY, 1982. [30] C. Frondel,v4m. Mineral 63 (1978) 17-27. [31] E. Evgenii and T. Wolfram, Orig Life Evol. Biosphere 30 (2000) 431-434. [32] R.D. Murphy and T.M. El-Agez, Indian J. Chem. 35 A (1996) 546-549. [33] B.L. Feringa and R.A. van Delden, Angew. Chem., Int. Ed 38 (1999) 3418-3438. [34] D.Z. Lippmann and J. Dix, in: Advances in BioChirality (Eds. G. Palyi, C. Zucchi and L. Caglioti) Elsevier, Amsterdam, 1999, 85-98. [35] H. Zepik, E. Shavit, M. Tang, T.R. Jensen, K. Kjaer, G. Bolbach, L. Leiserowitz, I. Weissbuch and M. Lahav, Science 295 (2002) 1266-1269. [36] W.A. Deer, R.A. Howie and J. Zussman, An Introduction to the Rock-Forming Minerals, John Wiley, NY, 1971. [37] E.S. Dana, A Textbook of Mineralogy, John Wiley, NY, 1949. [38] R. Parsons, Surface Sci. 24 (1964) 418-826. [39] J.A. Davis and D.B. Kent, Rev. Mineral. 23 (1990) 177-259. [40] H. Chm-chill, H. Teng and R.M. Hazen, Am. Mineral, in press. [41] M.C. Goldberg, E.R. Weiner and P.M. Boymel, J. Chem. Soc. Faraday Trans. 80 (1984) 1491-1498. [42] Y. Xiao and A.C. Lasaga, Geochim. Cosmoch. Acta 60 (1996) 2283-2295. [43] P.H. Ribbe, Feldspar Mineralogy, T'^ Edition, Mineralogical Society of America, Washington, DC, 1983. [44] A.W. Bailey and A.R. Pahner, Eds., The Geology of North America: An Overview, Geological Society of America, Boulder, Colorado, 1989. [45] D.W. S\xmmx,Am. J. Sci. 297 (1997) 455-487. [46] H. Lowenstam and S. Weiner, On Biomineralization, Oxford University Press, Oxford, UK, 1989. [47] S. Weiner and L. Addadi, J. Mater. Chem. 1 (1997) 689-702. [48] H. Teng, P.M. Dove, C.A. Orme and J.J. DeYoreo, Science 282 (1998) 724-727. [49] A.M. Cody and R.D. Cody, /. Crystal Growth 113 (1991) 508-519. [50] M. Babel,^rc/z. Mineral. 44 (1990) 103-135, plates 1-9. [51] A.G. Cairns-Smith, Clay Minerals and the Origin of Life, Cambridge University Press, 1986. [52] H. Hartman, G. Sposito, A. Yang, S. Manne, S.A.C. Gould and P.K. Hansma, Clays & ClayMin. 38 (1990)337-342. [53] J.P. Ferris, C.-H. Huang and W.J. Hagan, Jr., Orig Life Evol. Biosphere 18 (1988) 121-128. [54] L.E. Orgel, Orig Life Evol. Biosphere 28 (1998) 227-234. [55] K. Soai, S. Osanai, K. Kadowaki, S. Yonekubo and I. Sato, /. Am. Chem. Soc. 121 (1999) 11235-11236. [56] R.T. Downs and R.M. Hazen, J. Mol. Catal. in press. [57] R.T. Downs and M. Hall-Wallace, Am. Mineral 88 (2003) 247-250.

Progress in Biological Chirality G. Palyi, C. Zucchi and L. Caglioti (Editors) © 2004 Elsevier Ltd. All rights reserved.

Chapter 12 Implication of Polya's Urn Experiment in Biochirality and Cerebral Lateralization Noboru Hokkyo Senjikan Institute, Igarsi 2-847S-16, Niigata 950 -2102, Japan noboru@icn. ne.jp

1. Introduction Everything is dependently co-arisen Nagarjuna(BC250-150)[l] A large majority of chiral molecules in living organisms is present in only one enantiometric form, while the abiotic synthesis of organic compounds yield exclusively racemic mixtures. Presently we do not know exceptions to the homochirality rule in DNA or RNA. Yet, there is a rapidly increasing number of biochemical studies [2] detecting the sign of "a record of the past from a racemic world" in living organisms [3]. Examples of macroscopic manifestations of chiral feature in the spatial structures and behaviors of vertebrate are also known. Here we propose to understand the evolutionary origin of chiral purity and impurity of the biological world on the same footing from molecular level to macroscopic level in terms of Polya's urn experiment (1923), a toy experiment of stochastic population growth with feed back. The recent neurophysioiogical observations indicate the exclusive use of a hand in humans requires an active ("enantioselective") neuro-functional organization developed during postnatal brain maturation to prevent mirror movements in ("racemic") infants.

2. Polya's Experiment and Chiral Evolution of a Racemic Pair of Molecular Strands To be an Error and to be Cast out is a part of God's design W.Blake (1757-1827) [4] In 1923 Eggenberger and Polya [5, 6] proposed the following "urn scheme" as a model for the stochastic population growth and the development of contageous phenomenon. Consider a cell, or an urn, initially containing one black ball and one white ball from which one ball is drawn at random. Whatever its color, it is returned to the urn together with a fresh ball of the same color. Repeating the same procedure many times and observing the percentage of, for example, black balls in the urn, we find that the percentage can approach any value between 0 and 100 depending on the outcome of the first few draws.

154

Progress in Biological Chirality 100

50

1 1

500

1000

Figure 1. Percentage of black balls vs. number of draws in three separate Polya experiments

In Fig.l three separate experiments are plotted, each up to 1000 draws. In all of them an initial stage of fluctuations is followed by a stable behavior, which differs in each case. It can be shown that there is an equal probability for the behavior converging to any percentage of black balls. In Oparin's theory [4] early life began by the successive accumulation of more and more complicated molecular populations within a naturally occurring cell, such as a droplet of coacervate evolving into self-sustaining metabolic cycles. Polya's thought experiment indicates the corresponding model for the stochastic chiral evolution defined by the following assumptions. 1. Early life originated within an urn, such as a fine pore in rock [7], with small openings flooded by an aqueous solution of energy-rich monomers that diffuse in and out of the urn. 2. Early life originated from a racemic doublet of monomer strands, L and D, each serving as a template for self-replication by assembling monomers within the urn. 3. Each strand has a life time of a replication cycle. 4. Environmental factors (temperature, humidity, pH, etc.) and insufficient energy and monomer supply to the urn, as well as the geometrical constraint of the openings, make the simultaneous replication of L and D strands improbable. Fig. 2 shows the first four generations of a progeny of the racemic doublet LD. It is seen that the racemic doublet LD produces, in the first generation, a pair of chiral triplets, LLD and LDD, with equal probability 1/2, each of which, in turn, generates a pair of quartets, LLLD, LLDD and LLDD,LDDD with branching ratio 2/3,1/3, and 1/3, 2/3, and so on. The chiral asymmetry is beginning to be seen in the progeny in the 4th generation. 3. Polya Model of Cerabral Lateralization Any man who, upon looking down at his bare feet, doesn 7 laugh, has either no sense of symmetry or no sense of humour. Descartes (1596-1650) [8]

Implication of Polya' s Urn Experiment in Biochirality and Cerebral Lateralization

LLLLLD

LLLLDD

IXLDDD

LLLDDD

LLDDDD

155

LDDDDD

Figure 2. Development of chiral asymmetry in the progeny of the racemic doublet LD

The evolutionary and ontogenetic basis of the asymmetric specialization of the cerebral hemispheres is notfiiUyexplored. But if the neural hardware in each sphere is prespecified, it would be difficult to account for the ample evidence of the ability of each hemisphere to assume a specific mental operation of the other where the latter is damaged or totally excised. Mirror movements are involuntary movements on one side of the body that occur as reversals of an intended movements contralaterally. They are present physiologically in infants and commonly persists as normal phenomenon in young children. They usually disappear after the first decade of life, coinciding with the completion of the corpus callosum connecting the two hemisphere. It is therefore conceivable that the exclusive ("enantioselective") use of a hand in infants ("racemites") requires lateralized functional organization- developed during postnatal brain maturation to prevent mirror movements [9]. Fig. 3a compares the blood-oxigenization-level-dependent contrast functional magnetic resonance images (BOLD-fMRI of activated premotor cortex of a normal right-handed subject (left) and a left-handed subject (right), both instructed to perform self-paced grasp motions with right hands; Fig. 3b compares a native English speaker (left) and native Japanese speaker (right) instructed to read the same English text. We may define a Polya model for cerebral lateralization by the following assumptions to be examined by experiments. 1. Brain of vertebrate bifurcated into left and right hemispheres, L and R, to resolve conflicts between the evolutionary pressure of expanding cerebral cortex and the limited cranical capacity of the skull. 2. Human brain uses L and D as a binary decisions system detecting the difference in similar objects and the sameness in different objects, respectively. 3. Once activated, the decision system triggers successively higher level systems until a behavioral command to motor system is elicited in the premotor cortex. 4. The neuro-flinctional organization preventing the bilateral movements, developed during postnatal brain maturation by an exclusive use of a hand, suppresses the synchronous activation of L and D unless instructed. 5. The synchronous activation of L and D occurs in the process of perceptual binding ("creative synthesis") of what is known and what is to be known.

156

Progress in Biological Chirality

R

L

R

L

Figure 3. a: Brain activities of left-handed subjects, constrained to be right-handed (right) and unconstrained (left), using left hands; b: Native Enghsh speaker (left) and native Japanese speaker (right) reading the same English text [9]

4.

Conclusion It is often remarked that biochirality originates from a chiral molecule abiotically produced by chance in a droplet or a fine pore in terrestrial rock or comet of extra terrestrial origin. But the well-founded CPT symmetry of the physical universe between positive and negative charges (C), between left and right (P) and between past and future (T) makes it more likely that the chiral world of modern life evolved from an abiotically bom racemic pair of organic compounds. In the light of recent development of functional MRI evidences are being accumulated, attributing the origin of cerebral lateralization of humans to exclusive ("enanioselective") use of a hand suppressing ("racemic") mirror movements in infants during the postnatal development of the binary decision system, reminiscient of the Polya experiment in which only one or other out of a racemic mixture of L and D balls in the urn can be dawn at a time.

5.

References

[1] J. Garfield, The Fundamental Wisdom of the Middle Way, Oxford University Press, Oxford, 1995. [2] See papers in the Special Issue on the Origin of Chirahty and D-amino Acids in Biological World. Viva Origino 30 (2002). [3] G. Palyi, C. Zucchi and L. Caghoti, Eds., in: Advances in Biochirality, Elsevier, Amsterdam, 1999. [4] F. Dyson, / Mol EvoL 18 (1982)340-350. [5] F. Eggenberger and G Polya, Zeit. Angew. Math. Mech. (1932) 279-289. [6] F.J. Varela, Principles ofBiological Autonomy, North Holland, New York, 1979, p. 172. [7] H. Kuhn and J. Waser, in Biophysics, (Eds. W. Hoppe et al.) Springer, Berlin 1983. [8] A. Salam, J. Mol Evol. 33 (1991) 105. [9] T. Nakata, et al., Neuroscience Research 32 (1998) 355-362.

Implication of Polya' s Um Experiment in Biochirality and Cerebral Lateralization

157

Appendix to Chapter 12 Evolutionary Origin of Binary Decision System and Pharmaceutics

When the world ceases to be the stage for personal hopes and desires, where we, as free beings, behold it in wonder to question and to contemplate there we enter the realm of science and art. A. Einstein [1879-1955] In "What the Frog's Eye Tells the Frog's Brain" [Al] Lettvin, Maturana, McCuUoch and Pitts discovered two major populations of neurons in frog's visual system, sometimes called 'sameness neurons' and 'newness neurons' distinguishing between retinal images of known and unknown objects. The newness neurons are supposed to be excited to bring the animal into an attentive state in search of unknown stimuli. The known objects are further classified by higher level neurons into two groups, biologically favorable and unfavorable [A2]. Quickly moving small objects (preys) and slowly approaching dark dim objects (predators) are examples of favorable and unfavorable objects. It is conceivable that a 'creative synthesis' of the external image of upper Hmbs of amphibian falling in the field of vision with an internal image of an animal on land propelled its own motor behavior. Next, there must occur the partial release of motor activity from genetically determined reflexive movement to voluntary movement. Human brain can be regarded as a device to transform a stream of sensory information from dual world, external and internal, into improved behavioral response and world image. Depending on the urgency to respond, information is immediately converted to instinctive/intuitive (preconscious) response, while some information is momentarily stored as a short-term memory and elicits slowly developing change in the internal attitude conductive of emotional (subconscious) response. The information can acquire an adequate conceptual(conscious) meaning only as far as it requires a critical evaluation and updating of

Figure lA. Frog's binary decision system distinguishing between visual images of what is known and what is to be known

158

Progress in Biological Chirality

Figure 2A. Proto-semantic responses to sensory information biologically favourable (upper hemisphere) and unfavourable (lower hemisphere), divided into specific emotive responses according to three biological values, self-preserving, species-preserving and exploring. Emotionally neutral information felling on the equator does not ehcit conscious evaluation of the information conductive of behavioural response [A2]

our inner picture of the external world stored in the long-term memory reservoir reachable by introspection (meditation or transcendental consciousness). Since, in human, hypocampal formation within the emotive brain is innervated by both extro- and intero-ceptive nervous systems, it is in a position to store, compare, and synthesize externally and internally derived information. In mammals the reflexive and emotive brains still regulate autonomic (unconscious) activity of internal organs and reflexive behavioral responses, but somato-muscular organs in higher mammal (hand and speech) fall under voluntary control mechanisms involved in protosemantic communication within species. By means of such voluntary movement, we can use language in speech. The human use of symbols, which is not under genetic control, affords unlimited boundaries to create 'conceptual space' transforming physical reality into transphysical one (virtual reality).'Eureka pleasure' associated with a sudden leap of an innovative scientific idea is likely to be generated when the threat of unknown (newness), is converted to emerging deja vu feeling of reality (sameness). It is tempting to hypothesize that the dominant (left) cerebral hemisphere is devoted to detect difference in similar objects (articulation of the world), and the non-dominant (right) hemisphere to discover similarity in different objects (grasping of the world in its coherent whole). It is hoped that the chemical neuro-transmitter substances which are involved in each of the cerebral hemisphere will eventually be identified. The accuracy and stability of long-term memories which depend upon the specificity of those neuro-transmitters would therefore be affected in precise ways by changes in the chemical balance of the transmitters in a given individual. [Al] L. Lettvin et al. Proc. IRE, 47 (1959) 1940. [A2] N. Hokkyo, in: Biocomputers (Eds. T. Kaminuma and G. Matsimioto; tr. N. Cook) Chapman & Hall, London, 1991.

Progress in Biological Chirality G. Palyi, C. Zucchi and L. Caglioti (Editors) © 2004 Elsevier Ltd. All rights reserved.

Chapter 13 Theory of Hierarchical Homochirality Dilip K. Kondepudi Department of Chemistry, Wake Forest University, Winston-Salem, NC 27109, USA [email protected]

I dedicate this article to Ilya Prigogine whose vision of the fundamental nature of irreversible processes inspired so many.

1.

Introduction The ubiquity of chiral asymmetry in nature, from matter's most fundamental constituents to the morphology of mammals, makes it obvious to us that there are physico-chemical processes that generate and propagate chiral asymmetry. But vv^hat these processes are and what mechanisms propagate the asymmetry is far from obvious. Nevertheless, based on symmetry considerations, and using elementary group theory, one can construct a general theory of spontaneous generation of chiral asymmetry and elucidate general features of how asymmetry can propagate from one structure to another. In this article I shall present such a general theory that is based on the work I did with George Nelson [1,2]. In developing a general theory of chiral asymmetry, one must keep in mind the fact that structures on different scales are somewhat independent. In simple terms, the shape of a spiral staircase does not depend on the shape of bricks with which it is built. Similarly, spiral spatial patterns that arise as a consequence of chemical reactions and diffusion are unrelated to the structure of molecules. Conversely, the vorticity of fluid motion could have no observable effect on the outcome of chemical reactions involving chiral molecules. After all, we cannot feel the effect of the Coriolis force due to Earth's rotation while riding a bicycle: the scales of motion are widely separated and one has an insignificant effect on the other. This is because the physical laws that govern processes at various spatial scales are somewhat independent: nuclear states have very little effect on chemistry and the basic laws of hydrodynamics are quite independent of molecular structure. So it is only natural to think that the asymmetry we observe at the level of elementary particles is unrelated to that at the level of biomolecules and that the biomolecular asymmetry is unrelated to the morphological asymmetry we see in mammals. Must the asymmetries at different levels be always independent? Is it possible that there are indeed processes that connect asymmetries at various levels in certain situations? In the following sections I will outline a basic theory that enables us to see how asymmetry at various levels could be connected. This theory provides us with a framework for

160

Progress in Biological Chirality

understanding the connection between molecular asymmetry and morphological asymmetry and the connection between fluid vorticity and molecular aggregation that was recently reported [3].

2.

The Roots of Morphological Homochirality In the case of morphologies of mammals, the organ placement is asymmetric: in humans, the heart is a little to the left, the liver is on the right, the stomach on the left and so on. But this asymmetry, apart from rare exceptions [4], is the same in all humans; it is inherited. Why this "homochirality"? If the process that establishes asymmetry is independent of the asymmetry at the molecular level, and is established during morphogenesis of each human, we must expect statistically equal number of people with mirror-image morphologies - as we do in conglomerate crystallization. Since the homochirality of organ placement is inherited [5], might it not be coded in the genetic sequence of DNA just like all other inherited traits? Could it not be that a particular sequence of A, T, G and C codes for the morphological asymmetry? The answer is no: It is impossible to code handedness in a linear sequence of letters, ATTGGCATGA Richard Feynman [6] and Martin Gadner [7], have noted the fact that it is impossible to communicate what one means by "left" or "righf (without using basic parity violating laws of nature that distinguishes between the two) by sending and receiving messages, in other words with strings of letters. There is a simple proof for this assertion using reductio ad absurdum. Let us suppose there exists a physical process without any inherent bias towards the left or the rights that takes as its input a sequence ATTGGCATGA... and produces a chiral structure, say a right-handed helix. (If we do not require that the process is "without any inherent bias towards the left or right" then the problem does not exist: the bias is itself the code for what is meant by "left" or "right".) The requirement "without any inherent bias towards the left or right" means that the process is either achiral or chirally symmetric, i.e., it should be indistinguishable from its mirror image. Processes that are governed by the laws of mechanics and electromagnetism satisfy this requirements. The sequence ATTGGCATGA... is assumed to provide the right-handed bias that drives the system to generate a right-handed helix. Using the assumed process, one could communicate what is meant by a right-handed helix, as opposed to left-handed helix. Why is such a process not possible? Let us look at the process in a mirror, as show in Fig. 1, which is also realizable because the laws of mechanics and electromagnetism are mirror-symmetric. Since the mirror-image process is also realizable, we have to conclude that the same process must also produce a left-helical structure using the same input ATTGGCATGA... This contradicts our original assumption that the sequence ATTGGCATGA...carries a right-handed bias and produces right-helical structure. The mirror-image process tells us that the same sequence must also carry a lefthanded bias and produce a left-helical structure. Hence an achiral or symmetric process that produces only a right-handed helix with the input ATTGGCATGA... cannot exist. It is often tacitly assumed that computer codes can carry any type of information. Since computer codes are ultimately strings of zeros and ones, the above lemma shows that it is impossible to design a code that can specify what one means by "left": the chiral objects that we are able to display on a computer screen are left- or right-handed only because the display

Theory of Hierarchical Homochirality

161

A ATTGGCATGA

MIRROR

SYMMETRIC PROCESS

ATTGGCATGA

\J 0 Figure 1. Is it possible for a symmetric process to generate only a "right-handed" object using a linear sequence of letters such as ATTGGCATGA .... as input? If such a process existed, by looking at the processes in a mirror, we must conclude that the same process must also generate a "left-handed" object. Hence, a process that takes as input a linear sequence such as ATTGGCATGA .... and generates only a "right-handed" object cannot exist window is scanned, or pixels identified, with a particular convention, from left to right or vice versa. Whether an object on the screen is right- or left-handed depends on this convention, not the computer code. So the morphological asymmetry cannot be inherited through a genetic sequence. It must be coded in some three-dimensional structure in which a particular handedness is coded. What might this structure be? Can its chirality be traced all the way down to the homochirality of Lamino acids of proteins or the D-sugars? Recent investigations on the origin of morphological asymmetry in chick embryos suggest that the asymmetry originates in the asymmetric motion of cilia during development [8]. This asymmetric motion must clearly be due to the some chiral aspect of its macromolecular structure which is likely to have its origin in L-amino acids. If we have an understanding of how chiral asymmetry propagates from one level to another, at this stage we could at least see in principle how the morphological homochirality could have its roots in the homochirality of the amino acids.

3.

Theory of Spontaneous Chiral Symmetry Breaking Using group theory, a general theory of spontaneous chiral symmetry breaking can be formulated [1]. In this theory, the state of the system, be it chemical or other, is specified by a n-vector X=[xi, X2, ... x j . In the case of a chemical system, Xk would be the concentrations; for a macromolecules, XR would be the variables that specify its structure. In general, the time-

162

Progress in Biological Chirality

evolution of this system is described by a set of nonlinear equations: ^ = F[X,A] at

(3.1)

in which F is a nonlinear operator; >. is a parameter that represents a constraint on the system; in an open chemical system, for example, it could be a flow rate, or a concentration that is fixed. In general, for any given system there will be several such parameters but, to present the main features of the formulation, we may consider explicitly only one parameter that is relevant to chiral symmetry breaking; the other parameters can be thought of as implicit in the definition of F. The state of the system, X, also defines its chirality: X could be a chirally symmetric state or a chirally asymmetric state. We shall use P for the parity operator that transforms X to its "mirror-image state" (strictly speaking, mirror-image state is the result of P followed by a rotation). Thus, X and PX represent mirror-image states. If the state of the system is symmetric, then X = PX. In the case of a chemical system in which the state is represented by the concentrations of chiral compounds, X would be symmetric if the concentrations of the enantiomers are equal; under the operation P, the enantiomers are interchanged (L-enantiomer becomes D-enantiomer and vice versa). If X is not symmetric, then X ^ PX. The processes, such as chemical kinetics, difftision, forces between the constituents of a macromolecule etc., that govern the time-evolution of X are assumed to be chirally symmetric. In mathematical term this means: PF[X,;i]=F[PX,?i]

(3.2)

The parameter X is generally achiral (as for instance, a flow rate or a concentration of an achiral reactant) so P does not alter X. That the processes that determine the state of the system, X, are chirally symmetric is expressed by equation (3.2). Spontaneous "breaking" of chiral symmetry is in the sense that, tough the processes F(X,>.) that determine the state X are chirally symmetric, the state itself may not be symmetric (X ^ PX), i.e., chirally symmetric processes generate a chirally asymmetric state. As a consequence of (3.2), if X is a solution to (3.1) then PX is also a solution. Hence, if X ^ PX, there are two independent solutions. The general features of spontaneous symmetry breaking are as follows: for every value of the parameter X, and appropriate initial and boundary conditions, there corresponds a steady state Xs, for which dXs/dt = 0. (In some cases, the trivial Xs=0 may be the only steady state). This steady state is symmetric for a range of values of X, but at a "critical point" Xc, this steady state becomes unstable to small perturbations in X; as a consequence, an appropriate fluctuation can drive the system to an asymmetric state. In mathematical terms, the transition from symmetric to asymmetric states is because of "bifiarcation" of new states, which is described as follows. In the vicinity of the critical point, the asymmetric state can be represented by: X = Xs + aXA

(3.3)

in which XA is an asymmetric state and a its "amplitude". Using very general arguments, it is possible to obtain an equation for the time-evolution of the amplitude a of the form:

Theory of Hierarchical Homochirality

163

Figure 2. Diagram showing the bifiircation of asymmetric steady states (09^) from the symmetric steady state (a=0) at Xc. The symmetric steady state is unstable when X>Xc

da dt

G[a,A]

(3.4)

When X< Xc, the only steady state that is stable corresponds to a=0; when X>Xc, states with a^^O are also possible. Equation (3.4) is sometimes called the "bifurcation equation" because it tells us how a new solution (3.3) "bifurcates" from the solution Xs when X>Xc. The asymmetric state, XA, has the property PXA = -XA. As a consequence, since PXs=Xs and PX is also a solution of (3.1), PX = Xs - aXA is also a solution. Consequently, if a is a solution of G[a,A,]=0, then - a is also a solution. This symmetry property restricts the possible forms of G[a,^]. Using the general theory of bifurcation and group theory one can show that, in the vicinity of Ac, (3.4) is of the form: —-=-AQr^+B(A-X,)Qr dt

(3.5)

in which A and B are constants that depend on the particular features of the system, such as kinetic rate constants or force constants (as will be made clear in the example below). For a given F[X, X], the constants A and B can be calculated using general procedures[l]. dct The steady states of (3.5) (for which — = 0) as a fiinction of X shows the "bifurcation" dt of asymmetric states (Fig. 2). When X<Xc, it can easily be seen that a=0 is the only real steady-state solution; when X>Xc, two solutions are possible:

a = ±^B(X-Xc)/A

(3.6)

These general features can be illustrated in a simple chemical system which is an extension of a model proposed by Frank in 1953. The extension puts the model in a thermodynamic

164

Progress in Biological Chiialiw

framework so that spontaneous generation of asymmetric state can be understood as the appearance of a far-from-equilibrium dissipative structure [2, 9]. The model chemical system that shows spontaneous chiral symmetry breaking is: S+T

(3.7a) (3.7b) (3.7c) (3.7d) (3.7e)

X I , + XD

Here S and T are achiral reactants that produce chiral product X in the L- and Dconfirmations. The rate constants for the forward and reverse reactions are denoted by kif and kir and so on. The equality of the rate constants of reactions (3.7a) and (3.7c), and (3.7b) and (3.7d), reflects the chiral symmetry of the processes. As indicated by reactions (3.7b) and (3.7.d), the chiral species X is autocatalytic. In addition, the two enantiomers of X react to form an inactive species P. Reaction scheme: S + T -> S+T ^

S + T + Xj ;f^ 2Xj Xj,

S+T+Xj,^

XL+XD ^

?L=[S][T]

S T

2XD

P

a = ([XLl-[XD])/2 S T

XT

-

XD

A,
X>Xc

Figure 3. A chemical model to illustrate the general theory of spontaneous chiral symmetry breaking. With the reaction scheme shown on top. in an open system with inflow of S and T and an out flow of P, the system can spontaneously make a transition to a state in which XL^XD even though the kinetics of the two enantiomers are identical. The general behavior of all such systems is shown in the bifurcation diagram

Theory of Hierarchical Homochirality

165

As shown in Fig. 3, we consider an open system into which there is inflow of S and T and an outflow of P such that the concentrations of S and T are maintained at a fixed level in the reactor. With these flow constraints, the system could be described by a set of two rate equations:

™

dt

- kif[S][T]- k J X J + k,,[S][T][XJ-k,,[X,f -k3[XJ[XJ

(3.8a)

= k JS][T]-k J X J + kJS][T][XJ- k,,[XJ^ - k3[XJ[XJ

(3.8b)

in which [S] denotes the concentration of species S. These equations correspond to the general time-evolution equation (3.1). In this model reaction scheme, by setting the parameter X =[S][T], it is possible to obtain the generic equation (3.5) in which the "amplitude" of the asymmetric state a=([XL]-[XD])/2:

dt in which, Xc, the critical value of X. at which asymmetric solutions bifurcate is given by:

Xc

(2 s + p-4kifk?,f

(3.9)

2kL S=2k3,k, + 4(k^^k„)/(k3-k,,)

Explicit expressions for the coefficients A and B are:

k3(k2'^C-kir)

B = k2f _iS2L

V

2kifk2r kjAc klr.

(3.10)

The details of obtaining these expressions using the general formalism can be found in ref 1. The same formalism can also be used to describe structural transition from an achiral structure to a chiral structure. As a simple example, let us consider the following structural transition shown in Fig. 4. In this model, the potential energy between the two arms of the molecule depend on the distance X and angle 9 between them: V(A,<9)=A«9'*-B(/!c-;i)<9^

(3,11)

When the two arms of the "molecule" are far apart, ie., when X>Xc, the lowest energy or equilibrium state corresponds to G=0. As the distance between the arms decreases, due to the repulsive forces at close proximity, the molecule twists to the right or to the left. For this potential, the equations of motion in terms of 9 and d0/dt, would correspond to the

166 Progress in Biological Chirality time-evolution equation (3.1). By suitably defining A and B, these equations of motion can be written as:

dt

(3.12a)

' = (D

(3.12b) We see that equation (3.12b) is similar to the bifurcation equation (3.5). Because of the simplicity of the model, the time-evolution equations are the same as the equations that describe chiral symmetry breaking. The stable steady states of the system (d9/dt=dco/dt=0) are the minima of the potential (3.11). For X>}ic, the stable steady state is 0=0; for \
^±=VB(^-/1)/2A

The corresponding bifurcation diagram is shown in Fig. 4.

4.

Sensitivity of Chiral Symmetry Breaking Transitions With the above formalism, we can now address the general question of how asymmetry at one level (or one system) may influence another. For example, in the chemical model shown in Fig. 3, we can investigate the effect of a small chiral interaction that might enhance the reaction rate of one enantiomer compared to the other. This chiral influence could arise from a lower level of parity violating electro-weak interactions or p-radiation or other similar effects that usually have no observable chiral effects in chemistry. In the case of macromolecular

0

Figure 4. A simple example of structural chiral symmetry breaking. The potential energy between the two arms of the "molecule" has the form V^X.O) = A0^ - B{Xc - X)0^. When X>Xc the equilibrium stmcture is synunetric; when X<Xc the equilibrium structure is asymmetric

Theory of Hierarchical Homochirality

167

chiral structures, the chiral influence may be from the lower level of the building blocks of the macromolecule which posses an asymmetry. For example, if we consider an achiral polymer that can fold in a helical form, the transition from a linear structure to helix is a case of spontaneous chiral symmetry breaking. Making one of the monomers chiral will introduce a chiral influence that will bias the system to favor one of the two helical secondary structures. This in turn may influence the chirality of the tertiary structure an so on. In the following section, we shall discuss hierarchical nature of structural chirality in the framework of the theory presented in this section. The origin of the chiral interaction and its magnitude depends on the particular conditions of the system under consideration but, as in the previous section, it is possible to formulate a general theory of how a chiral interaction might influence a system when it makes a symmetry-breaking transition. To be sure, we expect it to be sensitive to even a small chiral influence. A quantitative theory should specify what exactly is meant by "small" and how to quantify the system's sensitivity. In analyzing the influence of a chiral interaction on chiral-symmetry-breaking transition, we found it necessary to distinguish between two types of time evolution as shown in Figure 5A. In the first type (I), the system is initially below the critical value Xc and evolves to a value larger than the critical value. We shall assume that X starts with some initial value A,o, which is below the critical point, and evolves to a value above the critical point at a rate y: ?i-?io + rt

(4.1)

During this process, the system makes a transition to the asymmetric state, a>0 or a<0. In the second type (II), the system is initially in a symmetric state a=0, at a value of ^>A,c and, because the symmetric state is unstable, makes a transition to one of the asymmetric states.

Figure 5. Evolution to an asymmetric state. A. Two types of evolutions, I and IL are shown. In type I, the system is initially below Xc and evolves to a higher value. During this process, it makes a transitionfroma=0 to a>0 or a<0 state. In type IL the system is initially at X>Xc and a=^0 and evolves to a>0 or a<0. B. Evolution in the presence of a chiral interaction and random fluctuations that favors a>0 state: the state a>0 is reached with a probability' P+

168 Progress in Biological Chirality

Systematic chiral interactions are usually very small. In this context, the notion of "small" is with respect to the size (root-mean-square value) of random fluctuations in a that are present in every physical and chemical system. When the random fluctuations and the systematic chiral interaction are included, the general equation changes to [1, 2]: — - - A a ' + B(>. - ^c )a + Cg + Vef(t) dt

(4.2)

in which C is a constant that depends on the system parameters, g is a parameter that gives us measure of the strength of the chiral interaction; for example, if the chiral interaction causes a difference in the reaction rates of the enantiomers, this difference may be expressed as:

ki,=ai+g)=^H-||]

(4.3)

in which AG is the difference in the Gibbs energy of the transition state of the enantiomers. 8 is mean-square value of the random fluctuations denoted by f(t) (one may assume f(t) to be Gaussian white noise for simplicity). The bifurcation diagram and a typical evolution of a when Cg>0 is shown in Figure 5B. The sensitivity of the system to the chiral interaction can be quantified by the probability, P+, that the system will make a transition to the asymmetric state favored by the chiral interaction or bias. At first sight, it might seem that the influence of the chiral bias will be small and unobservable if Cg<8^^^, i.e. the probability P+ will be nearly 0.5 if Cg<8*^^. This however turns out not to be the case for process I in which X evolves according to (4.1). In this case, careful analysis showed that[2]: ] _

1

ft^,,^/ ..2 /-y\A.. jExp(-x'/2)dx, I l_,A.|Jl — A. / ^ I \ J A . ,

A/VAT^—

,

^S

r

K

'By

(4.4).

This expression shows that rate, y, at which the system sweeps through the critical point is important; the smaller the rate, the higher the sensitivity to the bias. In fact, even when Cg<8^^^, if the value of y is small, P+ can be nearly 100%. The validity of this theory has been shown in electronic switches subject to large noise and a small bias "Cg" [10, 11] and through numerical simulations of nonlinear equations [12]; the theory has also been generalized to higher symmetry groups [13]. In the context of biomolecular asymmetry, Goldanskii and coworkers published a similar analysis in which the evolution of X is not that given in (4.1) [14]. The conclusions thus reached lead them to make several incorrect statements about the above theory which have been noted [15]. Nicolis and Prigogine also published a general analysis of symmetry breaking and pattern selection[16] but they did not arrive at the result shown in (4.4). For the a time evolution of type II (Fig. 5), when X»'kc, the probability P+ is given to a good approximation by N = . In this case, for Cg<s^^^, the value of N is 'JSijyK ~ KQ ) / 2 generally vary small and hence P+~ 0.5. Thus, if initially the system is well above the critical

Theory of Hierarchical Homochirality

169

point, it is not very sensitive to small chiral biases that are smaller than the root-mean-square value of the fluctuations.

5.

Homochirality From One Level to the Next In the framework of the general theory presented in the previous sections, we can analyze how homochirality may propagate from one level to the next through symmetry breaking transitions. In our earlier work [2], we have elucidated how small asymmetries due to electroweak interactions or other similar fundamental asymmetries could influence the outcome of a chiral symmetry breaking transition that occurs on a time scale of lO** years and on a volume scale of 10^ m^ (5km x 5km x 10m). Here I would like to note that the same theoretical framework would enable us to understand the molecular origins of morphological asymmetry. As we noted above, experiments have shown that the origin of morphological asymmetry in mammals is in the asymmetric motion of cilia [7]. But, when we look at the structure of cilia we find that it is very complex consisting of microtubules. The microtubules in turn are an assembly of proteins and deep with in these proteins are alpha helices made of L-amino acids. The relation between the structure and its asymmetric motion and the asymmetry of Lamino acids is far from obvious. The difficulty is that of scale: the macromolecular assembly that is a cilium is so complex, and has so many degrees of freedom, that it is hard to see how its chiral asymmetry originates at the level of amino acids. To see the connection between the chirality of a complex macromolecular structure, such as a cilium, and its frmdamental building blocks, I propose the following hierarchical approach. Level 1 Start at the level of L-amino acids which are obviously asymmetric. Replace each amino acid by a sphere and look at the protein primary structure as a string of spheres with achiral interaction that can fold the string into a helix or sheet. As a string of achiral spheres folding into a helix is clearly a chiral-symmetry-breaking transition. In such a transition, we can now introduce a chiral bias in the interaction between the spheres that corresponds to that due to Lamino acids. During the symmetry breaking transition, the bias due to the asymmetry of the L-amino acids forces the system to become a right-handed helix. This will give us the link between the right-handed a-helices and L-amino acids in terms of a biased symmetrybreaking transition. Level 2 On the next level, we replace each secondary chiral structure by an achiral structure. For example, we could replace each a-helical segment of a protein with a cylinder and a P-sheet with a ribbon (as is sometimes done in proteins diagrams). Now we have to understand the asymmetry of this structure. It is a string of cylinders and ribbons linked to each other by small polypeptide segments which has folded into a chiral structure (see Figure 6). The chirality of such a structure may not be obvious as that of a helix; nevertheless, we can look at small segments of the long string and clearly define its chirality. Each segment is a collection of cylinders and ribbons that has folded chirally. For each chiral segment, we shall assume that there is a variable a that can be defined for a string of ribbons and cylinders so that it is non-zero when the structure is chiral. (Such a measure of chirality is generally not unique, but uniqueness of the measure is not crucial for our theory.) As this string folds to a chiral

170

Progress in Biological Chirality

LEVEL N N

LEVEL N+1

I

Figure 6. Chirality at different structural levels in a polypeptide. At level N. we see the a-helix attached to psheets (which have a slight right-handed twist not shown in figure) at either end. At level N+1, the heUx is replaced by an achiral cylinder hnked to two achiral sheets. The chirality of the folded sheet-cylinder-sheet chain could be left- or right-helical. In the case shown above, this chain is right-hehcal. The selection of right-helix over a left-hehx could be due to the right-helicity of the a-helix and p-sheets at level N.

Structure, it breaks chiral symmetry. Now we can introduce the inherent asymmetry in the cylinders or the asymmetry of the neighboring segments as a small chiral bias and understand tertiary structure as a biased chiral symmetry breaking transition, the chiral bias arising from the asymmetry at the level of the secondary structure. At this level, the protein is a string of chiral objects, each object being a segment of cylinders and ribbons. We may refer to these structures as Level-2 chiral units. Indeed it has been noted by Cothia [17] that it is common to see strings of a-helices and P-sheets fold chirally in proteins (but this author argues that the chiral asymmetry is lost at higher structural levels). Level 3 The procedure that takes us to the next level is now clear. We simply replace each Level-2 chiral units with a simple achiral object, such as a sphere or a cylinder, and visualize the protein in terms of these objects and investigate its chirality and describe it as a result symmetry-breaking at Level 3 and introduce the effect of the lower level asymmetry as a chiral bias. Clearly, this processes can be continued till we reach a level at which the entire protein is

Theory of Hierarchical Homochirahty

171

reduced to a simple achiral object such as a sphere. We can now proceed to the higher level of collection of proteins as in the case of a microtubules and continue this process. At any stage in the above hierarchical procedure, it could happen that the asymmetry at one level is not significantly influenced by the asymmetry at the lower level. Indeed, there are examples of polymers of chiral molecules which have secondary structure that seems to be independent of the chirality of the building blocks [18]. It is important to note that, according to the general theory, whether the asymmetry of one level will influence the next higher level or not depends on the process that establishes the asymmetry at the higher level. The propagation of asymmetry from one level to the next is more likely to happen in type I processes (Figure 5) with a slow evolution through the transition point.

6.

Concluding Remarks Macroscopic systems often do not have the symmetries of the microscopic laws. This situation if often the result of spontaneous symmetry breaking a process that has profound consequences as was noted eloquently by Anderson [19]. It is truly remarkable that morphological asymmetry can have its origins in molecular asymmetry. With the above formalism we can begin to understand how molecular asymmetry can propagate up the hierarchy of chiral structures and finally cause the asymmetry in the motion of cilia. Working out the specific applications of the formalism presented here may give us a better understanding of when and how chiral asymmetry propagates and why it is ubiquitous.

7. [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19]

References D.K. Kondepudi and G.W. Nelson, Physica 125A (1984) 465-496. D.K. Kondepudi and G.W. Nelson, Nature 314 (1985) 438-441. J.M.Ribo, J. Cmsats, F. Sagues, J.M. Claret and R. Rubires, Science 292 (2001) 2063-2066. J. Bum, in: Biological Asymmetry and Handedness, Ciha Foundation Symposium, John Wiley, New York, 162(1991)282-296. I.e. McManus, in: Biological Asymmetry and Handedness, Ciba Foundation Symposium, John Wiley, New York, 162 (1991) 251-267. R. Feynman, The Character of Physical Law, MIT Press, Boston, 1990, Chapter 4. M. Gardner, The New Ambidextrous Universe, W.H. Freeman, New York, 1990, Chapter 18. D.M. Supp, S.S. Potter and M. Brueckner, Trends Cell Biol 10 (2000) 41-45. D.K. Kondepudi and LP. Prigogine, Modem Thermodynamics, John Wiley, New York, 1998, Chapter 19. D.K. Kondepudi, F. Moss and P. V.E. McClintock, Physica 21D (1986) 296-306. D.K. Kondepudi, F. Moss and P. V.E. McCUntock, Phys. Lett. I l l A (1986) 29-32 D.K. Kondepudi, I. Prigogine and G.W. Nelson, Phys. Lett. 114A (1985) 29-32. D.K, Kondepudi and M-J. Gao, Phys. Rev. A 35 (1987) 340-348. (a) V.I. Goldanskii and V.V. Kuz'min, Z Phys. Chem. 269 (1988) 216-274 and references therein, (b) VI. Goldanskii and V. Avetisov, Proc. Nat. Acad Sci. USA 93 (1996) 11435-11442. D.K. Kondepudi, Z Phys. Chem. 210 (1989) 843-844. G. Nicolis and I. Prigogine, Proc. Nat. Acad Sci. USA 78 (1981) 659-663. C. Cothia, in: Biological Asymmetry and Handedness, Ciba Foundation Symposium, John Wiley, New York, 162 (1991) 36-49. B.N. Thomas, B.N. Lindemann and N.A. Clark, Phys Rev. E59 (1999) 3040-3047. P.W. Anderson, Science 111 (1972) 393-396.

Progress in Biological Chirality G. Palyi, C. Zucchi and L. Caglioti (Editors) © 2004 Elsevier Ltd. All rights reserved.

Chapter 14 Possible Mechanisms for Production of Large Enantiomeric Excess David Z. Lippmann' and Julio Dix*" "" Center for Studies in Statistical Mechanics and Complex Systems, I University Station CI609, University of Texas, Austin, Texas 78712, USA lippmann@physics. utexas. edu ^ Department ofMathematics, Southwest Texas State University, San Marcos, Texas 78666, USA

1.

Introduction The origin of biological homochirality is puzzling because processes that produce chiral species always produce a racemic mixture unless there is some non racemic component in the process, such as an enantiomericly pure or nearly pure reactant or catalyst. But the enantiomericly pure reactant or catalyst must itself have been produced by a process that included an enantiomericly pure reactant or catalyst, and so forth, so we seem to be faced with an infinite regression. How was the first substance that was chiral but not racemic produced? Several asymmetries in nature that might have led to excess production of one of a pair of enantiomers have been suggested, but they could produce only a very small enantiomeric excess (defined to be ([S] - [R]) / ([S] + [R]) where S is the sinister isomer and R is the rectus isomer and the square brackets denote concentration). Furthermore, the enantiomeric excess tends to decay toward zero as the mixture approaches the equilibrium, racemic state. There are two conditions that must exist in order for a small initial enantiomeric excess to survive and increase. There must be a source of free energy that keeps the system far from equilibrium. Possible sources of free energy include sunlight, lightning, radioactivity, and temperature or concentration gradients. There must also be some mechanism that amplifies the initial enantiomeric excess. In our previous paper we discussed several possible processes that can amplify an enatiomeric excess [1]. In this paper we refine and extend those discussions. Any process that leads to large enantiomeric excess must begin with production of a compound or compounds that have high standard free energy and can react to form chiral molecules. We represent the lowfi-eeenergy starting materials by L and the high free energy materials by H. CO, CO2, H2O, CH2O, NH3, HCN, and H3PO4 are examples of L substances. (In this paper H does not represent hydrogen except in the formulas in the preceding sentence.) H could be small carbohydrates, simple amino acids, and other compounds that can react to produce polysaccharides, polypeptides, and other components of living organisms.

174 Progress in Biological Chirality The reactions are: production of H

L ^> H

and decay of H

H -> L

rate = ki[L]

(1)

rate = k2[H]

(2)

The H molecules might be chiral, but for simplicity we have assumed that chirality arises in a separate step: H^ S and H -^ R rate = kss [H] and rate = k3R [H] (3) There might be a consistent small difference between kss and ksR or they might be equal except for small, random fluctuations. S and R decay back to L: S^ L rate = k4[S]

and and

R -> L rate= k4[R]

(4)

The decay reactions (4) preserve the ratio [S] /[R] but tend to reduce any difference ([S] [R]) toward zero. Inversion of S and R, S -> R rate = ks ([R] - [S])

and and

R -> S rate = ks ([S] - [R]),

(5)

also tends to eliminate any deviation from a racemic mixture. Any process that amplifies an initial difference between [S] and [R] must overcome the effects of reactions (4) and (5).

2. Reactions That Can Amplify Enatiomeric Excess Autocatalysis by monomers: Frank showed that if S and R catalyze their own production, H + S->S + S rate = k,[H][S]

and and

H + R->R + R rate - k^MLR],

(6)

an initial difference between [S] and [R] will increase with time [2]. However, autocatalysis alone will leave the ratio [S] / [R] unchanged. Formation of oHgomers: n S ^ Sn rate = k7[S]"

and and

n R ^ Rn rate= k7[Rr

(7)

and and

Rn->nR rate = kgLRn]

(8)

and dissociation of oligomers: Sn^nS rate =^ kgLSn]

Possible Mechanisms for Production of Large Enantiomeric Excess

175

If [S] is greater than [R], Sn will be produced more rapidly than Rn, and the difference ([S] "^ i^[Sn]) - ([R] - n[Rn]) will increase with time. If the oligomers are more stable than the monomers, formation of oligomers will reduce the effects of reactions (4) and (5). If the oligomers are more stable than the monomers and are also catalysts for the production of their respective monomers, H + S„->S + Sn rate = k9[H][Sn]

and and

H + Rn->R + Rn rate = k9[H][R,,],

(9)

then catalysis by the oligomers will be more effective at increasing enantiomeric excess than autocatalysis by the monomers. Avetisov has pointed out that production of even a single large homochiral molecule is extremely improbable unless there is already a large difference between the concentrations of the enantiomeric monomers [3]. For example, if the mixture of monomers is racemic and if S and R add randomly to growing chains, the probability that a 50-mer will be homochiral, S50 or R50, is {Mlf^ = 2 X 10"^^ and the probability that a 100-mer will be homochiral is only 2 x 10'^^. If homochiral addition is faster than random addition, then the probability of forming homochiral oligomers is increased. Nevertheless, it is apparent that a large enantiomeric excess would have been necessary for production of homochiral oligomers as large as modern enzymes. The oligomers in reactions (7), (8), and (9) must have been small. There is a possible exception to the conclusion that the earliest catalytic oligomers must have been small. Suppose that the probability of forming a moderately large homochiral molecule, Sn or Rn, is so small that the rate of formation is not represented accurately by a continuous rate equation. Eventually a single Sn or Rn molecule is produced. Assume that it is Sn. If it is a catalyst for production of S from H, the rates of production of S from H and Sn from S increase initially, but as the concentration of H decreases the rates of reactions (7) and (9) decrease until they reach the limit imposed by the rate of production of H from L. The rate of uncatalyzed production of R from H decreases as the concentration of H decreases, and the probability that even one Rn molecule will form becomes even smaller than it was initially. The increased concentration of S and the decreased concentration of R increase the probability that additional homochiral Sn molecules will be produced. Kondepudi and his coworkers have carried out experimental studies of a system analogous to the one described in the previous paragraph [4, 5, 6, 7]. They started with an aqueous sodium chlorate solution. The individual sodium ions and chlorate ions in the solution are not chiral, but solid sodium chlorate crystals are chiral. Water was evaporated until the solution became highly supersaturated with respect to macroscopic crystals but only slightly supersaturated with respect to microscopic crystals. Eventually a single microscopic crystal formed spontaneously. It grew rapidly by incorporating ions from the solution. Rapid stirring caused microscopic daughter crystals to break away from the growing crystal, and the daughter crystals also grew and produced new generations of microscopic crystals. The stirring caused the decrease in concentrations of free sodium ions and chlorate ions to be uniform throughout the solution. As water continued to evaporate, the solution became and remained only slightly supersaturated with respect to macroscopic crystals but unsaturated with respect to microscopic crystals, so that only crystals with the same chirality as the first one could be formed. Repeated experiments produced either S or R crystals randomly, but only one enantiomer was produced in each experiment.

176

Progress in Biological Chirality

Mutual destruction of enantiomers: The most plausible mutual destmction process is formation, decomposition, and decay of mixed dimers. S + R -> SR rate = kio[S][R] (10) SR ~> S + R rate = kii[SR] (11) SR->2L rate = k,2[SR] (12) Frank demonstrated that a combination of autocatalysis and mutual destruction would cause enantiomeric excess to approach unity if the decay reaction (4), the inversion reaction (5), and the decomposition reaction (11) were negligible [2]. In our previous paper we showed that mutual destruction alone can cause the enantiomeric excess to approach unity if the supply of H is finite, even if the only difference between the rates of production of S and R is random fluctuation [1]. If the supply of H is infinite, random fluctuations will cause the enantiomeric excess to swing between positive values (excess S) and negative values (excess R). As [S] and [R] increase, larger absolute values of the enantiomeric excess will become more common and last longer. Mutual inhibition of chiral catalysts: Let Cs and Cr represent a pair of enantiomeric catalysts. Cs catalyzes production of S, and Cr catalyzes production of R. [Cs] and [Cr] represent concentrations if Cs and Cr are homogeneous catalysts or number of active sites if Cs and Cr are heterogeneous catalysts. [Cs] and [Cr] are equal initially. Then the following reactions occur: H + Cs ^ HCs rate = k,3[H][Cs]

and and

H + Cr -> HCr rate =k,3[H][Cr]

(13)

HCs -> H + Cs rate = k,4[HCs]

and and

HCr -> H + Cr rate = kM[HCr]

(14)

HCs —^ SCs rate = ki5[HCs]

and and

H C r - > RCr rate= k,5[HCr]

(15)

SCs —^ HCs rate= ki^SCs]

and and

RCr -> HCr rate-ki6[RCr]

(16)

SCs -> S + Cs rate = k n [SCs]

and and

RCr -> R + Cr rate = k]7[RCr]

(17)

S + Cs -> SCs rate = k,8[S][Cs]

and and

R + Cr -> RCr rate = ki8[R][Cr]

(18)

S + Cr ^ SCr rate = k,9[S][Cr]

and and

R + Cs -> RCs rate = ki9[R][Cs]

(19)

Possible Mechanisms for Production of Large Enantiomeric Excess

SCr -> S + Cr rate = k2o[SCr]

and and

RCs --> R + Cs rate = k2o[RCs]

177

(20)

If ki9 is larger than kig, or kio is smaller than kn, or both, each enantiomer will inhibit catalytic production of the other more than it inhibits its own production, and if there is an excess of either enantiomer, that enantiomer will be produced faster than the other.

3.

Limits to Enantiomeric Excess Any of the processes discussed in the previous section, or any combination of them, will amplify an initial difference between [S] and [R], but the effect will be limited. The reactions that increase the differences between [S] and [R] and between [Sn] and [Rn] become slower as H is consumed. The reactions that tend to equalize the concentrations of S and R and the concentrations of Sn and Rn by converting S, R, Sn and Rn to L accelerate as the concentrations of S, R, Sn and Rn increase. The inversion reactions (5) decrease the absolute value of ([S] - [R]). Catalysis may not be perfectly enantioselective, even if the catalyst is chiral [8]. We have not considered imperfectly enantiospecific catalysis explicitly because it is equivalent to the inversion reactions. Mutual inhibition of racemic catalysts is self limiting, because as more and more Cs and Cr are inactivated by conversion to RCs and SCr the rates of catalyzed production of S and R both decrease. The extreme case of this would occur if SCr and RCs never decomposed (k2o - 0). Even if there were a large difference between [S] and [R], eventually both Cs and Cr would be completely converted to RCs and SCr and the difference between [S] and [R] would decay. Even with the concentration of H maintained above its equilibrium value by an external source of free energy, enantiomeric excess can only increase to a maximum value less than unity and then either remain constant or decrease as a steady state is approached. However, an enantiomeric excess close to unity can be attained by successive mutations of a catalytic species. Suppose that an alteration in the structure of a single S, R, Sn, or R„ molecule increases the catalytic efficiency of that molecule or makes it more stable. Copies of the altered molecule will gradually replace the unaltered molecules and will become more numerous than the unaltered molecules were. Such favorable mutations will occur randomly, but the rate of mutation will be proportional to the concentration of the species that is modified by the mutation. Successive favorable mutations will cause one species to replace the original S, R, Sn and Rn almost completely. Darwinian evolution can occur, even at the prebiotic stages.

4.

Numerical Examples The sets of simultaneous differential equations that describe the time evolution of an initially racemic system were solved numerically for a number of different combinations of rate constants. Reactions that are the reverse of reactions (3), (4), and (6) were not included. These reverse reactions would have large positive standard free energy changes, so the rates of these reverse reactions would be negligible relative to the rates of their respective forward reactions. The algorithm used was NDSolve of Mathematica 4. These computations confirmed the qualitative arguments that were presented in the preceding sections. The resuhs of four of these computations are shown in Tables 1, 2, 3, and 4. The units of

178

Progress in Biological ChiraliU

concentration, time, and the rate constants are arbitrary, but they were consistent over all the calculations. Table 1. Number of amplification Rate Constants

Initial Concentrations

k,= 1000 k.= 1 1.01 k3S = 1 k3R = 1 k4 = 0.1 ks-

[L] = 1000000 [H]= 0 [S]= 1 [R]= 1

Steady State Concentrations [L] = 999.003 [H] =331895.0 [S] =334938 [R] =332170

Table 2. Catalysis by dimers Rate Constants

Initial Concentrations

k,= 1000 1 k2 = 1.01 k3S ~ 1 ksR = 1 k4 = 0.1 ks = k.= 1 1 k8 = ^10000

[L] =1000000 [H] = 0 1 [S] = IR] = 1 0 ISzl^ 0 lS2l = 0 [R2I-

Steady State Concentrations [L] = 50587.5 [H] = 0.0107 [S] 4340.56 [R] = 2935.81 [S2] = 325885 [S2] = 325885 [R2l = 145184

Table 3. Mutual destruction Rate Constants

Initial Concentrations

k,= 1000 k.= 1 1.01 k3S = 1 ksR = 1 k,= 0.1 ks = kio = 1000 1 kn = 1 ki2-

[L] =1000000 [H]= 0 tS]= 1 [R]= 1 [SR] = 0

Steady State Concentrations [L] = 999.003 [H]= 857821.0 [S]= 698.131 [R] = 1.3294 [SR] = 70241.3

Table 4. Catalysis by dimers, self protection by formation of dimers and mutual destruction Rate Constants

Initial Concentrations

1000 1 1.01 k3S = 1 k3R = 1 k4 = 0.1 ks = 100 k7 = 0.001 k8 = 100 k9 = 0.001 kn = 0.001 ki2 =

|LJ = 1000000 0 [H] = 1 [Sl = [R] = 1 0 tS2] = 0 1R2] = 0 [SR1 =

ki =

k.=

Steady State Concentrations [L] = 46.4344 [H]= 0.001164 [S] = 15.2067 [R] = 2.5538 [82] = 364793 [R2]= 34096.7 [SR] =101079

Possible Mechanisms for Production of Large Enantiomeric Excess

179

5.

Conclusions There are several mechanisms that can cause enantiomeric excess to increase in a system that is initially racemic or achiral. None of these can increase the enantiomeric excess to unity, and some of them produce only a transient increase that decays to a steady state having zero or very small enantiomeric excess. Mutual destruction is capable of producing enatiomeric excess that is close to unity, but mutual destruction causes the steady state concentrations of both S and R to be small. One of the most effective mechanisms for production of enantiomeric excess close to unity is random formation of a single homochiral oUgomer molecule that catalyzes production of its own monomers. Two or more mechanisms acting together are more effective than any one of them alone. Even a small or transient enantiomeric excess can lead to a large, permanent one because the chiral species can undergo favorable mutations, and the more abundant enantiomers will have greater probability of mutating. The relative importance of the several amplification mechanisms depends on the values of the rate constants. These values will have to be determined by experiment for specific reactions.

6.

Acknowledgements This work was partially supported by the Ilya Prigogine Center for Statistical Mechanics and Complex Systems, University of Texas at Austin and the Department of Mathematics, Southwest Texas State University.

7.

References

[1] D.Z. Lippmann and J. Dix, Possible mechanisms for spontaneous production of enantiomeric excess. In: Advances in BioChirality (Eds. G. Palyi, C. Zucchi and L. Caglioti) Elsevier, Amsterdam, 1999, pp. 85-97. [2] F.C. Frank, Biochim. Biophys. Acta 11 (1953) 459-463. [3] V.A. Avetisov, Origin of biological homochirality: directed selection or random motion? In: Advances in BioChirality (Eds. G. Palyi, C. Zucchi and L. Caghoti) Elsevier, Amsterdam, 1999, pp. 69-83. [4] D.K. Kondepudi, R.J. Kaufman and N. Singh, Science 250 (1990) 975-977. [5] D.K. Kondepudi, K.L. Bullock, J. A. Digits, J.K. Hall and J.M. Miller, J. Am. Chem. Soc. 115 (1993) 10211-10216. [6] D.K. Kondepudi and C. Sabanayagam, Chem. Phys. Lett. Ill (1994) 364-368. [7] D.K. Kondepudi, K.L. Bullock, J. A. Digits and P.D. Yarborough, J. Am. Chem. Soc. 117 (1995) 401-404. [8] V.A. Avetisov, in: Physical Origin of Homochirality in Life (Ed. D. Cline) AIP Press, Woodbuiy, New York, 1996, p. 141.

Progress in Biological Chirality G. Palyi, C. Zucchi and L. Caglioti (Editors) © 2004 Elsevier Ltd. Allrightsreserved.

Chapter 15 Sugar C-suifonic Acids Andras Liptak Research Group for Carbohydrates of the Hungarian Academy of Sciences, H-4010 Debrecen, P.O. Box 55, Hungary liptaka@tigris. kite, hu

1.

Introduction "Glycolipids are glycosyl derivatives of lipids such as acylglycerols, ceramides and prenols. The term glycolipid designates any compound containing one or more monosaccharide residues bound by a glycosidic linkage to a hydrophobic moiety such as an acylglycerol, a sphingoid, a ceramide (N-acylsphingoid) or a prenyl phosphate" [1]. This chapter focuses mainly on the occurrence of a special group of glycodiacylglycerolipids, namely the sulfoquinovosyldiacylglycerols (SQDG). The SQDGs differ from each other only in their fatty acid composition. The first representative of the SQDGs was isolated in 1959 from Chlorellapyranoidosa [2]. Additional sources were found in photosynthetic organisms, such as algae, cyanobacteria [20], sponge [11], starfish [12] and surprisingly in the bacteroid forms of Bradyrhizobium Japonicum [13], because this plant glycolipid is rarely found in bacteria. Structural studies [14-17] showed that these compounds were l,2-di-0-acyl-3-0-(6-deoxy6-sulfo-a-D-glucopyranosyl)-L-glycerols. These structures were confirmed by X-ray crystallographic studies [18] of the rubidium salts of the deacylated glycolipids which were also verified by the ^^C-NMR spectra [19]. The fatty acid distribution in the natural lipid from various sources has been examined [20-22] and as in most phospholipids and glycolipids [23], a wide spectrum of fatty acids was found. Hexadecanoic acid was found to be present predominantly at the C-2 OH group of the glycerol moiety and mono-, di- and trienic derivatives of octadecanoic acid occurred preferentially at the OH-1 of glycerol. 2.

Synthesis Miyano and Benson [17, 19] reported the first synthesis of 6-deoxy-6-sulfo-Dglucopyranose, as well as that of 3-0-(6-deoxy-6-sulfo-a-D-glucopyranosyl)-L-glycerol. The first preparation of 3-0-(6-deoxy-6-sulfo-a-D-glucopyranosyl)-1,2-di-O-hexadecanoyl-Lglycerol, the "sulfoquinovosyl diglyceride" was described by R. Gigg and his co-workers [24]. During the last decade, it turned out that the sulfoquinovosyl mono- (SQMG) and

182 Progress in Biological Chirality

Figure 1 diacylglycerols (SQDG) indicated strong inhibitory activities against eukaryotic DNA polymerase a and b [25, 26], They possessed extensive biological activities such as antitumour effects [27], P-selectin receptor inhibition [28], inhibition of HN-RT [25, 29] and AIDS-antiviral [10] properties. These biological indications initiated the syntheses of more than ten SQDG and SQMG compounds [24,30-32] and their biological investigations were achieved (Figure 1). The structure elucidation and the synthesis of 3-0-(6-deoxy-6-sulfo-a-D-glucopyranosyl)1,2-di-O-acylglycerol required a strong synthetic capacity and different processes were worked out for the preparation of 6-deoxy-6-sulfo-D-glucose. Partially or fully protected 6-0sulfonyl(tosyl [15, 16], mesyl, triflyl)-D-glucose or -glycoside were treated with nucleophilic reagents, such as NaHSOs [15, 16, 19, 34, 35, 37] or CH3COSK [24, 30, 31, 32, 38, 39]. In the first case 6-C-sulfonyl compounds were directly obtained, and in the second reaction the 6-Sacetyl intermediates were oxidized to 6-C-sulfonic acids by peracids, H2O2, Oxone or (NH4)6Mo7024.4H20. In another approach alkyl 6-deoxy-D-x);/o-hex-5-enopyranosides were treated either with Na2S03 or with CH3COSH under free-radical addition conditions to obtain directly the glucose-6-suIfonic acid or glucose-6-thioacetate which was oxidized to glucose-6C-sulfonic acid (Figure 2). Sulfoquinovose is the defining constituent of photosynthesizing organisms and an important component of the biological sulfur cycle [40]. The presence of other sugar sulfonic

Sugar C-sulfonic Acids 183 •

PO'

^SAc

SO3 Na O @

^^"Ai-*"*T
'OMe(H)

OP

OP

OP 15

16

14

\

aR= Ts bR= Ms cR= Tf

m-CPBA; H2O2 Oxone (NH4)6M07024 4H20 P= protective group(s)

SofNf^

Deprotection

Deprotection

CH2 15

NaHSO.;

po--'V^^-^^\ ^ ^ ' ^ • ^ "^T ^^^—0Me(H) OP 17 Figure 2

acid in Nature is very limited. Cell-wall hydrolysates of Halococcus sp., strain 24 contain 2amino-2,6-dideoxy-6-sulfo-D-glucose [41]. Its existence in Nature was questionable hecmse when O- glycoproteins are treated with alkali, the 0-glycosidic linkages between seftine or reducing end of the oligosaccharide chain and leaving the corresponding unsaturated amino acid in the peptide chain [40]. If this treatment is carried out in the presence of sodium sulfite; this reagent results in the release of sulfited oligosaccharides identified as Nacetylsulfohexosamine. The sulfohexosamine moiety is formed via unsaturated intermediates from 3- and 6-0-substituted 2-acetamido-2-deoxy-D-galactosyl residues furnishing 2acetamido-2,3-dideoxy-3-sulfo- and 2-acetamidO"2,6-dideoxy-6-sulfo-hexose [43].

184

Progress in Biological Chirality

The configuration of neither the sugar component of the cell wall hydrolysate of Halococcus sp., nor the 2-amino-2,3-dideoxy-3-sulfonic acid-D-hexose has been identified in the frame of a synthetic program. A simple synthesis was elaborated to prepare 2-amino-2,6-dideoxy-D-glucopyranose-6sulfonic acid. Oxidation of 2-acetamido-l,3,4-tri-0-acetyl-6-S-acetyl-2-deoxy-6-thio-p-Dglucopyranose [44] (19) with 30% hydrogen peroxide in acetic acid [45] gave a mixture of 20 and its 3,4-diacetate (21). The structures and conformations of both compounds were determined by ^H and ^^C-NMR spectra [45], and the structure of the a-methyl glycoside of compound 21 was verified by X-ray analysis [47].

^so?

so? H.O: 2^2 OAc

A c O - ' ^1 ^ 0 ^ \ iiin AcO^ .—^\. ^ + •^^^^OH •—-^T'V,,OH

AcOH

NHAc

kH3®

kH3®

20

19

21

Weber and Winzler [42] worked out a procedure for the preparation of 2-amino-2,3,(4)diedoxy-3-D-hexopyranose-3-sulfonic acid. The configuration of the compound could not be determined exactly. Fernandez-Bolanos and his co-workers [48] modified the method and assigned D-gluco configuration. They described the taurine analogues 22 and 23 by reduction of 2-amino-2,3 (and 2,6)-dideoxy-D-glucopyranose-3 (and 6)-sulfonic acids (24 and 25) with sodium borohydride. The structure of sulfo-aminohexitols (22 and 23) were assigned on the basis of analytical, IR, ^H ^^C NMR, and MS data. The structure of 23, crystallized as a monohydrate, was confirmed by X-ray analysis, too. ,—OH

I—OH —NH3"@

-NH3® NaBHi

—OH

—OH

—OH

I—OH -sof

mf

1—OH 24

HO—

22

25

23

Synthesis of sodium 2-acylamino (octanamido, dodecanamido, hexadecanamido)-2,6dideoxy-D-glucopyranose-6-sulfonates was reported by N-acylation of 2-amino-2,6-dideoxyD-glucopyranose-6-sulfonic acid. Oxidative transformation of l,3,4-tri-0-acetyl-6-S-acetyl-2deoxy-2-octanamido-6-thio-a-D-glucopyranose by hydrogen peroxide was also successful [46]. Oxidation of sulfoquinovose by hypoiodite gave 6-deoxy-6-C-sulfo-D-gluconic acid, isolated as the crystalline bis-cyclohexylammonium salt. Its ^H- and ^^C-NMR data confirmed the declared structure [50].

Sugar C-sulfonic Acids

185

3.

Synthetic, Unnatural Sugar C-sulfonic Acids Among the charged sugar oHgomers and polymers many biologically active compounds can be found. The charge can arise either from the sugar building blocks, such as uronic acids (plant polysaccharides, glycosaminoglycans), ulosonic acids (O- and N-glycoproteins, ligands for sugar binding proteins), muramic acids (bacterium cell-wall peptidoglycans) or from sugar esters, like sugar phosphates (DNA, RNA, nucleotide phosphates, sugar diphosphonucleotides) and sugar sulfates (proteoglycans, sulfated oHgosaccharides). The subject of this chapter is the group of sugar sulfonic acids, and the arising question is that how can the sugar sulfate esters be replaced by sugar sulfonic acids. The Introduction of this review explained that the chemistry of sugar sulfonic acids is equal to the chemistry of 6sulfo-D-quinovose. The chemistry of sugar sulfonic acids in a broader sense has to be elaborated. i. 1 Sugar sulfoulosonic acids Selectins are carbohydrate-binding transmembrane proteins expressed on platelets (Pselectins), leukocytes (L-selectins) and endothelial cells (E- and P-selectins). Their role is to mediate the first steps of the recruitment of leukocytes from the blood stream in a series of pathologic situations [51]. Carbohydrate ligands recognized by these selectins have been identified: E-selectin recognises the sialyl Lewis X (sLe^) tetrasaccharide [52] on the surface of leukocytes, P-selectin also binds sLe^ on leukocytes [53], and L-selectin weakly recognises sLe^ on endotheHal cells [54]. The selectin-carbohydrate interaction appears at the very early stage of inflammatory reactions or metastasis. A well-organized interplay among different proteins results in the recruitment of leukocytes, and the intracellular adhesion molecule with the endothelial cells allow the extravasation of neutrophils to the site of injury. When too many leukoc5^es are recruited, however, normal cells will also be damaged, causing inflammation. Control of this process by inhibiting the adhesion steps has been considered a new anti-inflammatory strategy [55]. Since one of the major natural ligands of selectins is the sialyl Lewis X tetrasaccharide, this tetrasaccharide (26) may be considered a central structure for the development of glycomimetics which structurally resemble and functionally mimic natural oligosaccharides. These compounds, designed as selectin receptor antagonists, are currently being evaluated as potential antiadhesive, anti-inflammatory and anti-metastatic drugs [56, 57]. Not only the sLe^ tetrasaccharide, but also the sulfated Lewis X trisaccharide (27) is a natural ligand of selectins [58]. This observation suggested the preparation of the carboxymethyl Lewis X (28) which is one of the most successftil drug candidates.

HO

OOH

0

loAi^^o"

NHAc

HO' AcHN

26

LR)

27

[R]=H03S-

28 [1]=-CH2-C00H

186

Progress in Biological Chirality

3.2 Sialyl Lewis X mimics To replace either the sialic acid moiety of the sLe^ tetrasaccharide (26) or the sulfate ester of the sulfated Lewis X (27) by sugar sulfoulosonic acid residue, the thioglycosides of three sulfoulosonic acids were synthesized [59]. The D-glucono-, D-mannono- and Lfliconolactone derivatives were reacted with ethyl methanesulfonate anion generated with nbutyllithium. Upon nucleophilic addition of the sulfonate ester carbanion to the lactone carbonyl, l-ethylsulfonyl-hept-2-uloses were obtained in a-anomeric forms. Reaction of hept2-uloses with ethanethiol or thiophenol in the presence of Lewis acids resulted in the formation of a-thioglycosides [60]. ^OBn OBn

OBn

OBn

BnO-V--'^, BnO \\)

^ 6 H

29

30

31aX = Et 31bX=Ph

Different aglycons, such as methyl 2,3,4-tri-O-benzyl-a-D-glucopyranoside, methyl and benzyl 2,6-di-O-benzyi-P-D-galactopyranosides and the partially protected pseudotrisaccharide (32) were glycosylated with thioglycosides of sulfoulosonic acid(s) (31a). OBn ,CH2-S03Et HO

OBn

OBn

33

.OH © ® -CHz-SOjN Bu4 HOA.. ^ OH HO OBn

OH

0^

1

^"^Zr^H 1 OH

34

OH

L^OBn OBn

Q^

Sugar C-sulfonic Acids 187 Deprotection of the pseudotetrasaccharide 33 gave the sulfoulosonic acid containing Lewis X mimic (34), its ID50 was in the mM region, which is comparable with the activity of sLe^. This result shows that the negative charge of the C-sulfonate residue might take over the role of the carboxyl group of the sialic acid residue [60]. 3.3 Mimics ofligands of Helicobacter pylori Very similar biological motivation initiated the synthesis of some sulfonated lactose derivatives. The microorganism is Helicobacter pylori, which is a causative agent in chronic active gastritis, gastric and duodenal ulcers, and presumably gastric malignancies [61]. Today therapy is based on antibiotic treatment in combination with a proton pump inhibitor. This treatment resuhs in a very low recurrence. Extension of this type of treatment is not recommended, owing to the resistance to antibiotics [62]. Efforts to develop vaccines till now are unsuccessful [63]. Many researchers believe in the success of an antiadhesion therapy based on carbohydrate receptor analogs. Some bacteria carry genes that encode more than one adhesion molecule, each capable of binding a different carbohydrate receptor, which is a molecular mimicry between the microorganism and the host. The carbohydrate-binding specificity of H. pylori is unusually high [64]. More than ten different kinds of carbohydrate-binding specificity have been observed and described. Among them there are sialic acid types of specificity, such as aNeu5Ac-(2->3)-p-D-Galp-(l->4)-D-Glc [65], a-Neu5Ac-(2->3)-p-D-Galp-(l->4)-DGlcNAc; sulfated mono- and oligosaccharides: 3-0-sulfate-D-Gal [66], 3-0-sulfate-(3-DGalp-(l->OCer) [67], 3-0-sulfate-Lewis X and 6-O-sulfate-D-GlcNAc [66]. Some 3'-C-sulfonated lactose derivatives were synthesized to obtain inhibitors of adhesion in the case ofH pylori. These syntheses started from the same compound, namely, from/?methoxyphenyl 2,3,6,2',6'-penta-0-benzyl-p-lactoside. In the first case Bu2SnO-mediated allylation [68] resulted in the 3'-0-allyl ether (88%), then sulfite addition led to the 3'-0(sodium sulfonato propyl)-lactoside derivative. The reaction was catalysed [69] with tertbutyl peroxybenzoate [70]. Kharasch and co-workers [71] established that this reaction can be interpreted on the basis of a free-radical mechanism, and the reaction goes anti-Markovnikov, resuhing in 1-alkane-sulfonates. Later, the mechanism of the reaction was investigated [72] and a mechanism was proposed. Removal of the benzyl protecting groups by catalytic hydrogenolysis, the end-product (35) was obtained [73] in an excellent yield (96%). HO

/OH

Na03S~CH2~CH,-CH,-0..\^\.

O^ x

-

x

o -/.^ )vc ^CH

Starting from/7-methoxyphenyl 2,3,6,2',4',6'-hexa-0-benzyl-P-lactoside [73], the 3'-ulose derivative was obtained by Dess-Martin periodination [74], which was reacted with ethyl methylene-sulfonate anion generated with n-butyllithium and two epimers were formed and separated in a ratio of 4:1. The gw/o-isomer (37), in which the -CH2-S03Et is equatorial, was

188

Progress in Biological Chirality

transformed into the tetrabutylammonium salt and was deprotected to obtain the free glycoside (39). BnO anu

^OBn won

^^^

BrO L

®CH SO B

^OBn <

<^^^

36

37

(71%)

Hj/Pd-C ftK)

^OBn

OBn

HO

-O-<0^*' ^ " ^

/OH

OH .0 HO

BU4N O 3 S - C H 2

)--<§)-OCH3

OH 38 (18%)

OH

39

The P-D-gulopyranosyl configuration of the ultimate monosaccharide of the major product was determined on the basis of the NMR CHi-H-T and CH2-//-4' carbon-proton coupling constants, that depend on the dihedral angle in a manner similar to ^JH,H. The small values of both coupling constants verified the equatorial arrangement of the CH2-S03Et group. The third type of lactoside derivatives were obtained by the glycosylation of pmethoxyphenyl 2,3,6,2',6'-penta-0-benzyl-P-lactoside with ethyl 3,4,5,7-tetra-O-benzyl-ldeoxy-1 -ethylsulfonato-2-thio-a-D-gluco-hept-2-ulopyranoside (31a) using NIS-TfOH activation. The reaction afforded a separable mixture of the two trisaccharides (40 and 41), but unfortunately, the ethyl ester could not be transformed into ammonium salts. Direct reduction of the esters of 40 resulted in complex mixtures and also the interglycosidic bonds were cleaved. The effort to isolate the free trisaccharide glycosides was unsuccessful. The sulfonate esters are rather stable under hydrolytic conditions [75], the isopropyl esters are more versatile for synthetic purposes [76,77], they can readily be removed by treatment with boiling methanolic ammonia [75]. OBn OH

OBn

OBn •OCH3

OBn 40(47%)

OBn BnOBnO-

OBn

-SOaEt

BnO 6

QBn

^g^^

41(11%)

\^=/

Sugar C-sulfonic Acid

189

4. Secondary Sugar C-suIfonic Acids Comparing sugar sulfonic acids with sugar sulfates, they can be divided into three groups: O Sugar—S—O—Na

A Sugar—0—S—O—Na O H O Sugar—C—S—O—Na

i iS The first type of compounds is sugar sulfonic acid, the second type of compounds is a sugar monoester. The two compounds are not in isosteric relation. The third type of compounds is a sugar-methylene-sulfonic acid. This last compound is in isosteric relation with the sulfate ester. In the first part of this chapter it has been mentioned that secondary sugar sulfonic acids had not been found in Nature. 2-Amino-2,3-dideoxy-3-C-sulfo-D-glucopyranose was synthesized by elimination-addition reactions, starting from 2-acetamido-2-deoxy-Dglucopyranose [48]. Most recently methyl 2-deoxy-2-C-sulfo-P-D-glucopyranose 49 and methyl 2-deoxy-2-Csulfono-a-D-mannopyranoside 45 have been prepared and characterised [78]. This approach is based on a stereospecific 1,2-alkyl/arylthio group migration [79-82]. Thioglycosides (42 and 46) having a good trans leaving group at position 2 react in the presence of a nucleophile to give 2-thioalkyl/aryl l,2-rraA/5:-glycosides (43 and 47). The acid sensitive 2-thioether group(s) can be oxidized directly to 2-deoxy-2-C-sulfo-glycosides (44 and 48) [78]. Three different thioglucosides were rearranged into 2-thio(/?-methoxybenzyl)-, 2-thio(2'naphthylmethyl)and 2-thiotrityl-a-D-mannopyranosides. The thiotrityl a-Dmannopyranoside derivative was converted into a methyl 2-thiotrityl-P-D-glucopyranoside surrogate. The one-pot oxidation of 2-thiotrityl analogs with Oxone [83] was the most successful procedure, and the 2-sulfoglycosides were obtained in high yields. SO3

STr

ph^ro

O

o

NaOMc

Q

STr

S(?3

O

^O BnO

BnO

BnO OMs

OMs

O

6 BnO

6

-HOOMe ^

BnO

•so?

•so?

190

Progress in Biological Chiralitv

This method is a valuable route to prepare 2-sulfonic acids of the a-manno-l^-gluco-, aallo-l^-altro-, a-ido-l^-gulo-^ and a-Za/o-ZP-ga/ac/o-epimeric pairs. Secondary sugar sulfonic acids can also be prepared by nucleophilic displacement reactions. Methyl 2,3,6-tri-O-benzyl-a-D-gluco- [84] (50) and -a-D-galactopyranoside [85] (51) were triflated, and the triflate esters were thioacetylated to invert configuration at C-4 to give methyl 4-S-acetyl-2,3,6-tri-0-benzyl-4-thio-a-D-galacto- [84] (52) and -a-Dglucopyranoside (53). The galacto configuration of compound 52, and the gluco configuration of compound 53 were confirmed by their ^H-NMR spectra. In the galacto isomer the ^J3,4 and ^J4,5 couplings are small, in the gluco-compounds the ^hA and ^J4,5 have high values. The resonance places of the C-4 signals were at high field, ^ 47 ppm. o II CH3-C-S

^OB

^OBn

HO--^ BnO^

^

Bno\

iv) Hj/Pd-C

ii) KSAc

50

HO

Na O3S iii) Oxone

i)Tf20

52

r^^OBn

\

Uo 54

^OBn iii)

i)

BnO^

HO^

,

BnO^

ii)

OBniMe

Na%S^

w -OHi, ^ Me

iv)

^B^Me

55

53

51

Oxidation of the thioacetyl groups was accomplished with Oxone, and the benzyl groups were removed by hydrogenolysis. Methyl 4-deoxy-4-C-sulfono-a-D-galacto- (54) and -a-Dglucopyranosides (55) were isolated as Na^ or NHEta^ salts. The ^^C-NMR data are summarized in Table 1. The tribenzylated methyl a-D-gluco- and a-D-galactopyranosides proved to be excellent starting compounds for the preparation of 4-C-sulfo-derivatives which possess very good solubility. These facts motivated the use of methyl 2,3,4-tri-O-benzyl-a-D-gluco- [85] (56) and -a-D-galactopyranoside [86] (57). The syntheses followed the route described above: triflation, thioacetylation, oxidation by Oxone and hydrogenolysis in the presence of Pd-on carbon catalyst. This reaction sequence furnished the methyl 6-deoxy-6-C-sulfo-a-D-

Table 1. '^C-NMR data of methyl 4-deoxy-4-C-sulfo-a-D-gluco- (55), -a-D-galacto- (54) and methyl 6-dcoxy-6-C-sulfo-a-D-gluco- (58) -a-D-galactopyranosides (66) Compounds Methyl C-1 C-2 C-3 C-4 C-5 C-6 OCH3 a-Glcp-4-S03 a-Glcp-6-S03 a-Galp-4-S03' a-Galp-6-S03 Measured in CD3OD

100.77 100.83 100.82 101.50

73.75 73.41 70.18 69.98

70.00 74.97 71.96 71.42

63.49 74.91 61.87 72.43

70.00 69.41 70.37 68.55

63.98 54.98 63.50 53.47

55.80 55.95 55.77 56.06

Sugar C-sulfonic Acids

191

glucopyranoside (58) in a good yield. This well-functioned reaction route failed in the case of triflated methyl 2,3,4-tri-O-benzyl-a-D-galactopyranoside (59). In a complex reaction mixture, the major product was methyl 3,6-anhydro-2,4-di-0-benzyl-a-D-galactopyranoside (60). S&,%a 1) ii) iii)

iv)

OBn^Me

0H6Me

56

58

^OH

^OTf

BnO

BnO Tf20

^

KSAc >•

inO^

^ OBn(')Me 57

BnO^

^

0B^6Me 59

60

During the displacement reaction an intramolecular reaction took place. This type of reaction was observed earlier in the case of benzylated [87], or even methylated [88] galactopyranoside derivatives having a good leaving group at position 6. It was also observed that the 3,4-O-isopropylidene ring can prevent the '*Ci->4C^ conformation transition [89]. Methyl 2-0-benzyl 3,4-0-isopropylidene-a-D-galactopyranoside (61) was treated with trifluoromethanesulfonic anhydride and the product (62) was converted into a thioacetyl derivative (63), whose oxidation with Oxone furnished the 6-deoxy-6-C-sulfo compound (64).

192

Progress in Biological Chirality

When the triflate 62 was treated with Na2S03 in ethanol-water solution, compound 64 was obtained. Acid hydrolysis of the isopropylidene group (->65) and catalytic hydrogenolysis of the benzyl residue gave the methyl 6-deoxy-6-C-sulfo-a-D-galactopyranoside (66). Successful nucleophilic displacement reaction proceeded in the case of methyl 6-deoxy2,3-0-isopropylidene-4-0-triflyl-a-manno- (67) and -a-L-talopyranoside (68). Their nucleophilic displacement reaction with CH3-COSK resuhed in the change of configuration at C-4, thus methyl 6-deoxy-2,3-0-isopropylidene-4-S-acetyl-4-thio-a-L-talopyranoside (69) was isolated from the L-manrjo-compounds (67), and methyl 6-deoxy-2,3-0-isopropylidene4-S-acetyl-4-thio-a-L-mannopyranoside (70) was obtained from the L-/a/o-derivative (68). The yield of the latter compound was lower, and a considerable amount of elimination product (75) was formed. Compounds 69 and 70 were oxidized with Oxone to prepare the protected sugar sulfonic acids (71 and 72) from which the isopropylidene groups were removed by acid hydrolysis and the end-product salts (73) and (74) were isolated and characterized [90]. 0CH3

67

69 OCH3

68,

71 OCH3

70

0CH3

0CH3

0CH3

73 OCH3

72

OCH3

74

OCH3

5.

Anomeric Sugar Sulfonic Acids The first representative of this group was presented in 1968 when Japanese authors isolated nojirimycin from the fermentation broths of several strains of streptomyces. The aqueous solution of the crude nojirimicyn was saturated with sulphur dioxide, and the stable bisulphite adduct readily crystallized from the solution [91]. The structure (76) was described by the equilibrium formula. Twenty years later the total synthesis of (+)-nojirimycin and (+)-deoxynojirimycin have been accomplished [92], the authors used also the bisulfite adduct for the crystallization of the synthetic nojirimycin and they desribed the structure in a "piperidinose" form (77). Recently, GlcNAc thiazoline [93,94] tri-0-acetate (78) or tri-0-benzyl ether (79) were oxidized with m-CPBA in the presence of ethanol into the 1-C-sulfonate esters (82 and 83).

Sugar C-sulfonic Acids

193

.SO?

HO^

h-OH HOH h-OH UNH3 CH2-OH 76

^OR

^.OR

RO V

^ J

I

^^QH(2-10equ.v.)^ RO CH,Cl„-15«C,30mm

V ^ ^/^^ J 0^\^OEt

78 R=Ac

80 R=Ac(35%)-

79 R= Bn

81 R=Bn(28%)

82% -^

82 R= Ac (38%) 83 R=Bn(30%)

The intermediate sulfinates (80 and 81) were formed as single diastereoisomers. An anomeric C-sulfonic acid was obtained from 2,3,4,6-tetra-O-benzyl-l-S-acetyl-l-thio-aD-glucopyranose [95] (84) by the Oxone-mediated oxidation. The perbenzylated sulfonic acid sodium salt (85) was deprotected by catalytic hydrogenation to give the free sodium salt (86). During the synthetic procedure the a-anomeric configuration did not change. In their ^^C-NMR spectra the chemical shift values of the anomeric (C-1) carbon were: 84: 83.46 ppm; 85: 86.40 ppm; and 86: 88.15 ppm.

Oxone 56%

84

SAc

BnO BnO

OBn O

OH PdonC quant.

OBn SOsNa 85

HO^V^^X OH I 86

SOaNa

The anomeric C-methylene-sulfonic acids were also synthesized; l-deoxy-2,3,4,6-tetra-0benzyl-l-exomethylene-D-glucono-l,5-lactone [96] (87) was treated [97] either with thioacetic acid or NaHSOa. In both cases anti-Markovnikov type addition occurred with complete P-stereoselectivity giving compounds 88 and 89. Oxone-mediated oxidation of 88 resulted in 89, whose reduction frirnished the 1-deoxy-l-C-methylene-P-D-glucopyranose sulphonic acid (90).

194

Progress in Biological Chirality ^OBn ^OBn

O O

OBn

OBn CHs-SOsNa

OBn

OBn^CHs

88

OBn

87

89

Oxone H./Pd-C

.OH HO HO

.CHs-SOgNa OH 90

6.

Carbohydrate-Derived Sultones The synthesis of sultones is generally carried out either by cyclization of the appropriate halo- or hydroxyalkane sulfonic acid, or by sulfonation of olefins with dioxane-sulfur trioxide [98, 99]. The preparation of a-sulfonate ester carbanions [100-102] opened a new epoch for obtaining the sultons. Treatment of either 1,2-ethylenedimethanesulfonate (91) or 1,3propylenedimethanesulfonate (92) in tetrahydrofuran at -78°C with one equivalent of butyllithium, followed by warming to 0°C for 30 min yielded sultones 93 and 94 in 68 and 88% yield [103], respectively. CH2 CH2-OS02~CH3 n-BuLi;THF H2C'''"'"^ \

i

H2-OSO2-CH3

•78'

SO2

0°C H:,C -0

93

91

CH2

CH2~OS02-CH3

X2C

n-BuLi; THF

XH2

CH2 CH2~OS02~CH3

•'^^

^

H2i^

io2 "0"

92

94

Alkanesulfonate esters of 1,2-halohydrins have also been applied for the synthesis of fivemembered ring sultones [103]. The first report about sugar-derived binuclear sultones was published in 1978 [104], when methyl 2-0-benzoyl-6-0-trityl-a-D-r/^o-hex-3-ulopyranoside (97) was mesylated with MeS02Cl-Et3N-CH2Cl2, and gave, in addition to the keto-sulfonate (98), the unsaturated sultone (99), whose hydrogenation resulted in the saturated sultone (100) exclusively. The authors supposed that the first step of the reaction was an aldol-type condensation and resuhed in the ammonium betaine (101), and then elimination gave 99.

Sugar C-sulfonic Acids

195

Ph CH2~C1

n-BuLi;

CH2-OS02-CH2Ph

THF

95

100

101 TsO

\OMs ^ X /(j)

CH^CU 98%

§02-0

ij^

103 TsO

.0.

1^

1

90%

o^s—0

104 ^OBs

A-

iJ^

105 O2S

NaHor o o

MsO"

106

m

^-V^OB 108

^^

196

Progress in Biological Chirality

This work initiated the examination of various mixed disulfonate ester derivatives of carbohydrates and showed that they undergo intramolecular displacement reactions, yielding sultones fused to the sugar ring. Some examples underline the importance of these investigations [105]. The procedure works well when a mesylate and a benzenesulfonate are used and can be applied for both pyranose and fiiranose sugar derivatives. The sugar sultones are very valuable synthetic intermediates and have been used especially in the field of nucleoside sultons. They were converted into different novel nucleotide analogues, where the phosphate residues were replaced by isosteric, non-charged sulphonate groups. Using the mixed sulfonate ester procedure [105], from compound 108a or 108b the thymidine sukone (109), and from 110a and 110b compound 111 was prepared [106]. Sultone 109 was generally much less susceptible to nucleophilic substitution at carbon-3' than sultone 111. CH.

CH.

NH

NH N^O

DMS0,2h 1^

0=S

108a R 108b R=Ms

\

109 CH.

CH

/

NH

NH N^O

RO—I

O Y

feMs yl N

EtgNH OsS

EtsNH'-OsS'^/^NJ

116

^

LiC^CH

DMS0.2h

/ 110a R=Ts 110b R=Ms

O.. %

^ P 111

Sugar C-sulfonic Acids

197

Th

EtjNH* 'OaS

TJ 117

LiNs 90''C DMF

NH4^ ' 0 3 8 ^ 0 O^V

These compounds, as novel isosteres of 5'-monophosphorylated nucleosides, may have valuable building stones to create modified DNA analogues, new type sulfonamides or sugarcontaining sulfopeptides.

7.

Miscellaneous: Sugar Sulfonic Acid Amide, Esters and Nucleotide Analogs The first sugar-sulfonic acid chloride was prepared by Lehmann and Reinshagen [107] who reacted it in various amines (aniline, 4-bromoaniline, 2-aminopyridine, 2aminopyrimidine and morpholine). The amides were isolated with a yield of 66-79% and all the compounds were crystalline. Spanish authors [108] prepared acid chloride either from glucosamine or from protected 1,2-0-isopropylidene-D-glucofuranose. Their reactivities were tested with primary amines or with simple alcohols (MeOH and isopropanol) as show in Figure 3. The fiiranosic acid chloride (124) resulted in a considerable amount of elimination products with primary amine or alcohol. I—SO2CI

AcOH

To prepare complex sugar sulfonic acid esters, the reaction between primary carbohydrate iodides and lithiated sulfonate esters opened a new route. The reaction of a-lithio sulfonate sters, mainly isopropyl, with primary carbohydrate iodides provided a facile route to complex sulfonate analogues [109].

198

Progress in Biological Chirality SO2-NHBU

SO2CI

SO2-OR R-OH

NHBz(|)Me

^ Ac0^r-^\

NHBz(!)Me

120 (70%)

NHBziMe 121 R=Me(41%)

119 NH2

1 2 2 R - (CH3)2-CH ( 5 4 % )

BU4NHSO4

NaHCOs

0 V NH—'^""X

^4Me

K 123 (40%) Figure 3

The lithiated sulfonate ester can even be a sugar derivative (131). The anion of the 3-0-mesylate of diacetone allose (131) could be allylated using the iodide O

o=i=o ^^

^ OMe

o

\

THF/ DMPU - 7 8 ° 0 — -20°C

125a = I 125b = OMs 125c = OTf

c

II

2) H'

X 126

O-n X'OH TMSO-\-*-''T^-V OTMS LiCH2-S02

128

129

Sugar C-sulfonic Acids

199

125a, product 132 reminds us to the oligonucleotide backbone of ribonucleic acids which is an isosteric and uncharged analogue. Similar bond-type could also be created by the addition of sulfonyl-stabilized HomerEmmons reagents to aldehydes. 3-0-Mesyl-l,2:5,6-di-0-isopropylidene-a-D-allofuranose (133) was converted to the Homer-Emmons reagent (134) by reaction with diethyl phosphorochloridate in the presence of KN(SiMe3)2. Reaction of 134 with the sugar aldehyde 135 followed by the reduction of the resulting unsaturated sulfonate with NaBH4 or [(Ph3P)CuH]6 gave the ester-type bonded disaccharide (136), which could be transformed into a dinucleotide [110].

X

0=CH

l)BuLi;-78°C

^ 2)NaBH4 "SO2

CH3-SOJ

CH2-PO(OEt)2

133

134

135

131 + 125a

OMe

References [1] lUPAC-IUB Joint Commission on Biochemical Nomenclature (JCBN), Eur. J. Biochem. 257 (1998) 293298. [2] A.A. Benson, H. Daniel and R. Wiser, Proc. Natl Acad. Sci. 45 (1959) 1582-1587. [3] V.M. Dembitsky, O.A. Rozentsvet and E.E. Pechenkina, Phytochemistry 29 (1990) 3417-3421. [4] A.L. Jones and J.L. Harwood, Biochem. Soc. Trans. 15 (1987) 482. [5] V.M. Dembitsky, E.E. Pechenkina and O.A. Rozentsvet, Phytochemistry 30 (1991) 2279-2283. [6] B.W. Son, Phytochemistry 29 (1990) 307-309. [7] B.W. Son, Bull. Korean Chem. Soc. 9 (1988) 264-266. [8] M. Katsuoka, C. Ogura, H. Etoh, K. Sakata and K. Ina, Agric. Biol. Chem. 54 (1990) 3043-3044.

200

Progress in Biological Chirality

[9 G.R. Pettitt A.L. Jones and L.H. Harwood. Phvtochemistry 28 (1989) 399-405. [10 K.R. Gustafson, J.H. Cardellina IL R.W. Fuller, O S . Weislow, R.F. Kiser, K.M. Snader, G.M.L. Patterson and M.R. Boy&, J. Natl. Cancer Inst. 81 (1989) 1254-1258. [11 H. Kikuchi, Y. Ksukitani, T. Manda, T. Fujii, H. Nakanishi, M. Kobayashi and I. Kitagawa, Chem. Pharm. Bull. 30 (1982) 3544-3547. [12 I. Kitagawa, Y. Hamamoto and M. Kobayashi, Chem. Pharm. Bull. 27 (1979) 1934-1937. [13 Y. Tang and R.I. HoUingsworth, Glycobiology 1 (1997) 935-942. [14 M. Lepage, H. Daniel and A. A. Benson, J. Am. Chem. Soc. 83 (1961) 157-159. [15 H. Daniel, M. Miyano, R.O. Mumma, T. Yagi, M. Lepage, I. Shibuya and A. A. Benson, / . Am. Chem. Soc. 83 (1961) 1765-1766. [16 M. Miyano and A. A. Benson, J. Am. Chem. Soc. 84 (1962) 57-59. [17 M. Miyano and A. A. Benson, J. Am. Chem. Soc. 84 (1962) 59-62. [18 Y. Okaya.y4c/« Cryst. 17 (1964) 1276-1282. (19 S.R. Johns. DR. Leslie. R.I. Wilhng and D.G. Bishop. Aust. J. Chem. 31 (1978) 65-72. [20 A.P. Tulloch, E. Heinz and W. Fischer, Z Physiol. Chem. 354 (1973) 879-886. [21 H P . Siebertz. E. Henz. M. Linscheid. J. JoyardandR. Douce, Eur. J. Biochem. 101 (1979)429-438. [22 H.D. Zepke, E. Heinz, A. Radurz, M. Linshceid and R. Pcsch, Arch. Microbiol. 119 (1978) 157-162. [23 R.H. Gigg, in: Rodd's Chemistry of Carbon Compounds, vol. IE (Ed. S. Coffey), Elsevier, Amsterdam, 1976. p. 349. [24 R. Gigg, A.A.E. Penghs and R. Conant J.CS Perkin 1 (1980) 2490-2493. [25 Y. Mizushina, I. Watanabe, K. Ohta, M. Takemura, H. Sahara, N. Takahashi, S. Gasa, F. Sugawara, H. Matsukage and S. Yoshida, K. Sakaguchi, Biochem. Pharmacol, 55 (1998) 537-541. [26 (a) K. Ohta, Y. Mizushina, N. Hirata, N. Takemura, F. Sugawara, A. Matsukage, S. Yoshida and K. Sakaguchi, Chem. Pharm. Bull. 46 (1998) 684-686. (b) Idem, ibid. 47 (1999) 111-116. [27 H. Sahara. M. Ishikawa. N. Takahashi. S. Ohtani. N. Sato. S. Gasa, T. Akino and K. Kihuchi, British J. Cancer 75 (1997) 324-332. [28 J. Golik. J.K. Dickey, G. Todderud. D. Lee, J. Alford. S. Huang, S. Klohr, D. Eustice, A. Aruffo and ML. Agler, J. Nat. Prod. 60 (1997) 387-389. [29 S. Loya. V. Reshef, E. Mizrachi. C. Silberstein, Y. Rachamia S. Carmeli and A. Hizi, J. Nat Prod. 61 (1998)891-895. [30 D M . Gordon and S.J. Danishefsky. J. Am. Chem. Soc. 114 (1992) 659-663. [31 S. Hanashima, Y. Mizushina, T. Yamazaki, K. Ohta S. Takahashi, H. Koshino, H. Sahara, K. Sakaguchi and F. Sugawara, Tetrahedron Lett. 41 (2000) 4403-4407. [32 S. Hanashima, Y. Mizushina, T. Yamazaki, K. Ohta. S. Takahashi, H. Sahara, K. Sakaguchi and F. Sugawara, Bioorg. Med Chem. 9 (2001) 367-376. [33 B. Helferichand W. Ost Z. Physiol. Chem. 31 (1963) 114-117. [34 R.L. Whistler and D.G. Medcalf, .4rc/z. Biochem. Biophys. 105 (1964) 1-6. [35 J. Lehmann and A.A. Benson. J. Am. Chem. Soc. 86 (1964) 4469-4472. [36 (a) J. Lehmann and W. Weckerle, Carbohydr. Res. 11 (1972) 23. (b) Idem, ibid 11 (1972) 317-324. [37 M. Hoch, E. Heinz and R.R. Schmidt. Carbohydr. Res. 191 (1989) 21-28. [38 V. Ulgar, I. Maya, J. Fuentes and J.G. Femandez-Bolanos, Tetrahedron 58 (2002) 7967-7973. [39 A. Liptak, E. Balla, L. Janossy, F. Sajtos and L. Szilagyi, Tetrahedon Lett 45 (2004) 839-842. [40 J. L. Harwood and R.G. Nicholls, Biochem. Soc. Trans. 1 (1979) 440-447. [41 R. Reistad, Carbohydr. Res. 54 (1977) 308-310. [42 P. Weber and R.J. Winzler. Arch. Biochem. Biophys. 137 (1970) 421-427. [43 A S . Edge and P. Weber, Carbohydr. Res. 126 (1984) 279-285. [44 W. Meyer zu Reckendoif and W.A. Bonner, J. Org. Chem. 26 (1961) 5241-5243. [45 J.S. Showell, J.R. Russel and D.J. Swem, J. Org. Chem. 11 (1962) 2853-2858. [46 J. Femandez-Bolanos, I. Maya Castilla and J. Femandez-Bolanos Guzman, Carbohydr. Res. 147 (1986) 325-329. [47 J.G. Femandez-Bolanos, J. Morales, S. Garcia, M.J. Dianez, M.D. Estrada, A. L6pez-Castro and S. Perez, Carbohydr. Res. 248 (1993) 1-14. [48 J.G. Femandez-Bolanos, S. Garcia, J. Femandez-Bolanos, M.J. Dianez, M.D. Estrada, A. Lopez-Castro and S. Perez, Carbohydr. Res. 282 (1996) 137-147. [49 J. Femandez-Bolanos. I.M. Castilla and J. Femandez-Bolanos Guzman, Carbohydr. Res. 173 (1988) 3340.

Sugar C-sulfonic Acids

201

[50] A.B. Roy and MJ.E. Hewlins, Carhohydr. Res. 302 (1997) 113-117. [51] E.E. Simanek, G.J. McGarvey, J.A. Jablonowski and C.-H. Wong, Chem. Rev. 98 (1998) 833-862 and references cited therein. [52] E.L. Berg, M.K. Robinson, O. Mansson, E.C. Butcher and J.L. Magnani, J. Biol. Chem. 266 (1991) 14869-14872. [53] D. Sako, K.M. Commess, K.M. Barone, R.T. Camphausen, D A . Cumining and G.D. Shaw, Cell 83 (1995) 323-332. [54] E. V. ChandrasekaiBn, R.K. Jain, R.D. Larsen, K. Wlasichuk and K.L. Malta, Biochemistry 34 (1998) 2925-2936. [55] U. Sprengard, G. Kretzschmar, E. Bartnik, C. Hiils and H. Kunz, Angew. Chem., Int. Ed. Engl. 107 (1995) 990-993. [56] M.S. Mulligan, J.C. Paulson, S. DeFrees, Z.-L.- Zheug, J.B. Lowe and P.A. Ward, Nature 364 (1993) 149-151. [57] D.J. Lefer, D.M. Flynn, M.L. Phillips, M. Ratchffe and A.J. Buda, Circulation 90 (1994) 2390-2401. [58] D. Tyrell, P. James, N. Rao, C. Foxall, S. Abbas, D. Dasgupta, M. Nashed, A. Hasegawa, M. Kiso, D. Asa, J. Kidd and B.K. Brandley, Proc. Natl. Acad Sci. USA 88 (1991) 10372-10376. [59] A. Borbas, G. Szabovik, Z. Antal, P. Herczegh, A. Agocs and A. Liptak, Tetrahedron Lett. 40 (1999) 3639-3642. [60] A. Borbas, G. Szabovik, Z. Antal, K. Fehar, M. Csavas, L. Szilagyi, P. Herczegh and A. Liptak, Tetrahedron: Asymmetry 11 (2000) 549-566. [61] P.M. Simon, P.L. Gogode, A. Mobasseri and D. Zopf, Infect. Immun. 65 (1991) 750-757. [62] K.-A. Karlsson, in: Carbohydrates in Chemistry and Biology, Part IT. Biology of Saccharides, vol. 4: Glycohiology of Helicobacter pylori and gastric disease (Eds. P. Ems, G.W. Hart, and P. Sinay) WileyVCH, Weinheim, 2000. [63] B.E. Dunn, H. Cohen and M.J. Blaser, Clin. Microbiol. Rev. 10 (1997) 720-741. [64] K.-A. Karlsson, Mol. Microbiol. 29 (1998) 1-11. [65] G. Faller, H. Steininger, B. Appelmelk and T. Kirchner, J. Clin. Pathol. 51 (1998) 244-245. [66] E.C.I. Veerman, C M C . Bank,F. Namavar, B.J. Appelmelk, J.G.M. Bolschevand A.V. Nieuw Amerogen, Glycobiology 7 (1997) 737-743. [67] T. Satoh, H. Natomi, W. Zhao, K. Okuzumi, K. Sugano, M. Iwamori and Y. Nagai, FEBSLett. 282 (1991)385-387. [68] S. David and S. Hanessian, Tetrahedron 41 (1985) 643-663. [69] M. Yoshikawa, S. Yamaguchi, H. Matsuda, N. Tanaka, J. Yamahara and N. Murakami, Chem. Pharm. Bull. 42 (1994) 2430-2435. [70] I. Kolker and A. Lapworth, J. Chem. Soc. Ill (1925) 307-315. [71] M.S. Khaiasch, E.M. May and F.R. Mayo, J. Org. Chem. 3 (1939) 175-192. [72] C.J. Norton, N.F. Seppi and M.J. Renter, /. Org Chem. 33 (1968) 4185-4165. [73] A. Borbas, M. Csavas, L. Szilagyi, G. Majer and A. Liptak, J. Carbohydr. Chem. in press. [74] D.B. Dess and J.C. Martin, J.Org. Chem. 48 (1983) 4155-4156. [75] B. Musicki and T.S. Widlanski, J. Org. Chem. 55 (1990) 4231-4233. [76] B. Musicki and T.S. Widlanski, Tetrahedron Lett 32 (1991) 1267-1270. [77] P A . Crooks, R.C. Reynolds, J.A. Maddiy, A. Rathore, M.S. Akhtar, J.A. Montgomery and J.A. Secrist III, J. Org Chem. 57 (1992)2830-2835. [78] A. Liptak, F. Sajtos, L. Janossy, D. Gehle and L. Szilagyi, Org. Lett. 5 (2003) 3671-3674. [79] K. J. Ryan, E M . Acton and L. Goodman, J. Org. Chem. 36 (1971) 2646-2657. [80] K.C. Nicolaou, T. Ladduwahetty, J.L. Randall and A. Chucholowski, J. Am. Chem. Soc. 108 (1986) 24662467. [81] V. Pozsgay, Carbohydr. Res. 235 (1992) 295-302. [82] H.M. Zuurmond, P.AM. van der Klein, G.A. van der Marel and J.H. van Boom, Tetrahedron Lett. 33 (1998) 2063-2066. [83] B.M. Trost and D.P. Curran, Tetrahedron Lett. 22 (1981) 1287-1290. [84] J. Elhalabi and K.G. Rice, Carbohydr. Res. 335 (2001) 159-165. [85] A. Liptak, I. Jodal and P. Nanasi, Carbohydr. Res. 44 (1975) 11. [86] M. Ek, P.J. Garegg, H. Hultberg and S. Oscarson, J. Carbohydr. Chem. 2 (1983) 305-311. [87] S. Winstein, E. Allred, R. Heck and R. Glick, Tetrahedron (1958) 1-13. [88] J. S. Brimacombe and O.A. Ching, J. Chem. Soc. (C) (1968) 1642-1646.

202 [89] [90] [91] [92] [93] [94] [95] [96] [97] [98] [99] [100] [101] [102| [103] [1041 [105] [106] [107] [108] [109] [110]

Progress in Biological Chiralit> A. Borbas, A. Biro, and A. Liptak, ACH-Models in Chem. 131 (1994) 455-465. A. Liptak, L. Lazar, M. Csavas and A. Borbas, Arkivoc in press. S. Inouye, T. Tsunioka, T. Ito and T. Niida, Tetrahedron 23 (1968) 2125-2144. T. lida, N. Yamazaki and C. Kibayashi, J. Org. Chem. 52 (1987) 3337-3342. S. Knapp and D.S. Myers, J. Org. Chem. 66 (2001) 3636-3638. S. Knapp and E. Darout Tetrahedron Lett. 43 (2002) 6075-6078. Z.-J. Li, P.-L. Liu, Z.-J. Li, D.-X. Qiu and M.-S. Cai, Synth. Commun. 20 (1990) 2169-2175. T.V. RajanBabu and G.S. Reddy, J. Org. Chem. 51 (1986) 5458-5461. J. Gervay, T.M. Flaherty and D. Holmes, Tetrahedron 53 (1997) 16355-16364. D.S. Breslow and H. Skolnik, in: The Chemistry of Heterocyclic Compounds (Ed. A. Weissberger) Interscience Publishers Inc., New York, 1966, Vol. 21, Part I, pp. 79-87; Vol. 21, Part II, pp. 774-780. F.G. Bordwell, R.D. Chapman and C.E. Osborne. J. Am. Chem. Soc. 81 (1959) 2002-2007. (a) E.J. Corey and T. Durst. J. Am. Chem. Soc. 88 (1966) 5656-5657. (b) Idem. ibid. 90 (1968) 55485552. W.E. Truce and D.J. Vrencur. Can. J. Chem. 47 (1969) 860-862. T. Durst and J. du Manoir. Can. J. Chem. 47 (1969) 1230-1233. T. Durst and K.-C. T i a Can. J. Chem. 48 (1970) 845-851. M.B. YunkerandB. Fraser-Reid. J.C.5. Chem. Comm. (1978) 325-326. Fraser-Reid. K.M. Sua R. Y.-K. Tsang. P. Sinav and M. Pietraszkiewicz. Can. J. Chem. 59 (1981) 260263. P A . Crooks, R.C. Reynolds. J. A. Maddr\, A. Rathoe, M.S. Akthar, J. A. Montgomery and J. A. Secrist III, J. Org Chem. 57 (1992) 2830-2835.' J. LehmannandH. Reinshagen, Synthesis (1973) 222-223. V. Ulgar, I. Maya, J. Fuentes and J.G. Femandez-Bolanos, Tetrahedron 58 (2002) 7967-7973. B. Musicki and T.S. Widlanski, J. Org. Chem. 55 (1990) 4231-4233. B. Musicki and T.S. Widlanski, Tetrahedron Utt. 32 (1991) 1267-1270.

Progress in Biological Chirality G. Palyi, C. Zucchi and L. Caglioti (Editors) © 2004 Elsevier Ltd. All rights reserved.

Chapter 16 Racemization of Amino Acids in Hydrothermal Environments: a Contribution of Temperature Gradient Koichiro Matsuno* and Atsushi Nemoto Department of BioEngineering, Nagaoka University of Technology, Nagaoka 940-2188, Japan kmatsuno@vos. nagaokaut. acjp

1.

Introduction Despite that hydrothermal environments in the primitive ocean might have played a significant role for prebiotic chemical evolution, there could have been some deleterious side effects also. One such effect is the degradation of optical purity of amino acid molecules available near hydrothermal vents through racemization driven by heat energy from the vents [1], let alone their dehydration, deamination and decarboxylation [2]. We shall investigate in the present article how the degradation of optical purity of amino acids would proceed in an experimental setup simulating a hydrothermal environment as focusing on the role of salt ions available there. Our specific attention will be on the ftmctional role of sharp temperature gradient near hot vents on the sea floor in the ocean spewing out hot water including various ions [3, 4] to racemization of L- and D-amino acids.

2.

Racemization in Temperature Gradient Thermodynamic consideration suggests that even if we start with either pure L-amino acid or D-amino acid solution, reaction kinetics in thermal equilibrium would eventually lead to a complete racemic mixture of L- and D-amino acid in the end when the temperature is elevated (see, for instance, [5]). Needless to say, hydrothermal environments also observe such reaction kinetics in thermal equilibrium. However, there is one significant point to be noted as an exception to this rule. That is the presence of sharp temperature gradient across the interface between hot and cold seawater near the hot vents on the sea floor. One would have to have some reservation with applying reaction kinetics in thermal equilibrium where there is temperature gradient. If the rate of racemization of L-amino acid differs from that of D-amino acid even only slightly in thermal equilibrium and if the reactants are subject to temperature gradient as experienced around hot vents in the actual ocean, a question arises as to whether the insignificant difference of the rates of racemization would remain as it is as in thermal equilibrium or be amplified. We shall experimentally examine the reaction kinetics of

204 Progress in Biological Chirality racemization proceeding in the presence of temperature gradient, while the reactants are allowed to visit both a high temperature and a low temperature region in a recycling manner. 3.

Materials and Methods We used a flow reactor simulating a hydrothermal environment in the primitive ocean [6, 7], and examined how an optically pure L-alanine or D-alanine could be racemized through the operation of the reactor. The reaction solution including L-alanine or D-alanine was injected at a flow rate of 8 mL per min from a 15 mL heated chamber at 230 °C and 24 MPa. The reactants passed through a nozzle (diameter 0.8 mm) into a larger, cooler chamber at 0 °C and approximately the same pressure. Cycle time for the total volume of fluid (500 mL) was 62.5 min, but with stirring the reactants recycled roughly once per minute. Each run lasted 2 hours and was started from room temperature, requiring approximately 15 min to reach 150 °C and 20 min to the designated 230 °C. The prepared reaction solution consisted of either Lalanine or D-alanine, or its racemic mixture, though the initial reactants of alanine were somewhat depleted with time through their decomposition and oligomerization [1, 8, 9]. Ionic sahs were further added in a controlled manner if required. Analysis of the products was conducted by sampling a small amount of the reaction fluid at a fixed time interval. The amount of each of L- and D-alanine remaining in the reaction solution was measured by an OPA-NAC method [10]. Derivatization of L- and D-alanine was accomplished by thoroughly mixing 20|il of the solution with 20|il of OPA-NAC reagent in a polyethylene microfuge tube. About 2 minutes after the mixing, 475|il of 50mM Na acetate

100

o o

E 80 0)

•

\

o g 60 m o 1^ 40

i

•

L

O •

D racemic

]

>,

J

'c )

S 20 C LU

i.

^

^

#

0

o

•^

t

0 i

1

20

1

40

60

__-l 80

J 1

100

LJ

120

Time [ min ] Figure 1. Time developments of enantiomeric excess for three cases of starting with 50 mM solution of pure L-alanine (solid circle), pure D-alanine (open circle), and complete racemic mixture of L- and D-alanine (solid square). The temperature of the high tenq)erature chamber reached the designated 230 °C about 20 minutes after the start of theflow-reactoroperation. Error bar denotes a half of the standard deviation of measured enantiomeric excess estimated out of three tofiveindependent experiments

Racemization of Amino Acids in Hydrotheraial Environments: a Contribution of...

205

(pH 5.2) was added. The specimen of 250JLI1 of the solution was then taken for the HPLC analysis. The fluorescence detector equipped with a lOjiil flow cell was used with an excitation wave at 320nm and an emission wave at 440nm. The column used was the reversed phase YMC ODS-M80 (4.6mm x 250mm). The mobile phase was 27.5% methanol and 50mM sodium acetate (pH 2.5).

4.

Results Typical time developments of enantiomeric excess for alanine are displayed in Figure 1, in which the concentration of alanine was maintained at 50mM. Enantiomeric excess is denoted as the numerical figure |[L]-[D]|xlOO/([L]+[D]) [%l where [L] and [D] represent the concentrations of L- and D-alanine. We showed three cases of starting with pure L-alanine, pure D-alanine, and complete racemic mixture of L- and D-alanine. The initial pH measured at room temperature was 5.7 for both pure L- and D-alanine, while the final pH after 120 minutes operation of the flow-reactor measured again at room temperature was 7.6 for the Lalanine case and 7.8 for D-alanine. There was observed a sizable difference of enantiomeric excess whether the reaction solution started with pure L-alanine or D-alanine. In order to see the contribution of the pH value on enantiomeric excess, we measured the similar time developments for the pH valued fixed at 4.0 by adjusting the amount HCl to be added (Figure 2a), and at 8.0 by adjusting the amount of NaOH to be added (Figure 2b). The pH value was constantly monitored at the low temperature chamber maintained at 0 °C. We

100 •O O

[

1

80

S E o c

e UJ

-

i

-

()

-

40

•is «B

L D

60

iS o 'C

• O

i

•

o

o

•

•

20

O

t

0 1

11

20

1

40

60

^

80

1

100

•u 120

Time [ min ] Figure 2a. Time developments of enantiomeric excess for two cases starting with 50 mM solution of pure L-alanine (solid circle) and pure D-alanine (open circle). The pH of the reaction solution at the low-temperature chamber maintained at 0 °C was constantly fixed at 4.0 by adjusting the amount of HCl to be added. Error bar denotes a half of the standard deviation of each measurement estimated out of three tofiveindependent experiments

40

60

80

Time [ min ]

Figure 2b. Time developments of enantiomeric excess for two cases starting with 50mM solution of pure L-alanine (solid circle) and pure D-alanine (open circle). The pH of the reaction solution at the low-temperature chamber maintained at 0 °C was constantly fixed at 8.0 by adjusting the amount of NaOH to be added. Error bar denotes a half of the standard deviation of each measurement estimated fi*om three tofiveindependent experiments

206

Progress in Biological Chirality 3 U

• *-

E'°

O D j

-^ ^

CA

#

§30

•

•-

\ o

H

•£ 20

i

§

o c UJ

Oj

^

"0 \

0

i_^.*.u4

0.001

0.01

^-L^^lJ

0.1

U^

1

10

1 1 im..l

100

Concentration [ mM ]

Figure 3. Contribution of ionic strengths to enantiomeric excess. The excesses were measured after 120 minutes operation of theflowreactor starting with either pure L-alanine (solid circle) or pure D-alanine (open circle) with the controlled amount of added sodium chloride. The concentration of alanine was 50mM for either case. Error bar denotes a half of the standard deviation of each measurement estimated out of three tofiveindependent experiments

could not however observe a significant difference in the contribution of the pH value to the racemization of alanine whether it started with pure L-alanine or D-alanine. One more candidate for influencing the racemization may be ionic strengths of the reaction solution. The effect of added sodium chloride on enantiomeric excess is then displayed in Figure 3, demonstrating that the reaction solution starting with L-alanine had slightly greater enantiomeric excess compared to the case starting with D-alanine. One remarkable difference between the two cases of starting with L-alanine and with D-alanine is that L-alanine is quite sensitive to the concentration of sodium chloride for its racemization, while D-alanine remains rather indifferent. In addition, there is one subtle point to be noted. When there was no added sodium chloride (see Fig. 1), the reaction solution starting with pure L-alanine had slightly smaller enantiomeric excess compared with the case starting with D-alanine. The conversion from enantiomeric excess in favor of D-alanine to L-alanine occurred in the narrow region of ionic strength between 0.0 and 1 /i M of sodium chloride.

5.

Discussion Although optical purity of amino acids [11, 12] could almost fully be lost through racemization in thermal equilibrium at high temperatures available near hydrothermal vents, the actual hydrothermal environments in the ocean are not in thermal equilibrium because of the presence of constant temperature gradient between the hot water inside the vents and the cold water in the surroundings. Such temperature gradient may provide a likelihood of enantioselectivity that could not be expected in thermal equilibrium. One indication of

Racemization of Amino Acids in Hydrothermal Environments: a Contribution of...

207

enantioselectivity of nonequilibrium origin thermally can be seen in the fate of the racemization of each individual amino acid molecule experiencing the temperature gradient. Our observation of enantioselectivity of L- over D-alanine in racemization in the flow reactor facilitating a sharp temperature gradient, as demonstrated in the case of low concentration of sodium chloride, manifests that the rate of an individual L-alanine molecule to be transformed into a D-alanine molecule is smaller than the similar rate of the reversed transformation. So long as the reaction environment is maintained in the presence of such a temperature gradient as in the actual hydrothermal environments, the survival probability of L-alanine relative to that of D-alanine could be enhanced with time because of its multiplicative, instead of merely additive, nature of the very survival in time. Of course, even racemization for both reaction solutions starting with L- and with Dalanine could be expected when these reactants visit the region at high temperatures as with our high temperature chamber maintained at 230 °C. The unevenness of enantiomeric excess between L- and D-alanine acquired as traversing the accompanied temperature gradient would certainly be mitigated when the reactants visit and stay in the high temperature region. What is significant to such hydrothermal environments is that there is expected a situation enhancing enantiomeric excess in favor of L-alanine in the presence of sodium chloride or in favor of D-alanine in its absence, even locally both spatially and temporarily. Prebiotic synthesis of amino acid molecules might have almost completely been racemic because of the random nature of energy sources available effectively at high temperatures. Nonetheless, enantioselectivity of L-alanine in racemization while experiencing temperature gradient in the presence of sodium chloride could have served as an evolutionary screen for enhancing the relative population of L-alanine in prebiotic or protobiotic organizations insofar as the environments in the vicinity of robust hydrothermal vents were persistently maintained on the sea floor in the ocean.

6.

Concluding Remarks The difference of the rates of transition from an L- to D-amino acid and its reverse in thermal equilibrium is almost insignificant because of the absence of energetic difference between the enantiomeric isomers. Only an extremely weak asymmetric influence of physical origin such as parity-violating jS -decay, electroweak force mediated by a neutral boson Z and magnetic field from a rotating neutron star [13] could account for the difference of the rates of racemization internal to the enantiomeric isomers. If so, it would be required to justify how such an extremely small difference of the rates of racemization imputed to parityviolating physical processes, if any, could be amplified up to the detectable level as in our observation with use of the flow reactor simulating a hydrothermal circulation of seawater through hot vents. A rationale for the enormous enhancement of the difference of the apparent rates of racemization between different enantiomers rests upon the rapid quenching of the reactants across temperature gradient [14]. When a tiny solution droplet including the reactants suddenly happens to suffer its temperature drop as traversing the interface separating between the hot and the cold region, the droplet would actualize only the quickest temperature drop out of possible alternatives. In this temperature drop, the reactants in the droplet would come to transform themselves so as to accommodate them to the sudden temperature drop outside. In fact, the transformation facilitating the quickest temperature drop comes to take over there

208

Progress in Biological Chiralin

since there is no room for the temperature to further drop once the quickest one has been in place. When amino acid molecules quickly traverse the narrow interface zone between hot and cold regions, there could be four possibilities of transformation and no-transformation such as L to L, L to D, D to L and D to D. Among those four possibilities, the alternative that may survive will be the ones dropping their temperature fastest while crossing the interface. Enantiomeric selectivity conceivable in the interface between hot and cold regions as expected in hydrothermal environments in the ocean is for those enantiomeric isomers substantiating the quickest temperature drop as traversing the interface from the hot to the cold, in spite of the fact that racemization carrying no such selectivity persistently would have proceeded outside the interface zone. In particular, L-alanine has been demonstrated to have less heat capacity compared with D-alanine [13, 15]). Less heat capacity means more rapid in quenching if other conditions are equal. This observation would come to suggest that Lalanine would be more favored in the hydrothermal environments in the presence of sharp temperature gradients. In summary, in a simulated hydrothermal environment allowing fluid circulation between hot and cold regions repeatedly, L- and D-alanine molecules were racemized differently depending upon the quantities of salt ions available there. When there were no added salt ions, the rate of racemization starting with pure L-alanine was slightly greater than that starting with pure D-alanine. When there were added salt ions, on the other hand, the rate of racemization starting with L-alanine was slightly smaller than that starting with D-alanine. Hydrothermal environments in the primitive ocean could have maintained the capacity of selectively retaining enantiomeric excess in favor of either L-amino acids or D-amino acids depending upon the quantities of salt ions available there.

7.

Acknowledgments Experimental work reported in this article has been done in collaboration with Minako Horie, Eiichi Imai, Hajime Honda and Kuniyuki Hatori.

8. [1] [21 |3] [4] [5] [6] [7] [8] [91 [10] [11] [12] [13] [14] [15]

References K. Kawamura and M. Yukioka, Thermochim. Acta 375 (2001) 9-16. J.L. Bada S.L. Miller and M. Zhao, Origins Life Evol. Biosphere 25 (1995) 111-118. M.J. Russell and A.J. Hall. J. Geol. Soc. London 154 (1997) 377-402. G.D. Cody, N.Z. Boctor, R.M. Hazen, J.A. Brandes, H.J. Morowitz and H.S. Yonder, Jr., Geochim. Cosmochim. Acta 65 (2001) 3557-3576. V.A. Basink, Ad\^ Space Res. 27 (2001) 335-340. K. Matsuno, Viva Origino 25 (1997) 191-204. E. Imai. H. Honda. K. Halori. A. Brack and K. Matsuno. Science 283 (1999) 831-833. A. Nemoto. E. Imai, H. Honda, K. Halori and K. Matsuno, Viva Origino 29 (2001) 168-173. A. Nemoto, M. Horie, E. Imai, H. Honda, K. Hatori and K. Matsuno, Origins Life Evol. Biosphere, in press. D.'W. Aswad, Anal. Biochem. 137(1984)405-409. JR. Cronin and S. Pizzarello, Science 275 (1997) 951-955. G.L.J.A. Rikken and E. Raupach. Nature 405 (2000) 932-935. W. Wang, F. Yi, Y. Ni, Z. Zhao, X. Jin and Y. J. Biol. Phys. 26 (2000) 51-65. K. Matsuno and R Swenson. BioSystenis 51 (1999) 53-61. R. Sullivan, M. Pyda, J. Pak. B. Wunderlich, J. R. Thompson, R. Pagi, H. Pan, P. Schwerdtfeger, and R. Compton,./. Phys. Chem. A 107 (2003) 6674-6680.

Progress in Biological Chiralit>' G. Palyi, C. Zucchi and L. Caglioti (Editors) © 2004 Elsevier Ltd. All rights reserved.

Chapter 17 The Theory of Chirality Induction and Chirality Reduction in Biomolecules Paul G. Mezey'^'' ""Albert Szent-Gyorgyi Professor, Department of Organic Chemistry, Eotvos Lordnd University of Budapest, H-1117 Budapest, Pdzmdny Peter Setdny 1/A, Hungary ^Canada Research Chair in Scientific Modelling and Simulation, Department of Chemistry and Department of Physics and Physical Oceanography, Memorial University of Newfoundland, Saint John % NF, CANADA A IB 3X7 [email protected], mezeypg@hotmailcom, [email protected]

1.

Introduction In this study two aspects of molecular chirality are discussed: chirality induction and chirality reduction. Both processes occur on both the local and global levels of molecules. Some of the fundamental interactions within molecules have the power to extend the chiral influence of any small region throughout the entire molecule, and the ability of these interactions to induce chirality in remote parts of the molecule have important practical consequences, especially in large molecules. Whereas much of the importance of the general relations are already evident in small molecules, as we shall demonstrate this in the next section, nevertheless, our main emphasis will be on the role of these relations in those chiral molecules which are perhaps of the greatest significance: chiral biomolecules. In this context, evolution will be one of the aspects often referred to, nevertheless, several of the conclusions we shall reach appear equally valid for organic as well as inorganic compounds. Whereas some of the discussions will rely on results obtained using somewhat special mathematical methods not routinely involved in biomolecular studies, here we shall not elaborate on the mathematical detail and the reader will be referred to the published literature for precise derivations. The fundamental principles of chirality induction and chirality reduction are very general and also universal: they provide constraints for practically all processes involving chiral molecules. Some of these constraints may be regarded as very natural, and are often taken for granted, yet they may also appear as being in conflict with some other, also natural expectations concerning biomolecules. In certain sense both chirality induction and chirality reduction may be regarded as natural trends, following either from overwhelming experimental evidence or from well-established natural laws, but the two opposing processes represent a conflict that appears to have essential connections to life processes. As a consequence of the wide scope of phenomena involving chirality induction and chirality

210

Progress in Biological Chirality

reduction, the examples include applications of the principles to parts of small molecules, to building blocks of larger biomolecules, and also to complex molecular systems involved in living organisms. Specifically, we shall discuss the following topics: (i) chirality induction in potentially achiral parts of electron density clouds within chiral molecules, and the theoretical basis of chirality reduction accompanying molecular decompositions; (ii) chirality induction by selection mechanisms and the combinatorial under-utilization of choices in the building processes of large biomolecules, influenced by type (i) of chirality induction; (iii) chirality induction and chirality reduction as a simple analogy for the entropy balance of living organisms, and the chirality induction and reduction processes of complex systems of biomolecules.

2. Some Fundamental Relations Governing Chirality Induction and Chirality Reduction in Molecules In the molecular context, both chirality induction and chirality reduction are processes involving interactions between various parts of molecules. It is of some interest to investigate, how large parts are involved in such interactions, and how the chirality of some molecular part is detected and processed by another molecular part, leading either to chirality induction or to chirality reduction (or, possibly, to no net change of chirality) there. Ahhough in the traditional viewpoint molecular chirality is often considered to require some minimal molecular size, nevertheless, in reality, even single atomic neighbourhoods, or even smaller, but positive volume parts of the electron density cloud may be chiral. Using the old concept of chiral carbon centers of organic chemistry, at least four different substituents about some carbon nucleus have been regarded as necessary for that type of chirality, that, of course, implied at least four nuclei for that type of chiral entity. As molecular modelling has become more refined and detailed electron density modelling and reliable quantum chemistry calculations have become more widely used, the focus has been narrowed and it has become increasingly possible and useful to consider the local features of chiral electron density clouds on a much smaller scale. Evidently, these clouds reflect all the chirality aspects of molecules, and even small parts of this cloud show evidently chiral features in any chiral molecule. Based on these developments, one natural question has been the following, how accurately can such local electron density ranges represent the chirality features of complete molecules, or at least, of those dominant parts of molecules which are primarily responsible for chiral, interactions with other molecules, or with polarized light in chemical or physical experiments. A simple answer can be given to the above question, relying on a general result of molecular physics: any small, positive volume part of the electron density cloud fiilly represents all chirality features of the entire molecule. (Strictly speaking, this statement has been rigorously proven only for molecules in their non-degenerate ground electronic states). The fundamental theorem that implies the above result for chirality underscores the importance of molecular electron density as the primary information carrier in molecules, and this includes chirality information. The origin of this fundamental theorem is the well-known foundation of density functional theory, a resuh on electron densities as established by the Hohenberg-Kohn Theorem [1], stating that for the non-degenerate electronic ground state of any quantum system of

The Theory of Chirality Induction and Chirality Reduction in Biomolecules

211

electrons, such as the electron density cloud of a molecule, the energy is a unique functional of the electron density. Since through the Hamiltonian operator of energy all other properties are also determined, it follows that for any molecule of a non-degenerate electronic ground state all properties are determined by the electron density. In fact, this result of the Hohenberg-Kohn Theorem is not surprising, since a molecule contains no other "material" than the atomic nuclei and the electron density cloud, and knowing that the nuclear distribution is fully reflected in the distribution of the electron density cloud, the latter must in fact determine all properties of the molecule. If the molecular electron density clouds were confined to a finite volume, that is, if molecular electron density clouds would be of finite size and they would have boundaries (what they actually cannot have), then an early result of J. Riess and W. Miinch [2] implies that local electron densities can be extended in a unique way to this finite and bounded density cloud. In a more recent development, both of these results have been extended to apply to local ranges of electron density clouds of actual, boundaryless electron densities of real molecules. This more recent result, as established by the proof of the Holographic Electron Density Theorem[3], states that the local electron densities of molecules are already sufficient to have a fully deterministic role in determining all molecular properties: any nonzero volume part of a molecular electron density in a non-degenerate ground state contains the complete information about all properties of the entire, boundaryless molecule. The scope of the Holographic Electron Density Theorem extends to both the exhibited and the latent properties [4] of molecules. In particular, various consequences [4] of this theorem provide new tools for Molecular Informatics and for the prediction of a whole range of latent molecular properties, such as those exhibited in different circumstances, in alternative conformations or in excited electronic states. The Holographic Electron Density Theorem also provide the justification and the tools for local molecular shape analysis as employed in shape activity correlations, where the activity often involves the complete molecules, or molecular parts not well identified. According to the "Holographic Electron Density Theorem" it is meaningful to use such local shape evaluations in correlations with observed molecular properties even if the actual mechanism and details of the molecular effect is not well understood, and if one cannot safely assign the observed property to any local range of the molecule. Recent reviews of various aspects of molecular shape analysis can be found in references [5-11], whereas the original developments and early applications, some in fact starting with knot-algebraic description of molecular chirality [12], can be found in references [12-42]. In fact, the methods originally developed for the electron density shape analysis of complete molecules [13-17] are now applicable to local molecular moieties [43-51]. Another development has coincided with the introduction of the above local shape analysis approaches: it has become possible to compute ab initio quality electron densities [53, 54] and other properties [55], such as electrostatic potentials [56], molecular energies [57], and forces acting upon individual nuclei [54] for truly large molecules, such as proteins of well over thousand atoms. This has been achieved using a linear-scaling algorithm (one that has computer time requirements directly proportional with molecular size, as opposed to conventional ab initio methods where the computer time requirement grows with the fourth power of the number of electrons). This method, the Adjustable Density Matrix Assembler (ADMA) method [53, 54], based on the Additive Fuzzy Density Fragmentafion (AFDF) approach, also provides efficient tools for the study of local ranges of macromolecules.

212

Progress in Biological Chirality

Whereas an early, numerical density representation approach was suitable only for macromolecular density computations [58-63], the more advanced ADMA approach [53, 54] provides the tools for the computation of many more macromolecular properties [53-57]. Ab initio quality quantum chemistry computations of macromolecular conformation changes, computational studies of forces acting on individual nuclei within a protein, the analysis of latent properties, detailed local shape analysis of complex macromolecules, including macromolecular shape - activity correlations in large biochemical systems, specifically, the study of chiral shape features of biomolecules, both on the global and on the local levels are now possible using the linear-scaling ADMA method and the additive fuzzy density fragmentation (AFDF) approach applied to large systems [53, 54].

3.

Chirality Measures as Symmetry Deficiency Measures In order to be able to quantify chirality induction and chirality reduction in the context of individual molecules, it is advantageous to be able to say whether chirality has increased or decreased in a given process. A natural approach is to use some type of measure for the degree of chirality. Whereas several such chirality measures have been proposed, here we shall focus on one type of measure, that is based on the concept of symmetry deficiency. It is clear that in the three-dimensional context, chirality is a manifestation of a special type of symmetry deficiency: the lack of mirror planes and of improper rotation axes of the S(2n) type. It is of special significance that the definition of chirality, in terms of nonsuperimposability of mirror images, is in fact equivalent to the definition relying on the lack of these two types of symmetry elements in three dimensions. Nevertheless, we may formulate the particular chirality measure in the more general context of symmetry deficiency, that is, in fact, applicable to any symmetry element R, not just those of mirror planes and of improper rotation axes of the S(2n) type, whose lack implies chirality. For simplicity in describing the concepts, first we shall discuss classical, macroscopic objects instead of quantum mechanical molecules. We say that a set D is an R-set if this set D has the symmetry element R, that is, if by carrying out the symmetry operation corresponding to the symmetry element R, the object D is transformed into an object that is indistinguishable from the original object D. Consider a three-dimensional rigid object A with a well-defined boundary and a particular symmetry element R that either is or is not present for the given object A. Consider all the R-sets which are contained within the object A, and also all R-sets which contain the object A. Since the classical, geometrical object A has a well-defined volume V(A), as well, all the above R-sets also have well-defined volumes, there must exist at least one set B that is a maximum volume R-subset of set A, as well as at least one set C that is a minimum volume R-superset of set A. Note that neither the maximum R-subset B nor the minimum R-superset C needs to be unique for a given set A, nevertheless, their volumes V(B) and V(C) are both unique, well-defined values for any given set A. These three volumes, V(A), V(B), and V(C) serve as the basis of two symmetry deficiency measures. The internal symmetry deficiency measure I(A,R) of set A with respect to symmetry element R is defined as: I(A,R) = (V(A)-V(B))A^(A), (1)

The Theory of Chirality Induction and Chirality Reduction in Biomolecules

213

whereas the external symmetry deficiency measure E(A,R) of set A with respect to symmetry element R is defined as: E(A,R) = (V(C)-V(A))/V(A).

(2)

Note that these quantities I(A,R) and E(A,R) are scaled by the volume V( A) of set A, that provides a natural normalization appropriate for symmetry deficiency measures. Clearly, if the object A actually has the symmetry element R, then both sets B and C are equal to set A, A = B = C,

(3)

hence both the internal and the external symmetry deficiency measures of set A with respect to symmetry element R give zero, I(A,R) = E(A,R) = 0,

(4)

that is, no R-symmetry deficiency occurs for object A. On the other hand, if sets A, B, and C are different, then (with the exception of degenerate, zero volume differences, such as that occurs when having a line protruding from set A that without this line would have the symmetry element R, where this line, of course, is not present in maximum R-subset B) the symmetry deficiencies of set A with respect to symmetry element R will have some positive values. Note that the internal and external symmetry deficiency measures are typically (but not necessarily) different for an R-symmetry deficient object A. The internal symmetry deficiency measure I(A,R) expresses how much of the volume of A must be taken away in order to achieve symmetry R, whereas the external symmetry deficiency measure E(A,R) expresses how much additional volume must be added to the object A in order to obtain an object that already has the symmetry element R. Here we assume that the volumes are scaled so that the volume V(A) of object A is considered as the reference. Consider now a molecule M and a particular symmetry element R that either is or is not present for the given molecule M. Since molecules are not classical objects with precisely defined boundary and volume, the simplest application of the above method can be achieved if one considers isodensity contour surfaces of the molecule, in order to provide a representation analogous to a macroscopic body with boundary surface and volume. For example, an isodensity contour surface A = A(M,0.01)

(5)

of molecule M, taken at the density value of 0.01 a.u. (atomic units) provides a suitable representation. Such isodensity contours do have volumes V(A), hence the above described approach is, indeed, applicable. Note that a more general approach, involving pointwise density differences and an integral measure extending over the whole space can also be used, effectively replacing the classical volume concept v^th a quantum mechanically more appropriate integral measure. For some alternative approaches, the reader may find information in references [64, 65], and for there broader context of chirality manifestations.

214

Progress in Biological Chirality

ultimately related to stmctural concepts, in references [66, 67]. If one is to consider several symmetry deficiencies simultaneously, one may generate new measures. Take, for example, a family F of symmetry elements, F = R(l),R(2),...R(k),

(6)

and evaluate the internal and external symmetry deficiency measures for the given object A according to each of these symmetry elements: I(A,R(1)), I(A,R(2)),

I(A,R(k)),

(7)

and E(A,R(1)), E(A,R(2)),

E(A,R(k)).

(8)

One can then compute, for example, the minimal, maximal, and average symmetry deficiency measures of object A with respect to the entire family F of symmetry elements: Imin(A,F) = min { I(A,R(1)), I(A,R(2)),

I(A,R(k)) },

Imax(A,F) = max { I(A,R(1)), I(A,R(2)), Iav(A,F) = [ I(A,R(1)) + I(A,R(2)) +

(9)

I(A,R(k))},

(10)

+ I(A,R(k)) ] / k,

(11)

Emin(A,F) = min { E(A,R(1)), E(A,R(2)), Emax(A,F) = max { E(A,R(1)), E(A,R(2)),

E(A,R(k)) },

(12)

E(A,R(k))},

(13)

+ E(A,R(k)) ] / k .

(14)

and Eav(A,F) = [ E(A,R(1)) + E(A,R(2)) +

Depending on what aspect of symmetry deficiency is of relevance, one may choose the appropriate symmetry deficiency measure from the alternatives (9) - (14). In the case of chirality, considered as symmetry deficiency of all mirror planes and all improper rotations of the S2n type, the family F of symmetry elements considered is precisely the set of all mirror planes and all improper rotations of the S2n type. According to our choice, the molecular object considered is the molecular isodensity contour A. This object A is achiral if any one of the symmetry elements from the family F is present. On the other hand, fi-om a given chiral isodensity surface A one may obtain an achiral subset or achiral superset the simplest way, that is, by the smallest volume change, if the symmetry element R(i) of the least actual symmetry deficiency measure is considered, either by removing or adding the smallest necessary volume. Consequently, for a chirality measure, taken as an "achirality deficiency measure", it is natural to consider the two minimal measures from the choices (9) - (14), that is, the measures Imin(A,F) and Emin(A,F).

The Theory of Chirality Induction and Chirality Reduction in Biomolecules

215

In terms of these symmetry family deficiency measures, the internal chirality measure Ic(M,0.01) and external chirality measure Ec(M,0.01) of molecule M, as represented by the isodensity contour surface at the density value of 0.01 a.u. are defined as Ic(M,0.01)-Imin(A,FX

(15)

Ec(M,0.01) = Emin(A,F),

(16)

and where F is the family of all mirror planes and of all improper rotations of the S2n type. These chirality measures are numerical expressions of "how much" relative volume change is required as a minimum, in order to obtain a subset or a superset that is already achiral. It is of some interest to pinpoint which actual "achirality inducing" symmetry element R is the one that is requiring the smallest volume change for the object in order to turn it into an achiral object.

4.

Chirality Induction and Chirality Reduction within Individual Molecules Whereas chirality induction and chirality reduction are perhaps of the greatest significance in the context of biochemistry and biology, a similar phenomenon of chirality induction and chirality reduction also exists on a smaller, non-biological scale within a given molecule. As a consequence of the interactions of local ranges of electron density clouds within a molecule, in a chiral molecule all, potentially achiral fiinctional groups or other groups of atoms are also chiral. A rigorous proof has been given earlier, asserting that local-only achirality cannot exist in any molecule: for example, a local mirror plane of a methyl group may exist only, if the entire molecule has a mirror plane: in fact, the entire molecule must have the same mirror plane. One may view this as chirality induction: if the molecule is chiral, that induces chirality also in such potentially achiral functional groups as a methyl group. On the other hand, if the molecule is decomposed, and the methyl group becomes an independent, non-interacting CH3 radical, then it becomes achiral. Hence, reduction of complexity, the decomposition of the original molecule may be accompanied by chirality reduction.

5.

Chirality Induction and Chirality Reduction in Biomolecules The structure, stability, function, and versatility of biomolecules are the resuh of molecular evolution. All of these features are influenced by various contributions to the chirality of biomolecules and are often fine tuned (but sometimes only roughly tuned) by evolution. Evolution is an opportunistic process where Nature usually takes advantage of variety, and make selections from a pool of all of the options available. However, the high dominance of homochirality in biomolecules implies that, somewhat uncharacteristically, evolution did not find advantageous to exploit the very high degree of combinatorial variability offered by mixing, for example, L- and D- amino acids in the largest family of biomolecules where such amino acids occur: peptides and proteins. There exists a very high degree of combinatorial under-utilization of choices. Apparently, structural regularities, that is, limitations to variety offered by homochiral amino acids when combined into larger molecular structures, have been more important in the course of evolution than the opportunities offered by the astronomical variety of potential peptide-like molecules generated fi-om a pool of non-

216

Progress in Biological Chirality

restricted combinations of L- and D- forms. This fact lends support to the idea that already in early evolution, the generation of complex systems necessary for life benefited from a rather general molecule-building mechanism that has relied on homochirality. Such a mechanism, in effect, provided chirality induction, by enhancing the overall chirality of a biomolecule built from chiral units of a given homochirality class. As the complexity of life forms have increased during evolution, this chirality induction process has become more dominant. Evolution of life has been accompanied by chirality induction. A similar chirality induction process accompanies the development of an individual multicell organism from a single cell to a complex, full-grown individual. On the other hand, death of an organism involves chirality reduction, since many of the ultimate decomposition products of biomolecules are small molecules, such as CO2 and N2, that are achiral. Of course, many chiral biomolecules become re-used by various life forms without decomposition to such small molecules, nevertheless, a significant amount of biomolecules lead to achiral decomposition products after the death of an individual organism. Death is usually accompanied by chirality reduction. As Professor Albert Szent-Gyorgyi has often stated, "life needs energy to fight entropy". The entropy of a live organism is lower than the system the same set of atoms will have after the death of the organism; in this sense an unusually low entropy is one characteristic of life. With respect to entropy, any (possibly inorganic) process that selectively favors homochiral arrangements also shows a reduction relative to the potential randomness when other, nonhomochiral building blocks are available. In this context, homochiral selectivity, even if demonstrated by some inorganic mechanism, may be regarded as a primitive model for life. In addition, many of the ultimate product molecules of metabolism generated by the normal life processes of a healthy individual are also achiral. Much of the intake of most terrestrial life forms is water, an achiral molecule, that predominantly plays the role of solvent, and much of this water is secreted without changing achirality. However, for most other molecules the situation is different. For many animals, metabolism involves the intake of chiral molecules and the uhimate generation of achiral ones, notably CO2. Whereas for many plants some important molecules in the intake are achiral, (again, one may think of CO2), that are used to build chiral biomolecules, nevertheless, the ultimate decomposition product of dead plant, if left to itself without biological utilization by other organisms, are also achiral. In analogy with Szent-Gyorgyi's statement, life needs mechanisms for chirality enhancement to fight achiral death. Within the context of chirality and life on Earth, we may think of life as chirality induction and death as chirality reduction.

6.

Acknowledgement The financial support of NSERC of Canada and the Albert Szent-Gyorgyi Award of Hungary are gratefully acknowledged.

7.

References

[1] P. Hohenberg and W. Koha Inhomogeneous electron gas. Phys. Rev. 136 (1964) B864-B871. [2] J. Riess and W. Miinch. The theorem of Hohenberg and Kohn for subdomains of a quantum system. Theor. Chim. Acta 58 (1981) 295-300. [3] P.G. Mezey, The holographic electron density theorem and quantum similarity measures. Mol. Phys. 96 (1999) 169-178.

The Theory of Chirality Induction and Chiralily Reduction in Biomolecules

217

[4] P.G. Mezey, The holographic principle for latent molecular properties. J. Math. Chem. 30 (2001) 299-303. [5] P.G. Mezey, Shape in Chemistry: An Introduction to Molecular Shape and Topology, VCH Publishers, NewYork/l993. [6] P.G. Mezey, Quantum chemistry of macromolecular shape. Internal. Rev. Phys. Chem. 16 (1997) 361-388. [7] P.G. Mezey, Shape in quantum chemistry. In: Conceptual Trends in Quantum Chemistry, Vol. 3 (Eds. J.-L. Calais and E. S. Kiyachko) Kluwer Academic Publ., Dordrecht, The Netherlands, 1997, pp. 519-550. [8] P.G. Mezey, Fuzzy measures of molecular shape and size. In: Fuzzy Logic in Chemistry (Ed. D.H. Rouvray) Academic Press, San Diego, 1997, pp. 139-223. [9] P.G. Mezey, Descriptors of molecular shape in 3D. In: From Chemical Topology to Three-Dimensional Geometry (Ed. A.T. Balaban) Plenum Press, New York, 1997, pp. 25-42. [10] P.G. Mezey, Shape analysis. In: Encyclopedia of Computational Chemistry (Eds. P. v. R. Schleyer, N.L. Allinger, T. Clark, J. Gasteiger, P.A. Kollman, H.F. Schaefer III, PR. Schreiner) John Wiley & Sons, Chichester, UK, 1998, Vol. 4, pp. 2582-2589. [11] P.G. Mezey, Topology and the quantum chemical shape concept. Advances in Molecular Similarity 2 (1998)79-92. [12] P.G. Mezey, Tying knots around chiral centres: chirality polynomials and conformational invariants for molecules. / . Am. Chem. Soc. 108 (1986) 3976-3984. [13] P.G. Mezey, Group theory of electrostatic potentials: a tool for quantum chemical drug design, Internal. J. Quantum Chem., Quant Biol. Symp. 12 (1986) 113-122. [14] P.G. Mezey, The shape of molecular charge distributions: group theory without symmetry. /. Comput. Chem. 8 (1987) 462-469. [15] P.G. Mezey, Group theory of shapes of asymmetric biomolecules. Int. J. Quantum Chem., Quant. Biol. Symp. 14 (1987) 127-132. [16] P.G. Mezey, Shape group studies of molecular similarity: shape groups and shape graphs of molecular contour surfaces. J. Math, Chem. 2 (1988) 299-323. [17] P.G. Mezey, Global and local relative convexity and oriented relative convexity; appUcation to molecular shapes in external fields. J. Math. Chem. 2 (1988) 325-346. [18] G.A. Arteca, V.B. Jammal and P. G. Mezey, shape group studies of molecular similarity and regioselectivity in chemical reactions. J. Comput. Chem. 9 (1988) 608-619. [19] G.A. Arteca, V.B. Janmaal, P.G. Mezey, J.S.Yadav, MA. Hermsmeier and T.M. Gund, Shape group studies of molecular similarity: relative shapes of Van der Waals and electrostatic potential surfaces of nicotinic agonists. /. Mol. Graphics 6 (1988) 45-53. [20] G.A. Arteca and P.G. Mezey, Shape description of conformationally flexible molecules: application to two-dimensional conformational problems. Internal. J. Quantum Chem., Quant. Biol. Symp. 15 (1988) 3354. [21] G.A. Arteca and P.G. Mezey, Molecular similarity and molecular shape changes along reaction paths: a topological analysis and consequences on the Hammond postulate. / . Phys. Chem. 93 (1989) 4746-4751. [22] G.A. Arteca and P.G. Mezey, Two approaches to the concept of chemical species: relations between potential energy and molecular shape./w/. J. Quant. Chem. Symp. 23 (1989) 305-320. [23] G.A. Arteca and P.G. Mezey, Analysis of molecular shape changes along reaction paths. Int. J. Quantum Chem. 38(1990)713-726. [24] P.G. Mezey, Point synmietry groups of all distorted configurations of a molecule form a lattice. J. Math. C/jm. 4(1990)377-381. [25] G.A. Arteca, G.A. Heal and P.G. Mezey, Comparison of potential energy maps and molecular shape invariance maps for two-dimensional conformational problems. Theor. Chim. Acta 76 (1990) 377-390. [26] P.G. Mezey, The Degree of similarity of three-dimensional bodies; apphcations to molecular shapes. J. Math. Chem. 1 (1991) 39-49. [27] P.D. Walker, G.A. Arteca and P.G. Mezey, A complete shape characterization for molecular charge densities represented by Gaussian-type functions. /. Comput. Chem. 12 (1991) 220-230. [28] X. Luo, G.A. Arteca and P.G. Mezey, Shape analysis along reaction patiis of ring opening reactions. Int. J. Quantum Chem. Symp. 25 (1991) 335-345. [29] I. Rozas, G.A. Arteca and P.G. Mezey, On the inhibition of alcohol dehydrogenase: shape group analysis of molecular electrostatic potential on Van der Waals surfaces of some pyrazole derivatives. Int. J. Quantum Chem., Quant. Biol. Symp, 18 (1991) 269-288. [30] G.A. Arteca and P.G. Mezey, Configurational dependence of molecular shape. /. Math. Chem. 10 (1992) 329-371.

218

Progress in Biological Chiralily

[31] G. A. Arteca and P.G. Mezey, Similarities between the efFects of configurational changes and applied electric fields on the shape of electron densities. J. Moi Struct. Theochem 256 (1992) 125-134 [special issue: Electrostatics in Molecules (Ed. G. Naray-Szabo and W.J. Orwille Thomas)]. [32] P.G. Mezey, Shape-similarity measures for molecular bodies: a 3D topological approach to QShAR. J. Chem. Inf. Comp. Sci. 32 (1992) 650-656. [33] P.G. Mezey, Dynamic shape analysis of biomolccules using topological shape codes. In: The Role of Computational Models and Theories in Biotechnology (Ed. J. Bertran) BUuwer Academic Publishers, Dordrecht, 1992, pp. 83-104. [34] P.G. Mezey, Dynamic Shape analysis of molecules in restricted domains of a configuration space. J. Math. Chem. 13 (1993) 59-70. [35] P.G. Mezey, Topological shape analysis of chain molecules: an application of the GSTE principle. J. Math. Chem. 12 (1993) 365-373. [36] X. Luo, G. A. Arteca and P.G. Mezey. Shape similarity and shape stability along reaction paths. The case of the ppo to opp isomerization,//?/. J. Quantum Chem. 42 {1992) 459-414. [37] P.D. Walker, G.A. Arteca and P.G. Mezey, Shape groups of the electronic isodensity surfaces of small molecules: the shapes of 10-electron hydrides. J. Comput. Chem. 14 (1993) 1172-1183. [38] P.G. Mezey, Methods of Molecular shape-similarity analysis and topological shape desiga In: Molecular Similarity in Drug Design (Ed. P.M. Dean) Chapman & Hall - Blackie Publishers, Glasgow, U.K., 1995, pp. 241-268. [39] P.G. Mezey, Molecular similarity measures for assessing reactivity. In: Molecular Similarity and Reactivity: From Quantum Chemical to Phenomenological Approaches (Ed. R. Carbo) Kluwer Academic Publ., Dordrecht The Netherlands, 1995, pp.57-76. [40] P.G. Mezey, Dynamic Shape group theory' of molecular nuclear potentials. In: Chemical Group Theory (Eds. D. Bonchev and D.H. Rouvray) Gordon and Breach Publ. Group, Reading, U.K., 1995, pp. 163-189. [41] P.G. Mezey, Semi-similarity of molecular bodies: scaling-nesting similarity measures. Int J. Quantum CAe'w. 51(1994)255-264. [42] P.G. Mezey, Quantum chemical shape: new density domain relations for the topology of molecular bodies, functional groups, and chemical bonding. Can. J. Chem. 72 (1994) 928-935 (Special issue dedicated to Prof J.C. Polanyi). [43] P.G. Mezey, Local shape analysis of macromolecular electron densities. In: Computational Chemistry: Reviews and Current Trends. Vol.1 (Ed. J. Leszczynski) Worid Scientific Publ., Singapore, 1996, pp. 109137. [44] P.G. Mezey, Functional groups in quantum chemistry. Adv. Quantum Chemistry 27 (19%) 163-222. [45] P.G. Mezey, Local electron densities andftmctionalgroups in quantimi chemistry. In: Topics in Current Chemistry, Vol. 203 "Correlationand Localization" (Ed. PR. Surjan) Springer-Veriag, Berlin, Heidelberg, New York, 1999, pp. 167-186. [46] P.G. Mezey, Holographic electron density shape theorem and its role in drug design and toxicological risk assessment. J. Chem. Inf Comp Sci. 39 (1999) 224-230. [47] P.G. Mezey, Combinatorial aspects of biomolecular shape analysis. Bolyai Soc. Math. Stud. 1 (1999) 323332. [48] P.G. Mezey, K. Fukui and S. Arimoto, A treatment of small deformations of polyhedral shapes of ftmctional group distributions in biomolccules. Int J. Quant Chem. 76 (2000) 756-761. [49] P.G. Mezey, Topological methods of molecular shape analysis: continuum models and discretization, DIMACS Series in Discrete Mathematics and Theoretical Computer Science 51 (2000) 267-278. [50] P.G. Mezey, Shape-similarity relations based on topological resolution. J. Math. Chem. 27 (2000) 61-69. [51] P.G. Mezey, Topological similarity of molecules and the consequences of the holographic electron density theorem, an extension of the Hohenberg-Kohn theorem. In: Fundamentals ofMolecular Similarity, (Eds. R. Carbo-Dorca, X. Girones and P.G. Mezey) Kluwer Academic/Plenum, New York, 2001, pp. 113-124. [52] P.G. Mezey, Theor>' and detailed computer modelling of biomolecules. In: Fundamentals o/(Eds. G. Palyi, C. Zucchi and L. Caglioti) Elsevier and Accademia Nazionalc di Scienze, Lettere ed Arti (Modcna), Paris, 2002, pp. 401-416. [53] P.G. Mezey, Macromolecular density matrices and electron densities with adjustable nuclear geometries. J. Math. Chem. 18 (1995) 141-168. [54] P.G. Mezey, Quantum similarity measures and Lowdin's transform for approximate density matrices and macromolecular forces. Inl J. Quantum Chem. 63 (1997) 39-48.

The Theory of Chirality Induction and Chirality Reduction in Biomolecules

219

[55] P.G. Mezey, Computational aspects of combinatorial quantum chemistry. J. Comput. Methods ScL Eng. (JCMSE) 1 (2001) 99-106. [56] T.E. Exner and P.G. Mezey, Ab initio quality electrostatic potentials for proteins: an application of the ADMA approach. J. Phys. Chem. A 106 (2002) 11791-11800. [57] T.E. Exner and P.G. Mezey, Ab initio quality properties for macromolecules using the ADMA approach. J. Comput. Chem. 24 (2003) 1980-1986. [58] P.D. Walker and P.G. Mezey, Molecular electron density lego approach to molecule building. J. Am. Chem. Soc. 115 (1993) 12423-12430. [59] P.D. Walker and P.G. Mezey, Ab initio quality electron densities for proteins: a MEDLA approach. J. Am. Chem. Soc. 116 (1994) 12022-12032. [60] P.D. Walker and P.G. Mezey, Realistic, detailed images of proteins and tertiary structure elements: ab initio quality electron density calculations for bovine insulin. Can. J. Chem. 72 (1994) 2531-2536. [61] P.D. Walker and P.G. Mezey, A new computational microscope for molecules: high resolution MEDLA images of taxol and HIV-1 protease, using additive electron density fragmentation principles and fuzzy set methods. J. Math. Chem. 17 (1995) 203-234. [62] P.G. Mezey and P.D. Walker, Fuzzy molecular fragments in drug research. Drug Discovery Today (Elsevier Trend Journal) 2 (1997) 6-11. [63] P.G. Mezey, Computational microscopy: pictures of proteins. Pharmaceutical News 4 (1997) 29-34. [64] P.G. Mezey, Generahzed chirality and symmetry deficiency. J. Math. Chem. 23 (1998) 65-84. [65] P.G. Mezey, Theory of biological homochirality: chirality, symmetry deficiency, and electron-cloud holography in the shape analysis of biomolecules. In: Advances in BioChirality (Eds. G. Palyi, C. Zucchi and L. Caghoti) Elsevier Sci. Publ, Amsterdam, The Netherlands, 1999, pp. 35-46. [66] R.B.King, Chirality and handedness: the ruch 'shoe-potato' dichotomy in the right-left classification problem. In: Chemical Explanation: Characteristics, Development, Autonomy (Ed. J.E. Emloy) Ann. N. Y. Acad Sci. 988 (2003) 158-170. [67] K. Balasubramanian, Enumeration of internal rotation reactions and their reaction graphs. Theor. Chim. Acta5^(\919) 129-146.

Progress in Biological Chirality G. Palyi, C. Zucchi and L. Caglioti (Editors) © 2004 Elsevier Ltd. All rights reserved.

Chapter 18 Transfer of the Chiral Information of Natural Amino Acids in Biomimetic Organic Syntheses Karoly Micskei,^* Csongor Hajdu,'*'' Tamas Patonay,'' Claudia Zucchi/ Luciano Caglioti^ and Gyula Palyi"* ^Department of Inorganic and Analytical Chemistry, ^Department of Organic Chemistry University of Debrecen, Egyetem ter I, H-4010 Debrecen, Hungary kmicskei@delfin. kite, hu ^Department of Chemistry, University ofModena and Reggio Emilia, Via Campi 183,1-41100 Modena, Italy [email protected] '^Department of Chemistry and Technology of Biologically Active Compounds, University "La Sapienza*', P.leA. Moro 5,1-00I85, Roma, Italy

1.

Introduction Biological processes are very important sources of organic compounds. These procedures called biosyntheses [1] are typically characterized by a very high level of chemo-, regio-, diastereo- and enantioselectivity. Biosynthetic processes are operating in aqueous medium and in narrow temperature and pH ranges. Their key elements are enzymes (biocatalysts) which operate using a vast variety of "chemical tools" for achieving an exceptional level of efficiency. This capability is a result of a very fme, concerted tuning determined by several factors such as hydrophilic/hydrophobic groups, macromolecular properties, complexed metal ions or chiral structural elements, H-bonds, TC-TT stackings, electrostatic interactions. The formation of a highly organized "chiral space" is considered essential for the exceptional level of enantioselectivity. Today quite much is known about these processes and about their structural requirements but we are only at the beginning of understanding biosyntheses. However, on the basis of the available information and experience biotechnological industry produces many organic molecules, some of which are practically unavailable through other methods. These "other methods" are the organic chemical syntheses [2] which were developed starting from observation and imitation of biological processes. But then the stereoselective syntheses took an independent way of evolution utilizing conditions which are very far from the biological ones. Such conditions are the use of non-aqueous, often strictly anhydrous solvents, strong acidic (protonating) or basic reagents, inert atmosphere, much wider temperature ranges than the biological conditions — to list only a few of them. These conditions allow several reactions impossible by biosyntheses but the transformations are characterized by lower yields and selectivities than those typical for biological processes. One

222 Progress in Biological Chiralib of the major successes of synthetic organic chemistry in the last decades is the use of transition metal containing catalysts [3]. These derivatives provided nice tools to avoid drastic reaction conditions by lowering the activation energy barriers dramatically and, consequently, making possible to obtain substantially higher selectivities. Prominent achievements in this field were obtained in the enantioselective catalytic synthesis of chiral, non-racemic organic compounds by the use of transition metal complexes of chiral organic ligands, recognized also by the Nobel Prize in Chemistry in 2001 [4]. Another outstanding result in this field is the discovery of chiral autocatalysis [5], which is also the topic of Chapter 29 of this book. Comparison of the biochemical discoveries with the most recent resuhs of synthetic organic chemistry prompted several groups to initiate a new direction in synthetic chemistry by amalgamating these two fields. This trend led to the development of so-called biomimetic chemistry [6] characterized by searching and finding synthetic methods which are using experience from biochemistry as much as possible and similar experimental conditions (aqueous media, near-neutral pH, more or less ambient temperature, atmospheric pressure, etc.), as well. Our groups in the last decade joined this trend by investigating the chemistry of transition metal complex reagents containing typical "bio-organic" molecules (mostly amino acids) as ligands and which could be used under the near-biological conditions listed above. We summarize these resuhs in the present paper, discussing also the possibilities of future developments.

2.

Precedents Previous studies have shown that various Cr(Il) complexes with aminopolycarboxylate ligands (e.g. IDA, NTA, EDTA, etc.) can be used as highly selective reagents for the (reductive) transformation of natural organic substances (monosaccharides, morphine alkaloids) [7, 8]. Later, these reagents were found to be useful in a modified variant [9] of the Nozaki-Hiyama reaction [10]. This variant allowed the formation of carbon-carbon bond in neutral aqueous medium. A similar reagent was also used for the generation of ferrocenylketyl radicals [11] with unusual structure and stability. These results prompted us to focus on the use of natural amino acids as ligands in the Cr(II) complex reagents [12]. Natural amino acids are one of the most abundant and, therefore, usually relatively cheap natural sources of chirality. However, in spite of these attractive features L-amino acids themselves are only rarely used as chiral ligands in transition metal complex-catalyzed or -assisted enantioselective transformations; in this field the best resuhs were obtained by using non-natural amino acids [13] which are much more expensive [14] or commercially not available. On the other hand, naturally occurring amino acids play an important role in the stereoselective synthesis in the form of their derivatives such as 1,2-amino alcohols, bisisoxazolines, oxazaborolidines, etc. [13, 15]. We believe, the major obstacle for the use of amino acids in their original form is that they and their complexes are only sparingly soluble in most of organic solvents (used generally as reaction medium). In aqueous media, however, most natural amino acids are nicely soluble and form many — well characterized — complexes with transition metal [16]. One of the most studied metal ion is chromium in its various oxidation states. Although chromium(II) is a broadly used reducing reagent in organic chemistry [17] nothing has been published on the utilization of the amino acid complexes in the synthetic organic chemistry.

Transfer of the Chiral Information of Natural Amino Acids in Biomimetic Organic Syntheses

3.

223

Amino Acid Complex Reagents in Aqueous Media In their complexes, natural amino acids are typically coordinated through the a-amino and the carboxyl(ate) groups as bidentate ligands forming stable five-membered rings. The stability, structure, reactivity and solubility of these complexes is influenced in a fairly broad range by the structure of the a-substituent ("side chain"). Several classifications based on the structure of these side chains has been done which are widely used in biochemistry to interpret (or predict) the behaviour of amino acids, as well as their oligo- and polycondensates. One of the most important features is the polarity of the side chain [18]. The following natural amino acids can be regarded typically apolar: Ala, Val, Leu, He, Pro, Phe, Tyr, Trp, Cys, Met. These representatives are characterized by low solubility and micelle-formation in water. The members of other group (Lys, Arg, His, Asp, Glu, Asn and Gin) bear polar, mostly dissociable groups in their side chain which assures a good solvation in water and (consequently) higher solubility. In spite of the fact that the majority (14) of the natural amino acids possesses groups with typical ligand properties in the side chain, in aqueous solution these are only weakly or not coordinated. Only three natural amino acids can be regarded as tridentate (His, Asp, and Cys) [16]. All natural amino acids with the exception of Gly contain at least one sterogenic center (the a-carbon atom) and belong to the L-series in an overwhelming majority of cases [19]. The structure of the side chains widely influences the applicability of their complexes as synthetic reagents. Until now, mostly reagents from amino acids with apolar side chains were studied [13, 15c] predominantly because of their solubilities in organic solvents. In aqueous media the use of non-derivatized amino acids (as inductors of chirality) is much less common [20, 21]. Recently, some uses of natural amino acids as sources of chirality in the enantioselective synthesis of carbon-carbon bonds were reported [22, 23]. The formation of amino acid complexes 1, 2 is accompanied by a fairly complicated network of equilibria (Figure 1), In aqueous solutions amino acids 4 undergo

R-CH

GOGH

coo"' NH2

NH2

R-CH

+

NHT

NH3

r

o

(1)

R-CH

R-CH

^ coo"

COOH

000"=

R-CH

{n-1)®

w

* C-O R-CH ;M

M"'

NH2

(2)

NH2

1 {n-1)(5

R-CH

;M

R-CH

coo^ NH2

NH^

O w C-0 R-CH NH2

(n-2)®

,NH2 CH-R O-C^

(3)

1 ,^n@

j^Q^^

[M(0H)^f-'^)®

(4)

Figure 1. Solution equilibria of the transition metal ion - amino acid complexes in aqueous medium

224

Progress in Biological Chirality [Cr2+]^Q^ = 1.0

10.00 m M

[ala-]

=

Cr2+

30.00 mM

^,.—

^•^|-I^i':ii^^:i|i-P'^::'i

Cr((

Cr(ala)2 0.8

0.6

0.4

^

qr(ala)'T^

0.2

0.0 10

12

PH

Figure 2. Formation of chromium(II)-L-alamne-complexes in aqueous medium

deprotonation/reprotonation processes (Eq. 1), resulting charged and zwitterionic species 3, 6 and 5, respectively. The deprotonated form 3 can react with the metal ion as a bidentate ligand, resulting a 1:1 complex 1 (Eq. 2). Then, the next ligand 3 can easily be attached to the metal forming a 1:2 complex 2 (Eq. 3). Under aqueous and basic conditions hydroxide ions as ligands can be coordinated (parallel or concurrently) to the metal (Eq. 4). Noteworthy, that deprotonation is crucial for the coordination since the carboxyl group itself gets much weaker ligated. All these aspects should be quantitatively considered in the planning of the reagent. Detailed knowledge of the equilibrium constants [24] and production of suitable computer programs [25] are crucial. The distribution of various complexes in the aqueous Cr(II)/Ala system are shown as an example in Figure 2, diagrams of other complexes are of similar shape. The most important practical conclusion from these studies is that the Cr(II)/amino acid reagents should be very carefully planned and prepared otherwise the composition of these systems cannot be controlled. The narrow operative ranges (as well as the fact that these ranges appear just between pH-5 and -9) and the sensitivity towards changes in the experimental conditions are very similar to the behaviour of enzymatic systems.

Chirality Transfer of Natural Amino Acids 4,1 Reduction ofC=0 double bond At the beginning of our joint research efforts we investigated whether the chirality of natural amino acids coordinated to Cr(n) could be transferred to simple organic compounds obtained by reduction with Cr(II) from prochiral precursors. Acetophenone (8) was chosen as a model substrate [26] (Scheme 1). Indeed, it has been found that treatment of ketone 8 with solutions of Cr(II)/amino acid 1:2

Transfer of the Chiral Information of Natural Amino Acids in Biomimetic Organic Syntheses

O

HO H CH3

225

H OH

Cr(ll)L* (2 equiv.) H2O/DMF. pH = 6-9

R-9

S-9 L* = natural L-amino acid Scheme 1

complexes (2, M = Cr) afforded 1-phenylethanol (9) with >95% conversion. The analysis of the enantiomeric ratio of the product 9 showed that the chiral information [27] of the ligand amino acids appeared in the product. The transfer of the chiral information resulted in different enantioselectivities and different absolute configuration of the major enantiomer depending on the structure of the amino acid used (Figure 3). The highest enantiomeric excess (75% e.e.) was obtained with Cr(L-Val)2 complex at pH-'9. While bidentate L-amino acids such as Ala, Pro, Asn yielded excesses of the R-9, typical tridentate L-amino acids (Asp, His) induced S preference. In accordance with the pH optimum in the formation of Cr(L-Ala)2 the increase of pH from 6-7 to 9 resulted in a slight increase of e.e. values. A few control experiments showed that: (i) without amino acid ligand there is no detectable reduction of 8, (ii) with achiral ligands (EDTA) there is no detectable enantiomeric excess of 9, and (iii) corresponding D-isomers of the amino acid ligands induced the formation of the enantioselective reduction of a broad scale of various ketones, with excellent chemical yields (Ala, His, Asp) and low-to-moderate e.e. (5-55%). Again, both enantiomers could be achieved by using natural L-amino acids. The structure

t

f«

ij-

Ala

D.AIa

Val

Leu

Met

Phe 0-Phe

^-^h-^

lyr

Tip

Ligand

Pro

i

I a IT

Hypro Asp

Asn

Lys

i His

D>His

dpH 6-7 • pH~9 Figure 3. Dependence of enantiomeric excess of alcohol 9 on the structure of the L- and D-amino acid ligand

226

Progress in Biological Chirality o

o

a^ 10

14

15

of the amino acid ligand has a decisive influence on the absolute configuration of the product. Bidentate ligands usually furnished S preference ahhough in the case of Leu both S and R enantiomers could be obtained depending on the structure of the substrate. Similarly, the tridentate ligand His gave S enantiomers as the predominant isomer for all substrates but in the case of Asp R enantiomers were the major products in the reduction of cyclic substrates 16,18,19. The remaining starting ketones afforded S preference, again. Our data revealed that the enantioselectivity of the reduction was also influenced by the substrate which could be divided into two classes on the basis of the conformational properties. In the series of alkyl aryl ketones 8-14 and the conformationally more flexible benzosuberone (17) His provided the best e.e.'s but it gave poor enantioselectivity in the reduction of the more rigid benzocyclanones 16, 18, 19. The reverse effect was observed for Ala. These observations clearly show that the structure of the starting ketone also influences the

ketones

Figure 4. Enantioselectivity of the Cr(II)/Z-amino acid/H20/DMF system in ketone reduction

Transfer of the Chiral Information of Natural Amino Acids in Biomimetic Organic Syntheses

227

Stereochemical outcome of the asymmetric reduction Optimization experiments with chromanone (18) supported the important role of pH, e.e.'s and conversions slowly decreased in the range of pH 10 and 8 followed by a sharp breakdown between 7 and 8. Solvent dependence has also been found. The best yields and enantioselectivities were found using DMF and formamide/water mixtures, other donor solvents gave poorer results. These differences could be interpreted in terms of the structure and reducing potential of the Cr(II)/amino acid complexes. Consequently, the efficiency of our system can be highly tuned by the modification of the composition of the reducing species and of the conditions used. Analysis of the preparative results, the study of the model system [11], spectroscopic (UVVIS and CD) studies of the reaction mixtures [29] as well as comparison of these results with some relevant literature reports [30, 31] allowed us to deduce a hypothesis (Scheme 3) about the possible steps of the reduction. Steps of this mechanism include the one-electron formation of a ketyl radical-anion (22) (Eq. 5), its (fast) protonation (Eq. 6), the reaction of the radical 23 with the chromium(II) complex 21 under the formation of an organochromium intermediate 24 (Eq. 7) and the subsequent proton-metal exchange (Eq. 8). The crucial step, where the evolvement of the chirality happens, is the formation of the intermediate 24. The discrimination leading to

)=0

+

Cr"(L*)n

20

SET

)5-0® +

y

21

C^"(L*)n

(5)

22 R^ > R (CIP priority order)

.0

H«

R (6)

-OH 23

22

)^OH

+

Cr"(L*)n

SET

21

23

R.

Cr"'(L*)n

R'

OH

R"X Ri

H--Y-Cr'"(L*)n TS1

TS2

©

®

OH

OH

(7) OH

S-24

R{ R

H I

CrJ"(L*)n

R

R-24 inversion

retention

Rl.

I R..

OH

R'

H

R-25 Scheme 3

inversion

R\ R H---y--CH"{L*)n OH (8)

Tsr Rl.

OH

A R

H

S-25

228

Progress in Biological Chirality NOH ^CH3

NH

"

H2N H Cr(ll)L*(2equiv.)^ ^

Cr(ll)L*(2equiv.)

H2O/DMF, pH - 6 *

H2O/DMF, pH - 6 26

Y

"CH3

L S-28

27 hydrolysis

HO

O CH3

L* = natural L-amino acid '

Cr(ll)L*(2equiv.)^ | ^ H2O/DMF, pH - 6 *

8

Y ^ ^ ^ 3

U ^

RS-9

Scheme 4

different amounts of R- and S-24 products is probably governed by steric factors (different steric demand of R, R^ groups and the amino acid side chain) and internal hydrogen bonding between hydroxyl functionality of the intermediate 24 and the amino acid ligands. The enantiopurity of the R- and S-25 depends on the discriminating power of these interactions since it is retained in the proton-metal exchange step taking place with inversion. On the other hand, we can not exclude the attack of a proton coming from the inner ligand sphere. This step may take place with retention leading to the opposite enantiomer and, therefore, decrease the enantiomeric excess. 4.2 Reduction of C=N double bond The next step in our joint research was to attempt the reduction of C=N double bonds by the Cr(II)/amino acid reagents. For practical reasons oximes were chosen as substrates. No literature precedents of this reaction are known by us [32]. As our first model reaction the reduction of acetophenone oxime (26) was investigated. In these experiments [33] we found that Cr(II)/L-amino acid complexes could be used for the enantioselective reduction of the C=N double bond. The reaction could be performed with medium-to-excellent yields (50 - >90%) but only with low-to-moderate (--5 - 50%) enantiomeric excesses (Scheme 4, Figure 5).

ee% 20

amino acid Figure 5. Enantiomeric excesses (e.e.) of the reaction of acetophenone oxime (26) and Cr(n)/L-amino acid complexes

Transfer of the Chiral Information of Natural Amino Acids in Biomimetic Organic Syntheses 229 The reduction of oximes represents a more complicated problem as compared to that of the ketones (Scheme 4). Our studies have indicated that there is a four-electron reduction and a hydrolytic side reaction of the intermediate imine 27 yielding the corresponding ketone 1. The reduction of the by-product 8 leads to alcohol 9 as described earlier (Scheme 1). This complication can be (partly) avoided by using much higher (-7) Cr(n)/substrate ratio than applied in the case of the ketone. This modification allows to obtain 1-phenylethyl amine (28) in almost quantitative (>90%) yield. The pathway outhned above is in accordance with the observation of Corey and Richman [34] who reported the formation of the corresponding ketones from oxime 0-acetates upon treatment with Cr(II) acetate in aqueous THF, clearly by the hydrolysis of the imine intermediate formed in the first reduction step. Model experiments have shown that: (i) aqueous Cr(II) solution (containing presumably [CR(II)(H20)6]) does not give any reduction product, and (ii) Cr(II) ions in the presence of achiral aminopolycarboxylate ligands such as IDA, NT A, EDTA) do not give any enantioselectivity. This latter system, however, could be advantageously applied for the preparation of amines (as racemic mixtures) from the corresponding oximes 29-34 shown by Scheme 5 in moderate (9-55%) yields and poor (-3:2) diastereomeric (33, 34) ratios [35]. Experiments on the enantioselective reduction of these latter substrates are in progress. NOH

29

NOH

NOH

NOH

NOH

NOH

30

However, stereochemical outcome of the reactions with bi- and tridentate L-amino acids is not so clearly separated as in the ketone reduction. The highest e.e. (-50%) was observed with tridentate L-Asp (7^-28). At the same time, now L-Ala gave the same preferred absolute configuration in the product 28 but in the case of the similarly bidentate, somewhat more apolar L-Val a small but reproducible excess of S-28 was found. The rational analysis of the preparative results, the spectroscopic (UV-VIS and CD) studies on the analogous reduction of a-oximino carboxylic acids (vide infra) (Figure 6), as well as some relevant literature reports [31, 36, 37] seem to support the most important steps of our suggested mechanism shown in Scheme 6. The reaction course of oximes (Scheme 6) is clearly much more complicated than the previous one of ketones shown by Scheme 2, this fact is also reflected in the problems to achieve better chemical and enantiomeric yields, as well as in the different influence of bidentate v^. tridentate ligands on the absolute configuration, in comparison with ketones. The first steps of the mechanism (Eq. 9-11) results in the imine intermediate 38 by a reductive dehydroxylation. Again, the key element of the enantioselective reduction is the formation of the a-alkylchromium intermediate 40 (Eq. 13) which gives the final product 41 in a hydrolytic step by inversion (Eq. 14). As before, we can not exclude the possibility of a retention. The UV-VIS/CD /spectra shown in Figure 6 indicate:

230

Progress in Biological Chirality

500

600

800

Wavelength (nm)

500

600 Wavatongth (nm)

Figure 6. (i) UV-VIS and (ii) CD spectra of (a) L-Ala; (b) Cr(n); (c) Crai)+L-Ala; (d) Cr(II)+L-Ala+a-oximino-phenylacetic acid in water solution

(i) an inner sphere coordination of the ligand amino acid prior to the addition of oxime which phenomenon is verified by the presence of low energy d-d bands, (ii) formation of the chiral organochromium intermediate 40 upon addition of oxime, and (iii) chiral perturbation of the central chromium ion in both cases as verified by the appearance of the low energy CD bands. The experience obtained in the asymmetric reduction of acetophenone oxime (26) prompted us to set ourselves a more ambitious aim, the enantioselective preparation of a-amino acids from achiral precursors using Cr(n)/L-amino acid complexes as reducing agent and source of chirality. This goal could be achieved [38] by reducing oximes of a-ketocarboxylic acids with Cr(II)/amino acid 1:2 complexes (42) in aqueous solution (or water/DMF mixture) in pH range 8.9-9.4. These reaction took place with:

Transfer of the Chiral Information of Natural Amino Acids in Biomimetic Organic Syntheses )=NOH +

Cr"(L*)n

35

SET

^ N *

+

CrJ"(L*)n + :0H^

231 (9)

36

21

R^ > R (CIP priority order) )=N*

+

CH'(L*)n

SET

(10)

21

36

)=N

37

+ H2O

>=NH

+

(11)

Cr"'(L*)n + :OH®

38

37 1.SET 38

2. H2O

)^NH2 R^^ 39

SET

R.

21

)^NH2

+

Cr"(L*)n

^

Cr'"(L*)n

R^

21

39

(12)

+

R1

Cr"'(L*)n

A

"^

NH2

R

R-40

NH2 S-40 inversion

inversion

retention

TS3

K >r"'(L*)n R---K p,i

NH2

TS4

R1 R H---y--CrJ"(L*)n

Ri R H--y:-CrV)n

NH2

H I

(13)

R

NH2

R^^H

R-41 Scheme 6

J

L

NH2 (14)

TS3'

R1 R

NH2 H

S-41

(i) nearly quantitative (>90%) chemical yields, and (ii) low-to-moderate (-5-50%) e.e. for such systems where (iii) the ligand amino acids were different from the products expected in the reduction of the substrate (Scheme 7). Transformation of oxime 43 by Cr(II) complexes of L-Asp, L-His, L-Val, L-Ala, L-Phe, afforded a-phenylglycine (46) in excellent chemical and various enantiomeric yields. The major enantiomer was always S, independently of the nature of the ligand (Figure 7). Transformation of oxime 43byCr(II) complexes of L-Asp, L-His, L-Val, L-Ala, L-Phe,

232

Progress in Biological Chirality

O COOH R-CH

(n-2)®

cr^® R-CH

NH2

h,-^^' NH2

4

42

CH-R '0^0 / w O

O 42 (4 equiv.; OH

OH

NOH

NH2

43-45

46-48

43,46

44,47

45,48

C3H5

C6H5CH2

CH3

Scheme 7

afforded a-phenylglycine (46) in excellent chemical and various enantiomeric yields. The major enantiomer was always S, independently of the nature of the ligand (Figure 7). Further experiments with other oximes such as a-oximino-ji-phenylpropanoic acid (44) and a-oximinopropanoic acid (45) led to similar resuhs (Figure 9). The only qualitative difference was that in the case of these substrates the absolute configuration of the major enantiomer depended on the structure of the L-amino acid ligands, again, in accordance with the highly complicated character of the reduction process. We are carrying on our efforts in the directions as follows: (a) to develop analytical techniques which render the reliable analysis of reaction mixtures possible in those cases when the ligand and product amino acids are the same, (b) to improve the enantiomeric yields, and (c) to find catalytic variants.

his

val amino acid

Figure 7. Enantiomeric excess (e.e.) in the reaction of a-oximinophenylacetic acid (43) and Cr(IiyL-amino acid complexes

Transfer of the Chiral Information of Natural Amino Acids in Biomimetic Organic Syntheses

asp

val

ala

his

233

phe

amino acid Figure 8. Enantiomeric excess (e.e.) obtained in the reaction of a-oximinophenylacetic acid (43), a-oximino-p-phenylpropanoic acid (44) and a-oximinopropanoic acid (45) with Cr(II)/L-amino acid complexes

At the present stage we believe that the most important result of this research is that the chirality of natural amino acid ligands could be transferred to a-amino acids prepared from prochiralprecursors by reducing these with complexes of the former amino acids. These results are highly relevant to problems of the so-called biological homochirality [39]. This is the phenomenon that all living organisms use chiral molecules of very high enantiomeric excesses which in the whole living nature (on Earth) are belonging to the same series of configuration (L-amino acids, D-sugars, etc.). The origin of this phenomenon is not yet clear. Several theories speculate that it arises from an autocatalytic amplification of any casually or accidentally [19d, 39] formed and originally small excess of these enantiomers. Such autocatalytic amplification, however, has been reported in a single reaction (alkylation of aldehydes by iPr2Zn in organic solvents [5]) so far and its conditions are fairly far from the usual operational mode of living organisms. We hope that the system based on the aqueousphase preparation of chiral, non-racemic and naturally occurring amino acids by metal complexes with natural amino acids as ligands might represent the first step in the development of a system which is autocatalytic in the sense of chirality of the product amino acids.

6.

Acknowledgements This joint research project was supported by the [Hungarian] Scientific Research Foundation (Grant OTKA, No. T33130, T32429), the [Hungarian] Ministry of Education (Grant: FKFP 0614/2000), the [Italian] Ministry of University and Research, the [Italian] National Research Council (CNR) and the Italian-Hungarian Erasmus/Socrates exchange

234

Progress in Biological Chirality

program. Help in the initial steps of the analysis of amino acids is gratefully acknowledged to Prof Miklos HoUosi, Dr. Gyula Szokan and Dr. Sandor Szabo (Budapest).

7. [I]

References

(a) R.H. Garrett and CM. Grisham, Biochemistry, Saunders College Publ., Fort Worth, 1995. (b) K. Drauz, H. Waldman, Eds., Enzyme Catalysis in Organic Synthesis, VCH, Weinheim, 1995. [2] B.M. Trost and I. Fleming, Eds., Comprehensive Organic Synthesis, Pergamon Press, London, 1991. [3] B. Comils and W.A. Hermann, Eds., Applied Homogeneous Catalysis with Organometallic Compounds, Vol. 1-2., VCH, Weinheim, 1996. [4] (a) W.S. Knowles, Angew. Chem., Int. Ed. 41 (2002) 1998-2007. (b) R. Noyori, ibid 41 (2002) 2008-2022. (c) K.B. Sharpless, ibid. 41 (2002) 2024-2032. [5] K. Soai, T. Shibata, H. Morioka and K. Choji, Nature 378 (1995) 767-768. (b) K. Soai and T. Shibata, in: Advances in Biochirality (Eds. G. P^yi, C. Zucchi and L. Caglioti) Elsevier, Amsterdam, 1999, pp. 125136. (c) K. Soai, T. Shibata, I. Ssio,Acc. Chem. Res. 33 (2000) 382-390. (d) K. Soai, in: Fundamentals of Life (Eds. G. Pdlyi, C. Zucchi and L. Caglioti) Elsevier and Accademia Nazionale di Scienze, Lettere ed Arti (Modena), Paris, 2002, pp. 427-435. [6] (a) B. Honig, K. Sharp and A.-S. Yang, J. Phys. Chem. 97 (1993) 1101-1109. (b) A. Lubineau, J. Auge and Y. (Jueneau, Synthesis (1994) 741-760. (c) P. A. Grieco, Organic Synthesis in Water, Thomson Science, London, 1997. [7] G. Kov^cs, J. Gyarmati, L. Soms^ and K. Micskei, Tetrahedron Lett. 37 (1996) 1293-1296. [8] K. Micskei, J. Gyarmati, G. Kov^cs, S. Makleit, C. Simon, Z. Szabo, J. Marton, S. Hosztafi, H. Reinke and H-J. Drexler, £:wr. J. Org Chem. (1999) 149-153. [9] K. Micskei, A. Kiss-Szikszai, J. C^armati and C. Hajdu, Tetrahedron Lett. 42 (2001) 7711-7713. [10] Leading references: (a) Y. Okude, S. Hirano, T. Hiyama and H. Nozaki, J. Am. Chem. Soc. 99 (1977) 31793181. (b) K. Takai, K. Kimura, T. Kuroda, T. Hiyama and H. Nozaki, Tetrahedron Lett. 24 (1983) 52815284. [II] Z. Ratkovic, L. Somsak, K. Micskei, C. Zucchi and G. Palyi, J. Organomet Chem. 637 (2001) 813-819. [12] K. Micskei, O. Holczknecht, C. Hajdu, M. Meo, T. Patonay, C. Zucchi and G. P^yi, 9^^ Internal Conf on Circular Dicroism in Chemistry and Life Sciences (Aug. 31-Sept. 4, 2003, Budapest, Hungary), Abstracts L20p.43. [13] A. Studer, 5yw//2g5;5 (1996) 793-815. [14] (a) Fluka, Chiral Complounds Chemistry, Fluka Chemie AG, Buchs (CH) 1994. (b) Sigma-Aldrich, Chiral Nonracemic Compounds, Sigma-Aldrich, Milwaukee (WI, USA), 1998. [15] (a) A. Togni, A, L. M. \Qr\2axn, Angew. Chem., Int Ed 33 (1994) 497-526 (b) L. Deloux, M. Srebnik, Chem. Rev. 93 (1993) 763-784. (c) K. Drauz, A. Kleemann and J. Usntns, Angew. Chem. 94 (1982) 590613. [16] T. Kiss, in: Biocoordination Chemistry: Coordination Equilibria in Biologically Active Systems (Ed. K. Burger) Elhs Horwood (Publ), New York, 1990, pp. 57-134. [17] Leading references (a) L.A. Wessjohann, G. Scheid, Synthesis (1999) 1-36. (b) A. Fiirstner, Chem. Rev. 99 (1999) 991-1045. [18] L. Stryer, Biochemistry, 4* ed., W. H. Freeman and Co., New York, 1995, p. 17-23. [19] See for example: (a) P. JoU^s, Ed., D-Amino Acids in Sequences of Secreted Peptides of Multicellular Organisms, Birkhauser Verl., Basel, 1998. (b) Y. Nagata, in: Advances in BioChirality (Eds. G. P^yi, C. Zucchi and L. CagUoti) Elsevier, Amsterdam, 1999, pp. 271-283. (c) G. KreU, ibid. pp. 297-304; Annu. Rev. Biochem. 66 (1997) 337-345. (d) G. P^yi, K. Micskei, L. Bencze and C. Zucchi, Magyar Kern. Lapja 58(2003)218-223. [20] U.M. LindstrOm, Chem. Rev. 102 (2002) 2751-2771. [21] A. Lubineau, J. Aug6 and Y. (Jueneau, Water-Promoted Organic Reactions, in: Houben-WeylMethods of Organic Chemistry (Eds. G. Helmchen, R. W. Hoflftnann, J. Mulzer and E. Schaumann) G. Thieme (Publ.), Stuttgart-New York, 1995,4* ed. Vol. E21d., pp. 741-760. [22] S. Otto, G. Boccaletti, J.B.F.N. Engberts, J. Am. Chem. Soc. 120 (1998) 4238-4239. [23] S. Otto and J.B.F.N. Engberts, J. Am. Chem. Soc. 121 (1999) 6798-6806. [24] (a) K. Micskei, F. Debreczeni and I. Nagypdl, J. Chem. Soc., Dalton Trans. (1983) 1335-1338. (b) K. Micskei and 1. NagypM, ibid (1986) 2721-2723.

Transfer of the Chiral Information of Natural Amino Acids in Biomimetic Organic Syntheses

235

[25] L. Z^kany and I. Nagypal, in: Computational Methods for the Determination of Formation Constants (Ed. DJ. Leggett) Plenum Press, New York, 1985, pp. 291-353. [26] J. Gyarmati, C. Hajdu, Z. Dinya, K. Micskei, C. Zucchi and G. P%i, J. Organomet. Chem. 586 (1999) 106-109. [27] G. Palyi, C. Zucchi, L. Bencze and L. Caglioti, in: Bioinformatics (Eds. J. Seckbach and E. Rubin) Kluwer Academic Publ., Dordrecht (NL) 2004, in press. [28] T. Patonay, C. Hajdu, J. Jekd, A. Levai, K. Micskei, C. Zucchi, Tetrahedron Lett. 40 (1999) 1373-1374. [29] J. Telek-Gyarmati, PhD Thesis, University of Debrecen, 2000. [30] (a) R.S. Nohr and J.H. Espenson, J. Am. Chem. Soc. 97 (1975) 3392-3396. (b) G.W. Kirker, A. Bakac and J.H. Espenson, ibid 104 (1982) 1249-1255. [31] J.H. Espenson, in: Advances in Inorganic andBioinorganicMechanisms, Ed. A.G. Sykes, Academic Press, London-New York, 1982, Vol. 1. pp. 1-63. [32] T. K^gl and L. Koll^ in: Encyclopedia of Catalysis (Ed. I.T. Horvath) Wiley-Interscience, Hoboken (NJ, USA), 2003, Vol. 2. pp. 509-534. (b) ibid. Vol. 6. pp. 697-770 [index], (c) J. McCleverty and T.J. Meyer, Eds., Comprehensive Coordination Chemistry II. Vol. 10, Applications of Coordination Chemistry, Elsevier, Oxford,UK, 2004, in press. [33] K. Micskei, C. Hajdu, V. Marchis, O. Holczknecht, T. Patonay, A. Uvai, M. Meo, C. Zucchi and G. Palyi, submitted for publication, 2004. [34] E. J. Corey, J. E. Richman, J. Amer. Chem. Soc. 92 (1970) 5276-5277. [35] T. Patonay, A. Levai, V. Marchis, K. Micskei, unpublished results. [36] (a) J.K. Kochi andD.D. Davis, J. Am. Chem. Soc. 86 (1964) 5264-5271. (b) A. Bakac, V. Butkovic, J.H. Espenson, R. Marcec and M. Ohranovic, Inorg. Chem. 25 (1986) 2562-2566. (c) P. Huston, J.H. Espenson and A. Bakac, Inorg Chem. 30 (1991) 4826-4830. (d) J.H. Espenson, ^cc. Chem. Res. 25 (1992) 222-227. [37] (a) G.H. Timms and E. Wildsmith, Tetrahedron Lett. 2 (1971) 195-198. (b) K. Takai, N. Katsura and Y. Kunisada, Chem. Commun. (2001) 1724-1725. [38] K. Micskei, O. Holczknecht, C. Hajdu, T. Patonay, V. Marchis, M. Meo, C. Zucchi and G. Palyi, J. Organomet Chem. 682 (2003) 143-148. [39] Reviews: (a) L. Keszthelyi, Quart Rev. Biophys. 28 (1995) 473-507. (b) Marko, L. Diss. Savariensis 24 (1998) 1-64. (c) G. P^yi, C. Zucchi and L. Caglioti, in: Advances in BioChirality (Eds. G. Palyi, C. Zucchi, L. Caglioti) Elsevier, Amsterdam, 1999, pp. 3-12. (d) L. Keszthelyi, ibid, pp. 99-103. (e) idem. Fundamentals of Life (Eds. G. P^yi, C. Zucchi and L. Caglioti) Elsevier and Accademia Nazionale di Scienze, Lettere ed Arti (Modena), Paris, 2002, pp. 379-387.

Progress in Biological Chirality G. P^yi, C. Zucchi and L. Caglioti (Editors) © 2004 Elsevier Ltd. All rights reserved.

Chapter 19 CD and Visual Science Koji Nakanishi,* Nathan Fishkin and Nina Berova* Department of Chemistry, Columbia University, New York, NY 10027, USA kn5@columbia. edu, ndhl@columbia. edu

1.

Introduction The cross section of a human eye is shown in Fig. 1. The estimated 7 million cone cells, responsible for color ("photopic") vision are clustered around the fovea where the incoming image is focused most sharply. The cone cells contain three kinds of rhodopsins (Rh) or visual pigments absorbing around 450 nm (blue light), 530 nm (green light) and 560 nm (red light). A yellow solution absorbs the blue light at 450 nm so that the light that enters our eyes is green and red, which when mixed gives a yellow color. On the other hand, the 100 million rod cells responsible for black and white ("scopotic") vision are mostly distributed in the peripheral area and absorb at 500 nm. Since solar energy is strongest at 400-700 nm, the human eye is taking full advantage of the solar energy distribution. The rod/cone cell ratio depends on the animal: in nocturnal rats it is 4000, humans 20, goldfish 15, frogs 1, while owls only have rod cells. Although not nocturnal, cows and sheep (4 million rods), horse, and dog only have rods and hence do not have color vision. retina

•

rods (bl/wh)

ros

hv

Na*

\

cones (color)

\ disks

\^miBmm\\uim^

rhodopsln Figure 1. Cross section of the eye

sensation yellow

purple

bl.-green

238

Progress in Biological Chirality

The outer segments of rod cells (rod outer segment or ROS) consist of about 1500 thin membranes called disks that are formed at the basis of the rod visual cells by pinching off sections of the plasma membrane. The disks, which embed the 40 kDa Rh molecules consistin of 348 amino acids, move towards the tip of the cell in about 15 days where they are phagocytosed. Phagocytosis of debris by the retinal pigment epithelial cells (RPE) leads to accumulation of orange fluorescent pigments, A2E and others, resulting in age-related macular degeneration (AMD). Visual transduction is the process by which visual cells convert light into a neural signal, which in turn is transmitted to the brain via the optic nerve. In the dark, Na^ ions flow from the rod inner segment (RIS) into the rod outer segment (ROS) (see Fig. 1). Upon absorption of one photon of light, the flow of >million Na^ ions is blocked. Namely, absorption of a single photon activates -1000 molecules of G protein (guanyl-nucleotide binding protein, in the case of vision the term transducin is used), which via activation of phosphodiesterase, hydrolyzes 100,000 molecules of cyclic GMP. This drop in cGMP closes the cation-specific ion channels and leads to a build-up of electric potential that is picked up by the optic nerve.

2.

The 11-cis Retinoid Chromophore The unique attributes of the 11-c/^-retinal that serves as the skeleton for all visual pigments are summarized in Figure 2. The C20 1 l-c/^-retinal, biosynthesized from the C40 P-carotene, is the basic structure of all visual pigments. The P-carotene produces C20 vitamin A (?i\\-trans retinol), the precursor of retinal. However, since humans cannot biosynthesize carotenoids, we have to depend on exogenous sources for vision. The visual chromophore of salmons is 11-c/^-retinal in fresh water but during their migratory period in the ocean, the

Why 11 'Cis-retinal ?

I

The polyene chain is linked to a protonated Schiff base linkage. This allows the positive charge of SBH"" to delocalize depending on: Distance of counter ion and local electrostatic field within binding site. The chromophore is nonplanar around bonds 6/7 and 12/13, the degree of nonplanarity depending on the receptor opsin. The extent of nonplanahty in turn enables the pigment to adjust its X^ax to suit environment. 9-Me is needed for hydrophobic binding and for changing the Rh shape after bleaching; the transducin activation of 9-desmethyl Rh is weak. Double bond or OH can be added to adjust >.rnax to suit environment. Retinal is readily available from carotenoids. A single chromophore, aided by its receptor, can cover the range from 350 to 680 nm! It is a most unique chromophore that is central to all vision. Figure 2. Attributes of 11-cis-retinal as the visual chromophore

CD and Visual Science

239

becomes 3,4-dehydro-ll-cz5'-retinal to induce a red shift in the vision more suited for the ocean where there is less light. The chromophore of the tadpole is also 3,4-dehydro-ll-c/6'retinal, which again becomes 11-c/i'-retinal in the frog. Squids have an additional 4-ahydroxyl group [1] while insects have 3p-hydroxy-11-c/^-retinal [2], the chromophore possibly responsible for their UV sensitivity.

3.

The Visual Transduction Cycle The rhodopsin visual cycle, the result of immense amount of studies over the years, is outlined in Figure 3 together with the involvement of vitamin A. Irradiation of Rh at 500 nm causes 11 -cis -^trans isomerization to yield photo-Rh by femtosecond order photochemistry during which the 11-cz^ becomes a highly distorted trans geometry [3],[4]. This is the only light-triggered change that occurs in Rh, the subsequent changes to batho-Rh and others being caused by relaxation of the protein to relieve the high strain caused by the ''trans'' 11 -ene of the chromophore/protein complex. The strain energy, estimated to be 30-36 kcal/mol, [5] is spectra measured in liquid He, a temperature at which it can be sequestered [7]. Microscopic rate constants coupled with time-resolved spectroscopy led to the finding that a blue-shifted intermediate (BSI, not shown) is in equilibrium with batho-Rh. After batho-Rh the pigment reflected in the red-shifted absorption maxima of photo-Rh (570 nm) and batho-Rh- (543 nm).

OCOR all-trans retinyl esters carotenoids caroten

\ 11-cis-retinal

1

bovine rhodopsin hv

CHO

498 nm

<200 fs enzymatic cascade

11

y®\2^

photo, 570 nm phot # < . / • ' a l l - t rans" highly tlistorted

[ rhodopsin kinase

min, >0°

497 nm

+ H \ ms _ meta-ll, 0° 380 nm * ^ m e f a - / , - 1 5 ° 1^^ ( S B uunprotonated) (SB 480 nm

batho, -140 °, 543 nm "distorted trans" Figure 3. Visual transduction cycle

240 Progress in Biological Chirality Batho-Rh was detected as early as 1958 [6] and more properly characterized by absorption relaxes to lumi-Rh-, meta-I- Rh, and at the meta-II- stage the conformation of extramembrane loops on the cytoplasmic side excites the G protein, thus initiating a cascade of enzymatic reactions.[8] One light-activated Rh molecule (quantum yield 0.67) activates 1,000 molecules of G protein, which in turn through activation of phosphodiesterase hydrolyzes 10^ molecules of cyclo guanosine monophosphate (cGMP) to GMP. The bleaching of Rh to the active meta-II Rh allows the protein cavity to become more solvent accessible, and following a reprotonation of the Schiff base (Meta-III Rh), imine hydrolysis releases all-/r^A75-retinal from opsin. All-/raw5-retinal is then reduced to transretinol or vitamin A by alcohol dehydrogenase and is carried to the retinal pigment epithelium (RPE), where it is esterified with fatty acids by lecithin retinol acyhransferase (LRAT). Hydrolysis of the retinyl esters provide the uphill 4 kcal/mole energy needed to activate the isomerase leading to 1 l-cis retinol which is oxidized to 1 l-c/^-retinal that forms Rh. [9]

4.

3D Picture and 2.8 A X-Ray Structure of Rh Rh is one of the most thoroughly investigated G protein coupled receptors (GPCR) [10] consisting of seven hydrophobic transmembrane helices A, B,... G (or 1,2,... 7) interconnected by hydrophilic extramembrane loops. Bovine Rh is the most thoroughly studied of all the GPCRs because it is readily accessible from bovine eyes; the majority of studies on Rh are from this source. The first photo-affmity studies of Rh, performed by Nakayama and Khorana using the photoreactive tritiated trifluoromethyl-diazirine-phenyl-ll-c/5-retinal showed that the diazirine carbon had cross-linked to several amino acid residues in helices C and F [11]. To further define the cross-linked sites, site-specific mutant genes expressed in monkey kidney cells (named COS-1) were used to replace specific amino acids and mutant Rhs' were studied

intradiscal side

Lys296 helix G Trp265 heHxF

C4ermfiinus — ^ n .C-terrminus to,

cytoplasmic side

cytoplasmic side

Figure 4. a) Diazoketo-ret6 bound to opsin b) 2.8 A X-ray structure of Rh

CD and Visual Science

241

for their spectroscopic properties and G-protein activation. The results showed that Trp-265 is located close to retinal and probably also Glu-122, Trp-126, and Tyr-268 [12] (Figure 4). The retinal analog used in these early results used a freely isomerizable 11-c/^ configuration and hence it was likely that the multiple binding sites involved those to which the aW-trans-YQtinsil also bound. To simplify the results, [15-^H]-3-diazo-4-oxo-10,13-ethanoll-cis retinal or diazoketo-ret6 (see Figure 4a) where the ll-cz.s' to trans isomerization is blocked was employed. Incubation of this diazoketo-ret6 with bovine opsin gave a pigment in 53 % reconstitution yield. Irradiation of the pigment at 254 nm resulted in a 23% cross-link as estimated from the radioactivity of the receptor after separation from the unbound retinal by size-exclusion HPLC. The cross-linked apoprotein in membrane suspension was cleaved by V8 protease at Glu239/Ser240 on the cytoplasmic side into large (V8-L) and small (V8-S) fragments; cleavage yield ca.50%. After detergent solubilization and carboxymethylation of cysteine residues, the cleaved V8-L (large) and V8-S (small) fragments were separated by sodium dodecyl sulphate - polyacrylamide gel electrophoresis (SDS-PAGE). All radioactivity resided in V8-S, while no radioactivity was associated with the V8-L portion of Rh, indicating that the cross-linking site resided in helix F or G. Cyanogen bromide cleavage showed the radioactivity to be in CNBr peptide-13, and Edman degradation of this peptide disclosed Trp265 and Leu-266 as the crosslinked sites [13]. The fact that both the labeled Trp-265/Leu-266 residues and the Schiff base linkage to Lys296 are in the middle of the lipid bilayer showed that the chromophore in ground-state Rh has the polyene chain essentially parallel to the membrane plane and that the entire chromophore resides near the center of the bilayer, with the polyene long axis slightly tilted relative to the membrane plane. This orientation of the chromophore favors maximal exposure of the polyene chain to the incident light. After many years of effort by various groups, the first X-ray crystallographic structure of Rh was secured in 2000 [14]. In Figure 4b the chromophore is seen with its ionone ring close to Trp265. The X-ray structure serves as the basis for future references to Rh structure and dynamics.

4.

2.6 A X-Ray Structure of Rh The X-ray diffraction pattern with 2.6 A resolution succeeded in showing the critical role played by 6 water molecules in the transmembrane region [15]. It was suggested that the water molecules located in the vicinity of highly conserved amino acid residues in the binding site participate in salt bridge stabilization. The water molecules (shown in grey, figure 4c) are involved in the regulation of Rh activity and spectral tuning, i.e., regulation of absorption maxima of visual pigments.

5.

UVA^IS and CD Spectra of Rh and Diazoketo Rh Diazoketo-ret6 (Figure 4a) that allowed the diazoketo moiety to be photoactivated while retaining the ll-cis geometry clarified the ground state mode of binding of the chromophore in Rh. The next objective was to follow, if possible, the entire transduction process through the various intermediates of the Rh bleaching process, namely batho-, lumi-, meta-I and metaII Rhs (Figure 3). For this purpose, (llZ)-3-diazo-4-oxo[15--^Hi]retinal (see Figure 5, structure with -CTO instead of -CHO) containing two photoactive groups, the diazoketo and

242

Progress in Biological Chirality

Figure 4 c 2.6 A X-ray structure of Rh

ll-c/5-ene, was selected as the retinoid. This unstable retinoid was first prepared by submitting the all-trans diazoketone to retinochrome, an enzyme contained in squid eyes that isomerizes all-/raw5-retinal into the W-cis form. [16] Because of the presence of this isomerase in the eye, [17] [18] the photocycle of squid vision does not require exogenous replenishment of retinal. Retinochrome indeed performed the expected isomerization but this route was abandoned because the scale was limited to microgram quantities.

+ Opsin CHO -|

336 nm

j CD

/ ^

497 nm

306 nm

ID-

CD/ \

5

462 nm

< • 1 /

371 nm

-H

250

1

1

aOD

35Q

364 nm

/ A

j) uy/*\

496 nm 1

4DD 450 5DD Wavelength (nm)

L

5-

467 nm

r

-ID-

5SD

25D

3Da

35D

400

450

500

Wavelength (nm)

Figure 5. UV/VIS & CD of Rh and diazoketo-Rh (in CHAPSO / HEPES, pH 7.0)

550

BOO

CD and Visual Science

243

An efficient and general protocol for c/^-hydrogenation of an 11 -yne group to 11 -CIS-QIIQ in retinoids was developed employing Cu/Ag activated Zn dust[19]. This route was then employed to make the light sensitive and unstable 1 IZ-diazoketo retinal. As depicted in Figure 5, the Rh analog incorporating diazoketo-ret6, diazoketo-Rh (DKRh) has similar UV/VIS and CD spectra to those of native Rh. This suggested that the fit of diazoketo-ret6 into the binding site, despite presence of the 3-diazo-4-keto functionalities, was similar to that of ll-c/^-retinal. The leniency of the ionone binding site was not entirely surprising as it was shown earlier that even adamantyl allenic retinoids could be readily incorporated into opsin to give the corresponding Rh analog [20].

6.

Bleaching Intermediates of DK-Rh The bleaching intermediates of native Rh are shown together with the corresponding intermediates of DK-Rh (UV difference measurements A-D, Fig. 6). DK-Rh was dissolved in 66% v/v glycerol and irradiated by UV in the glass-like state. Irradiation with 436 nm light at -196 °Ccaused a red-shift in the absorption leading to a batho-Rh intermediate. Gradual temperature elevation showed formation of the three additional intermediates, Lumi-, Meta Iand Meta II-Rh. The photocross-linking was performed with the four intermediates at the temperatures indicated. For example, photocross-linking of DK-lumi-Rh, optimal temperature -80 ""C was performed as follows: 1) the (IIZ)-DK retinoid was incubated in the dark in 67% glycerol

1

1

6c^^ 11-c/s, 498 nm

_ ^ _ ^

1

11

^^ ^"'NN

1

t1

, OHO

<200fs

aW'trans, 385 nm

opsin

Batho, -140t3C, 543 nm "distorted trans"

min, >0t3C

'\ [IS

Lumi,-40tX: 497 nm

+|_|+

^ meta-l,-15t3C ^" 430 nm

Diazoketo-Rh

- ^ "^jZjT

meta-ll, Ot^ 380 nm

U ^ ^ " ' ' ^ Meta I

\:?^

r^^^N^w.^'hv

(C) - 40 °C

(A)-196°C -0.04 0.04

T" Lumi

0.00

' ence f^bso bance (B)-80°C

-0.04

\r\ (D) 0 °C -

500

I 400

Wavelength (nm)

Figure 6. Rhodopsin photocycle and the corresponding bleaching intermediates of diazoketo-Rh

244

Progress in Biological Chirality

with opsin at room temperature for 10 min to give DK-Rh, 467 nm ; 2) DK-Rh was irradiated with 480 mn light, 5 min at -196 °C tpyield DK-batho, 510 nm; 3) the temperature was raised to -80 °C, the DK-lumi temperature; 4) the lumi-Rh conformation was then maintained by freezing to -196 °C.and irradiated with 254 nm light for photocross-linking. The glycerol was removed by precipitating the Rh with trichloroacetic acid. The Rh analog was submitted to cyanogen bromide cleavage and the fragments were submitted to HPLC [21 J.

7.

Photocross-Linked Sites Along the Visual Transduction Path The small peaks with vertical bars around 33 min and 58 min in the HPLC trace were radioactive and indicated the CNBr-cleaved peptide fragments that contained the tritiated retinoid. The radioactive peptide at 58 min resulting from Rh and batho-Rh comprised part of helices F / G, while that from lumi, meta-I and meta-II-Rh at 33 min resulted from helix D. Edman degradation of each radioactive peptide showed that the cross-linked site in Rh and batho-Rh was cleanly only one amino acid, Trp265, while that in lumi, meta-I and meta-II-Rh was again only Ala 169 [21]. As depicted in the schematic drawing in Figure 7, this suggested that in the transition from batho-Rh to lumi-Rh, helix C moves out and helix D rotates so that Alal69 that was on the outside is now inside the binding site; simultaneously the W-cis to trans isomerization of the retinoid is accompanied by the flip-over of the ionone ring as depicted so that C-3 now becomes close to Ala 169.

Rh(dark) & batho (-196°) CN-13 (Val258 - Phe287)

Intradiscal

QQ^

0 ^

..©0®©

© Q Q Q j@©QQ

®Q®o

Cytoplasmic

mAII 1600

0

!•''

[^:

' '• •

20

30

\

•^

^ ,,._-j.._-''-^.-- 'I 10

rv-

^ ^

^^Bk

, 1

'1

1

A169

'''

!

800

%f%Q

\Rh \ bat ho

1

1

1200

400

•ooc ©

lumi meta-I meta-ll

, '

' ,

^'"^'"f 40

50

•••

1

•• 60

'--^ 70 min

lumi (-80°), meta-I (-40°) & meta-ll (0°) CN-9 (Ala164 - Gly182)

Figure 7. Reverse phase HPLC profile showing the labelled peptide fragments at each stage of the visual transduction path

CD and Visual Science

Rhodopsin

liqht

245

Meta II

Co-trans ^somerization Flip-Over of ;i--iGnone rbig

Figure 8. 1 \-cis->trans isomerization of the retinylidene chromophore modelled in the X-ray structure of rhodopsin

8.

Rhodopsin to Meta-II Conversion The left panel in Figure 8 is an expanded X-ray picture of the binding site of Rh showing the chromophore (in red) being close to Trp265 (in yellow); Lys296 of the Schiff base is in blue [15]. The right panel (created from the X-ray structure by Professor Y. Shichida, unpublished) shows the flipped-over retinal with its ring close to Alal69 on the yellow helix D, which has now rotated from its position in ground state Rh.

9.

The 6a- and 6p-Locked (llZ)-Retinals The two (llZ)-locked retinals 1 and 2 in Figure 9 were synthesized in order to determine the absolute sense of twist around the C6-C7 bond in the ground-state pigment [22]. As shown in Fig. 9-A, the a-enantiomer 1 was smoothly incorporated into opsin and gave rise to a CD (Fig. 9-C) similar to that of native Rh (Fig. 9-D). In contrast, the P-enantiomer 2 did not bind (Fig. 9-B). From these resuhs, it was concluded that the 6-s bond was twisted as depicted in the stereo drawing of the lowest energy conformation of 1 (Fig 9-E) [23].

10. (llZ)-Locked Cyclopropyl Retinal and its Binding to Rh The chromophore of (llZ)-locked cyclopropyl retinal-7 (CPR ret?) consists of a cyclopropyl group flanked by polyene and enal moieties, and hence was expected to clarify

246

Progress in Biological Chirality

6s-a-Locked retinal 1 binds t o opsin t o yield 6 a locked rhodopsin (Rh).

(6s-a-locked retinal

6s-p-locked retinal

However, 6s-p-locked retinal 2 does not bind.

403nm

(1.5 h) 539 nm

The CD of 6a-locked Rh with the a and P bands are similar to that of native Rh.

Conclusion: These results show t h a t T

1

t h e absolute sense of t w i s t

1

around t h e 6-s-cis bond of t h e 11-cis chromophore m Rh is negative as shown

336(+15) 6a-locked Rh

329 (+6.5)

(inCHAPSO/ HffES. pH6.6)

Figure 9. Absolute sense of twist around S-s-cis bond of 1 l-c/5-retinal chromophore in rhodopsin

the conjugative properties of the little known cyclopropyl group (CPR). As seen in Figs. 10-A and 10-B, the CD were of the exciton coupled bisignate type thus showing that the conjugative effect of CPR was minimal. Calculations also showed that two distinct conformations around the C6-C7 bond exist. Only the depicted 1 ip,12P-CPR enantiomer (Fig, 10-A), but not the corresponding a-CPR, binds to Rh (Fig. 10-D). The UV and CD of CPR Rh (Fig. 10-C) showed their extrema at 270 nm, blue-shifted from those of the protonated Schiff base formed with butyl amine (PSB, Fig. 10-B). This is unlike the case of other retinoids that exhibit their pigment extrema at longer wavelengths than the PSB, e.g., the PSB between retinal and butylamine absorbs at 440 nm whereas Rh absorbs at 500 nm. It is likely that after entering the binding site, presence of the P-CPR moiety results in a conformation unlike that of other retinoids. Detailed NMR measurements by DQF-COSY, NOESY and HSQC, and calculated energies by B3LYP/6-31G** (Jaguar 4.1) led to solution conformations E and F (Fig. 10) in which the C12/C13 bond is slightly positively twisted. Namely, the opsin binding cleft accommodates the p-CPR retinoid with a slightly positive twist or with the 13-Me infrontof the retinoid plane (see the stereo view in Fig. 10-F). This suggests that the native ll-c/5retinal enters the binding cleft with a positive 12-s-bond twist and that its conformation in the binding site is also positively twisted around the 12-s-bond [24].

CD and Visual Science

247

340 nm (-1.8) 246 nm (-10.1) ^268 nm (s 25,000) UV 200

2S0

300

350

«)0

200

E

3SD

400

4SQ

500

250

30O

350

400

450

500

wavelength (nm)

CPR-rhodopsin F

A 315 nm 5g|^ ^-^^

250

300

Protonated Schiff base (PSB)

aldehyde l^\^

250

wavelength (nm)

wavelength (nm)

300

"P"cyclo-Pr Binding

350

4O0

wavelength (nm)

Figure 10. (1 lZ)-locked cyclopropyl retinal and its binding to Rh

11. The Chromophore Conformation in Rh Results with the 6a-locked (1 lZ)-retinal (Figure 9) and (1 lZ)-locked CPR ret? (Figure 10) define, respectively, the biologically relevant conformation of (llZ)-retinal within the Rh binding site, as depicted by the twist between planes A/B and B/C (Figure 11).

12. The Mode of Entry of 13-Desmethylretinal into Opsin The 1980 observation that 13-desmethylretinal (13dmRet, see Fig. 12) and opsin induced dark activation of phosphodiesterase [25] was followed by a kinase assay that showed the dark activity to decay with time, i.e., the dark activity is transient. [26] The following mechanism of this dark activity was proposed based on studies with ISdmRet, its 11,12dihydro analog, and a further analog, ret6 [27]. The opsin cavity is divided into three domains, site I for the ring moiety, site II for the polyene chain, and site III for the protonated Shiff base (PSB). Step I: The retinal ring occupies site I. Step 11: When SB formation is blocked by dimethylation of Lys296 there is no transient activity. [26] Moreover, there is also no transient activity when the SB becomes protonated (see Fig. 12-A). Step III: As the chromophore enters the binding site with its cyclohexene ring in site I and the unprotonated Schiff base in site III, the polyene enters site II and adjusts its torsional angles. The ring-locked and rigid ret6 cannot achieve this and hence induces no dark activity (see Fig. 12-B). Step IV. In the case of the native Rh CD, only the physiologically active meta-II intermediate

248

Progress in Biological Chirality

Figure 11. Chromophore confonnation in dark Rh

exhibits a negative CD at 280 nm as opposed to the positive 280 nm CD of Rh and other intermediates. [28] During the process of settHng into the binding site, ISdmRet shows a negative Meta-II like CD around 270 nm 3-7 min after incubation. It is at this intermediate stage when the flexible polyene chain is about to settle into site II that transient activity is detected (Fig. 13-C). No dark activity has been reported for 1 l-c/^-'retinal; an explanation for this is that since the binding of 11-c/.y-retinal is 10-fold faster than 13dm, the retinal and opsin rapidly attain the inactive state or Rh.

13. Entry of the ll-m-Retinal Chromophore into Opsin Figure 13 depicts several retinoids 1-6 that have been incorporated into bovine Rh. Ret-7 binds smoothly and gives UV/Vis and CD spectra closely resembling those of native Rh [29]. This smooth entry showed that opsin has a potential cleft that readily incorporates ret-7. In contrast, ret-8 2 was not readily incorporated; it bound to extramembrane Lys groups to give confusing results that took a few years to resolve. It was only after the extramembrane Lys groups were permethylated to leave only Lys296 unmethylated that a pigment was successfully obtained. [30] This suggested that the conformation of the 8-membered ring was hindering the entrance and hence the other surface-lying Lys residues had to be inactivated before the entry of ret8. As described above in Figure 10, incorporation of the rigid ret-7 PCPR retinoid 3 defined the shape of the binding site around the C12-C13 bond. Similarly, the binding of 6a-locked (llZ)-retinoid 4 defined the sense of twist around the 6-S-C/5'bond (Figure 9). Photocross-linking of diazoketo retinal 5 to Rh indicated that it is the ring side of

CD and Visual Science

249

Rigid polyene: inactive site I"

site I Schiff base or H-bond

K296 site II flexible polyene cytoplasmic side Figure 12. Mechanism of transient dark activity in opsin

the retinal that flips over during the critical batho -> lumi transition. Furthermore, the transducin bioassay showed that the rhodopsin pigment containing 4 exhibited 80% activity of the native chromophore implying that the conformation around the 6-s-cis bond remains

1 — ret-7: binds smoothly

^ ^^^

ret-67-a-locked:

^HO

2

^—^

CHO

ret-8: binds with difficulty

O (llZ)-diazol<eto ret

CHO ret-7-p-CPR: binds defines 11,12,13 conformation

CHO

P-locked does not bind cdamantyl allene ionone binding site is lenient (i)

this side flips oveP; (ii) chromophore conformation in Rh

CHO

opsin "binding cleft" Figure 13. Retinal analogs used to determine chromophore conformation in dark Rh and chromophore movement in the various bleaching intermediates

CHO

250

Progress in Biological Chirality

unchanged in the native Rh bleaching process. This may not be surprising in view of the observation that even adamantyl allenic retinal 6 could be bound to opsin [20]. The fact that ret-8 2 was difficult to incorporate, and only retCPr 3, but not its enantiomeric form yielded a pigment, show that as the retinoid enters the intradiscal (N-terminal) side of the receptor from the side indicated by arrows in Figure 13-i, the conformation of the bottom side of the 7-membered ring in 3 can fit into the potential cleft in opsin; however, this is not the case for its enantiomer. Namely, we infer that the methylene portion of the sevenmembered ring in 3 is most likely predisposed with the proper conformation that can be accomodated by the receptor binding pocket. This allows us to conclude that the retinoids enter the Rh binding site from the side of the 5-methyl and aldehyde groups (see arrows in Fig. 13-i). Finally, as described in Figure 9, the bound conformation of the chromophore in Rh can be represented as in Figure 13-ii.

14.

Acknowledgements These studies were supported by NIH grants GM-36564 and GM-34509 (to K.N. and N.B.) and an NIH Vision training grant EY 13933-03 (to N.F.)

15. References [I] Y. Katsuta, M. Ito, K. Yoshihara, K. Nakanishi, T. Kikkawa and T. Fujiwara, Synthesis of (+)-(4S)- and (-)-(4R)-(l lZ)-4-hydroxyretinals and determination of the absolute stereochemistry of a visual pigment chromophore in the firefly squid, Watasenia scintillans. J. Org. Chem. 59 (1994) 6917-21. [2] M. Ito, T. Hiroshima, K. Tsukida, Y. Shichida and T. Yoshizawa, A novel rhodopsin analogue possessing the conformationally 6-s-67.s-rixed retinylidene chromophore. J. Chem. Soc, Chem. Commim. (1985) 14431444. [3] Y. Shichida, S. Matuoka and T. Yoshizawa, Formation of photorhodopsin, a precursor of bathorodopsin, detected by picosecond laser photolysis at room temperature. Photochem. Photobiophysics 7 (1984) 221228. [4] R.W. Schoenlein, L.A. Peteanu, R.A. Mathies and C.V. Shank, The first step in vision: Femtosecond isomerization of rhodopsin. Science 254 (1991) 412-415. [5] A. Cooper, Energy uptake in the first step of visual excitation. Nature 254 (1979) 531-533. [6] T. Yoshizawa an^ Y. Kito, Chemistry of the rhodopsin sycle. Nature 182 (1958) 1604-1605. [7] T. Yoshizawa and G. Wald, Pre-lumirhodopsin and the bleaching of visual pigments. Nature 197 (1963) 1279-1286. [8] D. Kliger and J. Lewis, Spectral and kinetic characterization of visual pigment photointermediates. Isr. J. Chem. 35(1995)289-307. [9] R. Rando, The Biochemistry of the Visual Cycle. Chem. Rev. 101 (2001) 1881-1896. [10] Y. Shichida and H. Imai, Visual pigment: G-protein-coupled receptor for light signals. Cellular & Mol. Life Sciences 54 (1998) 1299-1315. [II] T. Nakayama and H.G. Khorana, Orientation of Retinal in Bovine Rhodopsin Determined by Cross-linking Using a Photoactivable Analog of 1 l-c/5-Retinal. J. Biol. Chem. 265 (1990) 15762-15769. [12] T.A. Nakayama and H.G. Khorana, Mapping of the amino-acids in membrane-embedded helices that interact with the retinal chromophore in bovine rhodopsin. J. Biol. Chem. 266 (1991) 4269-4275. [13] H. Zhang, K.A. Lerro, T. Yamamoto, T.H. Lien, L. Sastry, M.A. Gawinowicz and K. Nakanishi, The Location of the Chromophore in Rhodopsin: A Photoaffinity Study. J. Am. Chem. Soc. 116 (1994) 1016510173. [14] K. Palczewski, T. Kumasaka, T. Hori, C.A. Behnke, H. Motoshima, B.A. Fox, 1. Le Trong, D.C. Teller, T. Okada, R.E. Stenkamp, M. Yamamoto and M. Miyano, Crystal structure of rhodopsin: A G protein-coupled receptor. Science, 289 (2000) 739-45.

CD and Visual Science

251

[15] T. Okada, Y. Fujiyoshi, M. Silow, J. Navarro, E.M. Landau, and Y. Shichida, Functional role of internal water molecules in rhodopsin revealed by X-ray crystallography. Proc. Natl. Acad. Sci. USA 99 (2002) 5982-5987. [16] B. Borhan, R. Kunz, A.Y. Wang, K. Nakanishi, N. Bojkova and K. Yoshihara, Chemoenzymatic synthesis of 11-cis-retinal photoaffinity analog by use of squid retinochrome. J. Am. Chem. Soc. 119 (1997) 57585759. [17] T. Hara and R. Hara, Rhodopsin and retinochrome in the squid retina. Nature 2\A (1967) 573-575. [18] T. Hara and R. Hara, Retinochrome and rhodopsin in the extraocular photoreceptor of the squid, Todarodes. J. Gen. Physiol. 75 (1980) 1-19. [19] B. Borhan, M.L. Souto, J.M. Um, B. Zhou and K. Nakanishi, Efficient Synthesis of 1 l-c/^-Retinoids. Chem. Eur. J. 5 (1999) 1172-1175. [20] R.A. Blatchly, J.D. Carriker, N.V. Balogh and K. Nakanishi, Adamantyl allenic rhodopsin. Leniency of the ring binding site in bovine opsin. J. Am. Chem. Soc. 102 (1980) 2495-2497. [21] B. Borhan, M.L. Souto, H. Imai, Y. Shichida and K. Nakanishi, Movement of Retinal Along the Visual Transduction Path. Science 288 (2000) 2209-2212. [22] Y. Fujimoto, R. Xie, S.T. Tully, N. Berova and K. Nakanishi, Synthesis of 1 l-c/^-locked bicyclo [5.1.0]octanyl retinal and an enantioselective binding to bovine opsin. Chirality 14 (2002) 340-346. [23] Y. Fujimoto, J. Ishihara, S. Maki, N. Fujioka, T. Wang, T. Furuta, N. Fishkin, B. Borhan, N. Berova and K. Nakanishi, On the bioactive conformation of the rhodopsin chromophore: absolute sense of twist around the 6-s-cis bond. Chem. Eur. J. 1 (2001) 4198-4204. [24] Y. Fujimoto, N. Fishkin, G. Pescitelli, J. Decatur, N. Berova and K. Nakanishi, Solution and biologically relevant conformations of enantiomeric 1 l-cw-locked cyclopropyl retinals. J. Am. Chem. Soc. 124 (2002) 7294-7302. [25] T. Ebrey, M. Tsuda, G. Sassenrath, J.L. West and W.H. Waddel, Light activation of bovine rod phosphodiesterase by non-physiological visual pigments. FEES Lett. 116 (1980) 217-219. [26] J. Buczylko, J.C. Saari, R.K. Crouch and K. Palczewski, Mechanisms of opsin activation. J. Biol. Chem. 271 (1996)20621-20630. [27] Q. Tan, K. Nakanishi and R. Crouch, Mechanism of transient dark activity of 13-desmethylretinal/rod opsin complex. J. Am. Chem. Soc. 120 (1998) 12357-12358. [28] T. Okada, T. Matsuda, H. Kandori, Y. Fukada, T. Yoshizawa and Y. Shichida, Circular dichorism of metaiodopsin-II and its binding to transducin - A comparative study between meta-II intermediates of iodopsin and rhodopsin. Biochemistry 33 (1994) 4940-4946. [29] H. Akita, S. Tanis, M. Adams, V. Balogh-Nair and K. Nakanishi, Nonbleachable rhodopsins retaining the full natural chromophore. J. Am. Chem. Soc. 102 (1980) 6370-6372. [30] S. Hu, P.J. Franklin, J. Wang, B.E.R. Silva, F. Derguini and K. Nakanishi, Unbleachable rhodopsin with an 11-cis-locked eight-membered ring retinal: The visual transduction process. Biochemistry 33 (1994) 408416.

Progress in Biological Chirality G. Palyi, C. Zucchi aiid L. Caglioti (Editors) © 2004 Elsevier Ltd. All rights reserved.

Chapter 20 Volatile Compounds in Food Aroma: Biosynthesis and Biotransformations Franco Bellesia,^ Adriano Pinetti,^'* Ugo M. Pagnoni,^ Riccardo Rinaldi,^ Claudia Zucchi^ Luciano Caglioti'' and Gyula Palyi^ "^Department of Chemistry, University ofModena andReggio Emilia, Via Campi 183,141100Modena, Italy pinetti. adriano@unimo, it AITEL, Associazione Tecnici del Latte, Via Galli 18,1-41100 Modena, Italy "" Department of Chemistry and Technology of Natural Substances, University 'Xa Sapienza' Piazzale Aldo Moro 5,1-00185 Roma, Italy

Flavouring compound analysis has widely grown with the introduction of easy sampling methods such as solid phase microextraction (SPME) and purge and trap (P&T) [1]. The high quantity of achievable data makes possible the research of some manufacture and ageing markers, or the finding of information about the biosynthesis and biotransformations occurring during the shelf-life. In this case some suggestions about their use in the production of asymmetric compounds may arise. Enz-B H

Enz

^N^CH3

H^

H\

H ; ^ N ^ C H 3

Tyrosine-piridoxal adduct

^0PO^

H'' ^l/^CHj

Tyrosine

Phenol Scheme 1

254

Progress in Biological Chiralitv

Serine + PMP +NH3 H+

/

While studying the composition of the volatile organic compounds (VOCs), a mixture responsible for the flavour [2] and for the fat alteration products [3] in Parmesan cheese, we found a release of phenol, which could arise by the action of tyrosine-phenol liase (TPL) [4, 5, 6] through an intermediate complex between tyrosine itself and pyridoxal phosphate (Scheme

1). Such an intermediate is also formed in the nucleophilic displacement of the serine hydroxy group in many transformations (Scheme 2), occurring both in animal and plant kingdom, leading to very important compounds, for example to thyptophan and L-DOPA when an indolyl anion or catechol respectively act as nucleophiles; Lactobacilli and Streptococci, featuring components of the bacterial flora of cheeses, can thus be used to obtain the above mentioned compounds, or many others having pharmacological or industrial appeal, by fermentation methods, instead of using expensive gene-modified microrganisms as biocatalysts.

Rib H

Rib H

Rib

H 3 C ^ ^ N % " H ^ H 3 C " ^ N 1 Y ^

H 3 C ^ ^ N ^ ^ H

FADH2

FAD

Ambrettolide

'^^

'^CH^Q''^

Rib ^ ^OO H H3C^^^v^Ny^^0 H3(X^^N^j^.H HoO^ H3C^^N+Y""H

H I O OH

H3C^^r:J^N^o

Rib Scheme 3

Volatile Compounds in Food Aroma: Biosynthesis and Biotransformations O ^

255

—HOO)

r>Xr^ONH2 V5J

Methyl ketones

L-^1 ^ j

(H3.

A-o^V R ) _|

O

Methyl esters

NAD"" Scheme 4

O Acetic esters

Other biotransformations encountered in this research involve some Bayer-Villiger microbial oxidations [7, 8, 9] occurring on methyl ketones, arising from an altered fatty acid degradation pathway, and some compounds, as muscone, coming from the feeding of cows. While in the oxidation of the former compounds to methyl esters no chiral centre is formed, on the latter ones a new stereocentre is created (Scheme 3). So even in this case it would be possible to use the normal milk bacterial flora in some fermentations to produce, for example, flavouring lactones [10, 11] enantiomerically pure from cyclic asymmetric ketones. Only methyl esters of even carbon atom number short chain fatty acids are found within the VOC mixture, indicating that the oxidation of methyl ketones occurs with high regioselectivity, supporting a reaction pathway similar to the one depicted in Scheme 3, with the peroxy-intermediate linked to the enzyme co-factor. Other pathways, like the one involving the release of a free hydroperoxyde anion by molecular oxygen through a NADH intervention (Scheme 4), seem not to work with sufficient selectivity. During the detection of VOC-s in various samples of Parmesan cheese, many a discontinuity have been found. A first source of error arises from the sample size. For SPME or P&T methods, few milligrams of cheese are required, a quantity too low to be really representative. This problem can be avoided by collecting the samples from a larger amount (about 5 grams) of homogeneously finely grated cheese. A further discontinuity is due to the complex migration process of small molecules inside the form during ageing. This accounts for differences in composition found among the various parts of the bulk. The processes are most likely linked to the hydrophilicity-lipophilicity of the migrating compounds (the rim is more "greasy" than the inside parts). A previous study [12] has brought to the suggestion that an ideal surface of the same concentration of each migrating product (flavours, free amino acids, free fatty acids etc..) might have the shape of a "red blood cell" (Figure 1). As all the releases of these molecules are a frmction of ageing, probably the distance between the rim and the surface, relative to a particular compound concentration or to a certain ratio of compounds, may be taken as a ripening marker. This concept can also be applied to the levels of epimerization of L- to D-amino acids, whose D/L ratio could be represented by an analogous set of surfaces.

256

Progress in Biological Chirality

Figure 1

Acknowledgement Financial support is acknowledged to the town council of Serramazzoni (Modena, Italy), to the Comunita Montana del Frignano (Modena, Italy), to the [Italian] Ministry of Instruction, University and Research (MUIR) and to the [Italian] National Research Council (CNR).

References [1] (a) F. Bellesia, A. Bianchi, A. Pinetti and B. Tirillini, Fl. Fragr. J. 11 (1996) 239. (b) Idem, Riv. Ital E.P.P.O.S. 23 (1997) 41. (c) Idem, Fl. Fragr. J. 13 (1998) 56. (d) Idem, J. Ess. Oil Res. 10 (1999) 483. [2] F. Bellesia. A. Pinetti, U. M. Pagnoni. R. Rinaldi. C. Zucchi, L. Caglioti and G. Palyi, FoodChem. 83 (2003) 55. 13] C. Zucchi, F. Bellesia, A. Pinetti, L. Simon-Sarkadi, B. Weimer, L. Caglioti and G. Palyi, Featuring compounds of the greasy base of the Parmesan cheese aroma 3'''^ Italian-French Meeting on Organic Chemistry. Organic Chemistry towards Interfaces. Pisa, 20-23/11/2002 Abstract p. 60. 14] S.V. Plevnev, M.N. Isupov, Z. Dauter, K.S. Wilson, N.G. Faleev, E.G. Harutyunyan and T.V. Demidkina, Biochem. Mol. Biol. Int. 38 (1996) 37. [5] H.Q. Smith and R.L. Somerville, J. Bacteriol. 179 (1997) 5914. 16] H. Chen and R.S. Phillips, Biochemistry 32 (1993) 11591. 17] (a) V. Massey, J. Biol. Chem. 269 (1994) 22459. (b) C.T. Walsh and Y.-C. Jack Chen, Angew. Chem., Int. Ed Engl. 28 (1988) 333. 18] S. Rissom. U. Schwarzhnek, M. Vogel, V.I. Tishkov and U. Kragl. Tetrahedron: Asymmetry 8 (1997) 2523. 19] M.T. Bes, R. ViUa, S.M. Roberts, RW.H. Wan and A. Willetts. J. Mol. Cat. B 1 (1996) 127. 110] C Fuganti and G. Zucchi, Chim. Ind 79 (1997) 745. HI] G. Feron, P. Bonnarme and A. Durand, Trends FoodSci. Tech. 1 (1996) 285. [12] C Zucchi, F. Bellesia, A. Pinetti, L. Simon-Sarkadi, B. Weimer and G. Palyi, unpublished results.

Progress in Biological Chirality G. Palyi, C. Zucchi and L. Caglioti (Editors) © 2004 Elsevier Ltd. All rights reserved.

Chapter 21 Origin of the Biological Chirality N. M. Chemavskaya, D. S. Chernavsky, R. F. Polishchuk Lebedev Physical Institute of the Russian Academy of Sciences rpol@asc. rssi. ru

Life is a negentropy current providing self-correction by the genetic code and supporting itself by free energy affluent [1]. Origin of the biological chirality in frame of the biosphere is an example of the spontaneous breaking of symmetry due to the energetic advantage criteria. Several organic molecules have a right (D-) and left (L-) chirality (helicity) and turn a light polarization vector to the right or left direction respectively. Liquids and crystals have a form of the D-junctions or L-junctions. More complex organic junctions may include fragments with both chiralities. The inorganic part of nature has a racemic "mixture" of its elements with the both chirality forms. In organic nature this symmetry gets to be broken. In nature all the amino acids (except glycine) are present in the chiral L-form. But L-form of its main fragments may be combined with the helicity of other fragments. All molecules of the C-5 sugar units (ribose) in the main chain of the ribonucleonic acids have a right helicity. Most of the nucleotides of the lateral ribopolymer groups have a left helicity. Why is the racemic mixture absent in living nature? Because a racemic mixture needs a double composition of proteins. All the chemical reactions in organisms are catalysed by enzyme-proteins. Each of these proteins have an active centre that is complementary for the substrate. Biological organisms with the double compositions of proteins are too cumbersome for successfial competition for the maintenance of life. We must take in account also one fact that for the proteins, catalysing reactions with substrates of some chiral protein-antipodes are acids, inhibitors. These inhibitor effects for DNA (desoxiribonucleonic acids) are considered in [2]. It was demonstrated that in a racemic mixture of nucleotides the enantiopure DNA-chains only with 50-100 links can be formed. Fore the longer chains the probability of the additional antipodes is increasing by factorial law. In modern biosphere the DNA-chain which codes for one fijnctional protein (approximately 200 amino acids) must contain more than 600 nucleotides. That is why in work [2] are introduced biological (number of links is more than 600) and chemical (less than 50) levels of complexity. In a racemic world non-enantiopure DNA with only chemical complexity level can duplicate. The choice of chirality type may be connected with (weak) external asymmetric influence. Polarized neutrino flows promote D-sugars to be synthesed and L-amino acids [2]. But this

258

Progress in Biological Chirality

influence is too weak to provide observable asymmetry. It is a more realistic hypothesis to suppose the spontaneous breaking of symmetry [3]. In the work [3] the following informational dynamic system is proposed: dii^ Idt- 2/, / r, + Z,^ .b,.uM, - a,w ^ + Z), Aw,

Here w, is a number of the /-th element (a concentration) in a family of w possible types, i; is a characteristic of the autoreproduction time. The term btj Ut tij describes the antagonistic interaction of the different elements. The term a^w,^ corresponds to the "narrowness" effect; D, is a diffusion coefficient. This dynamical system is multistable. Its development is dividing into separate phases: the origin of clean regions (clusters) in phase space, its extension and then a "victory" of one cluster. In the case of the biological chirality the number of possible clusters is two. This model describes also a chemical reaction of the synthesis of molecules with one centre of chirality from a racemic mixture of molecules with the different chiralities with two conditions: (a) new chiral molecules stimulate the synthesis of new molecules with the same chirality (autocatalysis), (b) the meeting of two antipodes removes both of these antipodes from the reaction sphere (antagonistic interaction). If the racemic state is nonstable then one of the chiral molecules will form. Its chirality type is accidental. Near a bifurcation point the evolution of a dynamical system may be described by the following equation: dTj/dt=ST]-rj\

7] = (C. - C.)/(C, + C.)

where rj is a chirality parameter, C+ and C. are concentrations of the D- and L-enantiomers, respectively, ^ is a (small) bifurcation parameter. If s <0, then the racemic state is stable (77 = 0). FoTS >0 there are stable non-racemic states TJ = ±4s . For the biological systems we have s ~ 1 and 77 = ±1 giving non-racemic products. Non-catalytic reactions are excluded. A very important problem of the origin of biological chirality is the time of its appearance: is it in the biological period or already in pre-biological one? We are convinced of the appearance of biological chirality in the period of the primitive hypercycles. In fact, the length of DNA-molecules in this period was 30-50 units. These molecules did not perform at this period any biological coding function but they were performing the chemical function of heterogeneous catalysts. Probably these molecules were in the homochiral form. The enantiopurity of the proteins was important in this phase. The primitive hypercycle processes could take place in racemic nucleotide mixtures: external homochirality was not needed for this. Also non-autocatalytic synthesis of these DNA-molecules can practically be excluded. These reactions were of the middle-level complexity - between the biological complexity level and the chemical one. The two "antipode" forms of the enantiopure hypercycles were arising in this period with equal probability. Then one of these antipodes was squeezed out by an other one. In this case the antagonistic interaction was more strong than it was between the different biological codes: one of the antipodes was "eaten" by the other one and it was eliminating itself due to this poisoning.

Origin of the Biological Chirality 259

The products of biological organisms are nutritious substances for other organisms. Of course these substances do not necessarily have chiral symmetry. The production of the opposite chiral isomers often has a protection function. As a consequence of the above argumentation the non-stability of the symmetrical state is a principal cause of the biological asymmetry. An external influence is not needed here. There are however some intermediate variants, considering the influence of the neutrino flux, which are discussed in [4]. These variants are possible if the parameter s changes gradually from negative to positive value and if the concentrations of antipodes are sufficient. Near the bifurcation point {s = 0) even a little external influence can change the choice of the finite state. In this case the Lyapunov parameters (in exponential evolution law for the dynamical system) are small and even fast fluctuations may be averaged yielding small s >0. Then our equation has the following form

drj /dt=S + 6rj-7]\

ri«\

In this case there is a stable evolution branch where a bifurcation for s = 0 is absent and for s> 0 stable and non-stable (a separative) branches are appearing. For s « 1 these branches approach (as 5^'^ ) to the main stable evolution branch. Formally here an external factor is the cause of the evolution picture but this variant is not realistic. More real is the situation when the concentration of hypercycle is small but its relative fluctuations are large and the antagonistic interaction is strong. Here a symmetric state is strongly non-stable and the Lyapunov's parameters are large. Then fluctuations of concentrations can not be averaged to the fast evolution. Here external factors as a neutrino influence are not important and the choice of an evolving variant is determined by the initial fluctuations accidentally. References [1] R.F. Polishchuk, in: Fundamentals of Life (Eds. G. Palyi, C. Zucchi and L. Caglioti) Elsevier and Accademia Nazionale di Scienze, Lettere ed Arti (Modena), Paris, 2002, p. 141-151. [2] V.A. Avetisov and V.I. Gol'dansky, Uspekhi Fiz. Nauk 160 (1996) 873-890 (in Russian). [3] D.S. Chemavsky, Synergetics and Information, "Znanie", Moscow, 1990 (in Russian). [4] D.K. Kondepudi and G.W. Nelson, Physics A: Statistical Mech. Appl. 125 A (1984) 465-496.

Progress in Biological Chirality G. Palyi, C. Zucchi and L. Caglioti (Editors) © 2004 Elsevier Ltd. All rights reserved.

Chapter 22 Chiral Spaces in Encapsulation Complexes Alessandro Scarso, Alexander Shivanyuk, Osamu Hayashida and Julius Rebek, Jr. * The Skaggs Institute for Chemical Biology and The Department ofChemistiy, The Scripps Research Institute, MB'26, 10550 North Torrey Pines Road, LaJolla, CA 92037 (USA) jrebek@scripps. edu

Encapsulation complexes are assemblies in which small molecule guests are completely surrounded by large molecule hosts [1, 2, 3, 4]. The hosts are made up of subunits held together by intermolecular forces: hydrogen bonds, van der Waal's forces and metal-Hgand contacts. The assemblies are formed reversibly and are dynamic; they come together and dissipate on time scales ranging from milliseconds to days, long enough for their study by NMR methods. When multiple hosts can assemble from a given set of subunits, template effects can be expected and they have recently been reported [5, 6, 7, 8, 9], The synthesis of capsules with asymmetric spaces has also been accomplished, but enantioselection of guests within these hosts is inadequate, whether the capsules are held together with covalent bonds [10], hydrogen bonds [11] or metal/ligand interactions [12, 13]. Moreover, the syntheses are lengthy and problematic and in this chapter we report alternatives. The first involves creation of capsules from achiral precursors and imprinting them chiral guests. The second involves the space left in a large achiral host when a small chiral guest is encapsulated. While the enantioselectivity is still low, the methods may have wider applicability. We consider the capsule la-la (Figure la [5]), formed when two self-complementary subunits la dimerize in organic solvents. A seam of eight hydrogen bonds holds the subunits together. The dimer has only C2 axes and exists as a pair of enantiomers even though each subunit features a plane of symmetry. The cavity of the capsule is chiral and asymmetric guests generally prefer one enantiomer of the capsule to its mirror image. The enantiomeric capsules racemize by complete dissociation and recombination of their subunits - a slow process - but guest exchange does not require this dissociation. Rather, guests get in and out of these capsules through flaps opened by the breaking of hydrogen bonds - a faster process [14, 15, 16]. A related structure lb with additional hydrogen bonding possibilities was prepared [17]. It has a slower rate of racemization when it dimerizes into capsules lb-lb. Parts of the NMR spectrum of lb-lb alone and with added chiral guest, (+)-pinanediol (+)2 are shown in Figure 2 (Figure lb). The broad signals of Ib^lb in /?-xylene-J;o are characteristic of unspecific aggregation. On the addition of 3 equivalents (+)2 a sharp spectrum is obtained in which the

262

Progress in Biological Chirality O

Y 0 R = n-heptylphenyl

H. Y

Y

OH

?" OH

^-*Ny

H-2

(+)-2

Figure 1. (Top) Line drawing of the monomers 1. (Middle) Energy-minimized [18] model of the dimeric assembly lb»lb. Some hydrogens and the /i-heptylphenyl groups of the dimer are omitted for clarity. (Bottom) Guests used in the study

A B

tJJi^

'>A«**^*ww^^«x/^^v'>nA'

X

B BfB* >H*vwyw-.v-*f

8.65

8.50

8.45

WH#«

2.90

i

T

B+B-

A+AI 2.75

2.60

Figure 2. Selected NH resonances of ^H-NMR spectra showing of the dimer (8.65-8.45 ppm) and Ha peaks for the guest in the enantiomeric capsules (2.90-2.60 ppm). The labels A and A* are the thermodynamically more stable complexes of (+)2 and (-)2, respectively, while B and B* are the less stable complexes, a) Ib^lb alone in /7-xylene-c/io (0.825 mM); b) 4 min and c) 120 h after addition of 3 eq of (+)2; d) 17 min and e) 191 h after addition of 30 eq of (-)2 other than those of the acids

Chiral Spaces in Encapsulation Complexes

263

two diastereomeric complexes are observed in approximately equal amounts (Figure 2b). After a few days, the system reaches its thermodynamic equilibrium at a 50% diastereomeric excess of the favored isomer (Figure 2c). The half-life for equilibration --20 hours) is the lifetime of the assembly. When 30 equivalents of the enantiomer (-)2 was added to the equilibrated mixture, the ratio of the two diastereomers partially inverts. This indicates a situation - unique to reversibly formed assemblies - in which a temporary excess of the less stable diastereomeric complex exists (Figure 2d). This occurs because the guest exchange process is much faster than dissociation of the two halves of the capsule. The rate of guest exchange (ti/2 ~ 1 min) is in line with that observed in similar systems [16]. The new guest (-)2 is now present in the capsule that formed preferentially around (+)2. Eventually the mixture of diastereomers returns to equilibrium. The half-life is -- 10 hours and the original 50% d.e. is achieved (Figure 2e). The rate is twice as fast as that measured for the initial equilibration, and is likely due to the high concentration of the polar, hydrogen-bonding 2 in the exchange mixture [19]. Additional competition experiments established the situation scheme shown in Figure 3. The memory, or "ghost" persists through multiple guest exchanges. For example, a solution of lb»lb templated with (+)-2. The solvent and the excess guest was removed, leaving only lb»(+)2»lb. The complex was redissolved in benzene and excess 1-adamantanol, 3, was added. This displaced the guest (+)2 into the solution. The solvent was removed and extraction gave a solution free of (+)-2. However, we were unable to detect any optical activity by polarimetry of this solution. When this solution was treated with excess (-)2, the NMR spectrum again showed a 2:1 excess of the less stable diastereomeric complex. Accordingly, that the chiral memory was maintained even in the absence of chiral guests. The memory of the capsule reflects the relative rates of guest and monomer exchange. The

V^«10-20h

Figure 3. Cartoon representations of the exchange equilibria. Guest exchanges are the horizontal equilibria and monomer exchanges are the vertical equilibria. The cartoons not intended to imply the assign a stereochemistry to the assemblies

264

Progress in Biological Chirality

Figure 4. (Left) Cartoon for monomer exchange by dissociation; (center) guest exchange by disrupting hydrogen bonds; (right) energy-minimized structure of the intermediate showing the openings for guest exchange exchange of subunits l b requires the complete dissociation of the capsule (Figure 4a), while guest exchange proceeds through openings that only disrupt a part of the seam of hydrogen bonds that hold the capsule together (Figure 4b). The inversion of the pyridazinyl ring of l b creates openings in the dimer that are large enough to allow passage of incoming and outgoing guests (Figure 4c [15]). Many thousands of molecules enter and depart the capsule during its lifetime, but each guest will experience the imprinted asymmetric microenvironment of the capsule. Examples of nonracemic chiral encapsulation complexes are rare [20, 21, 22] and the imprinting [23, 24] described here provides another route to such systems. It is likely that similar behaviors will also emerge from capsule held together by metal-ligand interactions [8]. The second system showing chiral spaces requires the encapsulation of two different guests. Consider placing a chiral object in an achiral space, a glove in a box for example (Figure 5). Since the glove is chiral the space remaining in the box becomes chiral. Now if a second chiral object enters the box, some selection is expected, especially if the two objects interact. That space should distinguish between enantiomers of the second guest. Relevant here for the molecular scale is the notion of "peristatic" chirality [25]. Reversibly formed capsules capable of surrounding two guest molecules - coencapsulation - are recent inventions [26] that can show of new forms of stereoisomerism [27], suggest possibilities for data storage [28] and permit encapsulated bimolecular reactions [29]. The guests of a coencapsulation complex are confined in space and time; they have much longer contact times and more defined orientations than the diffusion complexes in bulk solution. Diastereomeric complexes can be formed with two chiral guests, and either guest alone in the capsule would leave a chiral remaining space, but capsules are not formed if they are not properly filled. Instead, capsules are known to select guests that provide good fits. A good fit for the liquid phase means filling -55% of the cavity [30]. We explored the concept with the

^j-

•

Figure 5. A chiral object in an achiral container leaves a chiral space

Chiral Spaces in Encapsulation Complexes H.

265

0

Figure 6. (Top) Line drawing of the snbnnit and the ball and stick representation of capsule 4«4. Long, peripheral pendant chains have been removed. Cartoon representation used elsewhere in this work is shown on the left. Bottom: Coencapsulation complexes. Size matching (left) Coencapsulation of the small cyclopropane occurs when the large p-xylene is available. The enantiomers of cyclohexane diol (center) are slightly favored over two identical molecules (right)

cylindrical capsule of Figure 6. Earlier we found that an enantiomeric pair of cyclohexane diols fills the space in a capsule slightly better than two molecules of the same handedness (Figure 6) [31]. This observation may be related to the preference in nature for centrosymmetric crystals or, alternatively stated, the higher melting points of racemates vs enantiopure compounds. This generality is far from absolute, as we shall relate below. The capsule is too short to accommodate two p-xylenes and cyclopropane alone is not encapsulated. But a mixture of cyclopropane and p-xylene, 4*4 results in the coencapsulation complex [32]. The NMR spectrum of this arrangement shows different signals for the methyl groups of the encapsulated xylene. The separate methyl signals result from two restricted motions of the p-xylene guest: tumbling is slow on the NMR timescale and the two guests are too large to slip past each other while within the capsule. Shorter aromatics tumble rapidly on the NMR timescale inside the capsule [33]. The dimensions of the capsule select appropriately sized combinations and its shape can be used to fix rigid molecules. Polar functions are attracted to the seam of hydrogen bonds that holds the capsule together. These inherent characteristics of the capsule were useful for the enantioselection. A combination of (R)-styrene oxide 5 and isopropyl chloride in a solution of 4*4 in mesitylene-di2 gives an NMR spectrum showing the coencapsulation complex that features

266

Progress in Biological Chirality

*+x

II

Krl, I —

-2.0

-2.4

-2.8

d X

,|(

1 '^ i

>^' -2.4

-2.6

-2.8

-2.0

IMJJ^ -3.0

-4.0

8 Figure 7. Upfield region of the ^H-NMR spectra (600 MHz, 300 K) of coencapsulation complexes of 4*4 in mesitylene-di2 and various liquid guests, a) (R) - styrene oxide with i-PrCl; b) phenylcyclopropane with i-PrCl; c) (R) - styrene oxide with (±) 2-BuCl: •, * the 4- and 1- methyl groups, respectively, of 2-BuCl encapsulated with (R) - styrene oxide; the signals marked x and • are the 4- and 1-methyl groups, respectively, of the capsule containing two 2-BuCl guests [34]; d) (S)-mandelic acid with (±)- 2-BuOH: • is the 4-methyl group of (R) 2BuOH and • is the 4-methyl group of (S) 2-BuOH coencapsulated with the acid; x represents the 1-methyl groups of both diastereomeric complexes; the signals maiiced * and • are the 1 and 4 methyl groups, respectively, of the capsule containing two 2-BuOH guests

diastereotopic methyl groups of the smaller guest (Fig. 7). The large upfield shifts place its methyls near the end of the capsule while the observed magnetic anisotropy place the epoxide function near the capsule's center. In contrast, the coencapsulation of phenylcyclopropane and isopropyl chloride showed a complex with only one doublet at -2.61 ppm (Fig. 7b). The chiral epoxide with coguests racemic 2-butyl chloride, 2- butanol or 2-pentanol gave capsules including one molecule of each guest. Two sets of signals in 1:1 ratio correspond to the methyl protons of the encapsulated halide (Fig. 7c), and the chemical shifts indicate that the ethyl group of the halide is near the end of the capsule. No diastereoselection was observed, even though the asymmetric features of both guests are near the capsule's center, because no attractive forces are in play between the two guests. Their interface lacks steric contacts and they stay as far apart as possible (Figure 8) [35]. We had expected that racemic 2-butanol or 2-pentanol with (R)-styrene oxide would provide hydrogen bonds between the coguests but no diastereoselectivity was observed. With these alcohols, other diastereomeric complexes are also present. These arise from the encapsulation of two molecules of 2-butanol or 2-pentanol (i.e. R+S, R+R and S+S couples). Even so, no stereoselection is detected, yet the asymmetric centers are near the middle of the capsule and intermolecular hydrogen bonding between the two alcohols should exist. We obtained some better results with the superior donor (S)-mandelic acid 6. The ^H-NMR spectrum showed the formation of two diastereomeric capsules (Figure 7d), and their ratios

Chiral Spaces in Encapsulation Complexes

267

Figure 8. (Left) High packing coefficients (large spheres) force the interdigitation of large, medium and small groups. (Right) Smaller coguests result in a more remote arrangement of asymmetric centers and lower enantioselection

varied with temperature: 1.1 at 303 K and 1.3 at 283 K. Coencapsulation of 6 with (R)-2butanol established the identity of the diastereomeric complexes. This enantiomer was the better guest for the coecapsulation with (S)-mandelic acid. The contacts between mandelic acid and butanol guests were found in the energy minimized structures shown in Fig. 9. This arrangement brings the asymmetric centers closer but they are still at too far away from one another.

Figure 9. Structures (obtained from the MM^ forcefieldcalculations [36]) of the coencapsulation complexes. (R)-styrene oxide with i-PrCl (left); (R)-styrene and (R)-2-Cl-butane (middle); (S)-mandelic acid and (R)-2butanol (right)

Table 1. Guest (±)-2-bromo-3-methylbutyric acid (±)-2-bromovaleric acid (±)-2-bromobutyric acid

Enant Ratio (298 K) 1.5 1.3 1.6

268 Progress in Biological Chirality *

-3.0

•

-4.0

-3.5

6

Figure 10. Upfield region of the 'H-NMR spectra (600 MHz) of encapsulation complexes of U and guests in mesitylene-di2 at 298 K. a) (±)-2-bromo-3-methylbutyric acid (* and • are 4 and 3-methyl groups of 2-bromo-3methylbutyric acid); b) (±)-2-bromovaleric acid (• are 4-methyl groups of 2-bromovaleric acid); c) (±)-2bromobutyric acid (• are 4-methyl groups of 2-bromo-butyric acid)

Carboxylic acids gave clearly interpretable spectra, but the two molecules of 3hydroxybutyric acid 7 were encapsulated without diastereoselectivity. This guest gave modest selectivities in complexes with aromatics at 273K: (S)-mandelic acid showed 18% de, while (R)-mandelonitrile 8 and (R)-l-phenylethanol 9 each showed 21% de. The isomeric 2-OH butyric acid, with the asymmetric center nearer the coguest, poorly resolved spectra. Instead, we found that small a-bromo acids were good guests. Two molecules are encapsulated and with modest selectivity (Table 1 and Figure 10). Little differences in the length or in the shape of the acids cause differences in stereoselection. We assume that interguest hydrogen bonding occurs and that leaves the stereocenters at some distance. Figure 11 shows the results of Molecular Modeling MM+ and emphasizes that the acids' centers are, on average, further apart than the alcohols'. The acid appears important for good interactions between guests and two acids are also effective. This is unexpected as the hydrogen bonded dimers place their asymmetric centers no closer than 6.7 A (Fig 11). The diastereoselectivities are poor: with partners that have weak

^ Br

0-I+-0.

V A

\/wjl 6.7

Figure 11. Asymmetric centers of hydrogen bonded alcohols can be closer to each other than those of the acids

Chiral Spaces in Encapsulation Complexes

269

attractions or with alcohols there is no selectivity; only with carboxylic acids can some (up to 25% de) be observed. The volume of the capsule translates into ~4 M concentration of each guest inside and the lifetime of the complex is on the order of 1 second. The guests are isolated in space and in time and a chiral guest has ample opportunity to provide an effective asymmetric magnetic environment for its partner. Perhaps the two-point connections and the stronger hydrogen bonds of the dimeric acids increase the lifetimes of complexes. As a final observation, 2-bromo-3-methyl butyric acid (commercially available in optically active and racemic forms) gave the best stereoselection with itself, rather than its mirror image. Are these attractions between dipoles? Steric and magnetic environments appear insufficient for diastereoselection; multiple attractive contacts between guests should be more effective and provide the challenge.

Acknowledgements We are grateful to the Skaggs Foundation and the National Institutes of Health (GM 50174) for financial support. References [I] [2] [3] [4] [5] [6]

M.M. Conn and J. Rebek Jr., Chem. Rev. 97 (1997) 1647-1668. T. Kusukawa and M. Fujita, J. Am, Chem. Soa 121 (1999) 1397-1398. T.N. Parac, D.L. Caulder and K.N. Raymond, J. Am. Chem. Soc. 120 (1998) 8003-8004. R.G. Chapman and J.C. Sherman, J. Am. Chem. Soc. 120 (1998) 9818-9826. J.M. Rivera, T. Martin, J. Rebek Jr., Science 279 (1998) 1021-1023. (a) M. Scherer, D.L. Caulder, D.W. Johnson and K.N. Raymond,^wgew. Chem. I l l (1999) 1690-1694. (b) M. Scherer, D.L. Caulder, D. W. Johnson and K.N. Raymond, Angew. Chem., Int. Ed. 38 (1999) 15881592. [7] F. Ibukuro, T. Kusukawa and M. Fujita, J. Am. Chem. Soc. 120 (1998) 8561-8562. [8] S. Hiraoka and M. Fujita, J. Am. Chem. Soc. 121 (1999) 10239-10240. [9] H. Hof, C. Nuckolls and J. Rebek Jr., J. Am. Chem. Soc. Ill (2000) 4251-4252. [10] (a) J. Canceill, L. Lacombe and A. Collet, J. Am. Chem. Soc. 107 (1985) 6993-6996. (b) J. Yoon and D.J. Cram, J. Am. Chem. Soc. 119 (1997) 11796-11806. (c) J.K. Judice and D.J. Cram, J. Am. Chem. Soc. 113 (1991)2790-2791. [II] (a) J.M. Rivera, T. Martin and J. Rebek Jr., J. Am Chem. Soc. 123 (2001) 5213-5220. (b) J. Tokunaga and J. Rebek Jr., J. Am Chem. Soc. 120 (1998) 66-69. (c) F. Hof, S.L. Craig, C. Nuckolls and J. Rebek Jr., Angew. Chem., Int. Ed. 41 (2002) 1488-1508. (d) R.K. CasteUano, B.H. Kim and J. Rebek Jr., /. Am. Chem. Soc. 119(1997)12671-12672. [12] (a) M. Fujita, K. Umemoto, M. Yoshizawa, N. Fujita, T. Kusukawa and K. Biradha, Chem. Comm. (2001) 509-518. (b) S. Hirakoa and M. Fujita, / . Am. Chem. Soc. Ill (1999) 10239-10240. (c) T. Kusukawa and M. Fujita, J. Am. Chem. Soc. 124 (2002) 13576-13582. [13] D.L. Caulder and K.N. Raymond, ^cc. Chem. Res. 32 (1999) 975-892. [14] T. Szabo, G. Hilmersson and J. Rebek Jr., J. Am. Chem. Soc. 120 (1998) 6193-6194. [15] X. Wang and K.N. Houk, Org Lett. 1 (1999) 591-594. [16] J. Santamaria, T. Martin, G. Hilmersson, S.L. Craig and J. Rebek Jr., Proc. Natl. Acad Sci. USA 96 (1999) 8344-8347. [17] J.M. Rivera, S.L. Craig, T. Martin and J. Rebek Ir., Angew. Chem., Int. Ed 39 (2000) 2130-2132. [18] F. Mohamadi, N.G.J. Richards, W.C. Guida, R. Liskamp, M. Lipton, C. Caufield, G. Chang, T. Hendrickson and W.C. Still, J. Comp. Chem. 11 (1990)440-467. [19] Even low concentrations of hydrogen-bonding molecules can have dramatic effects on the rates of assembly processes. R.K. CasteUano, S.L. Craig, C. Nuckolls and J. Rebek Jr., J. Am. Chem. Soc. Ill (2000) 78767822.

270 [20] [21] [22] [23]

Progress in Biological Chiralily

R.K. Castellano, B.H. Kim and J. Rebek Jr.. 1 Am. Chem. Soc. 119 (1997) 12671-12672. R.K. Castellano, C. Nuckolls and J. Rebek, Jr., J. Am. Chem. Soc. Ill (1999) 11156-11163. C. Nuckolls, F. Hof, T. Martin and J. Rebek Jr., J. Am. Chem. Soc. 121 (1999) 10281-10285. For examples of imprinting in covalent materials, see: (a) K.J. Shea, Trends Polym. Sci. 2 (1994) 166-173. (b) G. Wulff, Angew. Chem. 107 (1995) 1958-1979; G. Wulff, Angew. Chem., Int. Ed Engl. 34 (1995) 1812-1832. (c) K. Mosbach and O. Ramstroem, Biotechnology 14 (1996) 163-170. (d) H. Shi, W.-B. Tsai, M.D. Garrisoa S. Ferrari and B.D. Ratner, Nature 398 (1999) 593-597. (e) K. Polbom and K. Severin, Chem. Commun. 24 (1999) 2481-2482. (f) K. Dabulis and A.M. Klibanov, Biotechnol. Bioeng 39 (1992) 176-185. [241 For examples of memory effects in noncovalent materials, see: (a) Y. Furusho, T. Kimura Y. Mizuno and T. Aida J. Am. Chem. Soc. 119 (1997) 5267-5268. (b) E. Yashima, K. Maeda and Y. Okamoto, Nature 399 (1999)449-451. [25] E. Graf, R. Graff, M.W. Hosseini, C. Huguenard and F. Taulelle, Chem. Commun. (1997) 1458-1460. [26] (a) T. Heinz, D. Rudkevich and J. Rebek Jr., Nature, 394 (1998) 764-766. (b) M.K. Ebbing, M.-J. Villa JM. Malpuesta, P. Prados and J. de Mendoza, Proa Natl. Acad Sci. USA 99 (2002) 4962-4966. (c) A. Shivanyiik and J. Rebek Jr., Chem. Commun. (2001) 2424-2425. [27] A. Shivanyuk and J. Rebek Jr., J. Am. Chem. Soc. 124 (2002) 12074-12075. [28] A. Shivan>iik and J. Rebek h., Angew. Chem., Int. Ed 42 (2003) 684-686. [29] J. Chen and J. Rebek Jr., Org. Lett. 4 (2002) 327-329. [30] S. Mecozzi and J. Rebek Jr., Chem.-A Eur. J. 4 (1998) 1016-1022. [31] T. Heinz, DM. Rudkevich and J. Rebek h.. Angew. Chem., Int. Ed 38 (1999) 1136-1139. [32] A. Shivanyuk, A. Scarso and J. Rebek Jr., Chem. Comm. (2003) 1230-1231. [33] S.K. Komer, F.C. Tucci, DM. Rudkevich, T. Heinz and J. Rebek Jr., Chem. Eur. J. 6 (2000) 187-195. [34] Under these conditions, 2-Cl-butane, 2-butanol and 2-pentanol all form encapsulation complexes with 2 guests inside. The complexes are diastereomeric: one meso-form including two guests of opposite handedness and those with two molecules of R or S 2-butanol. The diastereomeric capsules are formed in equal amounts. [35] J. Rebek Jr., B. Askew, M. Doa and P. Ballester, J. Am. Chem. Soc. 109 (1987) 4119-4120. [36] Hyperchem TM. Release 7. Hypercube Inc. 2002.

Progress in Biological Chirality G. Palyi, C. Zucchi and L. Caglioti (Editors) © 2004 Published by Elsevier Ltd.

Chapter 23 Serum Albumin and Natural Products Lorenzo Di Bari, Silvia Ripoli and Piero Salvadori* AmbiSEN- Centro di Alta Tecnologiaper lo Studio degli Effetti di Agenti Nocivi, Dipartimento di Chimica e Chimica Industriale - Universitd di Pisa, Via Risorgimento 35, 1-56126 PISA, Italy psalva@dcci. unipi. it

1.

Introduction The fundamental role of serum albumin (from now on SA, HSA for the human protein, and BSA for the bovine [1, 2]) as a vector for both endogenous and exogenous compounds has been recognized for a long time. By non-covalent binding to poorly soluble organic molecules or sometimes to metal complexes, it may cause meaningful increase in the blood concentration of these systems, permitting absorption after ingestion or inhalation. Conversely, it buffers the concentration of soluble drugs, by reducing the amount of free molecules. This issue is particularly relevant for pharmaceutical research: lead molecules that show high activity in vitro may be less effective in vivo when binding to serum albumin is relevant. Knowing and predicting interactions between drugs upon co-administration is a key step for setting up therapeutic protocols; its understanding on a molecular basis may save time and experimentation on living organisms. For example, it was demonstrated that one of the reasons for the adverse effects resulting from co-administration of phenylbutazone and warfarin, derives from their competition for SA binding. The plasma concentration of free warfarin is normally kept very low because it is largely (up to 99%) engaged in albumin binding on site I (vide infra), thus, it displays only a fractional anticoagulant activity. When also phenylbutazone is present in the blood, it competes for SA site I, displacing warfarin, that becomes more available for its receptors, ultimately enhancing the risk of haemorrhage [3]. All of this matter has been enucleated and reviewed [4], but we shall address a specific issue, which does not appear to be completely focussed in the literature: what is the impact of substances commonly assumed in the diet on the binding sites of SA and consequently what may be the relevance of quality and quantity of food introduced during our meals towards the absorption of drugs. While progress has been made [4, 5] and can be envisaged in the field of pharmaceuticals (thus on drug-drug interactions), the role of molecules present in food needs to be stressed, which is the stimulus for this contribution. The characterization of the thermodynamics of ligand association to serum albumin has been faced with many experimental techniques, ranging from classical biochemical methods like ultracentrifugation to chromatography and to spectroscopy. Summarizing this enormous

272

Progress in Biological Chirality

wealth of work in a brief contribution is impossible and drastic choices are necessary: we shall limit our discussion to a few techniques, which respond selectively only to the mole fraction of SA-bound molecules, that is, which are intrinsically blind to a solution of the ligand alone. Since one of the most successful spectroscopies applied to investigate SA interactions is circular dichroism (CD), the issue of molecular chirality and even more specifically of induced chirality after the binding is of primary interest. Only a couple of other methods very recently introduced will be discussed at the end of the chapter in order to open up the perspectives for widening the knowledge in the field, through innovation. The text will be divided into two parts: sections, 2, 3 and 4 contain introductory information: the general framework of ligand-protein binding (and the chirality related issues), and some of the best characterized cases where albumin is involved in the transport of molecules contained in the food. In the second part, sections 5 to 7, the three experimental techniques, we chose to focus on, will be discussed by means of selected examples, which demonstrate some meaningful features of the methods themselves.

2.

Serum Albumin: Multiple Binding Sites Serum albumin (SA) is a globular heart-shaped protein made of three main domains, subject to some conformational freedom. It contains various grooves and pockets where smaller molecules can be hosted, ensuring its role of carrier protein [4]. These binding sites can exhibit more or less pronounced affinity to specific ligands: a given compound may fit particularly well in one pocket, leading to very stable non-covalent binding, but at the same time display secondary binding elsewhere. Moreover, different compounds may be allocated simultaneously in different sites: as a general reminder. Figure 1 depicts the matter.

cys34 metal ions

domain III

subdomain IIIA (Sudlow site II) ibuprofen digitoxin diazepam

domain I

subdomain IIA (Sudlow site I) warfarin phenylbutazone

domain 11

Figure 1. Structure of human serum albumin HSA with the two most important sites for drug binding [2]

Serum Albumin and Natural Products

273

Multiple binding deserves a brief discussion of the network of complex solution equilibria, which can take place when the ligands X and Y are in the presence of S A. If X or Y are mixed with SA separately, one can write Kx SA + X ^ ^ SA'X

..

Ky SA^Y ::;^=±^SA*Y While, when both are simultaneously present, on top of the one-to-one equilibria, one may expect also multiple binding, leading to the formation of the trimolecular adduct: SA+X^Y

:z^=^ SA^X^Y

^2)

Provided KXY >0, we can distinguish the three cases KxY = KXKY KXY > KXKY KXY < KXKY

Independent binding Cooperative binding Anticooperative binding

(3)

The latter two cases are examples of allosteric binding, which must be associated to a variation of the properties of one site upon the occupancy of the other one. This is usually made possible by a conformational rearrangement, following one of the interactions, whereby the macromolecule becomes better (cooperative binding) or poorer (anticooperative) fit to the other ligand. The extreme case KXY = 0, provided both KX,KY > 0, means that the ternary adduct SA^X^Y cannot exist, which indicates that X and Y insist on the same site, thus leading to competitive binding and to a displacement of one for the other ligand. Interestingly, multiple binding can be observed also for one ligand alone. A slight modification of the above equations holds ^; SA+X :;;=±r SA^X^ (4) ^2

SA+X—^SA^X^

(5)

Kl2 SA+X~^SA*XJ'X2 (6) SA*X^ and SA^X^ represent the adducts where ligand occupies site 7 or site 2 on SA, respectively. Once more the two binding processes can be independent or (anti)cooperative. In the case of multiple binding sites for one ligand, it is worth recalling that the two processes described by equations (4) and (5) are equilibria with common species (SA and X): therefore the ratio [SA^X'] K h -4 = —- = constant (7) [SA-X'] K, is independent of the concentration of X or of the molar ratio. This implies that any instrumental response where the signals of SA^X^ and of SA*X^ are superimposed will not change its appearance during the course of a titration and that even sophisticated methods like factor analysis will fail detecting the co-presence of different species: deviations from this

274

Progress in Biological Chirality

situation will be manifest only at the onset of the ternary adduct SA*X^*X\ whose concentration has a completely different dependence on fXJ. The approximation of independent binding should be used with great caution for SA, because its various sites may have very different affinities for a given ligand. Nevertheless, it is often employed through more or less questionable approaches. The most convincing method [6] is to plot the ratio of the moles of total protein (SAtot) to moles of bound ligand [By versus the reciprocal concentration of free ligand, l/[F\. If a linear relationship is found, statistical factors {viz. perfect independence and equivalence of the sites) are predominant in distributing the ligand among the n sites:

^

=^ ^ . i

(8)

[B] n [F] n All of these equations have been used for interpreting the spectroscopic titration data, in order to obtain the apparent affinity constant, or the number of sites, as well as for evaluating the kind of interaction between different compounds (independent/competitive/allosteric). 2.1 Binding kinetics A rough picture of the ligand/SA interaction process from a dynamic point of view is necessary to better understand some of the techniques used to investigate the free^ound equilibrium. Indeed, most spectroscopic techniques (notably optical absorption and emission spectroscopy) are surely/ay/ in the sense that the spectrum of a mixture containing the ligand and the protein are the weighted sum of contributions, arising from the free molecule and from the complex(es). By means of suitable methods, like stopped flow or T-jump, one can monitor the evolution of the spectra toward thermodynamic equilibrium and thus obtain the kinetic association/dissociation rates. On the contrary, NMR may be on the edge: depending on several factors discussed below, the dynamic process may appear fast or slow. The non-covalent ligand binding of SA is made ftilly reversible by a comparatively fast association/dissociation kinetics. Although dynamic data are much less abundant than those concerning equilibrium constants, in many cases dissociation rates in the range 1-20 s" (at room or physiological temperature) have been reported. This implies that the ligand can be released promptly from SA in proximity of a receptor with higher affinity, which ensures a role of neutral carrier for HSA. Among the methods for studying the dynamic association/dissociation, there are stopped flow [7], chromatography, NMR [8] and EPR [9]. Much more recently, the equilibrium process has been investigated through various types of sensors: they constitute a breakthrough in the field of drug/protein binding which is likely to bear major consequences in thefixture.A brief review of these applications will be found at the end of this chapter. The dissociation and association rate constants, A:"'"^ and F"", are linked together with the equilibrium constant through the relation: lass

K =^

(9)

' In the case of multiple binding the concentration of bound Ugand equals the sum of the occupancy of all sites [B] = ^[SA^X']-^^ [SA •X' •X^]-\-.. Here the first term refers to single occupation, the second term to the ternary adduct and - at least in principle - this should go on with all possible higher orders. For the concentration of free ligand, we have [F] = [X].

Serum Albumin and Natural Products

275

Assuming that the association process is controlled by diffusion, we can assess j^ass ^ io^M"^s\ which allows us to have a first estimate of A:^' ^' ^ from the knowledge of the equilibrium constant only [10]. However, it should be borne in mind that this argument might be valid only if binding does not involve a more complex process, like, for example a structural rearrangement of either the ligand or SA, which would sizably reduce F^^ NMR is surely one of the most popular spectroscopic techniques in these days, in spite of the fact that high field instrumentation may be very expensive. A given dynamic process is defined slow or fast on the NMR timescale if its rate constant is respectively small or large compared to half the difference in angular frequency (measured in Hz), nAv^ between the resonances corresponding to the same nucleus in the two exchanging forms. In the case of fast exchange, one observes an NMR spectrum with only one set of resonances, where each line is located at the weighted average between those of the two forms; on the contrary, for slow exchange, one has two distinct sets of lines, with integral rafios equal to the mole ratio of free and bound. A typical complexation shifl; (difference between the chemical shift of the bound and free forms) for a proton can be of the order of O.lppm, which at 500MHz means TtAv =J60Hz, implying that the free/bound exchange rate constants of about 100 - 200 s"^ are likely to fall in the so-called intermediate exchange, where neither of the two situations depicted above is quite correct and one observes extensive line broadening. Naturally, the boundary for intermediate exchange depends on the complexation shifts. In fact, the same process may in principle appear slow for a given nucleus and fast for another. Moreover, it should not be forgotten that the frequency difference quoted above is a function of the external magnetic field and that on increasing its strength, fast processes may appear intermediate and finally slow.

3.

Albumin and Food Although the role of albumin in the circulation of drugs in the human body was largely investigated [4], less attention has been addressed to the constituents of food. Here we shall briefly summarize what is known about the interaction between SA and organic compounds often taken up during a meal; in particular flavors, toxins and antioxidants. The ability of SA to bind reversibly small molecules and its abundance in the blood gives this protein the crucial role of bio-carrier of nutrients. This is particularly relevant for two classes of molecules: first, apolar ligands, that may be practically insoluble in water, and whose enhanced plasma concentration arises from protein binding (we remind that the SA pockets have a hydrophobic character); second, degradation-prone molecules, which may be protected - notably from oxidation - by the polypeptidic environment of SA. Another important active role of SA is to mask some properties of the guest, by buffering its concentration [11]. On a more academic ground, it must be considered that, since the binding properties and the structure of albumin are well described in the literature, this protein constitutes a very good model to reproduce food proteic matrix. For this reason, albumin was chosen in investigations of flavors-protein interactions [12, 13]. The rate of release of volatile components is one of the aspects (beyond their composition) determining the perception of a fragrance and a taste, and consequently also the acceptability of food. Proteins are generally tasteless but they do influence the flavor of food by modulating the concentration of free (volatile) odorants through binding [13]. A work of Burova et ah [14] very recently brought our attention onto two interesting aspects of the flavor-protein interactions, which will deserve

276

Progress in Biological Chirality CHO

CHO

4,5,6

vanillin vanilla essence

„ ketones ^'3' '!^'?'^ components of cheese or oil

benzaldehyde ^^^^ ^,^ J„^

Figure 2. Chemical structures of some flavours, whose HS A interactions were studied

future investigations: a) prediction of the protein conformational changes in solution can provide information about the accessibility to its hydrophobic pockets, influencing the reversibility of odorant binding; b) little is known about the competition between proteins in mixture for a particular odorant and vice versa about that of odorants for a single protein. These authors demonstrated the effects of the acidic and thermal denaturation and aggregation of proteins (processes typically occurring during the yogurt production), on the binding of volatile compounds. As specific examples, the studies about alkylketons can be considered. Their interaction with SA was studied by means of different techniques from optical spectroscopies to differential scanning calorimetry (DSC) or diffusion NMR (see also below) [12-14]. From these studies, it can be inferred that the hydrophobic nature of the ligand-S A binding is the driving factor in the case of alkylketons, so that 2-nonanone has major affinity for SA with respect to shorter chain or internal ketones [12]. A few compounds whose interactions with SA have been studied are displayed in Figure 2. Although it may appear surprising, toxins are very common in our food. They may have very different origins: they may be products of more or less desirable fermentations or of cooking processes, or be present as residues of food preservatives or of pesticides, just to make a few examples. A particular interest was devoted to the ubiquitous contaminants ochratoxins (Figure 3) [15]. These compounds, derivatives of isocoumarin and Lphenylalanine, are released by Aspergillus and Pennicillum fungi found in badly stored food or feed. Ochratoxins are responsible for toxic effects like carcinogenity, mutagencity or nephrotoxicity; a wide literature on this topic is cited in reference 15. The high affinity of ochratoxins for plasma proteins (Kaff >10''^ M"^) causes two very relevant effects [15b,c]. First, is the prolongation of their half-life time in plasma, which is an example of the protection operated by SA binding toward metabolic evolution (and degradation) of ligated species. Second, their passive absorption from the digestive system is active also when the plasma concentration of the toxins is higher than in the intestine, as a consequence of the fact that the largest fraction of these compounds in plasma is captured by SA and subtracted by this way to the free osmotic equilibrium through the intestinal epithelium. These findings very recently stimulated Simon et al. to study the interaction between HSA and ochratoxin A (OTA), the most toxic of this class, and its derivatives by means of optical (absorption and fluorescence) spectroscopy [15b-d]. The main results of these studies can be summarized: • OTA binds to HSA as a dianion; • only two sites of HSA are available for OTA-dianion and each site can host only one

Serum Albumin and Natural Products [ f ^

COOH O

OH

277

O

Ochratoxin A (OTA)

f r %

COOH O

OH

O

Ochratoxin B (OTB) f f ^

COOEt O

OH

O

Ochratoxin C (OTC)

Figure 3. Structures of ochratoxins

molecule. The affinity constants for the two sites differ by one order of magnitude (K^=5' 10^M-*;K^=1-10^M-^); • through Trp214 fluorescence quenching and through warfarin competition, subdomain IIA was identified as the primary binding site, sub-domain IIIA as the secondary one. Another example is offered by uremic toxins (for example indole-3-acetic acid, indoxyl sulfate or hippuric acid), which accumulate in plasma of patients with chronic renal diseases [16]. This accumulation was related to SA binding (some uremic toxins bind to site II but also site I can be occupied, that is the case of 3-carboxy-4-methyl-5-propyl-2-fljran propanoic acid) and explains why minor plasma concentration of drugs as warfarin and furosemide is found in these patients. The study of Sakai et al [16b] shows that, in addition to the competitive displacement, other mechanisms (allosteric interaction or site-site interaction) can take place. Vegetable food and fruit may be very rich in antioxidants like e.g. flavonoids (and more in general polyphenols, see Figure 4) or carotenoids, and through this way they enter our daily diet. The attention of scientific research was directed to antioxidants largely because of their potential impact on health [17]. The protective action of vegetable food against cancer and coronary diseases has been associated with the intake of these compounds, notably acting as radical scavengers [18]. In order to better understand the activity of these compounds, some main aspects must be taken into account: their bioavailability, their metabolism, and the masking of their antioxidant activity. Serum albumin plays a crucial role in the first and third points. It is known that, in spite of the presence of polar groups, flavonoids are very scarcely soluble in water and their transport in plasma is conveniently mediated by albumin [19]; the same happens for carotenoids, essentially water insoluble [20, 21]. This would Hmit their availability in aqueous medium and thus in plasma, without a good carrier as it is albumin.

278

Progress in Biological Chirality

flavonol-glycoside anthocyanidtn

Figure 4. Structures of main flavonoids

Moreover, it was shown that albumin or other proteins can decrease the antioxidant activity of flavonoids [19]. These findings throw the basis for more detailed investigations on the flavonoids-SA interaction in order to identify which factors promote flavonoids absorption without reducing their activity. Site I (see Figure 1) is responsible for the binding of flavonoid, as well as of natural coumarins, which indicates possible interference between these compounds, and warfarin, phenylbutazone and other drugs which have affinity that site. The equilibria in a ternary solution HSA/warfarin/quercetin (a flavonol) have been investigated by means of fluorescence spectroscopy and demonstrate that the interaction is not a pure competition, and indicate that in the same site warfarin and quercetin may be both simultaneously accommodated [17c]. In the case of carotenoids it was found that the presence of carboxylic groups promotes the binding to HSA. Carotenoids form structured aggregates where HSA can act as asymmetric template on which a chiral self-assembly of carotenoids grows with increasing the carotenoid/HSA ratio [22]. This result is promising for achieving the transport of higher amounts of carotenoid derivates in aqueous phases. More on carotenoids will be discussed in section 5.3 below. Another important class of diet-related compounds that bind S A are fatty acids. There are many sites on this protein where this interaction can take place, and the number of long alkyl chain molecules that can be allocated on SA is variable, depending primarily on the length of the chain. X-ray crystallography determined the location of these sites and the conformation of the fatty acid molecule, which appears often bent or folded [23]. This binding is of utmost relevance, because of possible competition with drugs and of allosteric interaction with the other sites. Wainer and Noctor demonstrated these effects by means of affmity HPLC,

Serum Albumin and Natural Products

279

showing that site II is the primary binding site for octanoic acid, thus interference with profens (by direct competition) has to be expected, while allosteric effects with site I (e.g. hosting warfarin) can also occur [24]. Moreover, the effect of fatty acid binding induces a conformational change in SA, which exposes Cys-34 SH-groups to oxidation. As shown in Figure 1, this is one of the sites where metal cations, and in particular Cu(II), are hosted (often called site V). It has been demonstrated that the oxidized SH-Cys34/Cu(II) is an effective catalytical oxidant towards, e.g., ascorbate. The role of radical scavenger recognized for SA is thus reverted into a prooxidant [25].

4.

Albumin and Chiraiity In a book on biochirality, the issue of the role of molecular dissymmetry must occupy a central role. In fact, as more or less explicit, many of the molecules transported by albumin are chiral and the other ones assume a temporary chiraiity when they are bound to the polypeptide. Surely, a majority of drugs, nutrients, flavors are dissymmetric and usually the two enantiomers exhibit different biological activity, because of selective interaction with the receptors as well as with the transporters. We may recall a couple of facts; 1. Usually albumin displays high selectivity towards enantiomeric ligands, the affinity being very different for the two enantiomers. Indeed, the two mirror image structures of a drug may find place on different sites, as is the case of (R) and (S)-warfarin [26]. Owing to the broad spectrum of organic and inorganic systems that can be accommodated in the various grooves, pockets or simply on the surface of this protein, the scope of chiral discrimination is unusually wide. This fact, coupled with the long term stability of S A justifies the use of this protein in liquid chromatography as a modifier to obtain chiral stationary phases. Indeed, HPLC [4] and related separation techniques are most common for studying SA binding, and this whole subject will not be treated in the present contribution. It is worth observing that two enantiomers are nothing else than two different molecules and that the proper context to describe their SA binding is that discussed in section 2, with the equations (1), (2) and (3). 2. Albumin can induce chiraiity on intrinsically symmetrical ligands, by providing a symmetry-breaking environment. In both cases, induced chiraiity follows, which is a subject that will be discussed below. Although in normal physiology albumin plays the role of a molecular chaperon, deputated not to alter the transported compounds, in some cases it may exhibit interesting properties in reactivity. It has been reported that BSA binds differently the two enantiomers of binaphthol and ketoprofen, altering their photochemistry. As the absorption spectra of the diastereomeric adducts BSA/enantiomer are different, it is possible to irradiate selectively only one species, causing its photodegradation. The other isomer remains virtually intact and by dissociating the HSA/drug complex, the unreacted enantiomer can be collected [27]. This finding does not seem to have met further application to date.

5.

Induced Circular Dichroism Circular dichroism (CD, measured through Ae) is a property of chiral nonracemic molecules, because it is equal to the difference in molar extinction toward left and right

280 Progress in Biological Chiralit> circularly polarized light, / and €^, respectively: only a species with well-defined handedness can discriminate between the two chiral components of the electromagnetic radiation. Such a difference can be found non-vanishing only when s^,6^>0, that is, in proximity of an absorption band. Moreover, it can assume positive or negative sign depending on the relative magnitudes o f / and 8^ [28]. These features will account for the large popularity of electronic CD (i.e. relative to the spectrum of electronic transitions) in the field of ligand/SA interactions [29]. SA displays itself a strong CD spectrum below 300 nm, being essentially transparent at longer wavelengths. On the contrary, many conjugated organic chromophores are endowed with absorption bands well above that limit, thus do not suffer from interference. Achiral ligands do not show any CD when they are in the free state; on the contrary, once they are bound to SA, they are surrounded by a chiral environment. In the case of achiral ligands CD responds selectively only to the mole fraction of bound drug. For chiral non-racemic molecules, we can notice that CD is generally more sensitive than isotropic absorption to variations of the molecular environment. Thus, by taking the difference between the spectra of the ligand in the presence and in the absence of S A one can extract the contributions of the bound form. Finally, the sensitivity of CD to conformational variations of proteins [30] opens the way to the simultaneous investigation of the binding process and of the structural rearrangements occurring on SA. We shall concentrate on the case of an achiral chromophoric ligand, discussing the various processes that may be responsible for optical activity of the SA-bound form. The three mechanisms detailed below are not mutually exclusive and in certain cases may concur in determining the observed spectrum. 5.7 Coupling between albumin and the host Whenever an intrinsically symmetrical chromophore is embedded in a chiral environment, its electronic transitions may exhibit a CD (induced CD or ICD). Various mechanisms can be invoked to account for this phenomenon, but the one which often appears dominating is the coupling between electric or magnetic dipole transitions of the chromophore and those (mainly electric) localized on its macromolecular surroundings. In this case, the main contribution must be expected stemming from the aromatic groups on the amino acids side chains of SA. If this mechanism holds on an isolated transition of the ligand, a monosignated Cotton effect must be envisaged." Ligands characterized by a complex absorption spectrum, featuring several transitions, are expected to give rise to separate - and possibly of different sign Cotton effects, where the maxima in absorption and CD practically coincide, as shown in Figure 5. This may be no longer true when an absorption band is the envelope of several contributions. A special case is offered by magnetic-dipole-allowed electric-dipole-forbidden transitions: they are usually endowed with very small extinction, to the point that they may not show up in the isotropic UV-VIS spectrum, but, once the ligand is surrounded by the chiral environment of SA, they may give rise to intense Cotton effects. Predicting the sign of " A fundamental conservation law of chiroptical spectroscopy predicts that the total integral through all the spectrum vanishes. In the case of the coupling between albumin aromatic side chains and chromophoric ligands considered here, two contribution of opposite signs canceling each other are expected: one is centered around the wavelengths where the Ugand absorbs, while the counterpart will be located together with the CD of the protein and it is often not observed, possibly being offset by the intrinsic contribution of SA.

Serum Albumin and Natural Products

281

CD [mdeg] absorption

wavelength [nm]

Figure 5. ICD (solid line) and absorption (dashed line) spectra of the 1:1 complex between a synthetic coumarin and HSA. A single Cotton efifect is apparent alUed to the 400 nm transition of the conjugated chromophore [31]

the SA-ligand coupling is in general hard: it may depend on the structure of the macromolecular pocket and the detailed orientations of the aromatic sidechains around it. Usually, a heuristic approach is taken, based on the observation of the onset of the spectrum: from the mere existence of an induced CD, one has the qualitative evidence of a ligand/protein interaction; thereafter, one can quantitatively analyze the magnitude of the Cotton effects as a fiinction of the Hgand concentration (or of the ligand/albumin mole ratio) [32]. The analysis of ICD data is a very powerful tool for analyzing the ligand/SA interactions. Especially by means of the probes indicated in Figure 1, it becomes possible to identify the site where the ligand X is allocated and its possible allosteric interactions. To this end, the ICD band due to some electronic transition located on X in a mixture X/SA is followed while a specific probe, Y, is added stepwise. If the two ligands have affmity for the same site, a competitive equilibrium will take place and one must ultimately expect the complete disappearance of ICD of the ^S^^-Xadduct, with the simultaneous growth of that of SA^Y. This case represents the displacement of X operated by Y. The position of this equilibrium in the various steps of a titration depends on the relative affinities of S A for the two ligands, i.e. on the ratio Kx/Ky. If X and Y bind on different sites, one of the three cases represented in Equations (3) will be found. To have a full picture, cross titrations can be performed: the 1:1 mixture SA/X is titrated with Y, and/or the mixture SA/Y is titrated with X. The resuhs are compared with the direct titrations with X and Y separately. The CD data must be followed at wavelengths characteristic for SA»X or SA*Y or appropriate difference spectra must be taken. The plots of ICD versus the concentration of X or Y in the cross and direct titration are compared: if they match, and independent binding must be envisaged, if deviations are observed, allosteric interaction (cooperative or anticooperative) must be invoked. Quantitative analysis of the magnitude of ICD as a function of the concentration of the added ligand may yield the association constant, in analogy with spectrophotometry. Owing to the fact that, at wavelengths where SA is transparent, ICD is selectively sensitive to the bound ligand, one can write ICD =k[SA*X] (10) with k a. proportionality constant dependent on the pathlengts and on the molar dichroism,

282

Progress in Biological Chirality

here treated as a phenomenological parameter. By observing that the total concentration of SA, CA, and of ligand, cx, must be conserved, we obtain fSA*XJ=-ICD/k fSAJ = CA'ICD/k [X] = cx-ICD/k

(11)

we can transform equation (1) into K-

(^^^ (c^-ICDIk)(c,.-ICDIk)

(12)

which yields ICD = | [ c , + c , + / : - ' - 4{c, + c, +K'y-

4c,c^ j

(13)

A titration curve of ICD versus cx (at constant CA) can thus be fed into a least square fitting software, which can derive the affinity constant K. It must be observed that when association is very strong, K^ may become vanishingly small compared to (CA+CX): in such a case the equation above reduces to ICD = kcx which describes a linear growth of the ICD amplitude and does not allow one to determine K. One can also rearrange equation (12) and neglecting the term (ICD/kf obtain ^A^X _ 1 ^ ( ^ t f x l ICD kK k

(14)

Thus, plotting the first term versus (CA^CX), a straight line with slope 1/k and intercept 1 (kK) is expected. This linearization is possible only if (ICD/k) = [SA*X] is small, i.e. when the association is not very strong. Therefore the two methods, namely the fitting through the ftill equation (13) and through its linearization (14) have the same limitation with respect to the relative size of AT, CA and cx. 5.2 Exciton coupling I: conformational chirality The most popular approach to understanding CD is exciton coupling, which is a typical feature of two (or more) interacting chromophores, as shown in Figure 6. In the extreme, paradigmatic case of two identical chromophores, this leads to a split dissignated CD band, with positive and negative components. It is called an exciton couplet and defined positive or negative according to the sign of the low energy (long wavelength) component. The origin of this phenomenon is a through-space dipolar interaction between electric dipole allowed transitions, which may thus be located on the same as well as on two different molecules. In this first part, we shall discuss the intramolecular case. In the context of SA binding, the most notable example is offered by bilirubin, one of the products of the degradation of the heme porphyrin.

Serum Albumin and Natural Products ^ 3 9 5 (+38)

-

283

CD

/ \ negative couplet

\ I \

i

\ 1

\l

L-

A

43t (-72)

UVA^is

Wavelength (nm)

Figure 6. A typical bisignate CD spectrum due to exciton coupling in a bischromophoric system, obtained by derivatizing ponasterone A with a coumarin derivative [33]. We can appreciate that the negative chirality defined by the chromophores is reflected in a negative couplet

Bilirubin

Bilirubin is an achiral molecule made of two symmetrical moieties with extended conjugation, held together by a conformationally flexible bridge. Owing to intramolecular hydrogen bond, it assumes a bent chiral conformation, referred to as ridge-tile, whose two enantiomers are in fast exchange. In the presence of HSA, a 1:1 adduct is formed with one enantiomeric conformation of bilirubin: this displaces the equilibrium in an asymmetric transformation and a neat ICD is observed, with the shape of a positive exciton couplet. This sign is unambiguously related to the sense of twist of the transition dipoles located on the two chromophores, as depicted in Figure 7, and this indicates that the P conformer of bilirubin has higher affmity for HSA [34]. A closely similar situation is also found for curcumin, the major yellow dye of turmeric [35]. The powdered rhizome of this plant is a common ingredient in Asian food and used in traditional oriental medicine.

Figure 7. Bilirubin: enantiomeric conformations in equilibrium (reproduced with permission from Ref [34b])

284

Progress in Biological Chirality

Curcumin features a P-diketone connecting two symmetrical and conjugated moieties. Since it is most stable in its tautomeric form, electron derealization covers the whole molecule, which determines a long wavelength absorption, centered at 460 nm at neutral pH.

Curcumin

It forms a 1:1 adduct with HSA, which at high pH (pH=9) displays an exciton-splitted CD, shown in Figure 8, whose crossover point is about the absorption maximum of neutral curcumin. This spectral feature clearly calls for a coupled dipoles mechanism and a superficial interpretation may indicate that two molecules of curcumin are hosted on the same HSA, leading to a supramolecular chiral aggregate, as will be discussed in the next section. Such a view clashes against the simple evidence of a 1:1 complex, whereby no intermolecular

31

30

29

28

350

27

26

Wavenumber/10''(cm") 25 24 23 22 21

400

20

19

18 17

600

Wavelength (nm)

Figure 8. pH-dependent CD and absorption spectra of curcumin bound to HAS (reproduced with permission from Reference [35b])

Serum Albumin and Natural Products

285

coupling can be invoked. The ultimate explanation of the CD spectra is that: [35] • upon binding to HSA in alkaline solution, curcumin must adopt a folded conformation where the P-diketone is preserved in order to ensure the CH2 hinge between the two molecular halves; • as in the case of bilirubin, this conformation is chiral and one enantiomer is preferred; • accordingly, in the HSA adduct, delocalization is restricted with respect to the free, linear and fully conjugated curcumin and a blue-shifted absorption maximum (and CD crossover) should be expected; • owing to the high pH a deprotonation of the phenol hydroxyls occurs, therefore the two interacting molecular halves must be represented by the corresponding anions; • the absorption spectrum of the anionic moiety is accidentally similar to that of whole, neutral curcumin. This points justify the position of crossover, as the result of the concurrent blueshift (operated by conjugation restriction) and redshift (caused by deprotonation). As depicted in Figure 6, a left-handed arrangement of the transition dipoles must be envisaged, in agreement with the positive couplet at 460 nm. The optimal conformation of curcumin bound to HSA at high pH, proposed on the basis of the CD spectra and from a docking investigation on the HSA site I is represented in Figure 9. Interestingly, at neutral pH, a completely different situation is found, where the ligand exhibits a very weak ICD. The interpretation of this spectrum is still unclear: Reddy et al [36] reported that two molecules of curcumin may find place on HSA on two sites with very different affinity: the one characterized by a K=2.0'10^M'^ that is detected only through fluorescence, the secondary site is revealed through ICD (which appears insensitive to the occupancy of the other one). This situation looks particularly intriguing and may deserve fiirther investigation, although it should be observed that in the determination of the affinity constant by ICD, a modified Hill equation was used, thereby deriving log K from a logarithmic plot. This requires great confidence in the data, because a small error in log units implies a very large variation in the assessed K. Moreover, Hill equation is a questionable choice for albumin, owing to observed the strong correlation between the occupancy of the various sites (allostery) and the large differences in affinity constants. 5.3 Exciton coupling 11: supramolecular chirality As we have already seen, albumin is a multisite carrier, able to bind several guests

. ^ i ^ ^ ^

J^ «*^

Figure 9. Most likely structure of curcumin bound to HSA site I (reproduced with permissionfromReference [35a] and [35b])

286

Progress in Biological Chirality

simultaneously. In such an occurrence two identical or diflferent ligands can be accommodated in a close neighborhood, with a well-defmed geometrical arrangement. It is well known, for example, that SA is a fatty acids carrier and indeed, X-ray diffraction of albumin saturated with a large excess of different fatty acids revealed that there are at least seven sites distributed onto all the three domains of SA [23]. These locations when occupied, induce a supramolecular chirality of the ligands. The saturated alkyl chains of fatty acids have vanishing absorption above 300 nm and cannot give rise to Cotton effects. Polyene chains, as found in carotenoids, may be considered as having more or less the same structural and binding characteristics of fatty acids, were it not for the pronounced flexibility of the latter ones as opposed to the stiffness of the former ones. From the UV-VIS spectroscopy point of view, carotenoids are excellent chromophores, with very intense and red-shifted absorption bands, due to electric-dipole allowed K- 7t* transitions. If two (or more) carotenoid molecules are hosted in nearby sites on SA, intermolecular exciton coupling can occur, giving rise to the bisignated couplet feature. This is what found for crocetin, one of the components of a common spice, saffron, obtained from Crocus sativiis.

Crocetin is a dicarboxylic acid, with a 16-carbon atoms conjugated chain connecting the two terminal groups; a good starting point for predicting its binding sites on SA is to look for the seven locations where XRD determines palmitic acid (also a hexadecanoic system) [37]. Observing that in the seventh site (in the classification of reference [23]), palmitic acid is in a strongly folded conformation, which is unrealizable for the stiff carotenoid crocetin, its role can be excluded. In the other six, Simonyi and coworkers docked crocetin molecules and, assuming that what dominates the ICD spectrum (Figure 10) is a pairwise interaction (i.e. neglecting higher orders), they evaluated the chirality of the geometric arrangements between all possible pairs. The result of such investigation is that two pairs of sites, namely (3,4) and (4,5) can be held responsible for most of the exciton-coupled ICD seen with crocetin/SA. Interestingly, on displacing most of the carotenoid with palmitic acid, the bisignated doublet vanishes, leaving place to a monosignated Cotton effect, centered on the absorption maximum of crocetin, i.e. where the doublet had its crossover: this is what one expects from one chromophoric molecule bound to SA, as described above.

6.

NMR Methods: Slow Tumbling and Diffusion One of the most dramatic changes brought about on a ligand by interacting with SA is about the motional dynamics. Both translational and tumbling motions of free small organic molecules can be slowed down by 1 to 3 orders of magnitude in SA complex. The consequences become apparent only in spectroscopies sensitive to the translational or orientational diffusion, like are some contemporary high resolution NMR techniques: a long rotational correlation time is associated with high relaxation rates (manifested through broad lines and short Ti); moreover, it is nowadays very common to equip NMR probes of high field instruments with gradient coils, which allows the easy access to diffusion coefficients.

Serum Albumin and Natural Products

287

Figure 10. Absorption and CD spectra of mixures of HSA and crocetin as a function of the molar ratio of the two species (reproduced with permission from Reference [37])

When the ligand is in fast (on the NMR timescale) exchange between a free (F) and a bound (B) state one can write for a given property ^

r^'-^(l-x)'f^x'i^.

(15)

where x is the mole fraction of the bound form. Taking advantage of this equation (when it can be correctly applied), and setting up appropriate experimental protocols for extracting the diffusion coefficients or relaxation rates (which are dynamic properties), one can assess the thermodynamic affinity constant. Translational diffusion coefficients, D, can be determined through pulsed gradient spin echo (PGSE) experiments, also referred to as DOSY (Diffusion Ordered SpesctroscopY) [38]. In this case, the limiting quantity D^ can be approximated by the one of the protein alone, i.e. assuming that the bound ligand assumes the same translational diffusion of SA. The mole fraction x of equation (15) is dictated from the equilibrium constant, e.g. through an equation closely related to (13). A global fitting of titration data as a function of total albumin and ligand concentration affords the affinity constant. Such an approach has been used for salycilate,"' which is known to have several binding sites on SA. In a first approximation, these /i-sites were considered fully equivalent and independent: this amounts to saying that the effective concentration of the host is w-times the stoichiometric albumin concentration CA, thus in equation (13) one should replace n*CA for CA and determine n, as well (or equivalently consider CA as a fitting parameter). The results reported for this analysis are not completely satisfactory: a very large number («>30) of sites with very low affinity is "* In the case of salycilate, D^ is found concentration-dependent, and must be described through an empirical polynomial of the form

^FC^X)-^

^p^x^

where Gp are coefficients to be determined by a fitting procedure and cx is the total ligand concentration.

288

Progress in Biological Chirality

found, while the experimental trend of the data is poorly reproduced by the fitting. Apparently, there are two orders of problems in this approach: first, the approximation on equivalent and independent sites is very crude for a molecule known to have sites of very high or rather low specificity and to undergo profound conformational rearrangements with ligand binding, which justifies its allosteric behavior; second, the marked concentration dependence of the diffusion coefficient of free salycilate demonstrates that this molecule is subjected to other solution equilibria (e.g. dimerization), which are completely overlooked. These observations do not diminish the importance and scope of diffusion measurements for protein binding, which will likely strongly increase in the fiiture, owing to the bursting number of NMR probes equipped with gradients and to the growing interest toward DOSY. Measuring proton longitudinal relaxation rates by inversion recovery is one of the easiest NMR experiments, still, on observing that Ti is a function of the molecular tumbling rate, we realize that it must assume completely different values for the free and bound ligand. In order to perform a rapid screening of pharmaceutical lead molecules, it has been proposed to use a competition essay for HS A with tryptophan as a probe. A set of solutions containing 1 part of HSA, a 12.5-fold excess of L-tryptophan, and the same amount (4-fold) of each lead was subjected to a selective inversion recovery (i.e. an experiment when only the Trp C2-H is inverted by means of a shaped pulse and after a fixed mixing time the whole spectrum is acquired using a hard, non selective 90° pulse). The inverted tryptophan resonance has a faster recovery (shorter Ti) in the presence of a larger fraction of bound Trp: the lead molecules which displace it more efficiently from HSA will induce a lengthening of Ti of the probe resonances. Consequently, they will appear increasingly negative, as shown in Figure 11 [39]. The slowed tumbling of bound ligand has been exploited also to derive information of the structural rearrangement a ligand may undergo on protein binding, with the so-called transferred-NOE measurements (e.g. tr-NOESY) [38]. It is known that magnetization transfer processes are much more effective in slow reorienting systems than in small free organic molecules, in fact for the latter ones small and positive (i.e. of opposite sign to the diagonal in a NOESY) NOE's are detected, while they are large and negative for the former ones. If two protons I and S are far enough (4A or more) in the free ligand, they are expected to have a

R«for«nc«

27$583 276708

K o a 11 ^ M ( 9 8 . 2 % ) KO=10MM(98.4%)

Ko « 9 M M (98.5%) Ko a $ MM (99.0%) KD s 5 M M (99.2%) Ko ~ 2 M M (99.716) KD « 1 M M (99.8%)

Figure 11. Recovery of the C2-H resonance of tryptophan after a selective inversion pulse followed by a delay T=0.96 S, in the presence of the same amounts of the lead molecules indicated by the company codes and of HSA. The dissociation constants of each lead from HSA are reported below. The leftmost spectrum, labelled with an asterisk was obtained in the absence of any lead and used as a reference. Reproduced with permission from Reference [39]

Serum Albumin and Natural Products

289

vanishing cross-relaxation rate ois^^^, manifested by a null steady state NOE or no cross peak in a NOESY-like spectrum. If upon binding to SA a ligand rearrangement occurs, bringing I and S nearer, ais^''""'*<0. Thus by comparing the spectra of the ligand alone and of a mixture containing also SA, one may be able to recognize the conformation of the bound ligand. The very large size of SA makes the T2 of its protons particularly short, so that their contribution to the proton spectrum can be filtered out by means of suitable delays or pulse sequences (T2 filtering). This has been used to highlight protons that have experienced NOE contacts with the proteins, through the NOE-pumping technique, successflilly applied to octanoic acid and other saturated non-branched carboxylic acids, as models of fatty acids [40]. An experiment joining features of transferred NOE and diffusion measurements, called diffusion-NOE-pumping, was demonstrated to be useful for characterizing the interactions between BSA and flavor molecules [41]. At the beginning of the sequence, a stimulated echo is used to filter out all the ligand signals, by exploiting the slow difEision of the protein. The only magnetization in the sample is thus stored in SA and it is now allowed to flow (through dipolar interaction, i.e. NOE) to the ligand during a mixing time, analogous to the NOESY Xm. Like in NOE pumping, the effect is to keep only intermolecular Overhaser contacts between a large and a small molecule: the two experiments differ in the dynamic process used for selecting the two molecular sizes, tumbling rate (thus T2 for conventional NOE pumping) versus diffusion. Binding of several homologue ketones and of vanillin to BSA was followed through this experiment. In the series of 2-alkanones a regular trend in the free energy of association is observed: for each additional methylene group in the chain, the affinity gains 0.28 Kcal/mol, which calls for hydrophobic interactions as the main BSA binding source, in analogy with fatty acids. Thus, 2-nonanone displays the strongest binding (dissociation constant of KD "= 833 M"^). This figure shows that the consequences of these interactions in the field of medicinal chemistry are likely to be very modest (alkanones appear too weak competitors for most drugs transported by SA). On the contrary, this kind information is relevant for food chemistry: the characteristic fragrance of food and beverages depends not only on their flavor composition, but also on their rates of release from nonvolatile components. Thus binding to proteins or to other macromolecules (e.g. epicatechin tannins) may lower the instant concentration of a given odorophore and give it a longer persistence.

7.

The Quest for High Throughput: Sensors Affinity biosensors are defined as devices incorporating immobilized receptor molecules that can reversibly detect receptor-ligand interactions with a high differential selectivity and in a non-destructive mode. In these sensors, a stoichiometric binding event takes place and the associated physico-chemical changes are detected by an appropriate transducer [42]. In a large number of applications, the biorecognition elements are antibodies, receptor proteins or nucleic acids. Mass biosensors are the subset where the transducer responds primarily to mass variations (neglecting other effects) on the surface where the receptor is immobilized: two most important devices today available are quartz crystal microbalance (QCM) and surface plasmon resonance (SPR). In both cases, a face of a quartz crystal is covered with a metal, usually gold, and on top of this metal layer, the receptor molecules are immobilized by suitable chemistry: this is the part of the system coming in contact with the rest of the environment and more specifically with a solution possibly containing the test ligand. QCM measures the piezoelectric resonance frequency of the quartz crystal, acting as an acoustic

290

Progress in Biological Chirality

resonator, whose frequency is function i.a. of the total mass of the composite element (quartz, metal, organic layer). This transducer gives a direct gravimetric response and the signal variations must be related to mass gain-loss at the organic layer. SPR responds to the refractive index at the gold/glass interface, whose changes (measured as resonance units, RU) appear in good approximation proportional to mass variations on the surface. In analogy, other measurements like impedance [43], or surface acoustic wave, have been put forward on dedicated instruments. Owing to the very fast response times of the transducers listed above, one can measure the onset and decay of the signal, during the course of an analyte injection in a flow of buffer. In this way, the association and dissociation processes can be followed in real time, affording kinetic rate measurements. The limitations of these techniques are to be found in the mass of the bound ligand, which has to be large enough to induce reliable signal variations. The small organic molecules discussed in the present work and characteristic of SA binding lead to small or very small signals, which until recently were considered unsuitable for quantitative evaluations. A few papers very recently appeared in the literature, where SA is immobilized on a quartz crystal of a QCM, [44] or the gold surface of a SPR [45] and a number of more or less standard ligands are injected in the flow cell. The results demonstrate that the accuracy attained by some of the newest instruments is sufficient at least for a gross evaluation of the affinity constant: the figures obtained with SPR compare fairly well with what reported in the literature on a rather wide selection of ligands, as shown in Figure 12. The great advantages of these biosensors reside in their: 1. sensitivity: very small amounts of analytes are used in the assay, which is anyway non destructive; 2. ease of use: once the surface is prepared with SA a very large number of assays can be performed in a very short time (typically one measurement requires a few minutes or less);

' Tipr

.? 99.9 r c • • Napr

£

^ Rito • • Suit

99 f

o

•mm Warf 1.2,3 # Digi m Deia

^ 95

• Keta

i 90 o

•o

80

^

60 40 10

Diaz _ Pyn • Rifa Pred

S i r 0.04

Sah ••Phen

0.16 0.08 0.12 Biosensor response (RU/Da)

0.20

Figure 12. Comparison between molefractionsof drug bound to HSA as determined by SPR and as reported in the literature for a variety of ligands (tipranivir, naproxen,ritonavir,sulfadimethoxin, warfarin, delavirdine, digitoxin, ketanserin, pyrimethamine, diazepam, coumarin, salicylic acid,rifampicin,phenytoin, prednisone, tolterodine, quinine, 5-HM, salbutamol) Reproduced with permission from Reference [45a])

Serum Albumin and Natural Products

291

3.

wide scope: virtually any ligand can be studied, independent of its response to any spectroscopic or radioactive detection, because these devices reveal primarily mass variations; 4. affinity constants as well as association dissociation rates can be determined, giving a deep insight into the binding process. These feature make biosensors ideal tools for pharmacological screening on wide libraries of potential candidates and progress in applications can be envisaged. Like in chromatographic applications, SA must be covalently bound to a surface: this chemical modification introduces a new structural feature in the protein, which might behave differently from the native counterpart. Evidence to date shows that the linker does not dramatically alter the affinity toward the most studied probes, like warfarin of phenybutazone. A piezoelectric device was used to study the BSA binding of an alkaloid, berberine, common in many medicinal plants of ayurvedic and Chinese traditional medicine and active as antimicrobial. In this work, the binding process could be followed simultaneously on different instrumental channels, by monitoring the changes in resonance frequency and capacitance PQCl (piezoelectric quartz crystal impedance analysis) [43].

0CH3

Berberine

The association constant and the kinetic rates (association and dissociation) were determined as 410'^ M'\ 70 M"^s"^ and 1.710"'^ s"\ respectively. The kass is very small compared to what expected for a diffusion-controlled process (see above), which must be regarded as a strong indication for a slow conformational rearrangement involved in association. On observing the rigid structure of berberine, one must conclude that this process must be essentially localized on BSA. Here we can appreciate the relevance of kinetic measurements made possible by biosensors.

8.

Conclusions We have examined several aspects of the binding of natural products to serum albumin, by choosing only a few cases, which demonstrate the application of specific experimental techniques. Of course, the literature in the field is enormous and an exhaustive treatment would be very hard and definitely beyond the scope of the present work. Our aim has been to clarify two points. First, there are strong interactions between this carrier molecule and food components, which need to be carefully addressed for a better understanding of the role of diet in the development of diseases and in pharmacological therapy. The efficacy of a medicine may be strongly influenced by what we may have eaten and even more importantly by food supplements. Vitamins, antioxidants, radical scavengers are often transported by albumin and may interfere in the solution equilibria of this protein and other drugs. Second, the number of experimental techniques that one can use is becoming increasingly large. We chose to neglect completely affinity chromatography and related methods, to concentrate on

292

Progress in Biological Chirality

more specialized and innovative approaches. In short time and with great sensitivity, tools like CD, but also some of the NMR experiments briefly reviewed, provide information not only on the fact that an association occurs, but also on the structure of the intermolecular adduct. The problem is interpreting how this information is encoded in the experimental data: we offered only a few examples of keys. Especially CD is the technique of choice when the object of investigation involves chirality, which is one of the most common structural features of biomolecules. Sensors appear today as a promising alternative/complement to the other measurements. They will have to be further developed, but appear to fulfill the requirements of a fast and easy response. Undoubtedly, we are witnessing a new attitude in understanding protein functioning and pharmacokinetics, where mechanistic observations at a molecular level drive the rational development of knowledge.

9.

Acknowledgments The authors wish to thank Dr. Gennaro Pescitelli for his critical comments and his help with some figures. MIUR is gratefully acknowledged for financial support (Program Fitoterapici: ottimizzazione delle caratteristiche tecnologiche e biofarmaceutiche prot. 2001037727).

10. [1] [2] [3] [4]

[5] [6] [7]

[8]

[9]

[10]

References Some common abbreviations fomid in the literature: HSA human serum albumin, BSA bovine serum albumin, OA ovalbumin, rHA recombinant HSA. (a) D.C. Carter and J.X. Ho, Smicture of serum albumin. Adv. Prot Chem. (1994) 153-203. (b) X.-M. He and DC. Carter, Atomic structure and chemistiy of human serum albumin. Nature 358 (1992) 209-215. S. Harder and P. Thurmann, Clinically important drug interactions with anticoagulants: an update. Clin. Pharmacokinet 30 (1996) 416-444. (a) J. Oravcova, B. Bohs and W. Lindner, New drug-protein binding studies trends in analytical and experimental methodology. J. Chromatogr. B 677 (1996) 1-28. (b) D.S. Hage and S.A. Tweed, Recent advances in chromatographic and electrophoretic methods for the study of drug-protein interactions. J. Chromatogr. B 699 (1997) 499-525. (c) C. Bertucci and E. Domenici, Reversible and covalent binding of drugs to human serum albumin: methodological approaches and physiological relevance. Current Medic. C/7em. 9(2002)1463-1481. G. Ascoli, C Bertucci and P. Salvador!, Stereospecific and competitive binding of drugs to human serum albumin: a difference circular dichroism approach. J. Pharm. Sci. 84 (1995) 737-741. I.M. Klotz, F.M. Walker and R.B. Pivan. The binding of organic ions by proteins. J. Am. Chem. Soc. 68 (1946) 1486-1489. (a) J. Aa. Jansen, Kinetics of the binding of salicylazosulfapyridine to human serum albumin. Acta Pharmacol. Toxicol. 41 (1977) 401-16. (b) Y. Keita, W. Womer, G. Veile, B.G. Woodcock and U. Fuhr, Influence of non-steroidal anti-inflammatory drugs on the binding kinetics of dansylsarcosine to human serum albumin - stereoselectivity, steric and inductive effects. Arzn. Forsch. Res. 46 (1996) 164-168. (a) JR. Roberts, J. Xiao, B. Schleisman, D.J. Parsons and C.F. Shaw, Kinetics and mechanism of the reaction between serum albumin and auranofin (and its isopropyl analogue) in vitro. Inorg. Chem. 35 (1996) 424-433. (b) Y. Xu, P. Tang, L. Firestone and T.T. Zhang, F-19 nuclear magnetic resonance investigation of stereoselective binding of isoflurane to bovine serum albumin. Biophys. 70 (1996) 532-538. E. Lozinsky, A. Novoselsky, R. Glaser, A.I. Shames, G.I. Likhtenshtein and D. Meyerstein, Effect of ionic strength on the binding of ascorbate to albumia Biochim. Biophys. Acta Biomembranes 1571 (2002) 239224. T. Peters and B. Meyer, NMR spectroscopy of proteins. Angew. Chem., Int Ed. 42 (2003) 864-890.

Serum Albumin and Natural Products

293

[11] F. A. De Wolf and G.M. Brett, Ligand-binding proteins: their potential for application in systems for controlled delivery and uptake of ligands. Pharmacol. Rev. 52 (2000) 207-236. [12] D M . Jung, J.S. De Ropp and S.E. Ebeller, Application of pulsed field gradient NMR techniques for investigating binding of flavor compounds to macromolecules. J. Agric. Food Chem. 50 (2002) 4262-4269. [13] S. Damodaran and J.E. Kinsella, Flavor protein interactions. Binding of carbonyls to bovine serum albumin: thermodynamic and conformational effects. J. Agric. Food Chem. 28 (1980) 567-571. [14] T. V. Burova, N. V. Grinberg, V.Y. Grinberg and V.B. Tolstoguzov, Binding of odorants to individual proteins and their mixtures. Effects of protein denaturation and association. A plasticized globule state. Colloids and Surfaces A: Physicochem. Eng. Aspects 213 (2003) 235-244. [15] (a) D.R. McMasters and A. Vedani, Ochratoxin binding to phenylalanyl-tRNA synthetase: computational approach to the mechanism of ochratoxins and its antagonism../. Med. Chem. 42 (1999) 3075-3086. (b) Y.V. Il'ichev, J.L. Perry and J.D. Simon, Interaction of ochratoxin A with human serum albumin. Preferential binding of dianion and pH effects. J. Phys. Chem, B 106 (2002) 452-459. (c) Y.V. Il'idiev, J.L. Perry and J.D. Simon, Interaction of ochratoxin A with human serum albumin. A common binding site of ochratoxin A and warfarin in subdomain IIA. J. Phys. Chem. B, 106 (2002) 460-465. (d) J.L. Perry, Y.V. Il'ichev, V.R. Kempf, J. McClendon, G. Park, R. A. Manderville, F. Riiker, M. Dockal and J.D. Simon, Binding of ochratoxin A derivatives to human serum albumin. J. Phys. Chem. B 107 (2003) 6644-6647. [16] (a) Y. Tsutsumi, T. Maruyama, A. Takadate, M. Goto, H. Matsunaga and M. Otagiri, Interaction between two dicarboxylate endogenous substances, bilirubin and an uremic toxin, 3-carboxy-4-methyl-5-propyl-2fiiranpropanoic acid, on human senmi albumin. Pharm. Res. 16 (1999) 916-923. (b) T. Sakai, K. Yamasaki, T. Sajo, U. Kragh-Hansen, A. Suenaga and M. Otagiri, Interaction mechanism between indoxyl sulfate, a typical uremic toxin bound to site IL and ligands bound to site I of human serum albumin. Pharm. Res. 18 (2001) 520-524. [17] (a) J.B. Harbome, T.J. Mabry and H. Mabiy, Eds., The Falvonoids Chapman and Hall Ltd, N.Y., 1975. (b) N.C. Cook and S. Samman, Flavonoids-chemistry, metabolism, cardioprotective effects, and dietary sources. Nutr. Biochem. 7 (1996) 66-76. (c) O. Dangles, C. Dufour, C. Manach, C. Morand and C. Remesy, Binding of flavonoids to plasma proteins. Methods Enzymol. 335 (2001) 319-333. [18] (a) P.C.H. HoUman, M.V.D. Gaag, M.J.B. Mengelers, J.M.R van Trijp, J.H.M. de Vries and M.B. Katan, Absorption and disposition kinetics of the dietary antioxidant quercetin in man. Free Radical Biol. Med. 21 (1996) 703-707. (b) P.C.H. Hollman and M B . Katan, Dietary flavonoids: intake, health effects and bioavailability. Food. Chem. Toxicol. 37 (1999) 937-942. [19] (a) J.T.J. Arts, G.R.M.M. Haenen, H.-P. Voss and A. Bast, Masking of antioxidant capacity by the interaction of flavonoids with proteins. Food Chem. Toxicol. 39 (2001) 787-791. (b) J.T.J. Arts, G.R.M.M Haenen, L.C. Wilms, S.A.J.N. Bectstra, C.G.M. Heijnen, H.-P. Voss and A. Bast, Interactions between flavonoids and proteins: effect on the total antioxidant capacity. J. Agric. Food Chem. 50 (2002) 11841187. (c) F. Zsila, Z. Bikadi and M. Simonyi, Probing the binding of the flavonoid, quercetin to human serum albumin by circular dichroism, electronic absorption spectroscopy and molecular modelling methods. Biochem. Pharmacol. 65 (2003) 447-456. [20] M. Simonyi, Z. Bikadi, F. Zsila and J. Deli, Supramolecular exciton chirahtv of carotenoid aggregates. Chirality 15 (2003) 680-698, [21] F. Zsila, Z. Bikadi and M. Simonyi, Induced chirahty upon crocetin binding to human serum albumin: origin and nature. Tetrahedron: Asymmetry 12 (2001) 3125-3137. [22] F. Zsila, M. Simonyi and S.F. Lockwood, Interaction of the disodium disuccinate derivative of mesoastaxanthin with human serum albumin: from chiral complexation to self-assembly. Bioorg. Med. Chem. Lett. 13(2003)4093-4100. [23] A. A. Bhattacharya, T. Griine and S. Curry, Crystallographic analysis reveals common modes of binding of medium and long-chain fatty acids to human serum albumin. J. Mol Biol. 303 (2000) 721-732. [24] T.A.G. Noctor and I.W. Wainer, Allosteric and competitive displacement of drugs from human serum albumin by octanoic acid, as revealed by high-performance liquid chromatography, on a human serum albumin-based stationary phase. J. Chromat. B 577 (1992) 305-15. [25] Y.A. Gryzuno, A. Arroyo, J.-L. Vigne, Q. Zhao, V.A. Tyurin, C.A. Hubel, R.E. Gandley, Y.A. Vladimirov, R.N. Taylor and V.E. Kagan, Binding of fatty acids facilitates oxidation of cysteine-34 and converts copper - albumin complexes from antioxidants to prooxidants. Archives Biochem. Biophys. 413 (2003) 53-66.

294

Progress in Biological Chirality

[26] C. Bertucci, A. Canepa, G. A. Ascx)li, L.F.L. Guimaraes and G Felix, Site I on human albumin: differences in the binding of (R)- and (S)-warfarin. Chirality 11 (1999) 675-679. [27] A. Ouchi, G. Zandomeneghi and M. Zandomeneghi, Complexation of chiral aromatic substrates and their chemistry in ground and excited states. Catalytic and chiraUty recognition properties of the protein in the cases of binaphthol, its photoisomers and ketoprofen. Chirality 14 (2002) 1-11. [28] (a) A. Rodger and B. Norden, Circular dichorism and linear dichroism, Oxford University Press, Oxford (UK) 1997. (b) N. Berova, K. Nakanishi and R.W. Woody, Circular dichroism principles and applications, Wiley-VCH, New York (US) 2000. [29] (a) A. Rosen, The measurement of binding constants using circular dichroism. Binding of phenylbutazone and oxyphenbutazone. Biochem. Pharmacol. 19 (1970) 2075-2081. (b) C. Bertucci, E. Domenici and P. Salvadori, Stereochemical features of l,4-benzodiazepin-2-ones bound to human serum albumin: difference CD and UV studies. Chirality 2 (1990) 167-174. (c) S. Gaudreau, J.F. Neault and H.A. Tajmir-Riahi, Interaction of AZT with human serum albumin studied by capillary electrophoresis, FTIR and CD spectroscopic methods. J. BiomoL Struct. Dynam. 19 (2002) 1007-1014. [30] S.M. Kelly and N.C. Price, The use of circular dichroism in the investigation of protein structure and function. Current Protein and Peptide Science 1 (2000) 349-384. [31] G. Pescitelli, D. Drochner, N. Berova and K. Nakanishi, unpublished results. [32] C. Bertucci, G.A. Ascoli, G. Uccello-Barretta, L. Di Bari and P. Salvadori, The binding of 5-fluorouracil to native and modified human serum albumin: UV, CD, and ^H and ' ^ NMR investigation. J. Pharm. Biomed. Anal. 13 (1995) 1087-1093. [33] L.-C. Lo, Y.-C. Liao, C.H. Kuo and C.T. Chen, A novel coumarine-type derivatizing reagent of alcohols: appUcation in the CD exciton chirality method for microscale structural determination. Org. Lett. 2 (2000) 683-685. [34] (a) DA. Lightner, W.M. Wijekoon and M.-H. Zhang, Understanding bilirubin conformation and binding. J. Biol. Chem. 263 (1988) 16669-16676. (b) C. Puzicha, Y.-M. Pu and DA. Lightner, Allosteric regulation of conformational enantiomerism. Bilirubin. J. Am. Chem. Soc. 113 (1991) 3583-3592. (c) S.E. Boiadjiev and DA. Lightner, Optical activity and stereochemistry of linear oligopyrroles and bile pigments. Tetrahedron Asymmetry 10 (1999) 607-655. [35] (a) F. Zsila, Z. Bikadi and M. Simonyi, Unique pH-dependent biphasic band shape of the visible circular dichroism of curcumin-serum albumin complex. Biochem. Biophys. Res. Commun. 301 (2003) 776-782. (b) F. Zsila, Z. Bikadi and M. Simonyi, Molecular basis of the Cotton effects induced by the binding of curcumin to human serum albumin. Tetrahedron Asymmetry 14 (2003) 2433-2444. [36] A.C.P. Reddy, E. Sudharshan, AG. A Rao and BR. Lokesh. Interaction of curcumin with human serum albumin - A spectroscopic study. Lipids 34 (1999) 1025-1029. [37] F. Zsila, Z. Bikadi and M. Simonyi, Induced chiraUty upon crocetin binding to human serum albumin: origin and nature. Tetrahedron Asymmetry 12 (2001) 3125-3137. [38] W.S. Price, Recent advances in NMR diffusion techniques for studying drug binding. Aust. J. Chem. 36 (2003) 855-860. [39] C. Dalvit, M. Flocco, B.J. Stockman and M. Veronesi, Competition Binding Experiments for Rapidly Ranking Lead Molecules for their Binding Affinity to Human Serum Albumin. Comh. Chem. High Throughput Screen. 5 (2002) 645-650 [40] (a) A. Chen and M.J. Shapiro, NOE Pumping: a novel NMR technique for identification of compounds with binding affinity to macromolecules. J. Am. Chem. Soc. 120 (1998) 10258-10259. (b) A. Chen and M.J. Shapiro, NOE Pumping. 2. A high-throughput method to determine compounds with binding affinity to macromolecules by NMR. J. Am. Chem. Soc. 122 (2000) 414-415. [41] DM. Jung, J. De Ropp and S.A. Ebeler, Application of Pulsed Field Gradient NMR Techniques for Investigating Binding of Flavor Compounds to Macromolecules. /. Agric. Food Chem. 50 (2002) 42624269. [42] M. Minuimi, Biosensors based on nucleic acid interactions. Spectroscopy 17 (2003) 613-625. [43] Y. Mao, W. Wei, D. He, L. Nie and S. Yao, Monitoring and kinetic parameter estimation for the binding process of berberine hydrochloride to bovine serum albumin with piezoelectric quartz crystal impedance analysis. Anal. Biochem. 306 (2002) 23-30. [44] EL. Lyle, G.L. Hayward and M. Thompson, Acoustic coupling of transverse waves as a mechanism for label-free detection of protein-small molecule interactions. .4«flf(v5/127 (2002) 1596-1600.

Serum Albumin and Natural Products

295

[45] (a) A. Frostell-Karlsson, A. Remaeus, H. Roos, K. Andersson, P. Borg, M. Hamalainen and R. JCarlsson, Biosensor analysis of the interaction between immobilized human serum albumin and drug compounds for ediction of human serum albumin binding levels. J. Med. Chem. 43 (2000) 1986-1992. (b) R.L. Rich, Y.S.L. Day, T. A. Morton and D.G. Myszka, High-resolution and high-throughput protocols for measuring drug/human serum albumin interactions using BIACORE. Anal. Biochem. 296 (2001) 197-207. (c) C. Bertucci and S. Cimitan, Rapid screening of small ligand affinity to human serum albumin by an optical biosensor. J. Pharm. Biomed. Anal. 32 (2003) 101-114.

Progress in Biological Chirality G. Palyi, C. Zucchi and L. Caglioti (Editors) © 2004 Elsevier Ltd. All rights reserved.

Chapter 24 Selection in the Abiotic Synthesis of RNA Using Metal Ion Catalyst and Template Hiroaki Sawai Department ofApplied Chemistry, Faculty ofEngineering, Gunma University, Kiryu, Gunma 376-8515 Japan sawai @chem.gunma-u. ac.jp

1.

Introduction It has been proposed that RNA could play the roles of information carrier and catalyst at an early stage of the origins of Ufe [1]. RNA could have been formed during the chemical evolution. Contemporary RNA is exclusively composed from a D-ribo-P-furanosyl 3'-5' phosphodiester linkage. However, many linkage isomers of RNAs are chemically possible. Typical examples of isomers of RNA are shown in Figure 1, D- and L-enantiomers, P- and a~anomers, and 3-5 and 2-5 linkage isomers. Moreover, six-membered pyranose ring systems are also possible. These isomers of RNA could have been formed in the initial prebiotic process, and D-ribo, P-fliranosyl 3-5 linked RNAs were selected later during chemical and Furanose Type Linkage isomers 3'-5' jinicage 2'-5' linkage

EnantiomeB D-form L-form

n^ B = Base

Pyranose Type 3'-4' linkage

2*-4' linkage

A-O-f-O-J

Figure 1. Isomers of RNA

Anomers p-anomer a-^nomer

298

Progress in Biological Chirality Template

Leaving group

Primer

" " ^ Enzymi

^^^ ion(IVI^:Mg 2*. Mn2*

Figure 2. Hypothetical mechanism of DNA synthesis by DNA-polymerase

biochemical evolution. We have studied a number of simulation reactions of the plausible prebiotic synthesis of RNAs and examined the properties of the resuhing RNA isomers. In this article, I describe the results of the nonenzymatic syntheses of RNA and discuss the selective advantage of the contemporary natural-type RNA from our studies. I will not touch on the pyranose-type RNAs that have been studied by Eschenmoser and coworkers [2]. In the biochemical system, synthesis of nucleic acids is conducted by an enzyme, polymerase. Figure 2 shows a possible mechanism of DNA synthesis by DNA polymerase [3]. Nucleoside triphosphates are used as substrates and chain-elongation reaction takes place between the terminal 3'-0H of a primer and a-phosphate of the nucleoside triphosphate. Template DNA controls the incoming substrate triphosphate. For the bond formation, a metal ion such as magnesium or manganese ion works as a catalyst [3, 4], RNA is also synthesized in a similar way by metal ion-dependent RNA polymerase [5]. As a starting monomer for the non-enzymatic RNA synthesis, we tried a nucleoside triphosphate, such as ATP at first. However, we could get no oligomer from ATP without an enzyme, and only hydrolysis of the triphosphate was observed. Thus, we used an imidazoleactivated nucleotide, nucleoside phosphorimidazolide, (Figure 3) as a monomer for RNA synthesis. A simulation experiment for prebiotic synthesis of nucleoside phosphorimidazolide

Urea. ATP, ADP

Mg-'

Adenosine + Inorganic Phosphate

^

A l .Imidazole

Mg2-

^<^4

I Oligo RNA

O-A

ImpA (Adenosine phosphorimidazolide)

'NSNV ^

r 1 .Metal Ion Catalyst "^ 2. Clay Mineral Catalyst 3.Template RNA

Figure 3. Formation and oligomerization of ImpA

Selection in the Abiotic Synthesis of RNA Using Metal Ion Catalyst and Template

299

was reported by Lohrmann [6]. Evaporation of an aqueous mixture of ATP or ADP, imidazole and a magnesium ion and subsequent heating gave adenosine phosphorimidazolide (ImpA). Inspired by the roles of metal ions in the enzymatic reactions, we tried oligomerization of ImpA by a metal ion catalyst in neutral aqueous solution. A clay mineral catalyst or a template nucleic acid also promotes the oligomerization of ImpA. The clay catalyst reactions have been carried out mainly by Ferris' group [7], and template-directed reactions by OrgeFs group [lb, 8, 9].

2.

Synthesis of OligoRNA by Metal Ion Catalyst Scheme 1 illustrates oligoadenylate formation from ImpA [10]. 2'- or 3'-0H of ImpA reacts with phosphorimidazolide of the adjacent ImpA forming a phosphodiester bond. Imidazole is a leaving group. Further reaction of ImpA to the resulting dimer gives a trimer and longer oligomers. The metal ion organizes the two molecules of ImpA by coordination and controls the regio-selectivity of the resulting internucleotide bond [10]. Intra-molecular bond formation gives cyclic products. Without a metal ion catalyst, simple hydrolysis of phosphorimidazolide bond takes place. Oligoadenylates up to pentamer were formed in neutral aqueous solution. Lead, zinc, manganese and cobalt ions were effective catalysts. The linkage of the resuhing oligomers was mainly unnatural 2'-5' [10]. OHgomerization of other nucleoside phosphorimidazolides took place in a similar way with a lead ion catalyst [11]. Oligomers up to pentamer were obtained from uridine, inosine, cytidine and guanosine phosphorimidazolide in total yields of 40 to 50 %. As for the role of metal ions, we think that they could work as a kind of template by coordination to organize the OH and phosphate groups in proximate positions promoting the phosphodiester bond formation. Activation of the OH group and charge neutralization of phosphate by coordination also enhance the bond formation. Thus, the simple metal ion catalyzes the oligomerization reaction and could play

NppN

fpN

pNpNpN

4) HO vOH (ImpN)

HO OH

(ImpNpN) HO OH

pNpN

ImpN H2O

pNpNpN -^

ImpN ImpNpNpN

H2O

(pN)n n=4-18

Catalyst: Pb^*, Zn^*, Co^*, U02^*

Scheme 1. Oligomerization of nucleoside 5'-phosphorimidazolide by metal ion catalyst

300

Progress in Biological Chirality

the role of the metal ion in the enzyme polymerase. Nucleotides coordinate to a metal ion with a base, phosphate and ribose hydroxyl groups [12]. Hard metal ions, such as alkali and alkali-earth metal ions, coordinate with a phosphate. On the other hand, soft metal ions such as silver and mercury prefer coordination with a base. Intermediate transition metal ions coordinate with both a phosphate and a base. The uranyl ion has a unique coordination property [12]. It coordinates with a phosphate and a ribose hydroxyl group. Previously, Egami suggested that transition metal ions in primitive seawater could be used as a primitive catalyst for the prebiotic synthesis and incorporated into the metalloenzymes [13]. Iron, zinc, molybdenum, manganese and copper ions are comparatively abundant in seawater and are essential for all living organisms. Cobalt and nickel ions are essential for some organisms. The uranyl ion is nonessential and toxic for organisms, ahhough it is more abundant than the cobalt ion in seawater. As the uranyl ion has a unique coordination character, we examined the uranyl ion as a catalyst for the oligonucleotide synthesis. Figure 4 shows HPLC profiles of uranyl ion-catalyzed oligoadenylate formation. After one day at room temperature, oligomers from dimer to hexadecamer were formed [14]. The main peaks are 2'-5' linked oligomers and the side peaks are linkage isomers containing 3'-5' linkage. When concentration of the catalyst uranyl ion was as low as 10 \xM, the chain length decreases to 4 as shown in Figure 4(B). pH 7.5 is the optimum of the solution for the oligomerization reaction. The intemucleotide linkage is mainly unnatural 2'-5'. We fiirther examined the uranyl ion-catalyzed oligomerization of chiral D-ImpA and racemic D+L-ImpA to assess the effect of chirality of ImpA on the oligomerization reaction [15]. HPLC profiles of oligoadenylate formation from D-ImpA and racemic D+L-ImpA are shown in Figure 5. Complex mixtures of isomeric oligoadenylates were formed fi'om the D+L-ImpA, while oligomerization took place selectively fi'om D-ImpA. The chain-length of the resultingoligomers was somewhat higher in the case of racemic D+L-ImpA. In the dimer region, formation of homochiral dimers, DD- and LL-dimer (Peak 2), was observed along with the heterochiral dimers, DL- and LD-dimer (Peak 2') from racemic D+L-ImpA (Figure 5(B)). The ratio of the homochiral dimer to the heterochiral dimer was nearly 2:1, indicating the preferential formation of the homochiral dimerfromthe racemic monomer. We prepared adenosine thiophosphorimidazolide or thiophosphorbenzimidazolide and examined their polymerization to gain information on the mechanism and stereochemical

(A)

2'-6'(pA)3 2'-5'(pA)4 2'-6'(pA)2

2'-6'(p/\)2 I

=LiL^Ajy 10

20 30 30 (min) (min) Figure 4. HPLC profiles of polymerization products from ImpA by m:anyl ion catalyst. Polymerization of ImpA (0.05 M) was conducted at room temperature for 1 day in neutral aqueous solution in the presence of 1 mM (A) or 0.01 mM (B) uranyl-ion catalyst. The number shows the chain length of the oligoadenylate

20

Selection in the Abiotic Synthesis of RNA Using Metal Ion Catalyst and Template 301

1 (A) 2

11 IS <

(B)

|3

1^ 11 ^

Lui'Uw

110

10 20 Retention Time (min)

10 20 30 Retention Time (min)

Figure 5. HPLC profiles of polymerization productsfromD-ImpA (A) and racemic D+L-ImpA (B) by manylion catalyst. Polymerization of ImpA was conducted at room temperature for 1 day in neutral aqueous solution in the presence of 1 mM uranyl-ion catalyst. The number shows the chain length of the oligoadenylate. The peaks 2 and 2' in (B) are, homochiral and heterochiral dimers, respectively

features of the oligomerization by the uranyl ion catalyst (Scheme 2) [16], Introduction of a sulfur atom in place of an oxygen atom at the phosphate gives chirality, Rp- and Sp thiophosphate. Ribose is a D-type. Thus, the nucleoside thiophosphorimidaolide was obtained as a nearly 1:1 ratio of the Rp- and 5^-diastereomeric mixture. The Rp and Sp diastereomers of thiophosphorbenzimidazolide could be resolved by HPLC. Figure 6 (A), (B) and (C) illustrates HPLC profiles of the oligomerization products from the adenosine thiophosphorbenzimidazolide. The Sp- monomer effectively promoted the formation of 2'-5' oligomers up to hexamer (Figure 6(A)). These oligomers were digested easily with venom phosphodiesterase, which indicates that they have Rp configuration. On the other hand, the main product from the 7^-monomer was 2'-5' linked Sp-dimer (Figure 6(B)). Both Rp- and 5^-oligomers were obtained from Sp and Rp diastereomeric adenosine thiophosphorbenzimidazolide (Figure 6(C)). i^-oligomers were mainly formed in this case. From the diastereomeric mixture of adenosine thiophosphorimidazolide, i^-oligomers up to pentamer were formed preferentially (Figure 6(D)). These results suggest that the oligomerization took place by an SN2 inversion mechanism. We think that the oligomerization takes place via a

Sp or Rp

t^

f

)e-f2Lo-| ^(S)A

HO

OH

_ypi>

sSporRp

HrtXfcJ

HO

(Im) (Bzim)

Scheme 2. Oligomerization of Rp- and 5/?-adenosine 5'-thiophosphorazolide (Xp(S)A)

302

Progress in Biological Chiiality

(A)

Figure 6. HPLC profiles of uranyl-ion catalyzed polymerization productsfi"omadenosine 5'-thiophosphorazolide (Xp(S)A). (A), 5p-Bzimp(S)A; (B), i?/?-Bzimp(S)A; (C), 5/7-+;?/7-Bzimp(S)A; (D), Sp'+Rp -Imp(S)A

polymeric complex formed from a uranyl ion and monomer nucleotides. This type of uranylnucleotide complex was proposed by Feldman and Kainosho [17]. The uranyl ion is a hard acid and prefers to coordinate to the hard phosphoryl oxygen, but not to the soft sulphur [12]. Orientation of the thiophosphoryl monomer imposed by coordination to the uranyl ion with the oxygen is likely responsible for this Sp monomer selective oligomerization. Nucleophilic attack of the 2'-0H group to the thiophosphoryl group gives a thiophosphodiester bond with inversion. Thus, the resulting oligomers from the 5^-monomer have Rp configuration. In the case of imidazolide, we found that interconversion ofRp- and Sp- monomers took place more quickly than the oligomerization. The oligomerization could take place preferentially from the ;S^-monomer to the i^-oligomer as shown in Scheme 3 [16].

(/?p)-lmp(S)A

(Sp)-lmp(S)A

(Sp)^p(S)A * - I (Rp)-Oligomers

{Sp )-configuration

Scheme 3. Hypothetical mechanism for the 5/7-Imp(S) A selective uranyl-ion catalyzed oligomerization

Selection in the Abiotic Synthesis of RNA Using Metal Ion Catalyst and Template

303

We examined the oligomerization of a-anomeric Imp A by the uranyl ion catalyst [18]. Both a- and P-anomeric nucleosides were formed in equal amounts in the prebiotic simulation of nucleoside synthesis [19]. Oligomers up to hexamer were formed from aanomeric ImpA. But the yield and the chain length of the a-anomeric oligoadenylates were low compared with those from P-anomeric ImpA under the same condition. In the case of aImpA, the base part is located near the hydroxyl group and suppresses the nucleophilic attack of the hydroxyl group for the internucleotide bond formation. Thus, the P-anomeric nucleotide is more advantageous for the oligomerization, although both anomers were formed in equal amounts in the simulation reaction of prebiotic synthesis of nucleoside.

3.

Synthesis of OligoRNA on a Complementary Template RNA The template-directed synthesis of oligoRNA is considered as a model reaction of replication or ligation of RNA. A large number of studies have been done on the templatedirected synthesis of oligoRNA, especially by Orgel's group [lb, 8, 9]. For example, ImpA lines up on a poly(U) template by complementary hydrogen bonding and the neighbors condense with each other forming oligoadenylates. We are interested in the template-directed synthesis using short-chained oligoRNA, because no long-chained RNA was present at first and short-chained RNAs with 2'-5' and 3'5' linkage are formed in simulation experiments of prebiotic RNA synthesis without a template. We examined the helix formation between the oligo(A) and oligo(U) systems by UV, because the template-directed synthesis can occur under the helix-forming condition between the complementary strands [20]. Figure 7 shows thermal denaturation curves of the mixture of monoadenylate (pA) and several chain lengths of oligo(U) monitored by UV at 260 nm. Hypochromicity was observed in the case of the helix formation between pA and oligo(U). Oligo(U) with chain-length 5, 6 and 7 could not form a helix with pA at 0 °C under increasing the chain- length from 8, 10, 12 to polymer. Table 1 shows the summary of the helix formation between oligo(A) and oligo(U) at 0 °C. Longer oligo(U) over octamer can

-(PU%-PA (pU)6-pA
-^TempCC) '

Figure 7. UV thermal denaturing curves of the mixture of pA (0.025 M) and oligoU (0.05M) in the buffer (pH 7.0) containing 0.2 M NaCl and 0.075 M MgCl2. The oligonucleotide concentration is expressed on a nucleotide residue basis

304

Progress in Biological Chirality Table 1. Helix formation of oligo(A) and oligo(U) OUgo(U) (pU)„ 1

2

3

4

1 2

-

3

8

'

± "«-

±

+

+

+

+

+

+

+

+

+

+

4 5

±

+ +

O

6

+

+

8

+

+

10

+

poly(A)

poly(U)

6

^ < <

+

10

12

5

-

-f

+

+

Helix formation between 0.025 M of oligo(A) and 0.05 M of oligo U at 0°C in the presence of 0.075 MgCl2 and 0.2 M NaCl in the bufer (pH 7.3). -, no helix formation was observed. +, helix formation including partial formation was observed.

form a helix with monomeric pA. However, monour idyl ate and diuridylate cannot form a helix with any oligo(A) or poly(A). Triurydylate can form a helix with oligo(A) with chainlength over 6. We conducted the oligomerization of Imp A in the presence of the oligo(U) template under the same condition as that for the helix formation described above. In accordance with the helix formation, longer oligo(U) over 8 works as a template and the yield and chain length of the resulting oligoadenylates increased with increasing chain length of the oligo(U) template as shown in Table 2 [21]. A lead ion catalyst enhanced the oligoadenylate formation on the oligo(U) template [21]. The presence of the oligo(U) template and the lead ion catalyst controlled internucleotide linkage of the resuhing oligo(A). The ratio of 3-5 linkage of the resuhing diadenylate was increased from 12 to 65 % by the presence of the oligo(U) template and the lead ion catalyst.

Table 2. Oligoadenylate Formation from ImpA on Oligo U Template TEMPLATE

(pU) (PU)4 (PU)6 (PU)8 (pU)io (PU)12

poly(U)

(PA)2

YIELD (%) (PA)3 (PA)4

2.0 2.7 1.9 3.2 10.4 13.5 19.7 26.7

0.1 0.2 0.5 0.6 2.6 3.1 5.9 14.1

1.0 1.4 2.9 6.6

(PA)5

0.3 0.8 5.7

Oligomerization of ImpA (0.025 M) was carried out in the presence of oligo(U) template (0.05 M based on a nucleotide residue) at 0 °C for 16days in the buffer (pH 7.0) containing 0.075 M MgCl. 0.2 M NaCl

Selection in the Abiotic Synthesis of RNA Using Metal Ion Catalyst and Template

305

We further studied the chiral selection of the oligoadenylate formation on a poly(U) template [22]. Poly(U) was enzymatically prepared and exclusively composed from Duridine. D-ImA olivgomerized effectively forming oligoadenylates up to hexamer on a Dpoly(U) template (Figure 8(A)). However, oligoadenylate formation from L- and racemic D+L-ImpA took place less effectively than that from D-ImpA, and oligoadenylates with short-chain were formed in low yields as shown in Figure 8(B) and 8(C). Previously, Joyce and Orgel reported an interesting chiral selection in the template-directed oligoguanylate formation on a poly(C) template [23]. They used 2-MeImpG as a monomer and a D-poly(C) as a template. The D-monomer oligomerized very efficiently forming up to 15mer. On the other hand, the L-monomer or racemic D+L-monomer only gave short oligoguanylates in small amounts on the poly(C) template. The results that oligoguanylate and oligoadenylate formations take place more effectively from the corresponding D-monomers than from the Lmonomers on the complementary D-poly(C) and D-poly(U) template, respectively, indicate that the homochiral system is preferable for the template-directed synthesis of oligoRNA.

(C)

(A)

M wL,

\AJ 10 15 20 25 Retantion Time (min)

10

15 20 25 Retention Time (min)

Sa

(B)

LAJV_ ~10 15 To 25 Retention Time (min)

30

Figure 8. HPLC profiles of oligomerization products from a chiral and a racemic ImpA on a D-poly(U) template. (A) D-ImpA; (B), L-ImpA; (C) racemic D+ L-ImpA. OUgomerization of ImpA (0.025 M) was carried out in the presence of poly(U) (0.05M) in the buffer (pH 7.3) containing 0.2 M NaCl and 0.075 M MgCl2. The ohgonucleotide concentration is expressed on a nucleotide residue basis. The numbers show the chain length of the resulting oligoadenylates

306

Progress in Biological Chirality

4.

Ligation of Oligo(U) on Poly(A) Template Table 1 indicates that oligo(U) with chain length of more than three can form a helix with poly(A) or long-chained oligo(A), but not di- or monouridylate at 0 °C. The results of the helix formation suggest that triuridylate or longer oligouridylate could condense on a poly(A) or a long-chained oligo(A) template yielding the corresponding long oligouridylates at 0 °C. Thus, we conducted the ligation of (pU)5, (pU)4 or (pU)3 on a poly(A) template at 0 °C in the presence of a water-soluble condensing agent to gain information on the minimum requirement of the template-directed condensation of pyrimidine nucleotides on a polypurine template. Figures 9(A) and 9(B) illustrate the HPLC profiles of the reaction products from pentauridylate, (pU)5, on a poly(A) template [24]. After I day, the formation of 5'phosphorimidazolide of pentauridylate and a small amount of (pU)io was observed in addition

(A)

vTUJ (B)

(C)

Uli (D)

0

10 20 30 Retention Time (min)

Figure 9. HPLC profiles of oligomerization products from a pentauridylate on a poly(A) template. The reaction was performed in 20 ^1 solution containing 40 mM (pU)5, 20 iiiM poly(A), 20 mM MgCl2,0.2 M NaCl and 1.0 M water-soluble cait)odiimide (ED AC) in 0.4 M imidazole-HCl buffer (pH 6.0). (A) Reaction at 0 °C for 7 d. (B) Reaction at 0 °C for Id. (C) Reaction at -20 °C for 60 days. (D) Control reaction without template. Peak identification: 1, 5'-dephosphoiylated and/or cyclic pentauridylate, 2, (pU)5, 3, 5'-phosphorimidazolide of pentaurdylate, 4, (pU)io, 5, (pU)i5

Selection in the Abiotic Synthesis of RNA Using Metal Ion Catalyst and Template

307

Figure 9(A). Figure 9(D) demonstrates that a very small amount of (pU)io was formed after 7 days in the control reaction where no template was used. The yield of (pU)io decreased to less than 1 % when the condensation reaction was carried out at 25 ''C even in the presence of poly(A). On the other hand, poly(A) template-directed condensation reaction at -25 °C under a eutectic condition resulted in increase in yield of (pU)io, although both condensation and hydrolysis reactions of the intermediate phosphorimidazolide of (pU)5 were retarded largely under the eutectic conditions. Thus the yields of (pU)io and (pU)i5 were 6.6 and 0.4 %, respectively, and the intermediate phosphorimidazolide survived in a large amount at -25 °C after 60 days (Figure 9(C)). The condensing agent, ED AC, promoted the reaction of 5'phosphate of (pU)5 with imidazole initially to convert the phosphorimidazolide of pentauridylate, which condensed each other on a poly(A) template to form decauridylate. The intramolecular reaction of phosphorimidazolide of pentauridylate may also take place to form cyclic pentauridylate. The resulting (pU)io was isolated by HPLC and digested with nuclease PI to determine the ratio of the linkage isomers. Nuclease PI degrades only the 3'-5' linkage leaving only 2'-5' linked oligouridylate. The ratio of the linkage isomers of decauridylate, pUpUpUpUpU2'p5'pUpUpUpUpU to pUpUpUpUpU3'p5'UpUpUpUpU, was 92 to 8. Thus, the 2'-5' internulceotide linkage was mainly formed in the template-directed reaction. The preferential formation of the 2'-5' linkage was also observed in the template-directed oligoadenylate formation on a poly(U) template [8]. The condensation of (pU)3 or (pU)4 was carried out in the presence of a poly(A) template at 0 °C under the same condition as that for (pU)5. The coupling products, (pU)6 and (pU)8, were obtained in small amounts; however, their yields were higher than those obtained in the control reactions where no template was used in the reactions. On the other hand, the presence of poly(A) did not promote the coupling reaction of (pU)2 or pU. The results that oligouridylates with chain-length of more than three are condensed on a poly(A) template are in accordance with the helix formation ability between oligo(U) and poly(A).

5.

Ligation of Short OligoRNA on Complementary OligoRNA Template The long chained-RNA could not be formed in the initial step of RNA synthesis. Shortchained OligoRNA could have been formed in an initial stage of prebiotic processes of RNA synthesis. Hybridization and condensation of short-chained RNAs at low temperature, and strand-separation at high temperature could form long-chained RNAs. Thus we conducted condensation of an oligoRNA on a complementary oligoRNA template as a model reaction of ligation [25]. 2'-5' linked diadenylate and 2'-5' or 3'-5' linked decauridylates were used as substrate and a template, respectively, for the model reaction of ligation. Table 3 shows the results of the template-directed ligation of diadenylate. The 2'-5' decauridylate, ([2'-5']UioX served as a template for the synthesis of tetra- and hexaadenylates, (pA)4 and (pA)6, from the 5'-phosphorimidazoHde of 2'-5' diadenylate (ImpA2'p5'A). Joining of [2'-5']Uio and ImpA2'p5'A also took place in substantial amounts to yield long-chained oligoribonucleotides in the template-directed reaction. An unusual CD spectrum, (not shown) ascribed to helix formation between [2'-5']Uio and [2'-5'](pA)2 was observed under the same conditions as that of the template-directed reaction. The 3'-5' linked decauridylate, ([3'5']Uio), also promoted the template-directed synthesis of oligo(A)s from ImpA2'p5'A but more slowly compared to [2'-5']Uio. The results indicate that short-chained RNA with a 2'-5' phosphodiester bond could lead to longer oligoribonucleotides by template-oligomers

308

Progress in Biological Chirality Table 3. Template-directed Oligomerization of ImpA2'p5'A, and Ligation of ImpA2'p5'A to the Template Strand [2'-5']Uio TEMPLATE

RODUCT

YIELD(%)^

TYPE OF FC)RMED LINKAGE (%) 2'-5' 3'-5'

[2'-5']Uto (PA)4 (PA)6 (U10A2) (U10A4)

21 1.5 22 10

81"

19b

45^

55'

(PA)4 (pA)6

8 0.5

94b

6'»

[3'-5']Uio

The reaction of ImpA2'p5'A (0.01 M) on the Uio template was run in the presence of 0.03 M MgCl2 and 0.2 M NaCl in N-ethylmorpholine buffer (pH 7.0) at 0°C for 28 days. ^ Yield was estimatedfromthe peak area of products in the HPLC; ^ Ratio of type of linkage between two [2'-5'](pA)2 units, pA2'pA-pA2'pA; ''Ratio of type of linkage between [2'-5']Uioand [2'-5'](pA)2

directed chain elongation. We also studied the ligation of linkage isomers of short-chained mixed- sequence RNAs [26]. 2'-5' and 3'-5' linked tetramer, ACUG and complementary decamer, CAGUCAGUCA, were used in this study as a model substrate and template. The ligation reactions were carried out at 0 °C for 14 days under the helix-forming condition using water-soluble carbodiimide as a condensing agent in an imidazole buffer. The 2-5 linked tetramer condensed each other efficiently in the presence of a 2'-5' linked decamer template forming the octamer in 39 % yield. When the 3'-5' linked tetramer was condensed in the presence of a 3-5 decamer template, the corresponding octamer was obtained in 17 % yield. Only very small amount of the octamer was formed without a template. The ligation reaction of the 3'-5' linked tetramer in the presence of a 2'-5' decamer template or the ligation reaction of the 2'-5' linked tetramer in the presence of a 3'-5' decamer template took place less efficiently. The results suggest that the homo-linkage system is preferable for the template-directed synthesis of oligoRNA. Conformation of a double helix composed from different linkage isomers of RNA may be unsuitable for the bond formation of the ligase reaction. The yield of the Hgation product from the 2'-5' linked substrate and 2'-5' linked template system was higher than that from the 3'-5' linked substrate and 3'-5' linked template system. However, the hybridization ability between complementary 3'-5' linked RNAs is higher than that of the corresponding 2'5' linked RNAs [26, 27]. The high hybridization ability may be one possible reason for the selection of 3'-5' Hnked RNA over 2'-5' linked RNA, because template-directed ligation of 3'-5' linked RNA can take place under higher temperature conditions than that of the corresponding 2'-5' linked counterpart.

6.

Discussion Finally, I would like to mention the selection of isomers of RNA. There are many possible isomers of RNA. An overall account of the data leads to interpretations on the selection of RNA. The homo-chiral and homo-linkage isomers of this type of molecule are more

Selection in the Abiotic Synthesis of RNA Using Metal Ion Catalyst and Template 309 advantageous than the hetero-chiral and hetero-Hnkage isomers. The P-anomer is more preferable than the a-anomer. Criteria for the natural selection of linkage isomers of RNA during the chemical evolution are: ease of formation, stability, helix-forming ability and conformational flexibility, all of which affect biochemical functions. Studies on the abiotic synthesis of RNA done by our and other groups demonstrate that homo-chiral and homolinkage RNA systems are more advantageous than hetero-chiral and hetero-linkage RNA systems. However, the question why nature chose D-type RNA instead of L-type RNA remains unanswered. The origin of the D-nucleic acid system is likely related to the origin of L-amino acid and the protein synthesis system.

7.

References

[1] (a) W. Gilbert, Nature 319 (1986) 618. (b) G.F. Joyce and L.E. Orgel, in: The RNA World (Eds. R.F. Gestelandand and I E . Atkins) Cold Spring Harbor Laboratory Press, N.Y., 1993, pp. 1-25. [2] (a) M. Beier, F. Reck, T. Wagner, R. Krishnamurthy and A. Eschenmoser, Science 283 (1999) 699-703. (b) A. Eschenmoser, Science 284 (1999) 2118-2124. [3] T.A. Steitz, Curr. Opinion Struct. Biol. 3 (1993) 31-38. [4] H. Pelletier, M.R. Sawaya, W. Wolfe, S.H. Wilson and J. Kraut, Biochemistry 35 (1996) 12762-12777. [5] (a) S.C. Tyagi, Biochemistry 31 (1992) 6447-6453. (b) I. Treich, M. Riva and A. Sentenac, J. Biol. Chem. 266(1991)21971-21976. [6] R. Lohrmann, J. Mol. Evol. 10 (1997) 137-154. [7] (a) G. Ertem and J.P. Ferris, Nature 379 (1996) 238-240. (b) K. Kawamura and J. P. Ferris Orig. Life Evol. Biosphere 29 (1999) 563-591. [8] R. Lohrmann and L.E. Orgel,^cc. Chem. Res. 1 (1974) 368-377. 19] L.E. Orgel,^cc. Chem. Res. 28 (1995) 109-118. [10] H. Sawai, J. Am. Chem. Soc. 98 (1976) 7037-7039. [11] (a) H. Sawai and M. Olino, Bull. Chem. Soc. Jpn. 54 (1981) 2759-2762. (b) H. Sawai and M. Ohno, Bull. Chem. Soc. Jpn. 58 (1985) 361-366. [12] R.M. Izatt, J.J. Christensen and J.H. Rytting, Chem. Rev. 71 (1971) 439-481. [13] F. Egami, J. Mol. Evol. 4 (1974) 113-118. [14] H. Sawai, K. Kurodaand and T. Hojo, Bull. Chem. Soc. Jpn. 62 (1989) 2018-2023. [15] K. Osawa, H. Urata and H. Sawai, submitted for publication. [16] (a) M. Shimazu, K. Shinozuka and H. Sawai, ^/7gew. Chem. Int. Ed. 32 (1993) 870-872. (b) M. Shimazu, K. Shinozuka and H. Sawai, J. Chem. Soc. Perkin Trans I (2002) 1778-1785. [17] (a) I. Feldman and K.R. Rich, J. Am. Chem. Soc. 92 (1970) 4559-4563. (b) M. Kainosho and M. Takahashi, Nucleic Acid Symp Ser. 12 (1983) 181-184. [18] H. Sawai, T. Itoh, K. Kokaji and K. Shinozuka, J Mol. Evol. 45 (1997) 209-215. [19] W.D. Fuller, R.A. Sanchez and L.E. Orgel, J Mol. Biol. 67 (1972) 25-33. [20] H. Sawai, S. Totsuka and K. Yamamoto, Orig. Life Evol. Biosphere 27 (1997) 525-533. [21] (a) H. Sawai, J Mol. Evol. 17 (1981) 48-51. (b) H. Sawai, J Mol. Evol. 17 (1981) 108-109. [22] K. Osawa, H. Urata and H. Sawai, submitted for publication. [23] G.F. Joyce, G.M. Visser, C. A. A. van Boeckel, J.H. van Boom and L.E. Orgel, Nature 310 (1984) 602-604. [24] H. Sawai and M. Wada, Orig. Life Evol. Biosphere 30 (2000) 503-511. [25] H. Sawai, S. Totsuka, K. Yamamoto and H. Ozaki, Nucleic Acids Res. 26 (1998) 2995-3000. [26] H. Sawai, M. Wada, T. Kouda, A.N. Ozaki and H. Ozaki, unpublished results. [27] (a) R. Kierzek, L. He and D. H. Turner, Nucleic Acids Res. 20 (1992)1685-1690. (b) H. Sawai, J. Seki and H. Ozaki, J. Biomol. Struct. Dynam. 13 (1996) 1043-1051.

Progress in Biological Chirality G. Palyi, C. Zucchi and L. Caglioti (Editors) © 2004 Elsevier Ltd. All rights reserved.

Chapter 25 Different Internal Gradients for L and D Homochiral Solutions in Inhomogenous Magnetic Fields Romulus Scorei,* Vily Marius Cimpoia§u University of Craiova, A.I. Cuza 13, 1100, Craiova, Romania scorei@central ucv. ro, romulus scorei@hotmail com

1.

Introduction It is now accepted that molecules sufficiently complex to support life v^ill also sufficiently complex to support chirality, that is, the existence of left- and right handed mirror image pairs, or enantiomers [1]. It is well known that essential biopolymers associated with the life are made up of chiral molecules characterized by unique homochiralities (D-Sugars in DNA and RNA and L-amino acids occur in proteins) as well as absolute chiral purity (in these biopolymers no enantiomeric L-sugar or D-amino acid monomer units are found) [2]. While the indispensable role of chirality has been largely ignored in most speculative scenarios for the origin of life, the general association of molecular asymmetry with living matter has been appreciated since Pasteur (1860) pointed it out over 130 years ago [3]. It is very clear that a symmetry-breaker is needed to achieve the homochirality necessary to get life started. Once the symmetry is broken, initial selection of the one hand biomolecules fixes the handedness of the rest of biochemistry through diastereomeric connections, such as that between the D-sugar and L-amino acids [4]. In the 1884 Pasteur first investigated the possibility that magnetic and gravitational fields might induce the formation of chiral molecules, but his experiments were unsuccessful. Later, many investigators have claimed positive results in producing chiral molecules with both electric, magnetic, gravitational and centrifugal fields [5], and also numerous chiral influence have been proposed, ranging from the Coriolis force, circularly polarized light and weak force [6, 7]. The small parity-violating energy difference between enantiomers, in the range 10" kTio \0'^^ kTai room temperature, produces a slight excess of the more stable enantiomer [12] that could select the enantiomers used in biochemistry. These small energy differences need amplifying, and two mechanisms have been proposed. One is the Yagamata cumulative [13] and other is The Kondepudi catastrophic mechanism, based on nonequilibrium statistical mechanics [14]. We believe that the solvent influence (through hydrogen bonding), in this case water with different anions, will affect the chemical shifts of the enantiomeric monomer. The handedness of elementary particles themselves means that L and D molecules are not really enantiomers but diastereoisomers [8, 9]. These diastereoisomers, L and D molecules should therefore differ slightly in all properties, including NMR chemical shift [10] and

312

Progress in Biological Chirality

especially energy [11]. For solutions, the most method chosen to study the three-dimensional structure of sugars is nuclear magnetic resonance, through the parameters represented by chemical shifts, coupling constants and relaxation time measurements. Relaxation time measurements add information on the mobility and the behaviour of molecules in solution.

2.

Brief Connection Between Magnetic Field and Chirality The magnetic field, applied to a photo-induced chemical reaction leads [26] to the conclusion that there is a connection between chirality-homochirality and the magnetic field. The contribution of parity-violating effects to the phase transition of the D to L-alanine crystals was confirmed by ^H CRAMPS solid state NMR, DC-magnetic susceptibilities and ultrasonic measurements. It was found that the spin relaxation mechanism of alpha-H nucleus of D-alanine molecule is different from L-alanine and the effect is stronger than that of Lalanine [28]. Surrounding molecules also influence the NMR chemical shifts for nuclei in a molecule. In the absence of specific molecular interactions, solvent molecules will have the most profound effect. Pople et al. [15] and Foreman [16] have reviewed NMR studies on the solvent effect. It is also knows that hydrogen bonding plays a very important role in many biological situation and intramolecular and intermolecular hydrogen bonding will have an effect on the chemical shift [17, 18]. Buchingam et al. consider the "long range" and "short range" forces affecting the chemical shift in solution and proposed four shielding contribution from the solvent [29]: CTsoivent'^cTb^ cJa^ (J^v^ GE , whcrc Gb is a "long range" term arising fi-om the bulk magnetic susceptibility of the medium; Ga arises from the anisotropy in the molecular susceptibility of the solvent molecules; a^ arises from van der Walls or dispersion forces between solute and solvent; <JE is the "polar" effect caused by the charge distribution in the neighbouring solvent molecules leading to permanent electric dipole and quadrupole fields acting on the solute. In the high resolution NMR, an essentially two static NMR parameters, namely the magnetic shielding tensor
Different Internal Gradients for L and D Homochiral Solutions in Inhomogenous Magnetic Fields

313

The principal goal of this study is to explain the differences in the observed spin-spin relaxation time T2 for L and D enantiomer of amino acids and sugars, at different interpulse delay in order to characterize the dynamical process like chemical exchange process and diffusion phenomena. We have been concerned with the following problem: how does the difference between the chiral molecules in their natural environment, water, appears? We based on the interaction process between water and chiral molecules. This complex process is shown by hydration phenomena, hydrogen exchange phenomena and diffusion phenomena. All these phenomena can be investigated using time domain proton nuclear magnetic resonance (TD- ^H-NMR). When a chemical solution is placed in a inhomogeneous magnetic field, the difference between the various components of the global susceptibility tensor of the dissolved molecules and the water are responsible for the great local magnetic field gradients. Thus, the diffusion of the protons through these local inhomogeneities leads to excess in transversal relaxation {T2) detected by TD-NMR technique. For the chemical systems, the local gradient of the magnetic field can be stray and we say that average of gradients connected to the local inhomogeneities of the field, become important quantities for the NMR measurements. Also, the shape of the distribution function for the local magnetic field gradient may be an important measure of process near the macromolecule if we can distinguish between them.

3.

Experimental Procedures The D-, L-glucose, D-, L-ribose, D-, L-tryptophan and D-, L-asparagine were all acquired from Sigma. Solution were made in distilled, deionized water and adjusted to pH with phosphate buffer. The time domain NMR measurements (TD- ^H-NMR) were performed on a 25 MHz ^H-NMR AREMI 78 pulse spectrometer (manufactured by the Institute of Physics and Nuclear Engineering, Bucharest-Magurele, Romania) equipped with an electromagnet (110 mm poles diameter) (corresponding to the magnetic field ^o==0.589T). The overall stability of the magnetic field is lO'^T/min and the magnetic field uniformity is lO'^T/cm in sample volume (1 cm^). For an accurate determination, we use a quadrature phase sensitive detector. The duration of the nil pulse is less than 2^s. For dynamical measurements we use a combination of gradients on OX and OZ axes. This gradient g(^x, gy, gz) is given by definition of the main magnetic field B: ^ ^^o+gx-^^gr^^ gz= ^^, gy =0, g^= g^ -k\n(x). The value of go is 1.8 G/cm and 0.8 G/cm. The value of gz is 0.2 G/cm with orientation depends on the orientation of main magnetic field through sample. All measurements were carried out at 25±0.1 "^C. For CPMG sequence, we use various echo times (0.8, 1.2, 1.6, 2, 2.4, 3.2, 4, 5, 6.4, 8, 11.2, 14.2, 18, 22.4,32, 40, 50, 64, 80 ms) corresponding to the various experimental points (16000 at 300). The repetition delay (RD) was set to 15 sec and the enhancement was 6 (36 scans with S/N ratio~70 dB). By quantitative analyses, we establish that all the protons in the sample (water protons, OH protons and aliphatic protons) contribute to the NMR signal and the observed value of T2 is characteristic to the overall protons in the sample. The sequence consists of an initial pulse of 7c/2, followed by a train of TC pulses, each of them separated by an echo time, ts from the previous pulse n. The echoes are formed between the 71 pulses.

314

Progress in Biological Chirality

4. Current Equation of Spin Diffusion in Presence of External Magnetic Field Gradient We shall focus on the dependence of the decreasing signal of the echo time, IE. In the condition of free diffusion, the below equation shows a strong dependence on ts. For the heterogeneous sample, the internal magnetic field gradients have a large distribution of values and therefore the measured diffiisivity in the sample does not correspond to the true diffusion coefficient. These two effects, the internal magnetic field gradients and the various values of the diffusivity of the wall surface lead to the difficulty of interpretation of the relaxation and diffusion model near the macromolecules. In this work we have used CPMG pulse sequence in order to investigate diffusion effects in presence of spatially variation of main magnetic field, g. In the conditions of free diffusion, the signal starts to decrease exponentially accordingly to the well-known expression [8]:

where / is A:-echo time (t^ktE), Z)is the diffusion constant, g is the gradient of the magnetic field, ;rgyromagnetic constant and T2 is the intrinsic spin-spin relaxation time. We shall focus on the dependence of the decreasing signal as a function of the echo time (E. Same strong dependence on IE is present when chemical exchange between sites was occur. In the condition of fast exchange between two sites the Carver and Richards model [25] for transversal relaxation rates in such system, modified by Ishima and Torchia [30] give us reason to apply the constant gradient approximation. This is based on the similar dependence on tE, and the influence of chemical exchange in the dynamical process represented by constant gradient g is only a fluctuation of this. How can the above observation be generalized? In the heterogeneous systems, the magnetic field variations are even more complicated than those described by a constant gradient. In the absence of an utilizable theory of the inhomogeneous gradients, the practical conclusion of utilizing some local gradient g, as an average of specific local gradients, is obvious. Therefore, the damping of the echo, due to the diffusion of the spins in the molecular fields' gradients can be modelled as a distribution of local gradient/fgj of the sample [9]:

M{t)^M,\

f{g)e''^^ '^

^dg

5.

Mathematical Method We use mathematical algorithm presented in our work [22] in order to extract the probability density function/(^gj for magnetic field gradient in sample [29]. The distribution function is normalized by relation 1 = \f{g)dg.

The intrinsic spin-spin relaxation time T21

0

was calculated after the extrapolation of the curve T2^(tE) at very short tE (100p.s). This expression has a major inconvenience: it is very complicated in order to estimate simultaneously the difHision coefficient D, the distribution flinction/fgj and the local gradient g. We can introduce the constant value for D= 2.5 lO'^mVs corresponding to the pure water at

Different Internal Gradients for L and D Homochiral Solutions in Inhomogenous Magnetic Fields

315

25 °C, in order to reduce the mathematical complexity and to obtain a better understanding. 00

After extraction process, we can calculate the average g = \gfig)^g

Thus, the difference

0

between D and L solution, in averages, become: Sgj^^ = So ~ SL ^^ ^^^ ^^^ section we use this difference in order to characterize the link between origin of sample and the chirality. 6.

Results and Discussion Besides Bo inhomogeneity, the NMR linewidth is determined by the spin-spin relaxation time T2: Avi/2=(7cT2)'\ We use the inhomogeneity of B in order to differentiate the samples. The particular inhomogeneity of main magnetic field interact with chirality of the sample and thus the difference in distribution function/(gj, represented by g, appear. The origin of this interaction may be the molecular chirality gradient, defined as response of the chiral molecule at external magnetic field gradient. This molecular vector has same absolute value and an opposite orientation for the L and D enantiomers. In Figure 1, 2, 3, 4 we present the probability density function in presence of two experimental gradient A) ga=l.S G/cm and B) g(f=O.S G/cm. It is obvious that value of this gradient influences the shape of this function. The average value of internal gradient of the sample is presented in Table 1. The value of go gradient is critical because our first set of investigation reveals that greater value of 6G/cm and less that 0.2G/cm destroys the difference between D and L enantiomers. As you see in Figure 1 A, 2A, 3 A and 4A, the most interesting further investigation is centred by the investigation of the shape of the probability density Table 1

go

Glucose Ribose Tryptophan Asparagine 1

G/cm

G/cm

go=1.8 G/cm

g<)=0.8 G/cm

G/cm

G/cm

7.30 5.94 2.02 4.91

2.21 2.87 1.91 1.99

6.91 6.21 2.35 3.98

1.74 3.42 2.12 1.79

0.39 -0.27 -0.33 0.93

0.47 -0.55 -0.21 0.2

1

A

— — D(+) tTTptophan L(-) bTptpphan

J

^DL

gL

m

1

g (G/cm)

g (G/cm)

Figure 1. The probability density function f(g) for: A) D, L solution of tryptophan (pH 7.2, 12.5 mg/ml) at go=1.8G/cm and B) go=0.8G/cm

316

Progress in Biological Chirality

g(G/cm)

g (G/cm)

Figure 2. The probability density function f(g) for: A) D, L solution of asparagine (pH 7.2, 30 mg/ml) at go=1.8G/cm and B) go=0.8G/cm

D(-) ribose L(+) nbose

B

A '

Tg) 0.8

-

06

-

D(-) nbose L(+) ribose

-

J -

1\

0.4

0.3

II II

rh-^A 1"'

.

-

^"^^.^ g (G/cm)

g (G/cm)

Figure 3. The probability density function f(g) for: A) D, L solution of ribose (pH 7.2, 208 mg/ml) at go= L8G/cm and B) g
fCg) • D(+) ^ucose • L(-)^cose

g (G/cm)

m

— D(+) ^ c o s e " ' L(-) ^ c o s e

g (G/cm)

Figure 4. The probability density function f(g) for: A) D, L solution of glucose (pH 7.2, 208mg/ml) at go=1.8G/cm and B) g(rO.SG/cm

DifiFerent Internal Gradients for L and D Homochiral Solutions in Inhomogenous Magnetic Fields

317

function in order to establish the particular peak of process implied in the dephasing of the spins (diffusion, chemical exchange, susceptibility difference). The application of the peak decomposition algorithm to the distribution functions may be allows us to obtain correlations between proton populations around terminal group of chiral molecule and spin diffusion in the neighbourhood magnetic gradients. The difference between the averages ^^ and gi^^oi ~ SD~ SL^ become the expression of chirality in connection with all dynamical processes of water protons with solute protons. For example, at ribose, Sgj^^^h negative. We interpret <^^^'*'^as being connected with the common process of optical rotation of the polarized light, and we can for glucose, naturally connected to the specific rotation {+) or (-) and indicates, through differences of internal gradients, the molecular chirality. This connection between sign of Sg^j^ may be interpreted like physical condition for magnetochiral phenomena. In case of amino acids the connection indicated above is anti-correlated with the optical rotation and our further investigation will point out this fact. To explain the observed effects we propose a possible mechanism to do this. First, the water molecule diffuses and exchanges in echo time fe, second, it collides with solute molecule and loss of spin phase coherence (the phase difference increase) and finally resuhs the summation of the phase differences ACO^^ITTA V, expressed by g. The link between g and A v is give by statistical average over entire molecule: Av=^ yg Xs, where Xs is average value of the molecular structural dimension. Next, we try to answer to the two main questions: The possible observed magnetochiral phenomenon is an effect of parity violating phenomenon? The structure of magnetic field gradient g, near to molecule, is a very complex function and we can insert in this structure a term from distortion of molecular magnetic field generated by small weak force (Z-force) between electron and nucleus. When the electron and the nucleus interact via a Zo exchange, the parity violating weak force can be viewed as producing a slightly chiral orbit, or a spiral. Thus, the electrons in helical orbits also produce rotation, like rotation of linearly polarized light by right-handed sugar molecules. This rotation is very tiny (on the order of a microradian) because of small amount of Z-force (6 order of magnitude weaker than the Coulomb force), but is large enough to see it with current experimental techniques. The difference of average magnetic gradient between the enantiomers Sg^^^ also leads to an energy difference between the states of the enantiomers in the chosen solvent, by means of the relation: ^^L = ^ ^ ^ ^hy Sg^j^X^. With we choose a reasonable value for A^^ 10' -10' ^m and if we equalize the AEDL with PEVD (parity violating energy difference -10" -10" kT per molecule), results that the theoretical value for Sgj^^ between the L and D enantiomers is on the order of a 10"^-10"^ G/cm, close enough to the experimental values shows in Table 1. The most important remark is that: the energy difference between the states of the enantiomers in solution, in the presence of the magnetic field gradient, can generate differences in the chemical reactivity and overall stability, and therefore results a process of chiral selection.

318

Progress in Biological Chiralitv

7.

Conclusions The CPMG spin - spin relaxation curves are used to extract the distribution function of the internal gradient between the solvent and the hydrated solute. For the enantiomers of glucose, ribose, asparagine and tryptophan, the distribution functions are slightly different in shape and we can use the average value of internal gradient for characterize the sample. The difference between these calculated average values named by us 3g^^ is connected to the molecular chirality, and especially if we consider them from the point of view of the specific rotation, these measurements have a unitary character. We have proposed a possible linkage between the parity violation phenomenon and magnetochirality. This is based on the accumulation of the weak energy converted in differences of internal gradient between enantiomers, during relaxation process. The difference in energy between the states of the enantiomers in solution may be source of the homochiralities of life.

8

References

[1] A.J. MacDeimott and G.E. Tranter, Circumstellar Habitable Zones: Proceedings of the First International Conference (Ed. L.E. Doyle) Travis House Publications, 1996, pp. 364-370. [2] W.A. Bonner, Orig. Life Evol Biosphere 25 (1995) 175-190. [3] L. Pasteur, Rev. Scientifique 71 (1894) 1. [4] G. Melcher, J. Mol Evol. 3 (1974) 121-140. [5] W.A. Bonner, Orig life Evol. Biosphere 21 (1991) 59-111. [6] W.A. Bonner, Orig Life Evol. Biosphere 20 (1990) 1-13. [7] L.D. Barron, Chem. Soc. Rev. 15 (1986) 189-223. [8] L.D. Banon, Mol. Phys. 43 (1981) 1395-1406. [9J G. Melcher, J. Mol. Evol. 3 (1974) 121-140. [10] A.L. Barra, J.B. Robert and L. Wiesenfeld, BioSystems 20 (1987) 57-61. [11] R.A. HegsUom. D.W. Rein and P.G.H. Sandars, J. Chem. Phys. 73 (1980) 2329-2341. [12] A.J. MacDermott and G.E. Tranter, Croatica ChemicaActa 62 (1989) 165-187. [13] Y. Yamagata, J. Iheor. Biol. 11 (1966)495-498. [14] D.K. Kondepudi, BioSystems 20 (1987) 75-83. [15] J.A. Pople, W.G. Schneider and H.J. Bernstein, High Resolution Nuclear Magnetic Resonance, McGrawHill, New York, 1959, Chapter 16. [16] M.I. Foreman, Nucl. Magn. Reson. 1 (1972) 295-320. [17] R.R. Shoup, H.T. Miles and E.D. Becker, Biochem. Biophys. Res. Commun. 23 (1966) 194-201. [18] L. Katz, and S. Penman, J. Mol. Biol. 15 (1966) 220-231. [19] AD. Buckingham, T. Schaefer and W.G. Schneider, J. Chem. Phys. 32 (1960) 1227-1233. [20] T.L. James, Nuclear Magnetic Resonance in Biochemistry. Principles and Applications, Academic Press, New York, 1975, pp. 56. [21] B.A. Goodman and J.B. ^j^ynQV,Adv. Inorg. Chem. Radiochem. 13 (1970) 135-362. [22] Gy. Steinbrecher, R. Scorei, V.M. Cimpoiasu and I. Petrisor, J. Magn. Reson. 146 (2000) 321-334. [23] A.L. Barra, J.B. Robert and L. Wiesenfeld, Phys. Lett. A 115 (1986) 443-447. [24] W.H. Pirkle and D.J. Hoover, Topics in Stereochemistry (Eds. N.C. Allinger, E.C. Eliel and S.H. Wiler) Wiley, New York, 1982, vol. 13, pp. 263-231. [25] J.P. Carver and R.E. Richard, J. Magn. Reson. 6 (1972) 89-105. [26] G.L.J.A. Rikken and E. Raupach, Phys. Rev. E 58 (1998) 5081-5084. [27] G.C. Borgia, R.J.S. Brown and P. Fantazzini, Phys Rev. E 5\ (1995) 2104-2114. [28] Wenqing Wang, Fan Bai and Zhi Liang, http://arxiv.org/abs/physics/0211099. [29] V.M. Cimpoiasu and R. Scorei, The classical Hausdorf momentum problem applied to the LRP-NMR measurements; stable reconstruction of the T2 distribution and magnetic susceptibility difference

Different Internal Gradients for L and D Homochiral Solutions in Inhomogenous Magnetic Fields

319

distribution. In: Magnetic Resonance in Food Science: Latest Developments (Eds. P.S. Belton, A.M. Gil, G. A. Webb and D. Rutiedge) 3 (2003) 77-84. ISBN 0 85404 886 [30] R. Ishima and D.A. Torchia, J. BiomoL NMR 15 (1999) 369-372.

Progress in Biological Chirality G. Palyi, C. Zucchi and L. Caglioti (Editors) © 2004 Elsevier Ltd. Allrightsreserved.

Chapter 26 Tryptophanase Activity on D-Tryptophan Akihiko Shimada,^'* Noriko Fujii,^ and Takeshi Saito^ ^Institute ofApplied Biochemistry, University of Tsukuba, Tsukuba, Ibaraki 305-8572, Japan Research Reactor Institute, Kyoto University, Kumatori, Osaka 590-0494, Japan shimada@ipe. tsukuba.ac.jp

1.

Introduction The problem of chiral breaking in biological amino acids has intrigued scientists since Louis Pasteur first discovered molecular chirality some one hundred and fifty years ago. By the time catalytically active primitive proteins were present, the switch to exclusive use of Lamino acids as building blocks might already have been made. Accordingly, the origins of chiral breaking have been actively discussed as abiotic chemical or physical processes [1], a general consensus has yet to be reached, thus indicating that not all of origins of homochirality are involved in their. It is well known that enzyme stereospecificity produces and supports today's asymmetric biological world. Therefore, if a mechanism for chiral homogeneity arose via abiotic processes in a primitive racemic environment, the mechanism might have been incorporated into early polypeptides, whose descendants may be traced to present enzymes. Early enzymes might have much possessed lower stereospecificity than extant enzymes. The higher stereospecificity of extant enzymes likely reflects the historical consequences of enzyme evolution. In this context, a mechanism for extant enzyme stereospecificity is considered the key to solving the puzzle of chiral homogeneity [2]. Nevertheless, there has been little related discussion because enzyme stereospecificity is believed to be absolutely immutable. Of course, it is natural to beheve this because if the balance were unstable, life would cease to function when exclusive selection of L-amino acids became impossible. However, recent progressive D-amino acid analytic techniques have demonstrated that D-amino acids are more widely distributed in the biological world than was previously believed. For example, D-alanine, D-aspartic acid and D-serine are present in eubacteria, archaea, bivalve, and crayfish as well as in the human eye and brain [3-6], In fact, D-Ala in frog's skin is a secreted neuropeptide that has physiological activity [7]. Such widespread distribution of D-amino acids reinforces our doubts regarding the exclusivity of enzyme stereospecificity. This paper reports the flexibility of enzyme stereospecificity. Tryptophanase is an enzyme with very strict stereospecificity, cleaving L-tryptophan but not D-tryptophan, and was used in this study. We attempted to change the stereospecificity, and demonstrated tryptophanase activity toward D-tryptophan in the presence of ammonium phosphate. Circular dichroism (CD) and florescence spectroscopy was used to monitor the conformational changes inducing

322

Progress in Biological Chirality

activity toward D-tryptophan in ammonium phosphate solution.

2.

Enzyme Activity Tryptophanase is a pyridoxal 5'-phosphate (PLP)-dependent enzyme that catalyzes the hydrolytic P-elimination of L-tryptophan (L-Trp) to indole and ammonium pyruvate as shown below, consisting of four identical 52 kDa momomers. Each monomer contains one molecule TPase

Q-

+ NH4^

of PLP, which forms an aldimine bond with a lysine residue. TPase is one of enzymes with the highest stereospecificity for optical isomers. In this reason, it has long been believed that tryptophanase is absolutely inactive to D-tryptophan (D-Trp). However, tryptophanase (TPase) activity toward D-Trp was firstly discovered in highly concentrated ammonium sulfate solution in 1992 by Shimada [8], We subsequently found that some of ammonium phosphate salts had higher activity than ammonium sulfate [9]. Ammonium phosphate has three species; ammonium dihydrogenphosphate (MAP), diammoniumhydrogen phosphate (DAP) and triammonium phosphate (TAP). Both DAP and TAP allow TPase to exhibit activity, while MAP denatures and deactivates TPase. This activity is reversibly restored to the native state upon removing DAP or TAP from reaction mixture, dependent on ammonium

Reaction temperature CC) Figure 1. Reaction temperature-dependent TPase activity for D-Trp in 50% saturation MAP, DAP and TAP solutions. TPase activity was determkied by absorbance at X=570nm. D MAP, A DAP, O TAP, x No ammonium phosphate

Tryptophanase Activity on D-Tryptophan 323 phosphate concentration, PLP and reaction temperature. Increasing temperature sharply increases TPase activity (Fig. 1), and higher activity levels are seen in TAP solution than in DAP solution. Optimal reaction conditions are 1.1 mM PLP, 50% saturation concentration TAP and 55°C. As tryptophan tends to racemize in alkaline environments, it is important to note that it is more labile than other amino acids [10, 11]. Therefore, we investigated whether D-Trp was chemically racemized. Racemization is also potentially possible by reprotonation of the quinonoid intermediate from the face opposite the PLP complex at the active site of TPase. The quinonoid intermediate formed from D-Trp may alternatively partition between elimination and racemization [12]. The possibility of nonenzymatic or tryptophanasecatalyzed racemization was analyzed, but D-Trp was not chemically or enzymatically racemized, even when exposed to 50% saturation concentration of TAP and DAP and to temperature as high as 55''C. The reaction with D-Trp is therefore caused by the changes in TPase stereospecificity, not by chemical or enzymatic racemization of D-Trp.

3.

Kinetics Kinetic analyses were performed to determine the reaction pathway for TPase activity toward D-Trp. A reaction mixture was prepared by combining the required D-Trp (245-980 |iM) and DAP concentrations (0 ~ 70% saturation) with fixed TPase (0.04 units) and PLP concentrations (380 ^M). The reaction pathway was proposed based on kinetic parameters, which were obtained as shown below (Fig. 2) [13]. KSD-TIP, KDAP or KioAP are the substrate constant for D-Trp, dissociation constant between DAP and TPase, and inhibition constant for DAP, respectively. DAP acts on TPase as activator below 50% saturation, but as inhibitor at saturation concentrations above 50%. Maximal activity toward D-Trp is obtained with a reaction temperature of 55°C in 50% saturation TAP solution including 1.1 mM of PLP and 0.2 |LAM of TPase. Kinetic constants of TPase for L- and D-Trp were determined from reciprocal plots of initial velocities and substrate concentrations. These values were then compared (Table 1). For L-Trp, Vm was 0.66 units/mg and Km 0.33 mM. For D-Trp, Vm was 0.13 units/mg. Because the reaction pathways for the D-Trp degradation reaction are complex, the Michaelis constant cannot easily be determined. Km here is defined as the concentration of D-Trp giving v=Vm/2. Km was determined to be 0.54 mM based on this definition. The ratio of Vm (D-Trp) to Vm (LTrp) was 0.6% and the ratio of catalytic efficiency (Vm/Km) for D-Trp to L-Trp was 0.3% at 37°C. TPase activity toward D-Trp sharply increases with increasing reaction temperature.

TPase

^ '

TPaseD-Trp

KD

TPaseDAP ^=^ TPaseDAPD-Trp — Indole + TPase +DAP KiDAP

KioAP L KS]>Trp

TPase(DAP)ex ==^ TPase(DAP)exD-Trp Figure 2. Reaction pathway for TPase activity toward D-Trp

324

Progress in Biological Chirality Table 1.1ECinetic constants of TPase with L- and D-Trp Substrate

L-Trp

(fimoles/min/mg) Km (mM) Catalytic efficiency (Vm/Km)

0.66 0.33 2

Vm

D-Trp/37 °C

0.004 0.67 0.006

D-Trp/55°C

0.13 0.54 0.24

Increasing rate of the acti\aty on D-Trp between 37 and 55°C 32.5 0.8 40

When reaction temperature is 55°C, Vm for D-Trp increases to 20% of Vm for L-Trp, and catalytic efficiency for D-Trp increases to 12% of that for L-Trp. Our reaction conditions thus produce substantially increased activity toward D-Trp 4.

Conformational Changes It is possible to speculate on the mechanism behind the activity for D-Trp based on the obtained kinetic data. DAP or TAP is believed to influence TPase by bringing D-Trp closer to the enzyme. It is not yet clear how these compounds actually act on TPase, but TPase must undergo a conformational change, thus allowing it to recognize D-Trp. These conformational changes would be very small, judging from the partially reversible changes indicated by the kinetic data. However, direct evidence for these conformational changes is limited. Therefore, other analytical methods must be applied in order to detect these subtle changes. In this study, circular dichroism (CD) and fluorescence spectroscopy were used to detect conformational changes induced by ammonium phosphate. The CD spectra of TPase were monitored in MAP, DAP and TAP solutions of 50%) saturation. Although actual measurement of CD spectra among TAP, DAP and MAP were carried out in the wavelength range of 200-600 nm, the region that gave rise to difference to the CD spectrum in phosphate buffer (PB) was 200-250 nm. In Fig. 3, the CD spectra in the 200-250 nm region were shown. It seems likely that spectral changes in the 200-250 nm region arise from structural changes of a- helix and P-sheet in TPase because there peaks are in 205-225 nm region. The CD spectra in TAP and DAP were almost close or similar to the CD spectrum in PB, whereas that in MAP had a little difference. MAP causes irreversible full loss of the catalytic activity, which probably gives larger structural changes to TPase in MAP than in TAP or DAP. It is reported elsewhere that TPase exhibits CD spectrum with maxima at 420 and 337 nm which is due to internal PLP-lysine aldimine bond, and rupture of the PLPlysine aldimine bond is accompanied by a markedly decrease of the 420 nm CD peak [14]. The CD spectral changes in Fig. 3 were relatively very small as compared with the spectral changes of the 420 nm CD peak. Additionally, no difference was detected in 250-600 nm region. These indicate that ammonium phosphates give an influence not on the internal PLPlysine aldimine bond, but on the steric structure, i.e. a- helical or P-sheet structure, of TPase. Conformational changes induced by TAP and DAP are so subtle that they have difficulty in being detected by CD. If we want to obtain conformational changes large enough to be detected by CD, severe treatment such as denature of TPase will be necessary. Measurement of the CD spectra reveals that the appearance of TPase activity on D-Trp is caused by subtle conformational changes of TPase, but CD is not very efficient for detecting them. Fluorescence spectrophotometry is also useful for detecting small conformational changes

Tryptophanase Activity on D-Tryptophan

325

KKP 6

^"^ "o B

^

3

S o

CCl

-a .Br

^pq ^o

K % u w

0

iii'i

C

-3

v'

" u^.

///

VV\

Ky

V' \

-6

'V

V^',

^''A7

V ^ ijL=XiS'^:j2-a<J>=-'^

-9

200

210

220

230

240

250

Wavwlength (nm)

(

Figure 3. Circular dichroism spectra of TPase in PB, MAP, DAP and TAP solutions. PB , 0.1 M phosphate buffer of pH 8.3), MAP ( , ammonium dihydrogenphosphate), DAP ( , diammoniumhydrogen phosphate), TAP ( , triammonium phosphate)

in TPase. Bis-ANS (4,4'-dianilino-l,r-binaphthalene-5,5'-disulfonic acid, 8 = 23 xio^ cm'^M" at 395 nm) was used to probe the changes in hydrophobicity. The excitation wavelength was 395 nm and emission scanning was conducted between 420 and 650 nm [15]. Bis-ANS-bound TPase was prepared and fluorescence intensity was measured at liTC in 50% saturation solutions of MAP, DAP, TAP and PB. Bis-ANS-TPase fluorescence intensity was highest in MAP solution followed by DAP and TAP solutions (Fig. 4). Bis-ANS-bound TPase fluorescence intensity differed slightly between the TAP and DAP solutions, but was markedly different in the MAP solution. No shift in maximum wavelength was seen among >> 2.0

4{K)

450

500

550

600

650

Wavelength (nm) Figure 4. Bis-ANSfluorescencespectra of TPase in MAP, DAP, TAP and PB solutions. PB (—), MAP ( \ DAP ( ) and TAP (—)

326

Progress in Biological Chirality

the solutions. The effect of the increasing exposure to hydrophobic amino acids is known to be closely related to aggregation of proteins [16], and any TPase unfolding undoubtedly leads to the loss of all activity. If increasing exposure to hydrophobic amino acids is connected with the unfolding, then decreasing exposure to hydrophobic amino acids is more important for increasing activity toward D-Trp. As shown in Fig. 4, fluorescence intensity is the largest in the MAP solution, indicating that the conformational changes are too great to activate TPase. Fluorescence intensity was very low in the DAP solution and even lower in the TAP solution. Exposure to smaller amount of hydrophobic amino acids gives TPase higher activity, and thus TAP is ideal for providing TPase with decreased exposure to hydrophobic amino acids, thereby decreasing random coiling or unfolding. Only very subtle conformational changes allow activity toward D-Trp. Fluorescence results indicate that very subtle changes prevent TPase from unfolding and are significant in its expression of activity toward D-Trp.

5.

Conclusion The presence of PLP, DAP, TAP and high temperatures are key factors in enhancing activity toward D-Trp by inducing subtle but vital changes to bring TPase in closer contact with D-Trp. TPase activity toward D-Trp reaches 20% of its activity toward L-Trp when TPase reacts with D-Trp under optimal conditions. The present reaction system is appropriate to further investigate the mechanisms behind TPase stereospecificity. Once this mechanism is clarified, we will be closer to solving the problems associated with the homochiral origins in early metabolism. Our resuks indicate that TPase stereospecificity may be artificially changed. The influence of DAP and TAP on TPase was spectroscopically examined using CD and fluorescence spectrophotometry. The data show that these conditions induce very subtle conformational changes that enable closer contact between TPase and D-Trp.

6. Acknowledgment This work was supported in part by the Project Research Program of the Kyoto University Research Reactor Institute. 7.

References

[1] W.A. Bonner, Ohg. Life Evol. Biosphere 25 (1995) 175-190. [2] A. Shimada, H. Shishido, H. Kogure, I. Nakamura, N. Fuji and M. Akaboshi, Tryptophanase-catalyzed Dtiyptophan degradation Reaction and its Significance for Chiral Homogeneity. In: The Role ofRadiation in the Origin and Evolution of Life (Eds. M. Akaboshi, N. Fuji and R. Navarro Gonzalez) Kyoto University Press, Kyoto, 2000, pp. 243-257. [3] Y. Nagata, T. Fujikawa, K. Kawaguchi, Y. Fukumori and T. Yanianaka, Biochim. Biophys. Acta 1379 (1998)76-82. [4] H. Abe and N. Yoshikawa, Viva Origino 30 (2002) 221-228. [5] A. Hashimoto, T. Oka and T. Nishikawa, J. Neurosci. 1 (1995) 1657-1663. 16] N. Fujii, Y. Momose, M. Yamasaki, T. Yamagaki, H. Nakanishi, T. Uemura, M. Takita and N. Ishii, Biochem. Biophys. Res. Commun. 239 (1997) 918-923. [7] L.H. Lazams and M. Attila, Progr. Neurobiol. 41 (1993) 473-507. [8] A. Shimada, and I. Nakamura, Viva Origino 20 (1992) 147-161. [9] A. Shimada, H. Shishido and I. Nakamura, Amino Acids 11 (1996) 83-89. [10] S.L. Zhao and Y.-M. Liu, Electrophoresis 22 (2001) 2769-2774. [11] Y. Yokoyama, H. Hikawa, Y. Murakami, J. Chem. Soc. Perkin Trans.-l (2000) 1431-1434.

Tryptophanase Activity on D-Tryptophan 327 [12] R.S. PhilUps, B. Sundaraiaju and N.G. Faleev, J. Am. Chem. Soc. Ill (2000) 1008-1014. [13] A. Shimada, H. Shishido and I. Nal^amura, Viva Origino 23 (1995) 169-178. [14] T. Erez, G.Y. Gdalevslty, Y.M. Torchinsky, R.S. Phillips and A.H. Parola, Biochim. Biophys. Acta 1384 (1998) 365-372. [15] G. Musci and L.J. Berliner, Biochemistry 24 (1985) 3852-3856. [16] T. Kortemme, M. Ramirez Alvarado and L. Serrano, Science 281 (1998) 253-256.

Progress in Biological Chirality G. Palyi, C. Zucchi and L. Caglioti (Editors) © 2004 Elsevier Ltd. All rights reserved.

Chapter 27 Deviation from Physical Identity Between D- and L-tyrosine Meir Shinitzky,^* Avshalom C. Elitzur^ and David W. Deamer'' ''Department of Biological Chemistry, The Weizmann Institute of Science, 76100 Rehovot, Israel meir. shinitzky@weizmann. ac. il ^ Unit for Interdisciplinary Studies, Bar-Ilan University, 52900 Ramat-Gan, Israel ""Department of Chemistry and Biochemistry, University of California, Santa Cruz, CA 95064, USA

1.

Introduction Symmetry under reflection - nature's indifference to the "right" and "left" directions (P symmetry) - is one of physics' most profound and well-established principles, and is unlikely to be violated by a biological process. Nevertheless, experimental studies and intensive survey of the literature have convinced us that such a violation may occur in natural amino acids. In one such case, tyrosine, we have detected significant deviations from macroscopic identity between the L- and D-enantiomers [1]. We reported that a supersaturated solution of Ltyrosine in water forms crystalline precipitates much more slowly than the analogous solution of D-tyrosine. Furthermore, supersaturated solutions of DL-tyrosine in water formed a precipitate of predominantly D- and DL-tyrosine, resulting in excess L-tyrosine in the aqueous layer. We suggested that the minute energy difference between these enantiomers, in combination with the highly cooperative process of crystallization of tyrosine, could account for this unexpected observation. Here we summarize a series of additional experiments that seem to support this unexpected phenomenon. We discuss its possible origins and suggest a mechanism for violation of chiral identity which, in practice, can lead to chiral enhancement. 2.

Kinetics versus Equilibrium Tyrosine is the least soluble of the natural amino acids in water, even less soluble than the hydrophobic phenylalanine. This surprising property indicates that the strength of intermolecular tyrosine-tyrosine coupling in crystals is greater than the interaction of tyrosine with water. In such coupling the phenol hydroxyl group forms hydrogen bonds to neighboring hydroxyl or carboxylate groups. In aqueous supersaturated solutions tyrosine coupling produces the nanocrystalline unit cells which then serve as nucleation centers for the formation of macroscopic crystals. Crystal formation in supersaturated solutions is a typical cooperative process. As such, the rate of formation of unit cells is the rate-limiting step in the overall process of crystal build-

330

Progress in Biological Chirality

up. We have carried out extensive determinations of the rates of tyrosine crystal separation from supersaturated aqueous solutions. In all of our experiments, D- and L-tyrosine, of a declared purity of 99% or higher from various vendors, were the starting samples. Each sample was recrystallized 3 times from double distilled water under argon and in the dark. No trace of impurities was detected by HPLC, optical rotation, and chromatography. Furthermore, analysis of chiral contamination in the samples of purified D- and L-tyrosine was carried out for us by Prof Schurig from the university of Tubingen. Both samples were found to contain less than 0.5% of impurity of the opposite enantiomer, which is the limit of sensitivity. During crystallization, supersaturated solutions approach equilibrium, at which point the rates of crystal formation and crystal solubilization are equal. In this respect, a critical aspect of the state of equilibrium is whether it is identical for D- and L-tyrosine. In our published experiments [1] we allowed crystallization to proceed for up to 7 days. By asymptotic extrapolation we estimated a significant difference in the apparent saturated concentration of D- and L-tyrosine. It is possible to question this conclusion by arguing that the systems have not yet reached a true state of equilibrium. In fact, we observed marked differences in the rates of crystal formation of D- and L-tyrosine, independent of their purity, and occasionally lack of reproducible results with a single sample. However, in experiments to be reported here, the accumulated data of the precipitation profiles confirmed our initial conclusion that D-tyrosine precipitates faster than L-tyrosine, and reaches a lower saturated concentration. The differences between D- and L-tyrosine in the kinetics of crystal formation and approach to equilibrium imply that a similar difference should be expected when tyrosine crystals are equilibrated in pure water. In an initial set of experiments we introduced tyrosine crystals into water and determined the concentration by absorbance (A275 nm). However, the reproducibility was unsatisfactory. Because it was necessary to detect relatively small differences between the saturated concentrations of D- and L-tyrosine, we improved the method as follows. D- and L-tyrosine of 99% initial purity were obtained from Merck, Sigma, Aldrich and Fluka and recrystallized 3 times (see above). D- or L-tyrosine solutions (10 mM, initial temperature 100^ C) were mixed with a large excess of pre-washed silica gel powder (Merck). The mixture was allowed to evaporate to complete dryness at 60^C under nitrogen with constant stirring. Knowing the average size of the silica particles, we estimated that the layer of tyrosine coating the particles was less than a micron thick. Samples of 200 mg tyrosine coated silica were placed in Eppendorf tubes containing 1ml distilled water and then shaken at 18^C for various periods. After centrifiigation, the tyrosine content in the aqueous layer was determined at A275 nm. An apparent equilibrium was reached after 2-3 minutes and remained constant thereafter. The results obtained with the independent samples of D- and L-tyrosine, each tested 8-10 times in quadruplicate, are presented in Fig. 1. The range of saturated concentrations for both enantiomers was rather large. Nevertheless, the average point of saturation of D-tyrosine was significantly lower, p<0.001, than that of L-tyrosine. Furthermore, the average values of the saturated concentrations (see Fig.l) approximately matched the extrapolated values obtained in the precipitation experiments published previously [1]. Autocatalysis in the process of crystallization is clearly reflected in supersaturated solutions of achiral substances which possess chiral crystals, as in the case of sodium chlorate [2, 3]. The above processes clearly emphasize the potency of chiral enhancement in racemic mixtures and warrants closer attention with respect to the origin of homochirality.

Deviation from Physical Identity Between D- and L-tyrosine

0.42

r 2.9 2.8

0.40

2.7

^

0.38

2.6 2.58 m M

?

331

0.361

S

I

I- 2.5 4 2.48 m M

O

r 2.4

0.341

2.3 P<0.001

0.30

s C/3

0.32

h 2.2 L-Tyrosine

D-TjTOsine 2.1

Figure 1. Saturated concentrations of D- and L-tyrosine in water at 18°C after coated on silica particles (see text for experimental details). Each point represents the average of three rephcate measurements and each sample was assayed separately 8-10 times. D-tyrosine is approximately 4% less soluble than L-tyrosine (p<0.001) Two additional experimental approaches provided clues to a possible explanation for the observed solubility differences of D- and L-tyrosine. When D2O was used as a solvent in analogous precipitation experiments, the difference in the kinetics of D- and L-tyrosine crystallization was markedly reduced. Figures 2 and 3 show precipitation profiles of D- and L-tyrosine in H2O vs. D2O. The marked difference in the precipitation profiles suggested a mechanism that will be proposed in the Interpretation section below. The second approach concerns the role of the methylene group of tyrosine. Most natural amino acids can be considered to be derivatives of glycine, often with a methylene bridge between the chiral center and the specific side chain. (Exceptions to this rule include alanine, valine, threonine, isoleucine and proline.) We therefore tested the crystallization kinetics of D- and L-(/?ara-hydroxyphenyl)glycine, which is equivalent to tyrosine lacking the methylene bridge. We first noticed that this compound is much more soluble in water than tyrosine, and more importantly that the precipitation profiles and saturation concentrations of D- and Lenantiomers were virtually identical (not shown). The fact that the D- and L-enantiomers are equally soluble in water suggested that the methylene bridge contributes to the differential solubility of D- and L-tyrosine, as will be discussed later.

3.

Evidence from Circular Dichroism The solubility difference between D- and L-tyrosine described above predicts that there should be a detectable structural difference between the enantiomers in the hydrated state. We carried out an extensive set of spectroscopic determinations to test this possibility, including spectrophotometric and fluorimetric titrations, excimer formation, self-quenching of fluorescence and fluorescence quenching by iodide. In all of these experiments the results for

332

Progress in Biological Chirality

100H

• \ji

L-Tyrosine in H2O

0 L-Tyrosine in D2O

80-

C/3

a

60-

rv ^>s^^^B

.

40-

.

\^o

• o o

0

20-

0-

1

—1

20

40

60

1—'

1—I

180

80

1

\

200

220

1 — '

240

Time (h)

Figure 2. Precipitation profiles of L-tyrosine in H2O and D2O at 18 °C starting with supersaturated solutions of 10 mM and followed up to 6 days. Note the large differences between H2O and D2O, as well as the differences between L and D-tyrosine in H2O, as shown in Fig. 3

100-1 1

• D-Tyrosine in H2O o

^ D-Tyrosine in D2O

80-

\

9 60-

40-

•

VL

•

0 0

0

0

0

20-

0-

•~i

20

—1

1

—

1

80

1 —

100

1—1

180

1

1

200

220

1 — '

240

Time (h)

Figure 3. Precipitation profiles of D-tyrosine in H2O and D2O under identical conditions to those described in Fig. 2

DeviationfromPhysical Identity Between D- and L-tyrosine

333

D-tyrosine were identical to those obtained for L-tyrosine. However, in circular dichroism spectra (CD) we could detect small differences in the CD spectra of D- and L-tyrosine in H2O and D2O (Figs. 4 and 5). The differences originate from a Cotton effect, which corresponds to their absorption spectra and consists of a peak and a trough in opposite directions. The mere difference between H2O and D2O CD spectra provides a clear indication that the subtle molecular angles in tyrosine are affected by the solvation layer, which complies with the data presented in Figs. 2 and 3.

-•- D-tyrosine in H2O »»L-tyrosine in H2O

3

^^^^^^n 300

Wavelength, nm Figure 4. CD spectra of D- and L-tyrosine in H2O (3 mM in 1 mm cell). Each point is an average of 20 scans

-3 J

Wavelength, nm Figure 5. CD spectra of D- and L-tyrosine in D2O determined in the same manner as the spectra in H2O (Fig. 4)

334 Progress in Biological Chiralit>

7.4 7.2 7.0

O 6.8

• D-tyrosine in H2O A L-tyrosine in H2O

j J1

a U 6.6 6.4

r

•

4

^

i

t

•

i ^

1

6.2

T i

I

P<0.05 6.0

Figure 6. The difference in millidegrees between the peaks and troughs at 275 and 240 nm corrected for absorbance. (CD/A275) recorded in a series of samples of D- and L-tyrosine in H2O (3 mM in 1 mm cell). The difference is of moderate significance (P<0.05) Careful examination of the CD spectra revealed a small but consistent difference in the angular gap between the peak at 275 nm and the trough at 240 nm (or vice versa) between Dand L-tyrosine in H2O. Since this difference is dependent on concentration, we corrected it for the measured absorbance, which varied by less than 5%. Evaluation of the alternative parameter of the ratio of these values, which in principle should be independent of small differences in absorbance, was hampered by variations in base line between different sets of experiments. The accumulated results in H2O are presented in Fig. 6. Despite the apparent spread in CD/absorbance values, the averages indicate a small but significant difference between D- and L-tyrosine in H2O (P<0.05). The correlation between CD spectra and molecular three-dimensional structure is in most cases quite complex. For the case of tyrosine the only conclusion that can be drawn is that the orientation of the phenol ring with respect to the chiral center is affected by the hydration layer and therefore changes to some extent when H2O is replaced by D2O. As indicated by the data presented in Fig. 6, the effect on the orientation of the phenol ring by H2O is slightly different in D- and L-tyrosine. At this stage it is difficult to assess the magnitude of this effect, but such a difference is consistent with the different rates of crystal nucleation implied by the crystallization profiles presented in Figs. 2 and 3. 4.

Violation of Chiral Identity in Other Amino Acids There have been two other reports of physical differences between D- and L-enantiomers of natural amino acids in their condensed phase. Wang et al. [4-6] have tested and compared a series of physical parameters, including magnetic susceptibility, Raman spectral shifts, NMR spectra and optical rotation around an experimental phase transition in single crystals of D-

DeviationfromPhysical Identity Between D- and L-tyrosine 335 and L-alanine. Clear differences in all three parameters were detected, and it was proposed that these were consistent with predictions made by Salam [7] that certain effects related to the weak force asymmetry may extend to macroscopic scales. A second report concerns chiral effects observed in amphiphilic serine derivatives. Nstearoyl D- or L-serine are classical micelle forming compounds in water. In organic solvents they exhibit CD spectra similar in shape and sign to those of the analogous N-acetyl serines. In water, however, they form tight micelles, and a strong absolute CD band is displayed with an opposite sign to that appearing in organic solvents. Under these conditions the micellar CD band of N-stearoyl D-serine was consistently stronger, by about 50%, than that of N-stearoyl L-serine [8]. This unusual CD band indicated that in these micelles chiral surfaces are formed which presumably are constituted by spines of serine moieties bridged by hydrogen bonds. The difference in magnitude of the CD bands between the enantiomers strongly suggested that in the micelles of N-stearoyl D-serine the tightness of the chiral surface is greater than in Nstearoyl L-serine [8]. In recent publications Cook and coworkers [9, 10] discovered octameric structures of L serine which could, in principle, serve as templates for chiral selection, a possible step in the genesis of homochirality. It is tempting to speculate that, similar to the results obtained for micelles containing headgroups of D- and L serine octamers of D-serine may be tighter than those of L-serine and their selectivity could therefore differ. This suggests another feasible route for chiral enhancement from racemic mixtures, and it would be interesting to look for selectivity differences between L- and D-serine octamers.

5.

Proposed Interpretation: a Two-Stage Violation of Chiral Identity In the following we propose that violation of chiral identity can operate on both microscopic and macroscopic levels. Their combined effect presumably underlies the macroscopic differences in solubility which were outlined above for D- and L-tyrosine. The unified theory which links the weak force with electromagnetism implied that the fiindamental asymmetry of the nuclear weak force can be extended to the outer electron cloud [7]. It was thus calculated that when the electron cloud is confined to an asymmetric center the energy exerted by the nuclear weak force on L-enantiomers is different by a very small magnitude (-10-17 e.v.) than that exerted on the D-enantiomer [7, 11-14]. Recently, more advanced calculations have confirmed the magnitude of such differences [15, 16]. It should be mentioned, however, that relativistic considerations may question this effect [17]. This parityviolating energy difference could be amplified when associated with an autocatalytic process. In racemic mixtures which undergo a simultaneous autocatalytic process the tiny difference in the parity-violating energy of the enantiomers can be amplified to a level which can be expressed in macroscopic parameters. The process of crystallization is typical autocatalytic in this regard. It seems feasible that the apparent macroscopic difference in the rates of crystallization between D- and L-tyrosine originates in part by the parity-violating energy difference between these enantiomers. The exceptionally high cooperativity in their crystalHzation process could partially account for the extension of this difference into a macroscopic and measurable scale. However, it is still questionable whether this route alone can account for the observations related to the significant difference in solubility between D- and L-tyrosine in H2O at equilibrium (Figure 1)

336

Progress in Biological Chirality

and the marked reduction of the difference in their crystalHzation kinetics in H2O when determined in D2O (see Figs. 2 and 3). There is one additional source of physical asymmetry that has been overlooked in this context, yet may exert a considerable discrimination between enantiomers in their process of crystallization. A unique interpretation along this line [18] is outlined in the following. Water molecules can exist in two distinct forms with respect to the spin orientation of their hydrogen atoms [19], either parallel ("orZ/yc?-water", 0-H2O) or antiparallel ("/7am-water", p-RjO). oH2O has 3 degenerate states and therefore these two forms exist not in a 1:1 but in a 3:1 ratio. One could argue, however, that the rapid proton exchange in bulk water should rapidly mix these populations so that from the thermodynamic viewpoint, water should be considered homogeneous. An unexpected observation related to this consideration was recently published by Tikhonov and Volkov [20] who demonstrated that 0-H2O and /7-H2O molecules preserve their configuration for orders of magnitude longer (-10^) than expected on the basis of the proton shuttling rate. Fractions rich in 0-H2O or in/7-H20 could be thus separated by simple macroscopic fractionation. An immediate conclusion corresponding to this finding is that proton exchange between 0-H2O and /7-H2O is actually forbidden, an intriguing question for future quantum mechanical attention. We propose that 0-H2O, due to its magnetic field corresponding to the net spin orientation, has some preference toward one of the tyrosine enantiomers. In this respect, the greater solubility of L-tyrosine in water (Figs. 1-3) could be explained by preferred solvation around the asymmetric center by 0-H2O which, as stated above is 3 times more abundant than/7-H20. It is of interest to note that the methylene group - CH2 - of tyrosine and other amino acids, also has ortho ^nd para configurations [18]. As mentioned above, (/7-hydroxyphenyl)glycine lacks the methylene group, and its enantiomers have identical solubility profiles, in direct contrast to the differences noted in tyrosine. The majority of natural amino acids have the methylene bridge, and one may speculate that interactions between 0-H2O and the methylene bridge of amino acids during chemical evolution on the early Earth may have guided the process that led to the first forms of chiral life. In summary, as suggested by others, chiral enhancement in amino acids can be triggered by the parity-violating energy difference between the D- and L-enantiomers when combined an exceptionally high cooperativity. Crystallization of tyrosine may provide an experimental model for such autocatalytic processes. We propose here an additional augmenting route for chiral enhancement, which is based on different selectivity of 0-H2O and /7-H2O in the hydration layer around the asymmetric carbon of D- and L-amino acids [18]. The results presented for D- and L-tyrosine in H2O vs. D2O are consistent with this hypothesis. 6.

References

[1] M. Shinitzky, F. Nudelman, Y. Barda, R. Haimovitz, E. Chen and D.W. Deamer, Unexpected differences between D- and L-tyrosine lead to chiral enhancement inracemicmixtures. Origins Life Evol. Biosphere 32 (2002) 285-297. [2] D.K. Kondepudi, R. Kaufman and N. Singh, Science 250 (1990) 975-977. [3] D.K. Kondepudi, K.L. Bullock, J. A. Digits, J.K. Hall and J. Miller, Kinetics of chiral synmietry breaking in crystallization. J. ylw. Chem. Soc 115 (1993) 10211-10216. [4] W.Q. Wang, X.G. Sheng, H.F. Jin, J. Wu, B. Yin, J. Li, Z.X. Zhao, H.S. Yang, F.M. Lou, Z.Z. Zhuang, G. Y. Yu, L. Shi and Z. Chen, Susceptibility behaviour and specific heat anomaly in single crystals of alanine and valine../. Biol. Phvs. 22 (1996) 65-71.

Deviation from Physical Identity Between D- and L-tyrosine [5] 16] [7] 18] [9]

[10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20]

337

W.Q. Wang, F. Yi, Y. Ni, Z. Zhao, Y. Jin and Y. Tang, Parity violation of electroweak force in phase transitions of single crystals of D- and L-alanine and valine. J. Biol. Phys. 26 (2000) 51-68. W.Q. Wang, NMR and parity violation: low-temperature dependence in ^H CRAMPS and ^^C CP/MAS ssNMR spectra of alanine enantiomer. Biophys. Chem. 103 (2003) 289-298. A. Salam, The role of chirality in the origin of life. J. Mol Evol 33 (1991) 105-113. M. Shinitzky and R. Haimovitz, Chiral surfaces in micelles of N-palmitoyl or N-stearoyl L- (or D-) serine. J.Am. Chem. Soc. 115 (1993) 12545-12549. K.J. Koch, F.C. Gozzo, S.C. Nanita, Z. Takats, M.N. Eberlin and R.G. Cooks, Chiral transmission between amino acids: chirally selective amino acid substitution in the serine octamer as a possible step in homochiiogmQsis. Angew, Chem. Int. Ed 41 (2002) 1721-1724. Z. Takats, S.C. Nanita and R.G. Cooks, Serine octamer reactions: indicators of prebiotic relevance. Angew. Chem. Int. Ed 42 (2003) 3521-3523. D.K. Kondepudi and G. W. Nelson, Weak neutral currents and the origin of biomolecular chirahty. Nature 314(1985)438-441. R.A. Hegstrom, D.W. Rein and P.G.H. Sandars, Calculation of the parity nonconserving energy difference between mirror-image molecules. J. Chem. Phys. 73 (1980) 2329-2341. S.F. Mason and G.E. Tranter, The parity violating energy differences between enantiomeric molecules. Mol. Phys. 53 (1984) 1091-1111. A. Szabo-Nagy and L. Keszthelyi, Determination of the parity violation energy difference between enantiomers. Proc. Natl. Acad Sci. USA 96 (1999) 4252-4255. A. Bakasov, T.K. Ha and M. Quack, yl^ initio calculation of molecular energies including parity violating interactions. J. Chem. Phys. 109 (1998) 7263-7285. R. Zanasi and P. Lazzeretti, On the stabihzation of natural L-enantiomers of a-amino acids via parityviolating effects. Chem. Phys. Lett. 286 (1998) I^O-IM. R. Wesendrup, J.K. Laerdahl, R.N. Compton and P. Schwerdtfeger, Biomolecular homochirahty and electroweak interactions. 1. The Yamagata hypothesis. J. Phys. Chem. A 107 (2003) 6668-6673. D.W. Deamer and M. Shinitzky, Spontaneous chiral enhancement in racemic mixtures of amino acids. Submitted for publication (2003). A. Farcas, Orthohydrogen, Parahydrogen and Heavy Hydrogen, The University press, Cambridge UK, 1935. V.I. Tikhonov and A. A. Volkov, Separation of water into its ortho- and para- isomers. Science 296 (2002) 2363.

Progress in Biological Chirality G. Palyi, C. Zucchi and L. Caglioti (Editors) © 2004 Elsevier Ltd. Allrightsreserved.

Chapter 28 Occurrence of D-Amino Acids in Food Livia Simon Sarkadi Department of Biochemistry and Food Technology, Budapest University of Technology and Economics, Miiegyetem rkp. 3, H-Ull Budapest, Hungary sarkadi@mail. bme.hu

1.

Introduction Knowledge of chemical composition of food is important to the health, well-being, and safety of the consumer. The food industry wishes to provide healthy foods and therefore needs to know what the optimal composition of these foods should be. Consumers would like to receive reliable information to base decisions about the merits and/or risks associated with exaggerated intakes of specific foods or components. The importance of amino acids is widely recognised in various fields, particularly in the field of nutrition. Their usefulness has two aspects: the first is in the role of supplementary nutrients. The second area of usefulness is for their physiological or pharmacological functions. Organisms vary widely in their ability to synthesise amino acids. Many bacteria and most plants can synthesise all of their nitrogenous metabolites. Mammals are able to biosynthesise about half of the amino acids in quantities needed for growth and for maintenance of normal nitrogen balance. Those amino acids that the body cannot manufacture are called essential amino acids. Essential amino acids (arginine, histidine, lysine, leucine, phenylalanine, isoleucine, tryptophan, valine, threonine, and methionine) must be obtained in the diet, and all within the same meal, since free amino acids are not stored. In both humans and rats Arg and His are classified as essential amino acids, but nutritional studies show that they are required in the diet only during the growth of juveniles. The amino acid score has been used to evaluate the nutritive value of food proteins. Dietary proteins are considered to belong to two different groups, depending on the amino acids they provide. Complete proteins, which constitute the first group, contain ample amounts of all of the essential amino acids. These proteins are found in meat, fish, poultry, cheese, eggs, and milk. Incomplete proteins, which constitute the second group, contain only some of the essential amino acids. These proteins are found in a variety of foods, including grains, legumes, and leafy green vegetables. The closest amino acid composition of a certain protein is given when the animal eating this protein, the higher the nutritional quality of the protein. The amino acids contained in foods are either in the free form or in the protein-form

340

Progress in Biological Chirality

COO®

L-amino acid

COO®

D-amino acid

Figure 1. Structures of L- and D- amino acids

(proteinogenic). Since the amino acid composition of individual proteins of foods is fixed hereditarily and the protein-form usually accounts for a major part of the total amino acids, the amino acid composition of a food is not greatly variable in principle because of the limited variation of protein composition of the food. However, the free amino acid composition changes considerably according to the condition of production, preservation and processing of the food. With the exception of glycine, all the genetically coded proteinogenic amino acids are optically active, and of identical chirality at the a-carbon atom. Chirality was first observed in 1848 by L. Pasteur, who noticed that the crystals of tartaric acid that formed around the neck of a wine bottle were different from those that appeared at the bottom of the bottle. The two forms were later identified as mirror images of each other. Such compounds, which are known as optical isomers or enantiomers, have the same chemical structure but opposite handedness [1]. The D and L notation is used for monosaccharide and for amino acids (Figure 1). An Lamino acid if the amino group is on the left and a D-amino acid if it is on the right. Unlike monosaccharide, where the D isomer is the one found in nature, most amino acids found in nature (especially in higher organisms) have the L configuration. D-amino acids have been found in nature as constituents of bacterial cell walls and several peptide antibiotics. Biological effects of enantiomers are often very different. \n pharmaceutical products, it is common to find that only one enantiomer is the effective therapeutic agent, while the second enantiomer has no activity, or fiinctions in a very different manner. In food, the chirality of a compound may determine its nutritional value, sensory properties or biological activity. Free L-amino acids are largely responsible for the taste of many foods. The taste of most of L-amino acids can be classified as either bitter or sweet. Among thefi"eeamino acids Arg, Phe, His, Val, Trp, He, Leu, Met are responsible for bitter tastes, Ala, Gly, Pro, Ser, Thr, Lys for sweet tastes, and Glu, Asp for sour tastes. Many pleasant odours of cooked food are due to the interaction products of amino acids and sugars (Maillard reaction, the interaction of sugars and amino acid). The occurrence and the role of D-amino acids in food are less known. However, there are few reviews concerning general considerations on D-amino acids including chemistry, nutrition and microbiological aspects [2, 3]. This paper is intended to summarise the most important data concerning D-amino acid content in a variety of foods.

Occurrence of D-Amino Acids in Food

341

2.

Processed Food Fermentation is an ancient preservation method to increase the self-life of various products. Microorganisms can be used to transform raw food into cheese, wine, beer, vinegar, sauerkraut, and other alcoholic products. Bacteria of the gQUQm Acetobacter, Bifidobacterium, Brevibacterium, Lactobacillus, Micrococcus, Propionibacterium, and Streptococcus, which are used as so-called starter cultures for the large-scale production of fermented foods and beverages in food biotechnology, have been investigated for the chirality of their amino acids [4]. In all bacteria D-amino acids were detected; those in the highest relative amounts were DAla and D-Asp (occurring in all bacteria) and, in several cases, D-Glu. Lower, but significant amounts of other D-amino acids such as D-Ser, D-Pro, D-Val, D-Thr. D-Ile, D-Leu, D-Met, D-Phe, D-Tyr, D-Om, and D-Lys were also detected in certain bacteria. 2.1 Dairy products All dairy products result from similar manufacturing techniques. Several authors have reported that raw milk from ruminants (cows, goats, and sheep) contained D-Glu, D-Ala, DAsp, D-Lys, and D-Ser [5, 6, 7]. D-amino acid contents of raw milk [8] analysed shortly after milking are shown in Table 1. D-amino acids found in cow's milk may be originated from the digestion and autolysis of rumen bacteria cell wall proteins (peptidoglycan). The enantiomeric purity of D-amino acids was monitored in milk samples stored at 4 "^C over a period of several weeks. D-Ala was the only free amino acid to show variation in D/(D+L) % ratio. The presence of D-Ala at a percentage grater than 4% can be considered as an indicator of milk contamination by psychrotropic bacteria [6]. Casal et al. [9] found the following D-amino acids in UTH cow's milk: D-Pro (0.36 mg/100 ml), D-Leu (0.04 mg/100 ml) and D-Ala (0.01 mg/100 ml). In pasteurised fermented milk with Bifibobacterium bifidum high concentrations of D-Ala, D-Asp, D-Lys, D-His, DGlu, D-Leu, D-Val and D-Phe were identified (Table 1). The effect of mastitis (an infection of the udder caused by bacteria) on D-amino acid content in milk has been studied by Csapo et al. [10]. Contents of D-amino acids were highly associated with the California Mastitis Test score and udder inflammation. Higher concentrations of free D-amino acids in bulk tank milk could be due to inclusion of foremilk and milk from cows having subclinical mastitis. The use of various microorganisms to fermented milk into a variety of dairy products significantly enhances their D-amino acid content. Bacteria, which ferment sugars to produce Table 1. D-amino acids in milk and yoghurt produced by different fermentation process Raw milk [8] UTH milk [9] Fermented milk Yoghurt [11] UTH yoghurt (%) (%) [9](%) (%) [9](%) Glu Ala Asp Ser Lys His Leu Val Phe Tyr

14.2 4.6 4.0 3.9

3.4

10.0 45.9 43.9

17.4

36.3 26.8 9.6 5.0 0.9

15.6 56.8 28.8

44.6 60.2 9.6 29.3 7.7 12.1 2.9 15.8 11.1

342 Progress in Biological Chirality Table 2, Occurrence of D-amino acids in cheese Asp(%) Glu(%) Ala(%) Leu(%) Ardrahan [13] Camembert [13] Danish blue [13] Emmental [13] Gouda[13] Parmesan [13] Mozzarella [13] Cantal[14] Cheese [8]

27.2 14.0 31.1 26.8 28.5 20.8 28.9 7.1 6.0

13.1 14.8 20.2 26.6 22.7 10.6 24.0 15.4 6.5

27.1 16.1 42.4 45.6 38.4 37.3 33.3 32.4 44.0

Ser(%)

2.0 4.4

lactic acid can be classified into distinct groups depending on whether they produce D or L lactic acid. Streptococcus thermophilus, a homofermentative organism, produces only L-lactic acid, while Lactobacillus bulgaricus, a heterofermentative bacterium, produces only the Denantiomer. Yoghurt are produced by a special starter in which the two major bacteria (Streptococcus thermophilus, Lactobacillus bulgaricus) are present. Jin et al. [11] published high concentrations of D-Ala, D-Asp, and D-Glu in yoghurt. Casal et al. [9] had similar results analysing low-fat UTH yoghurt (Table 1). Cheese is one of the oldest human foods. All cheese result from a lactic acid fermentation of milk. Proteolysis of casein during cheese ripening leads to increase of free amino acids. The amount and type of amino acids formed mainly depends on the type of microorganisms present and also on the nature of the commodity as well as the conditions of fermentation. Significant amounts of D-Ala, D-Glu, and D-Asp have been found in ripened cheeses such as Appenzeller, Emmental and Gruyere cheese and Gouda [12, 13]. In French pressed cheese (Cantal) investigated by Bruckner et al. [14] the sum of D-amino acids of 435 mg/kg was determined. The main D-amino acids found were D-Asp, D-Glu, DAla and D-Leu (Table 2). Moretti et al. [8] published average D-amino acid concentrations of cheese (Table 2). They established significant correlation between bacterial count and D-amino acid concentration in cheese-making milk, but not in cheese samples. Marchelli et al. [15] studied the presence of free D-amino acids in Parmigiano-Reggiano and Grana Padano cheese. They detected significant increase in D-amino acids during ripening related to the age and type of cheese. Free L- and D-amino acids were determined in Pecorino cheese manufactured using raw milk and heat-treated milk with added native starter cultures [16]. D-amino acids found in milk were D-Glu, D-Asp, D-Lys, D-Ala and D-Tyr. The D-amino acids represented in ripened cheeses were D-Glu (22.18 mg/lOOg), and D-Lys (17.18 mg/lOOg). Innocente and Palla [17] have reported that D-Ala percentage of total D- plus L-isomers correlated with ripening levels of cheese. All cheese samples contained also D-Lys, D-Om, and D-Tyr. Results concluded that D-Ala ratios may be reliable markers for ripening of Montasio cheese, and basic (D-Lys, D-Orn) and aromatic (D-Tyr) amino acids may be of use for quality assessment of the cheese. Parmigiano Reggiano, San Dalmazio and Santa Rita cheeses produced by different

Occurrence of D-Amino Acids in Food

343

technology (classical and biological) between 1997-2001 were analysed. The major free AAs detected in cheeses were glutamic acid (which represented 16.7% of the total free AAs in cheese), proline (10.8 %) lysine (10.0 %), leucine (8.0%), valine (7.0%) and serine (6.7%). The results of D-amino acid determination showed that D-amino acid content of older cheeses were approximately twice as much high as of younger ones, and the profile showed a slight decrease related to the sample distance to the centre of the cheese. On the basis of D/Laspartic acid ratios one can make better differentiation among cheeses. Parmigiano Reggiano had the highest D/L-aspartic acid ratio followed by Santa Rita and Dalmazio had the lowest D/L-aspartic acid ratio (Simon-Sarkadi and Csapo, 2003; unpublished data). 2.2 Fermented fish product Besides the fermentation of dairy products, a variety of meat and fish can be fermented. Free D-amino acids were determined in 60 fermented fish sauces from various outlets in Southeast and East Asia [18]. D-Ala, D-Asp, D-Glu was found as the major D-amino acids. D-Ala seems to be used as a molecular marker of bacterial activity in fermented fish products. 2.3 Crustaceans Seafood quality has traditionally been discussed and estimated based on degree of spoilage of the raw material or the product. This very perishable commodity requires immediate refrigeration or chilling to retard spoilage. Okuma et al. [19] determined distributions of free D-amino acids in tissues of crustaceans. Crustaceans studied were: kuruma prawn {Penaeus japonicus), crayfish {Procambarus clarkii), rock lobster (Jasus lalandii), snow crab {Chionoecetes opilio), woolly-handed crab (Eriocheir japonicus), marsh crab {Holometopus dehaani), and shore crab (Hemigrapsuspenicillatus). D-Ala (3.2 - 16.8 ^mol/g) was the most abundant and widely distributed among every crustaceans species studied. The second most abundant D-amino acid was D-Arg (2 fimol/g). D-Asp and D-Glu were present only in tissues of prawns, crayfish, and lobsters. 2.4 Wine Wine is the results of a complex interaction between yeast, grape must and physical conditions. Wines are known to contain many biologically active compounds. The amounts and compositions of these compounds depend on the type of grapes and their degree of ripeness, climate and soil of the viticultural area, as well as vinification techniques. Amino acid composition is great importance in wine production. Amino acids represent the main source of nitrogen both yeast and malolactic bacteria during wine fermentation and also serve as substrate for volatile aroma compounds in wine. The presence and the relative amounts of 1-3% of D-Asp, D-Glu, and D-Ala are typical in white, red and rose wines. In French rose wine 2.5 mg/1 total amounts of D-amino acids were detected [14]. The relative amounts of D amino acids in French rose wine are shown in Table 3. Portuguese wines (Roupeiro white) bottled during the period 1978-1989 were analysed for their D-amino acid contents. The D-amino acids detected were D-Ala, D-Val, D-Thr, D-Leu, D-Ser, D-Asn, D-Met, D-Phe, and D-Gln. The results showed that D-amino acids may be differences may be due to the different fermentation processes. Red wine is produced from whole grapes whereas white wine is produced from grape juice without skins. It means red wine is liable to be contaminated by D-enantiomer producing microorganisms.

344 Progress in Biological Chirality Table 3. Occurrence of D-amino acids in wines

Asp Glu Ala Tyr Asn Arg Val Leu

French rose [14]

Madeira wine [14]

Wine [11]

(%)

(%)

(%)

5.8 2.4 2.9 3.2

20.5 20.4 17.8

10.1 4.0

17.8 5.3 3.2 6.7

In fortified wines (Madeira, Sherry, Port) where alcohol is added to the vigorously fermenting must in order to stop the fermentation process, followed by curing and maturing of wines under typical conditions for several years, generally high amounts of D-Asp, D-Glu, and D-Ala are found [14]. The relative amounts of D amino acids in Madeira wine are shown in Table 3. 2.5 Beer Beer is defined as an alcoholic beveragefi"omstarch-containing raw materials serving as a source for maltose and glucose, which are fermented by brewers yeast. Although barley malt is the most important cereal, wheat, wheat malt, com, rice and millet are also used as starchcontaining adjuncts or extenders and sources for fermentable sugars. Since beer is generally consumed in greater amounts than wine, it has been suggested that beer might be more of a hazard to the consumer concerning D-amino acid content. The amounts and pattern of D-amino acids in beer is influenced by raw materials used in the brewing process, malting technology, wort processing, and the conditions during fermentation. In beer production alcoholic fermentation takes place by the action of selected strains of the yeast. Beers are classified in two groups, top- and bottom-fermented based on whether yeast floats or sinks by the end of fermentation. Besides Saccharomyces cerevisiae (top fermenting) and Saccharomyces carlsbergensis (bottom fermenting), various wild yeasts together with lactic acid bacteria are involved in the brewing process of special local beers. Erbe and Bruckner [21] established that raw materials contribute to a minor (grains, malt) or negligible (hops) extent to the D-amino acid content of beer. Possibly, D-amino acids are also formed in the course of the non-enzymatic browning or Maillard reaction on heating malt, mash and wort. Beers produced from different raw materials showed very deviating amino acid patterns and ratios of enantiomers. The highest amounts of D-amino acids were determined in wheat beer (Table 4). In comparison to special beers, the common lagers, ales and Pilsener beers contain lower absolute and relative amounts of D-amino acids [21]. Very high amounts of D-Pro were detected in Berliner weisse, as well as high relative amounts of D-Ala, D-Asp and D-Glu in all beer (Table 5). Bottled German Export beer produced with the aid of the bottom-fermenting yeast Saccharomyces carlsbergensis contained 11 mg/1 total amount of D-amino acids. Relative amounts of D-Asp, D-Glu and D-Ala were 29.0 %, 5.9 %, and 3.9 %, respectively [14].

Occurrence of D-Amino Acids in Food

345

Table 4. D-amino acids in beers produced by different fermentation process [21] Bottom fermented beers Top fermented beers Lambic Altbierand Wheat beers Pilsenerand Black beers Strong beers beers ales lagers •-amino acids (mgA) •-amino acids

6.4-41.6

6.5-41.6

11.3-96.3

8.3-29.8

8.9-24.9

13.2-50.8

1.0-5.3

0.9-2.7

1.3-11.3

0.7-1.3

1.2-1.8

0.9-1.2

(%) Table 5. D-amino acid content in beers [211 Berliner Weisse (mgA) Ala Pro Asp Phe Glu Tyr Lys S

(%)

22.5 51.9 5.2 1.3 6.6

30.1 21.1 14.7 3.2 10.9

1.3 88.8

4.0 11.7

Pilsener (mgA)

(%)

Pale ale (mg/1)

Bavarian wheat beer

(%)

(mg/1)

Strong beer

(%)

(mgyl)

(%) 2.6 0.3 9.7 3.2 4.8

1.2

3.1

2.4 1.5 4.3 1.5 2.5

3.2 0.4 13.5 7.5 5.7

2.6 0.8 3.1 1.2 4.6 1.7

3.0 0.2 8.5 1.4 6.2 3.0

6.7 2.0 7.3 3.4 5.9

0.7

12.2

1.5

24.0

1.3

25.3

4.2 1.7 3.9

1.5 0.3 8.3

3.1

12.9

A study by Ekborg-Ott and Armstrong [22] on enantiomeric composition of three amino acids in 25 different beers showed that proHne had the highest average absolute concentration and the lowest percentage of the D-enantiomer in most samples. In some cases the relative amounts of D-Phe and D-Leu exceeded 10% of the individual amino acids. The enantiomeric composition of the amino acids in different samples did not vary as extensively as the absolute concentrations. Casal et al. [9] identified 20.9 % D-Asp, 18.3 % D-Glu, 12.9 % D-Trp, 12.0 % D-His, 10.8 % D-Phe, 9.5 % D-Val, 8.3 % D-Met and 5.3 % D-Ile in black beer The D-enantiomers of Ala, Asp and Glu found in beers are considered as chemical markers for microbial activity [14, 23]. It can be concluded that knowledge of amino acids concentrations and enantiomeric compositions appears to be usefiil in characterizing specific beers and brewing processes. The quantities, relative amounts and pattern of amino acid enantiomers can serve as indicatives for authenticity and quality of beers. 2.6

Vinegar Vinegar can be made from any plant which contains enough sugar to ferment into the alcohol needed to make acetic acid. At the beginning of the fermentation the yeast Saccharomyces cerevisiae converts sugars into ethanol, which is then fermented to acetic acid by species of Acetobacter. As with wine, the formation of D-amino acids is due to the action of microbial enzymes. The pattern and the enantiomeric distribution of amino acids in vinegars were strongly dependent on the origin of the vinegar and raw materials (grape must, wine, cider, and spirits) used for their production [23].

346

Progress in Biological Chirality Table 6. Occurrence of D-amino acids in balsamic vinegars [23] Age

Ala Val Pro Leu Ser Asp Phe Glu Tyr Om Lys

5 years

10 years

12 years

25 years

(mg/1)

(%)

(mgA)

(%)

(mgA)

(%)

(mg/1)

(%)

9.5 0.5 18.9 2.1 1.5 5.4 1.3 3.4 0.5 2.1 0.4 45.6

9.1 1.4 5.8 4.7 4.5 12.1 4.0 11.8 3.2 5.8 1.7 5.0

15.3 0.5 30.8 2.2 1.5 8.7 1.9 6.7 0.9 7.1 1.1 76.7

7.7 0.7 4.5 2.9 3.8 9.9 3.7 7.8 3.0 8.0 2.1 4.3

4.4 0 30.6 1.7 1.5 5.6 2.2 4.0 1.5 5.3 0.8 57.6

5.0 0 3.2 2.3 2.0 7.8 2.8 5.1 2.0 6.3 1.3 3.0

40.2 0.9 191.9 9.6 8.1 41.6 16.8 21.0 9.1 14.7 3.6 361.3

20.2 1.3 26.1 15.2 25.0 35.3 24.3 27.9 20.0 20.3 12.2 20.2

Table 7. Occurrence of D-amino acids in wine and cider vinegars [23] Red wine vinegars White wine vinegars Cider vinegar

Ala Val Pro Leu Ser Asp Phe Glu Om Lys E

(mg/1)

(%)

(mg/1)

(%)

(mgA)

(%)

1.5-5.8

8.6-19.2

0.3-0.6

6.4-10.0

0.7 - 0.9

22.0-25.9

0.4-3.8

0.3-5.2

0.4-1.9

5.6-11.4

0.2-0.4

3.4-5.4

0.7-1.5 0.5-0.6 0.1-0.2 4.7-11.2

6.7-15.2 1.7-3.3 0.5-2.1 1.7-3.9

0.2-0.7

3.4-5.0

0.2-0.5 0.1-0.1 0.3-0.4

6.7-10.0 6.7-7.1 5.9-6.7

0.7- 1.9

0.5-0.5

1.6-1.7

3.7-4.6

Italy's famous balsamic vinegar is specially aged wine vinegar. Balsamic vinegars (aceto balsamico tradizionale di Modena) contained highest amounts of D-amino acids (46-361 mg/1). All balsamic vinegars had significant amounts of D-Leu, D-Phe, D-Tyr D-Orn, and DLys and most of them also contained D-Val and D-Ser. The amounts of D-Pro and D-Ala significantly increased in the course of maturation (Table 6). In a related study [9] showed, that balsamic vinegar from Modena (6*") had 37.7 % D-Glu, 28.8 % D-Lys, 28.3 % D-Ala, 16.2 % D-Asp, 16.1 % D-Met, 5.2 % D-Leu, 1.8 % D-His and 1.1%D-Phe. In comparison to balsamic vinegars, red wine vinegars contained much lower amounts of D-amino acids (Table 7). White wine vinegars contained only D-Ala, D-Asp and D-Glu and no D-Pro was detected. Cider vinegars (apple-based product) contained only low amounts of D-amino acids, the relative amounts was the highest for D-Ala (22.0 - 25.9 %). Vinegars made from sherry contained the highest absolute amounts of D-Ala, D-Asp and D-Glu of all the wine vinegars examined (Table 8). In spirit vinegars only few D-amino acids were detected. Bruckner etal [14] also investigated vinegar made from Sherry wine (Old Sherry wine vinegar, from Jerez de la Frontera). Relative amounts of 25.4 % D-Asp, 13.4 % D-Glu, 4.7 %

Occurrence of D-Amino Acids in Food

347

Table 8. Occurrence of D-amino acids in Sherry and Spirit vinegars [23] Sherry vinegar Spirit vinegars (m^)

(%)

(mgn)

(%)

Ala

6.0-6.8

22.2-30.3

0.1-0.1

10.0-25.0

Val Pro Leu Ser Asp Phe Glu Om Lys S

0.3-0.6 4.3-4.4 1.0-1.2 0.2-0.3 2.2-2.3 0.1-0.2 2.7-3.1 0.3-0.6 0.1-0.1 17.7-19.1

1.8-5.3 2.1-4.8 4.5-8.8 2.1-2.6 10.3-20.2 1.2-1.3 13.7-15.3 0.7-1.7 0.4-0.8 4.0-6.7

0.1-0.1

9.1-33.3

0.6-1.4

0.1-0.2

0.8-1.7

0.2-0.4

D-Ser, 24.4 % D-Ala and 9.4 % D-Leu were found. The results with vinegars indicate that the D-Pro content could serve as an indicator for the ageing process and thus for the quality and authenticity control of balsamic vinegars [24, 25]. D-Ala, D-Asp and D-Glu can be used as chemical markers to distinguish among fermented and synthetic vinegars. 2.7 Sauerkraut Sauerkraut has been very popular in many European countries due to its sensorial properties and favourable nutritional value. Sauerkraut or sour cabbage is produced from wilted, shredded white cabbage. Fermentation process can be carried out using either spontaneous fermentation (with relies on the lactic acid bacteria occurring naturally on vegetables) or controlled fermentation (using a starter culture of Lactobacillus species). Among microorganisms contributing to sauerkraut production Leuconostoc mesenteroides is of special importance in initiating the lactic acid fermentation. The next phase is characterised by the activity of homofermentative, no-gas-producing lactic acid bacteria with higher occurrence of Lactobacillus plantarum. The last phase of fermentation is dominated by heterofermentative lactic acid strains like Lactobacillus brevis, Pediococcus and Enterococccus [26]. Bruckner et al. [14] analysed the lactic acid fermented cabbage and cabbage juice. They found several D-amino acids with different quantities (Table 9). The main microorganisms involved in cabbage fermentation are Leuconostoc mesenteroides, Lactobacillus brevis and Lactobacillus plantarum. Table 9. Occurrence of D-amino acids in sauerkraut [14] Fermented cabbage Fermented cabbage (%) juice Qimol/I) Asp 4.7 181 Glu 12.3 20 Ser 2.5 49 Ala 8.2 718 Leu 7.2 59 Lys 4.9 2.4

348

Progress in Biological Chirality

2.8 Bread Bread is one of the most ancient of human foods. In bread making, yeast growth is carried out under aerobic condition. The yeast Saccharomyces exiguous with a Lactobacillus species produces the characteristic acidic flavour and aroma of such breads. Use of lactic acid bacteria and yeast in the fermentation of sourdough introduce D-Ala and D-Glu into the dough. Baking of the dough into bread induces a 44% decrease in the total free D-amino acid content [27]. 2.9 Coffee Coffee and cocoa beans are roasted to produce an optimal quality product. Roasted coffee contained 10-40% of D-Asp, D-Glu, and D-Phe [28]. A correlation among the roasting degree and amino acid racemization has been proposed. The extent of roasting can be calculated by determining the degree of racemization (D/D+L %). Casal et al. [9] analysed Robusta green coffee. They have found 9.3 % D-Lys, 0.9 % DPhe and 0.4 % D-Asp in coffee. 2.10 Honey Honey is an ever more widely used product in diets. Contamination in honey can occur indirectly from products used in agricultural practices through contaminated pollen. On the other hand, direct contamination can also occur through bees. Significant levels of D-amino acids were detected in a number of commercially available honeys. In white fir honey (Table 10) D-Asp, D-Glu, D-Ser, D-Ala, D-Phe, and D-Leu detected [14]. The D/L ratios of Leu, Phe, and Pro could serve as indicators of age, processing, and storage of honey [29]. 2.11 Miscellaneous Yeast extracts are widely used as seasonings in the food industry, so that these are potential sources of D-amino acids in many foodstuffs. In yeast extracts amounts of 2.7 % D-Asp, 2.3 % D-Glu, 4.9 % D-Asn, 1.2 % D-Ser, 2.3 % D-Ala, 1.9 % D-Tyr and 1.7 % D-Phe were determined by Bruckner et al. [14]. They have also found free D-amino acids (D-Ala, D-Val, D-Ser, D-Asn and D-Glu) in baker's yeast, brewer's yeast and wine yeast. The sodium salt of the naturally occurring nonessential amino acid monosodium glutamate is often added at levels of 0.2 - 0.9% to foods to improve flavour and palatability. An extensive survey by Rundlett and Armstrong [30] showed that a variety of processed foods contain significant amounts of D-Glu. The D-Glu for crackers, sauces, vinegars, sauerkraut juices, tomato products and milk products were 2.9 %, 7.9 %, 18 %, 36 %, 1.6 %, 6.2 %, respectively. Table 10. Occurrence of D-amino acids in honey [14] White fir honey (%) Fir honey (%) Asp Glu Ser Ala Leu Phe

5.2 4.1 1.8 3.3 6.9 1.8

3.3

5.7 3.8

Occurrence of D-Amino Acids in Food 349 Food industry often applied gelatines obtained by enzymatic or heat pre-treated raw materials to produce confectionery and canned meat. Liipke and Bruckner [31] determined several D-amino acids in gelatines. The relative amounts were for D-Asp (5.8-34.1 %), D-Glu (2.1-4.5 %), D-Ala (1.4-3.0 %), D-Ser (1.2-5.8 %), D-Leu (1.5-3.4 %), D-Phe (2.0-4.8 %), DPro (1.7-4.3 %), and D-Lys (1.8-3.0 %).

3. Fresh Food Since the presence of D-amino acids in food is due to the microbial action one may suppose that unprocessed and unfermented food do not contain D-amino acids. Nevertheless some study showed that both fruits and vegetable contain significant amounts of D-amino acids. 3. J Fruits and vegetables Fruits (apples, grapes, oranges) and vegetables (cabbage, carrots, garlic, tomatoes) as well as the corresponding juices contain measurable amounts of D-amino acids including D-Ala, D-Arg, D-Asp, and D-Glu [14, 32, 33]. Gandolfi et al. [33] followed the development of D-amino acids in grapefruit juice samples inoculated with bacteria {Lactobacillus plantarum), or yeasts (Sacharomices cerevisiae). Significant amounts of free D-Ala (8-31 mg/1) were found in juices affected by bacterial contamination. D-Ala may be considered as a marker of bacterial contamination occurring before or during juice processing. Jin et al. [11] found D-Glu (1.3 %) in tomato juice, tomato puree (2.5 %), and tomato ketchup (2.1%). The data in Table 11 demonstrate that D-enantiomers of certain amino acids occur in the free state in fruits. Relative amounts were in the low percentage range [34].

4.

Methods for the Determination of D-Amino Acids For qualitative and quantitative determination of D-amino acids in food various analytical methods are available. The various methods differ in sensitivity, selectivity, ease of sample preparation, and speed of determination. Table 11. Occurrence of D-amino acids (DAA |imol/l) and their relative amounts (%) infreshlyfruits [34] Apple Pineapple Watermelon Papaya Mango Yellow passion fruit Ala Val Leu Ser Asp Asn Glu Ghi Arg

DAA

(%)

2.9

2.7

3.4 3.2 14.6 3.4

1.7 0.4 0.7 0.5

DAA

(%)

2.3 2.5

1.1 2.4

3.6 24.7 3.0

0.9 0.8 0.4

DAA

(%)

DAA

(%)

DAA

(%)

DAA

(%)

4.0

0.6

3.6

0.9

8.4

0.4

12.0

0.8

8.8

0.5

3.1 3.1 33.9

0.8 1.4 0.7

3.9

1.5

17.1 16.2

0.6 0.5

1.7 5.5

0.4 0.9

0.4 0.3 0.3

0.5

0.2 0.4

4.9 7.3 6.8

17.1

8.5 12.5

350

Progress in Biological Chirality

The chiral analysis of amino acids usually carried out using a chiral thiol compound to form diastereomeric derivatives, which can be separated by gas chromatograhy (GC), highperformance liquid chromatography (HPLC) or capillary electrophoresis (CE). Although the chromatographic methods are reliable and very sensitive, they are sometimes too slow for the food industry. In routine analysis and quality control there is a need for simple and easily applicable analytical methods. The biosensors could provide good tools for simple determination of enantiomers. Another different determination possibility arises from the combination of biosensors with HPLC or the use of hyphenated techniques (GC-MS, HPLCMS). 4.1 HPLC Although many derivatization reagents have been developed for the separation of DLamino acids by HPLC, only a few reagents are applicable to determine DL-amino acids. The o-phthaldialdehyde (OPA) together with various chiral thiols such as thio sugars [35, 36, 37]. N-acetylpenicillamine [38], D-3-mercapto-2-methylpropionic acid [39], and Nacetylcysteine [40, 41] is suitable for primary amino acids, but secondary amino acids such as Pro are not detected directly without any treatment. Another disadvantage is the relatively low stability of the resuhing isoindole derivatives. One of the frequently used HPLC methods to determine D- and L-amino acids in foods and beverages is the precolumn derivatization with OPA combines with the chiral thiol Nisobutyryl-L(or D)-cysteine. The resulting diastereomeric isoindole derivatives were resolved on an octadecylsilyl stationary phase using a linear gradient formed from sodium acetate buffer (pH 5.95) and methanol-acetonitrile. For detection of the isoindoles, their fluorescence at 445 nm when excited at a wavelength of 230 nm was used [14]. A two-dimensional column liquid chromatographic method was discribed for determination of D- and L-amino acids in food by Merbel et al. [42]. Amino acids were separated initially on an ion-exchange column by gradient elution with a sodium citratesodium chloride buffer. Enantioseparation was by subsequent injection of individual amino acids onto a second column with a chiral crown ether stationary phase. Finally, fluorescence detection is carried out after post-column labelling of the amino acids using OPA-2mercaptoethanol reagent solution. The (+)-/(-)-l-(9-fluorenyl)ethyl chloroformate (FLEC) seems to be excellent for chiral resolution of DL-amino acids, but it has also some difficulties in reproducibility. The reagents R(-)- or S(+)-4-(3-isothiocyanatopyrrolidin-1 -yl)-7-(N,N-dimethylaminosulfonyl)-2,1,3benzoxadiazoles (DBD-PyNCS) react rapidly and quantitatively, under mild condition with both primary and secondary amino acids and give stable derivatives [11]. 4.2 GC In contrast to the procedure used for HPLC, for gas chromatographic determination the amino acids must be converted into volatile derivatives. Due to the wide differences in chemistry of the amino acid side chains it is not easy to achieve reproducible sample reparation and derivatisation. An additional problem in the analysis of amino acid enantiomers is the low thermal stability of the chiral stationary phases, with maximum operating temperatures allowed near 200 ""C [9].

Occurrence of D-Amino Acids in Food

351

The major derivatisation procedures for amino acid analysis by GC based on the preparation of N-(0)-trifluoroacetyl alkyl esters, N-perfluoroacyl alkyl esters and Nalkyloxycarbonyl-trimethylsilyl derivatives are time-consuming [43, 44]. For investigation of D-amino acids in juices of vegetables and fruits by gas chromatography, amino acid enantiomers were converted into their N-(0)pentafluoropropionyl amino acid 2-propyl esters and resolved on a Chirasil-L-Val fused silica capillary column [31]. Recently similar method using GC-MS coupled techniques have been dicrcibed by Bruckner and Westhauser [33]. These methods made possible the determination of D- and L-Pro in analytes. Gas chromatographic quantification of amino acid enantiomers in food by their N-(0,S)ethoxycarbonyl heptafluorobutyl ester derivatives was solved on Chirasil-L-Val column [9] 4.3 CE Boniglia et al [45] developed a capillary electrophoresis method for the chiral separation of D-L Val, He and Leu. The separation of derivatized amino acids with 9-fluorenylmethylchloroformate (FMOC) was performed by micellar electrokinetic capillary chromatography using P-cyclodextrin as chiral selector. A new micellar electrokinetic chromatography-laser-induced fluorescence (MEKC-LIF) method was developed to detect D-amino acids in orange juice. The derivatisation reagent was fluorescein isothiocianate (FITC) [46]. 4.4 Biosensors A novel biosensor electrode system, which aims to provide a rapid means of detecting amino acid enantiomers in food, has been designed based on the action of the enzyme amino acid oxidase. Some advantages as high selectivity and specificity, relative low cost of construction and storage, potential for miniaturization, facility of automation and simple and portable equipment construction for a fast analysis and monitoring in platforms of raw material reception, quality control laboratories or some stage during the food processing. A major drawback of sensors using non-specific amino acid oxidases is their differing sensitivity to different substrates, which implies that determination of the total amino acid content is not possible [47]. Another different determination possibility arises from the combination of biosensors with HPLC followed by electrochemical detection [48, 49]. In this technique the separation efficiencies of HPLC are combined with the specificity of enzymes and the sensitivity of an electrochemical detector. Simple sample pre-treatment is required, isolation and derivatization steps can be omitted. Varadi et al. [50] developed a rapid method for determination of the different form of free amino acids. Two enzymes (L- and D-amino acid oxidase) were immobilized in a thin-layer Plexi-cell on natural protein membrane and the enzyme-cell was built into a FIA system. The hydrogen peroxide generated during the enzymatic reaction was determined by an amperometric detector. Voss and Galensa [49] developed a coupled technique (HPLC with enzyme reactors) for the enantiomeric determination of L- and D-amino acids in different food. After separation on a Li cation-exchange column the amino acids are converted into keto acids and H2O2 under catalytic influence of L- or D-amino acid oxidase and H2O2 is detected amperometrically.

352

Progress in Biological Chirality

5. Conclusion D-amino acids commonly occur in the diet, in particular in fermented foods, as well as in fresh fruits and vegetables. Although lacks of experimental evidence on the risks of D-amino acid intake the production of foods with low amounts of D-amino acids is important to human health. The concentration and the kinds of D-amino acids occurring in foods seem to depend on both the manufacturing process and starter cultures. The carefully selected starter culture and controlled technology may resuh lower D-amino acid amine content in fermented foods. Detection of an unnatural enantiomer, or the ratio of enantiomers can be used to determine authenticity of a product, as a marker for the extent of processing, and may be usefril for assessing food quality. Alternatively, the presence of higher amount of a D-amino acid may indicate bacterial contamination of the product. To gain better knowledge on dietary D-amino acids and to set up a useful database systematic investigation of D-amino acid content of different foods is needed. In order to achieve this goal further development of the recently available methods and as well as new rapid methods for enantioselective analysis are expected in the future.

6. [1 [2 13 14 [5 [6

U [8 [9; [lo; 111 [12 [13 [14 115 [16 [17 [18 [19 [20 [21 [22 [23 [24 [25 [26 [27;

References R. Marchelli, A. Dossena and G. Palla, Trends Food Sci. Technol. 1 (19%) 113-119. J. Zagon, L.I. Dehne and K.W. Bogl, Nutr. Res. 4 (1994) 445-463. M. Friedman, J. Aghc. FoodChem. 47 (1999) 3457-3479. H. Bruckner, D. Becker and M. Liipke, Chirality 5 (1993) 385-392. J. SchormOller and J. Weder, Nahrung 6 (1963) 622-630. I. Gandolfi, G. Palla, L. Delprato. F. De Nisco, R. MarcheUi and C. Salvadori, J. Food Sci. 57 (1992) 377379. S.H. Lee, K.H. Kim, Y.C. Lee and S.T. Kim, J. Korean Chem. Soc. 39 (1995) 257-265. V.M. Moretti, A. Albertini, F. Bellagamba. T. Mentasti. F. Valfre and P. Poldori, Industrie Alimentari 36 (1997) 896-898 (in Italian). S. Casal, M.B. Oliveira and MA. Ferreira, J. Chromatogr. A 866 (2000) 221-230. J. Csapo, Z. Csapo-Kiss, J. Stefler, T.G. Martin and S. Nemethy, J. Dairy Sci. 78 (1995) 2375-2381. D. Jin, T. Miyahara, T. Oe andT. Toyo'o\^2L,Anal Biochem. 269 (1999) 124-132. H. Briickner, R Jack, M. LangerandH. GOAQI Amino Acids 1 (1992)271-284. E. Varga-Visi and J. Csa.p6,ActaAgraria Kaposvariensis 1 (2003) 47-61 H. Briickner, M. Langer, M. Lupke, T. Westhauser and H. Godel, J. Chromatogr. A 697 (1995) 229-245. R. Marchelli, G. Palla, A. Dossena, G. Galavema, R. Corradini and S. Clementi, Scienza e Tecnica Lattiero Casearia 48 (1997) 21-32 (in Italian). F. Requena-Rondon, M. Trabalza-Marinucci, D. Casciotti, C. Cavallucci and O. Olivieri, Milk Sci. /«rer/7a/. 53(1998)313-316. N. Innocente and G. Palla, J. Dairy Res, 66 (1999) 633-637. H. Abe, P. Jung-Nim, Y. Fukmnoto, E. Fujita, T. Tanaka, T. Washio, S. Otsuka, T. Shimizu and K. Watanabe, Fisheries Sci. 65 (1999) 637-641. E. Okmna, E. Fujita, H. Amano, H. Noda and H. Abe, Fisheries Sci. 61 (1995) 157-160. H.J. Chaves das Neves and J.P. Noronha, Fresenius'J. Anal. Chem. 352 (1995) 783-787. T. Elbe and H. Bruckner, J. Chromatogr A 881 (2000) 81-91. K.H. Ekborg-Ott and D.W. Armstrong, Chirality 8 (19%) 49-57. T. Elbe and H. Briickner, Z Lebensm. Unters. Forsch. A 207 (1998) 400-409. D.W. Armstrong, CD. Chang and Y.L. Wei, J. Agric. FoodChem. 38 (1990) 1674-1677. E. Chiavaro, A. CaUgiani and G. Palla, Ital. J. Food Sci. 10 (1998) 329-337. A. Halasz, A. Barath and W.H. Holzapfel, Z Lebensm. Unters. Forsch. 208 (1999) 434-438. M. Gobbetti, M.S. Simonetti, J. Rossi, L. Cossignani, A. Corsetti and P. Damiani, J. Food Sci. 59 (1994) 881-884.

Occurrence of D-Amino Acids in Food [28] [29] [30] [31] [32] [33] [34] [35] [36] [37] [38] [39] [40] [41] [42] [43] [44] [45] [46] [47] [48] [49] [50]

353

G. Palla, R. Marchelli, A. Decennia and G. Casnati, J. Chromatogr. 475 (1989) 45-53. M. Pawlowska and D. W. Armstrong, Chirality 6 (1994) 210-276. K.L. Rundlett and D.W. Armstrong, Chirality 6 (1994) 277-282. M. Liipke and H. Bruckner, Z Lebensm. Unters. Forsch. A 206 (1998) 323-328. H. Bruckner and T. Westhauser, Chromatographia 39 (1994) 419-426. I. Gandolfi, G. Palla, R. Marchelli, A. Decennia, S. Puelli and C. Salvadori, J. FoodSci. 59 (1994) 152154. H. Bruckner and T. Westhauser, Amino Acids 24 (2003) 43-55. S. Einarsson, S. Folestad and B. Josefsson, J. Liq. Chromatogr. 10 (1987) 1589-1598. A. Jegorov, V. Matha, T. Tmka and M. Cemy, J. High Resolut. Chromatogr. 13 (1990) 718-720. J. Csapo, J. Schmidt, Zs. Csapo-Kiss, E. Varga-Visi, G. Pohn and E. Csokona, Chromatograhia Supplement. 56 (2002) 169-171. R.H. Buck and K. Krummen, J. Chromatogr. 387 (1987) 255-265. A.L.L. Duchateau, H. Knuts, J.M.M. Boesten and J.J. Guns, J. Chromatogr. 623 (1992) 237-245. D. Aswad, Anal. Biochem. 137 (1984) 405-409. N. Nimura and T. Kinoshita, /. Chromatogr. 352 (1986) 169-177. N.C. van de Merbel, M. Stenberg, R. Oeste, G. Marko-Varga, L. Gorton, H. Lingeman and U. A.T. Brinkman, Chromatographia. 41 (1995) 6-14. C.W. Gehrke, K.C. Kuo and R.W. ZumwaU, Amino Acid Analysis by Gas Chromatography, CRC ress, Boca Raton, FL, 1987. F.G. Kitson, B.S. Larsen and C.N. McEwen, Gas Chromatography and Mass Spectrometry: A Practical Guide, Academic Press, San diego, CA, 1996. C. Boniglia, B. Carratu andE. Sanzini, J. FoodSci. 67 (2002) 1352-1355. C. Simo, C. Barbas and A. Cifuentes, J. Agric. FoodChem. 50 (2002) 5288-5293. F. Scheller and F. Schubert, Biosensors, Elsevier, Amsterdam, 1992. G. Schwedt and K. Stein, Z Lebensm Unters Forsch 199 (1994) 171-182. K. Voss and R. Galensa, Amino Acids 18 (2000) 339-352. M. Varadi, N. Adanyi, E.E. Szab6 and N. Trunmier, Biosensors Bioelectronics 14 (1999) 335-340.

Progress in Biological Chirality G. Palyi, C. Zucchi and L. Caglioti (Editors) © 2004 Elsevier Ltd. All rights reserved.

Chapter 29 Asymmetric Autocatalysis, Absolute Asymmetric Synthesis and Origin of Homochirality of Biomolecules Kenso Soai Department of Applied Chemistry, Faculty of Science, Tokyo University of Science, Kagurazaka, Shinjuku-ku, Tokyo, 162-8601 Japan soai@rs. kagu. tus. ac.jp

1.

Introduction One of the most characteristic features of the present living organisms on the Earth is that all of the living organisms use highly L-enriched amino acids and highly D-enriched sugars, />., biologically relevant molecules are homochiral. When and how did biomolecules become highly enantiomerically enriched? Because the chiral homogeneity of biomolecules is closely related to the origin and evolution of life, the origin of homochirality of biomolecules has attracted much attention [1]. Several mechanisms have been proposed as the origins of chirality of organic compounds. However, enantiomeric imbalances of organic compounds induced by the mechanisms such as circularly polarized light (CPL) and quartz have been usually very low (<2% ee). Therefore, amplification process of very low enantiomeric imbalance of organic compounds to very high enantiomeric excess (ee) is inevitable to explain the homochirality of organic compounds. Asymmetric autocatalysis is a reaction in which the chiral product acts as a chiral catalyst for its own production (Scheme 1) [2, 3]. In other words, the reaction is an automultiplication of a chiral compound. In asymmetric autocatalysis, the efficiency is high because the process is automultiplication. In an ideal asymmetric autocatalysis, no decrease in the amount of catalyst and no deterioration of the catalytic activity should be observed because the amount of catalyst increases during the reaction. In addition, there is no need to separate the catalyst rAsymmetric autocatalysis 1 The same structure & configuration A

+ B

Asymmetric autocatalyst P

Scheme 1

• P (Product)

356

Progress in Biological ChiraliU'

from the product because their structures are identical. Although Frank proposed kinetic model of asymmetric autocatalysis [4], neither actual compound nor actual reaction was mentioned. We describe here highly enantioselective asymmetric autocatalysis with amplification of ee [2] and asymmetric autocatalysis initiated by chiral substances such as quartz and sodium chlorate. We also describe absolute (spontaneous) asymmetric synthesis in conjunction with asymmetric autocatalysis. 2.

Asymmetric Autocatalysis Enantiomerically enriched ^ec-alcohols are formed by the enantioselective addition of dialkylzincs to aldehydes using P-amino alcohols as chiral catalysts [5]. During our study on the enantioselective addition of dialkylzincs to nitrogen containing aldehydes [6], we found for the first time that chiral 3-pyridyl alkanol acts as an asymmetric autocatalyst in the addition of diisopropylzinc (/-Pr2Zn) to pyridine-3-carbaldehyde [7]. (*S)-3-Pyridyl alkanol with 86% ee acts as an asymmetric autocatalyst in the enantioselective addition of /-PriZn to 3-pyridinecarbaldehyde, affording the same compound with 35% ee. After searching for various systems of asymmetric autocatalysis [8], we found that chiral 5-pyrimidyl alkanol 2 [9], 3-quinolyl alkanol [10] and 5-carbamoyl-3-pyridyl alkanol [11] are highly enantioselective asymmetric autocatalysts for the addition of /-Pr2Zn to pyrimidine-5carbaldehyde 1, 3-quinolinecarbaldehyde and 5-carbamoyl-3-pyridinecarbaldehyde, respectively (Scheme 2). Among these, chiral 5-pyrimidyl alkanol is the most significant The same structure & onfiguration

N

CHO /-Pr2Zn

Asymmetric autocatalyst 2a (93% ee) 2b (>99.5% ee) toluene, 0 °C

1a (R= H) 1b (R= Me)

2a (R=H) 90% ee 2b (R=Me) 98.2% ee

The same structure & ;onfiguration

Asymmetric autocatalyst 2c (>99.5% ee) N " ^

CHO /-Pr2Zn

Product 2c 1st round: >99%, >99.5% ee 10th round: >99%, >99.5% ee Obtained alcohol was used as an asymmetric autocatalyst for the next round. Scheme 2

Asymmetric Autocatalysis, Absolute Asymmetric Synthesis and Origin of Homochirality of Biomolecules 357 asymmetric autocatalysis. When (*S)-2-(2-/-butylethynyl)-5-pyrimidyl alkanol 2c with >99.5% ee was used as an asymmetric autocatalyst, (S)-2c with >99.5% ee was formed in a yield of >99% [12]. One of the avantages of asymmetric autocatalysis is that the structures of the asymmetric autocatalyst and the product are the same. Thus, the product and the initial autocatalyst of the first round was used as an asymmetric autocatalyst for the next round. Again, the product (S)-2c and the initial autocatalyst had ee of >99.5% and the yield of the newly formed (S)-2c was >99%. Even after tenth round, the yield of 2c was >99% and the ee was >99.5%. The amount of (5)2c has automultipHed by a factor of ca. 60 million times during these 10 rounds. Thus, 2alkynylpyrimidyl alkanol 2c was found to be an asymmetric autocatalyst with almost complete enantioselectivity and very high catalytic activity. 3.

Amplification of Enantiomeric Excess by Asymmetric Autocatalysis Surprisingly, the ee of pyrimidyl alkanol was found to increase in asymmetric autocatalysis [9a]. When pyrimidyl alkanol with low ee was used as an asymmetric autocatalyst, i.e., the ee of the product (including the original autocatalyst) was higher than that of the original catalyst. To take advantage of asymmetric autocatalysis with amplification of ee over nonautocatalytic amplification of ee [13], the product of one run was used as an asymmetric autocatalyst for the next run. Thus, extremely low ee of pyrimidyl alkanol was amplified to very high ee by consecutive asymmetric autocatalysis. Asymmetric autocatalysis of (*S)-pyrimidyl alkanol 2a with 2% ee gave (iS)-2a with 10% ee (Scheme 3). When, (*S)-2a with 10% ee was used as an asymmetric autocatalyst for the next asymmetric autocatalysis, pyrimidyl alkanol 2a with an increased ee of 57% was obtained. The subsequent consecutive asymmetric autocatalysis using the product as an asymmetric autocatalyst for the next round gave (^-pyrimidyl alkanol 2a with 81 and 88% ee, respectively. Thus, the overall process is the asymmetric autocatalysis of (*S)-2a starting from 2% ee with significant amplification of ee to 88% and with the increase in the amount without the need for any other chiral auxiliary [9a]. This stands as the first example of an asymmetric

Asymmetric autocatalysis with amplification of ee

2a-c 2c:R=f-Bu-^

lowee ca. 0.00005% ee (2c)

[Asymmetric autocatalyst]

2a-c high ee >99.5%ee(2c)

1 +

/-Pr2Zn Scheme 3

358

Progress in Biological Chirality 1

-yA

1

|i(S)-2c

==-

^(R)-2c

(Hi^ 600000 H

o

1 400000 H

1 200000 A

J yT

n i'mmmty. initial conditions

fmmsss'. r run 1 57% ee

/v..

_ 1

iffly.

WL)

run 2 run 3 99% ee >99.5% ee

ca. 0.00005% ee Fig. 1. Asymmetric autocatalysis with amplification of ee autocatalysis with amplification of ee. One-pot asymmetric autocatalysis of pyrimidyl alkanol 2b also significantly increased the ee from 0.28% ee to 87% ee [14a]. We also found the efficient amplification of chirality by using 2-alkynylpyrimidyl alkanol 2c from as low as ca. 0.00005%o ee to almost enantiomerically pure (>99.5% ee) product in only three consecutive asymmetric autocatalysis (Scheme 3, Fig. 1) [14b]. The first asymmetric autocatalysis with (S)-lz of ca. 0.00005% ee gave (^-2c in 96% yield with an enhanced ee of 57%. The second asymmetric autocatalyses with the autocatalyst of 57% ee afforded (*S)-2c with 99% ee, and the ee reached >99.5% ee by the third asymmetric autocatalysis. During the three consecutive asymmetric autocatalyses, the initially major (5)enantiomer of 2c has automuhiplied by a factor of 630,000 times, whereas the initially minor (i?)-enantiomer of 2c by a factor of only less than 1,000 times. As described, pyrimidyl alkanols 2 act as highly efficient asymmetric autocatalysts with amplification of ee. 3-Quinolyl alkanol [15a] and 5-carbamoyl-3-pyridyl alkanol [15b] were also found to act as asymmetric autocatalysts with amplification of ee.

Asymmetric Autocatalysis, Absolute Asymmetric Synthesis and Origin of Homochirality of Biomolecules

359

4. The Role of Asymmetric Autocatalysis as a Linkage Between the Origin of Chirality and Homochirality of Biomolecules 4.1 Asymmetric autocatalysis triggered by organic compounds with low enantiomeric excess It is known that circularly polarized light (CPL) is one of the origins of chirality of organic compounds [16]. Asymmetric photolysis of racemic leucine by right circularly polarized light (CPL, 213 nm) affords L-leucine with only 2% ee [16a]. Hexahelicene with low (<2%) ee is synthesized by asymmetric photosynthesis using CPL [16b,c]. However, these low enantiomeric excesses induced by CPL have not been correlated with high ee's of an organic compound. In the presence of L-leucine with only 2% ee as a chiral trigger, the reaction of 2methylpyrimidine-5-carbaldehyde l b with /-PrzZn afforded (i?)-pyrimidyl alkanol 2b with an enhanced ee of 21% (Scheme 4) [17a]. On the contrary, when D-leucine with 2% ee was used as a chiral trigger, {S)-2h with an increased ee (26%) was obtained. As shown in the preceding section, the ee of the obtained pyrimidyl alkanol can be fijrther amplified significantly by the consecutive asymmetric autocatalyses. In the presence of (P)-hexahelicene with very low (0.13%) ee as a chiral trigger, the reaction between aldehyde Ic and i-VvzZn gave (*S)-pyrimidyl alkanol 2c with 56% ee [17b]. When (M)-hexahelicene with 0.54% ee was used instead of (P)-hexahelicene, (JR)-2C with 62% ee was formed. Thus, a slight enantiomeric imbalance in compounds induced by CPL was correlated for the first time to an organic compound with very high ee by asymmetric autocatalysis with amplification of ee. In addition, various chiral organic compounds serve as chiral triggers in asymmetric autocatalysis [17c-e]. 4.2 Asymmetric autocatalysis initiated by chiral inorganic crystals Quartz and sodium chlorate are chiral inorganic crystals. They exist as either dextrorotatory (d) or levorotatory (/) enantiomorphs. Quartz has been considered as one of the origins of chirality. However, apparent asymmetric synthesis using quartz is not known [18], although a very small asymmetric induction is known in an asymmetric adsorption of chiral compound by quartz [19]. Asymmetric photolysis r-CPL DL-leucine

N ^ Me^^N 1b

ca. 2% ee toluene, O X [Ryib

21% ee Scheme 4

360

Progress in Biological Chirality

We thought that asymmetric autocatalysis triggered by quartz affords pyrimidyl alkanol with high ee (Scheme 5). When 2-alkynylpyrimidine-5-carbaldehyde Ic was reacted with /'Pr2Zn in the presence of the powder of tZ-quartz, (5)-pyrimidyl alkanol 2c with 97% ee was obtained in a yield of 95% [20a]. On the contrary, in the presence of/-quartz, (Ryic with 97% ee was obtained in a yield of 97%. These results clearly show that the absolute configurations of pyrimidyl alkanol formed were regulated by the chirality of quartz. A small enantiomeric imbalance of the initially formed (zinc alkoxide of) pyrimidyl alkanol induced by chiral quartz was amplified significantly by one-pot asymmetric autocatalysis to afford pyrimidyl alkanol 2c with very high ee. Thus, a chiral organic compound with high ee is formed, for the first time, using chiral inorganic crystals as chiral auxiliary in conjunction with asymmetric autocatalysis. Sodium chlorate (NaClOs) also acts as a chiral trigger. In the presence of J-NaClOs, (^pyrimidyl alkanol 2c with 98% ee was formed in >90% yield (Scheme 5) [21]. On the other hand, in the presence of /-NaClOa, {Ryic with 98% ee was formed. It should be noted that crystallization of a stirred solution of NaClOa affords either d- or /-enriched form [22]. As described, a chiral organic compound with high ee is formed using chiral inorganic crystals in conjunction with asymmetric autocatalysis. 4.3 Absolute (spontaneous) asymmetric synthesis Based on the theory of statistics, it is considered that small fluctuations in the ratio of the two enantiomers are present if chiral molecules are produced from achiral starting materials under conditions under which the probability of formation of the enantiomers is equal. So-

(Chiral inorganic crystals ) d-Quartz d-NaClOa

CHO Zn/-Quartz /-NaClOa

ee Scheme 5

Asymmetric Autocatalysis, Absolute Asymmetric Synthesis and Origin of Homochirality of Biomolecules

361

called racemate doesn't usually contain the exact numbers of (S) and (/?)-enantiomers, although the ee is below the detection level by contemporary analytical methods. However, if the small fluctuation of chiralty is amplified by asymmetric autocatalyses, one would obtain enantiomerically enriched compound with well above the detection level [23]. The significant ability of amplification of enantiomeric excess by the asymmetric autocatalysis of pyrimidyl alkanol prompted us to examine the absolute asymmetric synthesis, i.e., the reaction of achiral pyrimidine-5-carbaldehyde with diisopropylzinc without adding any chiral substance in combination with the subsequent asymmetric autocatalysis with amplification of ee. Pyrimidine-5-carbaldehyde was reacted with diisopropylzinc, and the resulting pyrimidyl alkanol was used as an asymmetric autocatalyst for the next asymmetric autocatalysis. The subsequent consecutive asymmetric autocatalysis afforded pyrimidyl alkanol of either S or R configuration with enantiomeric enrichment well above the detection level.

[Absolute (Spontaneous) Asymmetric Synthesis]

-^jJ^

+ /-Pr2Zn

1c

Scheme 6

100 Enantiomeric excess (% ee)

Absolute configuration and ee of pyrimidyl alkanol Fig. 2. Absolute (spontaneous) asymmetric synthesis of pyrimidyl alkanol 2c in combination with asymmetric autocatalysis

362

Progress in Biological Chirality

The effect of the substituent on the 2-position of 2-alkynylpyrimidine-5-carbaldehyde and the effect of solvent are significant. Reaction of 2-alkynylpyrimidine-5-carbaldehyde with /Pr2Zn in a mixed solvent of ether and toluene or in a mixed solvent of ether and dibutylether and the following one-pot asymmetric autocatalysis with amplification of ee gave enantiomerically enriched pyrimidyl alkanol well above the detection level [24]. The absolute configurations of the pyrimidyl alkanol formed show an approximate stochastic distribution of ^ and R enantiomers (19 times formation ofS and 18 times R). The approximate stochastic behavior in the formation of pyrimidyl alkanols form one of the conditions necessary for absolute (spontaneous) asymmetric synthesis [25].

5.

Conclusions We found that chiral 5-pyrimidyl alkanol 2, 3-quinolyl alkanol and 5-carbamoyl-3-pyridyl alkanol are highly enantioselective asymmetric autocatalysts for the addition of /-Pr2Zn to pyrimidine-5-carbaldehyde 1, 3-quinolinecarbaldehyde and 5-carbamoyl-3pyridinecarbaldehyde, respectively. Among these, 2-alkyny 1-5-pyrimidyl alkanol 2c is a highly efficient asymmetric autocatalyst with >99.5% enantioselectivity. Moreover, asymmetric autocatalysis with amplification of ee from extremely low ee to >99.5% ee was realized for the first time by consecutive asymmetric autocatalysis without the need for any other chiral auxiliary. Kinetic analysis of pyrimidyl alkanol suggested that the reaction proceeds via second order of the zinc monoalkoxide of pyrimidyl alkanol [26]. For the mechanism of the amplification from very low ee, the presence of additional mechanism as well as the second order mechanism of the zinc monoalkoxide of pyrimidyl alkanol is postulated [26b]. The elucidation of the actual reactive species remains as a future subject. Chiral organic compounds with low ee which are induced by CPL serve as chiral triggers of the asymmetric autocatalysis. The overall process correlates, for the first time, the chirality of CPL with an organic compound with very high ee. Very recently, chirality of CPL was directly correlated with the chirality of pyrimidyl alkanol with high ee by asymmetric photolysis of racemic pyrimidyl alkanol in combination with asymmetric autocatalysis [27]. Chiral inorganic crystals such as quartz and sodium chlorate act as chiral triggers and regulate the sense of the asymmetric autocatalysis. The process correlates, for the first time, the chirality of inorganic crystals with an organic compound with very high ee. Absolute (spontaneous) asymmetric synthesis is described in the formation of enantiomerically enriched pyrimidyl alkanol from the reaction of pyrimidine-5-carbaldehyde and /-Pr2Zn without adding chiral substance in combination with asymmetric autocatalysis. The approximate stochastic distribution of the absolute configurations of pyrimidyl alkanols strongly suggests that the reaction is an absolute (spontaneous) asymmetric synthesis. As described, asymmetric autocatalysis is closely related with the origin of homochirality of organic compounds.

6.

Acknowledgements Special gratitude is expressed to the coworkers whose names appear in the papers. Financial support from the Ministry of Education, Culture, Sports, Science and Technology, New Energy and Industrial Technology Development Organization (NEDO) and the Japan Space Forum is gratefully acknowledged.

Asymmetric Autocatalysis, Absolute Asymmetric Synthesis and Origin of Homochirality of Biomolecules

7.

363

References

[1] (a) K. Mislow and P. Bickart, Isr. J. Chem. 15 (1976/1977) 1. (b) I S . Siegel, Chirality 10 (1988) 24-27. (c) D.K. Kondepudi and G.W. Nelson, Nature 314 (1985) 438-441. (d) I. Weissbuch, L. Addadi, L. Leiserowitz and M. Lahav, /. Am. Chem. Soc. 110 (1998) 561-567. Reviews; (e) K. Mislow, Collect. Czech. Chem. Commurt. 68 (2003) 849-864. (f) A. Eschenmoser, Science 284 (1999) 2118-2124. (g) L. Keszthelyi, Q. Rev. Biophys. 28 (1995) 473-507. (h) W.A. Bonner, Orig. Life Evol. Biosphere 21 (1991) 59-111. (i) S.F. Mason and G.E. Tranter, Proa R. Soc. Lond. A 397 (1985) 45-65. (j) V. Avetisov and V.I. Goldanskii, Proc. Natl. Acad. Set. USA 93 (1996) 11435-11442. (k) M. Avalos, R. Babiano, P. Cintas, J.L. Jimenez and J.C. Palacios, Tetrahedron: Asymmetry 11 (2000) 2845-2874. (1) B.L. Feringa and R.A. van DQlden, Angew. Chem., Int. Ed. 38 (1999) 3418-3438. (m) S. Mason, Chem. Soc. Rev. 17 (1988)347-359. [2] (a) K. Soai and T. Shibata, in: Advances in Biochirality (Eds. G. Palyi, C. Zucchi and L. Caglioti) Elsevier, Amsterdam, 1999, Chapter 11. (b) K. Soai, T. Shibata and I. Sato,^cc. Chem. Res. 33 (2000) 382-390. (c) K. Soai, Enantiomer 4 (1999) 591-598. (d) K. Soai and T. Shibata, Yuki Gosei Kagaku Kyokaishi [J. Synth. Org. Chem. Jpn.] 55 (1997) 994-1005. (e) K. Soai and T. Shibata, in: Catalytic Asymmetric Synthesis, Second Ed. (Ed. I. Ojima) Wiley, New York, 2000, Chapter 9. (f) K. Soai, T. Shibata and I. Sato, Nippon Kagaku Kaishi [J. Chem. Soc. Jpn., Chem. Ind Chem.] 3 (2001) 141-149. (g) K. Soai, I. Sato and T. Shibata, Chem. Record 1 (2001) 321-332. (h) K. Soai and I. Sato, Chirality 14 (2002) 548-554. (i) K. Soai, in: Fundamentals of Life (Eds. G. Palyi, C. Zucchi and L. Caglioti) Elsevier and Accademia Nazionale di Scienze, Lettere ed Arti (Modena), Paris, 2002, pp. 427-435. (j) K. Soai, I. Sato and T. Shibata, Yuki Gosei Kagaku Kyokaishi [J. Synth. Org Chem. Jpn.] 60 (2002) 668-678. (k) K. Soai and I. Sato, Viva Origino 30 (2002) 186-198. [3] Reviews: (a) C. Bolm, F. Bienewald and A. Seger, Angew. Chem., Int. Ed Engl. 35 (1996) 1657-1659. (b) M. Avalos, R. Babiano, P. Cintas, J.L. Jimenez and J.C. Palacios, Chem. Commun. (2000) 887-892. (c) H. Buschmann, R. Thede and D. BQWQT, Angew. Chem. Int. Ed 39 (2000) 4033-4036. (d) M.H. Todd, Chem. Soc. Rev. 3\ {1001) IW'lll. [4] F.C. Frank, Biochim. Biophys. Acta 11 (1953) 459-463. [5] Reviews: (a) K. Soai and S. Niwa, Chem. Rev. 92 (1992) 833-856. (b) R. Notori and M. Kitamura, Angew. Chem. Int. Ed. Engl. 30 (1991) 49-69. (c) L. PuandH.-B. Yu, Chem. Rev. 101 (2001) 757-824. [6] K. Soai, S. Hon and S. Niwa, Heterocycles 29 (1989) 2065-2067. [7] K. Soai, S. Niwa and H. Hon, J Chem. Soc. Chem. Commun. (1990) 982-983. [8] (a) K. Soai, T. Hayase and K. Takai, Tetrahedron: Asymmetry 6 (1995) 637-638. (b) K. Soai, T. Hayase, C. ShimadaandK. Isobe, Tetrahedron: Asymmetry 5 (1994)1^9-191. [9] (a) K. Soai, T. Shibata, H. Morioka and K. Choji, Nature (London) 378 (1995) 767-768. (b) T. Shibata, H. Morioka, T. Hayase, K. Choji and K. Soai, J Am. Chem. Soc. 118 (1996) 471-472. [10] T. Shibata, K. Choji, H. Morioka, T. Hayase and K. Soai, Chem. Commun. (1996) 751-752. [11] T. Shibata, H. Morioka, S. Tanji, T. Hayase, Y. Kodaka and K. Soai, Tetrahedron Lett. 37 (1996) 87838786. [12] T. Shibata, S. Yonekubo and K. Som, Angew. Chem., Int. Ed 38 (1999)659-661. [13] Reviews: (a) C. Girard and H.B. Y.digtm, Angew. Chem., Int. Ed. 37 (1998) 2922-2959. (b) M. Avalos, R. Babiano, P. Cintas, J.L. Jimenez and J.C. Palacios, Tetrahedron: Asymmetry 8 (1997) 2997-3017. (c) D.R. Fenwick and H.B. Kagan, in: Topics in Stereochemistry (Ed. S.E. Denmark), Wiley, New York, 1999. Vol. 22, pp. 257-296. (d) C. Bolm, Advanced Asymmetric Synthesis (Ed. G.R. Stephenson) Blackie, Glasgow, 1996, pp. 9-26. (e) M.M. Green, J.-W. Park, T. Sato, S. Lifson, R.L.B. Selinger, J.V. ^QligQx, Angew. Chem., Int. Ed 38 (1999) 3138-3154. [14] (a) T. Shibata, T. Hayase, J. Yamamoto and K. Soai, Tetrahedron: Asymmetry 8 (1997) 1717-1719. (b) I. Sato, H. Urabe, S. Ishiguro, T. Shibata K. SQ2ii, Angew. Chem., Int. Ed 42 (2003) 315-317. [15] (a) T. Shibata, K. Choji, T. Hayase, Y. Aizu and K. Soai, Chem. Commun. (1996) 1235-1236. (b) S. Tanji, Y. Kodaka, A. Ohno, T. Shibata, I. Sato andK. Soai, Tetrahedron: Asymmetry 11 (2000) 4249-4253. [16] (a) J.J. Flores, W.A. Bonner and G.A. Massey, / . Am. Chem. Soc. 99 (1977) 3622-3625. (b) A Moradpour, J.F. Nicoud, G. Balavoine, H. Kagan and G. Tsoucaris, /. Am. Chem. Soc. 93 (1971) 2353-2354. (c) W.J. Bernstein, M. Calvin and O. Buchardt, J. Am. Chem. Soc. 94 (1972) 494-498. (d) Y. Inoue, H. Tsuneishi, T. Hakushi, K. Yagi, K. Awazuand H. Onuki, Chem. Commun. (1996) 2627-2628. Review: (e) O. Buchardt, ^wgew. Chem., Int. Ed Engl. 13 (1974) 179-185. (f) See also [Ih]. (g) Y. Inoue, Chem. Rev. 92 (1992)741-770. [17] (a) T. Shibata, J. Yamamoto, N. Matsumoto, S. Yonekubo, S. Osanai and K. Soai, J. Am. Chem. Soc. 120 (1998) 12157-12158. (b) I. Sato, R. Yamashima, K Kadowaki, J. Yamamoto, T. Shibata and K. Soai,

364

118]

[19] [20]

[21] [22] [23]

[24] [25] [26]

[27]

Progress in Biological Chirality Angew. Chem., Int. Ed. 40 (2001) 1096-1098. (c) I. Sato, S. Osanai, K. Kadowaki, T. Sugiyama, T. Shibata and K. Soai, Chem. Lett. (2002) 168-169. (d) I. Sato, Y. Matsueda, K. Kadowaki, S. Yonekubo, T. Shibata and K. Soai, Helv. Chim. Acta 85 (2002) 3383-3387. (e) I. Sato, A. Ohno, Y. Aoyama, T. Kasahara and K. Soai, Org. Biomol. Chem. 1 (2003) 244-246. G.M. Schwab and L. Rudolph, Natunvissenschaften 20 (1932) 363. A.P. Terent'ev and E.I. Klabunovskii, SbomikStatei Obshchei Khim. 2 (1953) 1598; Chem. Abstr. 49 (1955) 5262g. The reproducibility of these reports of the very low asymmetric induction using quartz was disproved in later careful experiments by Amariglio et al. A. Amariglio, H. AmarigUo and X. Duval, Helv. Chim. Acta 51 (1968) 2110-2132. W.A. Bonner, P.R. Kavasmaneck, F.S. Martin and J.J. Flores, Science 186 (1974) 143-144. (a) K. Soai, S. Osanai, K. Kadowaki, S. Yonekubo, T. Shibata and I. Sato, J. Am. Chem. Soc. 121 (1999) 11235-11236. (b) For the use of helical silica as a chiral trigger. I. Sato, K. Kadowaki, H. Urabe, J. Hwa Jung, Y. Ono, S. Shinkai and K. Soai, Tetrahedron Lett. 44 (2003) 721-724. I. Sato, K. Kadowaki and K. ^02i. Angew. Chem., Int. Ed 39 (2000) 1510-1512. D.K. Kondepudi, R.J. Kaufman and N. Singh, Science 250 (1990) 975. (a) K. Soai, T. Shibata and Y. Kowata, Japan Kokai Tokkyo Koho 9,268,179, (1997). (apphcation date: February 1, 1996 and April 18, 1996). (b) DA. Singleton and L. K. Vo, J. Am. Chem. Soc. 124 (2002) 10010-10011. K. Soai, I. Sato, T. Shibata, S. Komiya, M. Hayashi, Y. Matsueda, H. Imamura, T. Hayase, H. Morioka, H. Tabira, J. Yamamoto and Y. Kowata, Tetrahedron: Asymmetry 14 (2003) 185-188. (a) For an excellent comprehensive review and commentary, see ref. le. (b) Commentary: J.S. Siegel, Nature 419 (2002) 346-347. (a) I. Sato, D. Omiya, K. Tsukiyama, Y. Ogi and K. Soai, Tetrahedron: Asymmetry 12 (2001) 1965-1969. (b) I. Sato, D. Omiya, H. Igarashi, K. Kato, Y. Ogi, K. Tsukiyama and K. Soai, Tetrahedron: Asymmetry 14 (2003) 975-979. (c) D.G. Blackmond, C.R. McMillan, S. Ramdeehul, A. Schorm, J.M. Brown, J. ^w. Chem. Soc. 123 (2001) 10103-10104. S. Ishiguro. T. Saito, S. Komiya, I. Sato, K. Soai, H. Nishino and Y. Inoue, The 83th National Meeting of The Chemical Society of Japan, 1H5-08, March, 2003, Tokyo.

Progress in Biological Chirality G. Palyi, C. Zucchi and L. Caglioti (Editors) © 2004 Elsevier Ltd. All rights reserved.

Chapter 30 Charophyte Gyrogonites, the Result of Enantioselective Influence 250 Million Years Ago Ingeborg Soulie-Marsche Laboratoire de Paleobotanique, Institnt des Sciences de VEvolution, Universite Montpellier II, CP. 062, Place E. Bataillon, 34095 MontpelUer-Cedex 5, France marsche@i sent, univ-montp2.fr

I dedicate this paper to Jacques Vautier, who unfortunately departed in December 2003, as a sign of gratitude for his inestimable help with the writing of this article and as an expression of my sincere thanks for his continuous encouragement of my work on charophytes.

1.

Introduction Charophytes (stoneworts) are macrophytes growing entirely submerged in freshwater and brackish water. They become fossilised through their calcified oospores termed gyrogonites. These reproductive organs are an equivalent to the seeds of land plants. The gyrogonites of all modern species fit to a single basic structure consisting in five enveloping cells, which are twisted clockwise around an inner ovoid egg cell. This simple but characteristic pattern was recognised to be unique in the plant kingdom and also proved to be absolutely constant. One has to go back into the Palaeozoic Era to find different morphological types of gyrogonites. The structure of these ancestral gyrogonites was much more complex than the modern type. Based on the fossil record, the charophytes constitute a polyphyletic group that appeared in the Late Silurian (c. 400 Mya) with already two very distinct morphologies of calcified fructifications: a group with vertical and even transversally segmented enveloping cells and a group with numerous counter clockwise twisted enveloping cells. Both types were originally described fi-om the "Uppermost Silurian" in Podolia, Ukraine [1, 2]. It can be assumed that charophyte plants existed already much earlier but did not provide fossil remains as these ancestors probably did not yet form calcified fructifications. The hitherto earliest known gyrogonites with clockwise twisted enveloping cells came from the Middle Devonian of Canada [3]. The Devonian to Permian periods then saw the coexistence of both counter clockwise and clockwise twisted cells. At the end of the Palaeozoic, however, the counter clockwise twisted forms got extinct and since 250 Million years the morphological structure of the gyrogonites has been reduced to the single modern type with strictly five clockwise twisted cells. The aim of the present paper is to highlight the important change in chirality that occurred

366

Progress in Biological Chiralit>

Mirror line

Figure 1. Outline of a modem Charophyte gyrogonite and its theoretical mirror image. The gyrogonites of all living species fit to this unique pattern composed of five sinistrally (clockwise) twisted spiral cells surrounding an inner ovoid egg cell

for the charophyte gyrogonites at the Permian-Triassic boundary (PTB) and to discuss the possible causes of the loss of one type of handedness.

2. Structure of the Gyrogonites Charophytes provide striking examples of biological chirality not only in the female reproductive organ, the gyrogonite, but also in the plant's architecture. At macroscopic scale, the bud displacement follows a helical along the main axis. At microscopic scale, conspicuous asymmetry characterises the gyrogonites. Both structures can be explained by the ontogenetic development of the cells that follows the rules of 2/5^^ spiral phyllotaxis, well known for Embryophytes [4]. 2.1 Coiling direction: a definition The gyrogonites of all living Charophyte species are twisted clockwise with the cells ascending from right to left. Usually less than 1.5 mm in size, they represent homo-chiral bodies at microscopic scale (Fig. 1). The coiling direction of the spiral cells is determined by the growth dynamics ascending from the base to the apex. The polarity of the spiral can be accurately determined because the basal pole always shows a clearly determined basal structure in the form of a pentagonal opening or a pentagonal basal plate. Detailed analysis of the structure allowing the orientation of the gyrogonites was described earlier [4]. This coiling pattern of the gyrogonites is conventionally called sinistral in accordance with the definition traditionally employed by botanists. The botanical term "sinistrorse" points to plants "climbing upwards in a spiral form from right to left or clockwise"[5]. The Palaeozoic gyrogonites could be either sinistral or dextral. On the latter, the spiral cells are twisted counter clockwise, ascending from lower left side to upper right. The dextral or sinistral coiling thus is best seen in lateral views whereas in the basal and apical views of gyrogonites, the handedness of the spiral appears as it were exactly the opposite (Fig. 2). Martin-Closas pointed to the confusion that might result from the botanical definition being exactly the opposite of dextrogyrous and levogyrous as commonly used in other fields

Charophyte Gyrogonites, the Result of Enantioselective Influence 250 Million Years Ago

To the left = sinistral

^Si^^,^ ^ ^

^^^^ ^ ^

367

^^ ^^^ "Sh* = dextral

Figure 2. Comparison of a sinistral and dextral gyrogonite in basal view. a. Eochara wickenendi Choquette; b. Moellerina greenii Ulrich. Bothfromthe Devonian. Arrows indicate coiling (hrection startingfromthe basal opening. Scale bar 100 fim

[6, 7]. However, for more than a century, the gyrogonites have been described with the botanical meaning of "left" and "right". We will use these terms in this conventional sense throughout the text. 2,2 Morphological types of Palaeozoic Charophytes Palaeozoic Charophytes have been little studied and relevant literature mostly dealt with systematics and their stratigraphical value for the continental realm. In the light of phylogeny, the Charophytes represent a polyphyletic group with two well differentiated morphological types being present among the oldest known finds. A branch that displayed fructifications with numerous, up to 18, vertically arranged enveloping cells and a second branch with up to 12 slightly dextrally twisted gyrogonites [8, 9]. The existence of dextral gyrogonites was long known although not particularly underlined for its interest for the study of asymmetry. Abundant and well preserved gyrogonites from the Ural region of Russia described in 1906 [10] and from Ukraine definitely solved the question of whether or not these microfossils should belong to the Charophytes [11]. Amazingly, the very first description of dextral gyrogonites from Indiana, USA, published as Moellerina greenii Ulrich 1886 [12], was long disregarded because the material had been attributed to the "Foraminifera" and had been figured incorrectly as a sinistrally coiled body. The types of this material were preserved in the United States National Museum. Reexamination by R. Brown finally revealed the spirals to be dextral [13]. The original author, "Dr. Ulrich explained that the sinistral spirals were an inadvertence, he having drawn the illustrations by camera lucida directly on lithographic limestone. Consequently, in printing, a reversed image of the original resulted" [13, 14]. According to the rules of botanical nomenclature. Peck and Morales rehabilitated Ulrich's Moellerina greenii as the type species. Most of the dextral forms are currently attributed to genus Moellerina [15]. A first synthetic chart by L. Grambast showed the great variety of morphological types [8, 9]- The general evolufionary trends in the Palaeozoic Charophytes were suggested to proceed from a primitive type with vertical enveloping cells, the Sycidiaceae Peck 1934, to the

368

Progress in Biological Chirality

dextrally (counter clockwise) coiled gyrogonites whereas the first known sinistrally coiled forms come from the Middle Devonian of Alberta, Canada [3]. This scheme has to be modified according to later finds from eastern Europe. Indeed, fructifications with vertical and dextrally coiled cells were already present concomitantly in the "Uppermost Silurian", in the same geological formation, the Skaly unit in Podolia, Ukraine, and thus appeared simultaneously [1, 2]. One type is genus Primochara^ represented by Phmochara calvata Ishchenko and Saydakovskiy 1974, whose gyrogonites are illustrated with counter clockwise helicity. The other type, represented by Praesycidium sjluricum Ishchenko and Ishchenko 1982, belongs to the order of Sycidiales, characterised by the presence of vertical and even segmented enveloping cells, a structure out of purpose for the present paper. Both types also occurred together in the Devonian of northern Russia (St. Petersburg and Ural region) [10]. The forms with vertical cells became extinct in the Early Carboniferous and were considered a non adaptive lineage [9]. The dextrally coiled Moellerinales persisted up to the Late Permian and coexisted with the sinistrally coiled forms. The global trend for dextral as well as sinistral forms was a reduction of the variability and a reduction of the number of enveloping cells forming roughly a chronological suite [8, 9, 16]. Dextral and sinistral Devonian gyrogonites were approximately the same size (Fig. 3). These non ornamented species display large variation in shape to the point that individual specimens can be found that appear the mirror image one to the other suggesting they could represent "enantiomorphic", dextral and sinistral variants of a same species [4]. The current classification of the Palaeozoic Charophytes makes use of categories of the number of spiral cells [15]. The dextral forms are divided into three main types: (/) a goup with 8 to 12 cells (genus Moellerina Ulrich 1886), (//) a group with 8-9 cells (genus Gemmichara Wang Zhen 1984), (///) a group with 5-7 cells (genus Pseudomoellerina Wang Zhen 1984). The Palaeozoic sinistrally coiled gyrogonites comprise three categories as follows: (/) a group with 8-13, mostly 10-12 enveloping cells (genus Eochara Choquette 1956), (//) a group with 6 or 7 cells (genus Palaeochara Bell 1922), (///) a group with strictly five spiral cells, (genus Leonardosia Sommer 1954 and the large family of the Porocharaceae Grambast 1962).

Figure 3. Comparison of Devonian gyrogonites in lateral view. a. sinistral Eochara wickenendi Choquette. clockwise twisted, ascending from lower right to upper left {bot.: sinistrorse); b. dextisA Moellerina greenii Ulrich, counter clockwise twisted, ascending from lower left to upper right (bot.: dextrorse); Scale bars 100 ^m

Charophyte Gyrogonites, the Result of Enantioselective Influence 250 Million Years Ago

369

The Porocharaceae, defined by the presence of an apical opening, provide the basis of all post-Palaeozoic Charophytes. Depending on the geographical area and on the classification used by individual authors, these early Charales were split into a number of different genera [17]. Given the tight resemblance of these forms, the distinction of so many taxa, however, seems not always justified and the systematics of these forms claims revision. At the end of the Palaeozoic, all dextral forms got extinct. Only the Charophytes with five sinistrally coiled cells came to survive into the Mesozoic Era. 2,3 Evolution of asymmetry in the Charophytes Asymmetry of the gyrogonites appeared with anti-clockwise arrangement of a variable number of enveloping cells. Given the high number of cells (8-12), only a slight torsion was necessary to cover the internal egg. The shape of each single cell corresponds roughly to a reversed S as shown by Eochara wickenendi and Moellerina greenii (Fig. 3). Progressive reduction of the number of cells increased the degree of twisting. The sinistrally coiled forms underwent the same process in the opposite direction and rapidly remained with a fixed number of five cells. A number of specimens with opposite handedness can be found that represent the mirror image one to the other suggesting they could represent enantiomorphous pairs [4]. It was also suggested that a single mutation could have been responsible for the reverse coiling direction [9]. The general evolutionary trend was toward reduction and fixation of the number of spiral elements leading to increased spiralisation. Whereas the dextral five-celled gyrogonites, classified as Pseudomoellerina, made only a short appearance in the Devonian, the five-celled sinistral gyrogonites appeared in the Carboniferous and became dominant in the Permian. From the PTB onwards no dextrally coiled gyrogonite has ever been found. The review of the distribution of the Palaeozoic Charophytes makes evidence that the sinistral forms expanded rapidly whereas the dextral ones regressed. Up to the Late Permian, dextral species with 8-9 cells and sinistral forms with five cells still coexisted in North China and grew even at the same place in Liaoning Province [18, 19]. Although the number of their spiral cells is different, their outline, size and shape are very close (Fig. 4). From the botanical standpoint the sinistral forms seem to have had better fitness or an adaptive advantage over the dextral forms. Possibly, plants of the dextral Gemmichara with a

Figure 4. Comparison of the latest coexisting dextral and sinistral Charophyte species, a. Gemmichara sinensis Z. Wang; b. Leonardosiajinxiensis Z. Wang. Both from the Late Permian in Liaoning Province, North China. Redrawn from photographs [18,19]

370 Progress in Biological Chirality variable number of cells were less inter-fertile than the strictly five celled species. Indeed, it cannot be established if the gyrogonites with 8 cells and those with nine cells could be produced by the same plant or if they correspond to different plants of the same species. In contrast to examples of chiral structures in animal species Hke Foraminifera or snails, the coiling direction of the gyrogonites suffers no exception. The five-celled sinistrally coiled structure of the gyrogonites obviously represents an inheritable coiling pattern. However, the molecular background of the phenomenon has not yet been investigated.

3. Palaeogeographic Distribution of Palaeozoic and Triassic Charophytes During the Devonian period, the dextral forms were largely dominant. Their known distribution concentrates on the palaeo-equatorial latitudes (Fig. 5) The type species of the earliest dextral Charophytes, Moellerina greenii wasfi-equentin the northern states of the USA [14, 22-24] where it was already present in the Late Silurian [25]. Abundant material was also described from northern and eastern Europe including Russia and Ukraine [10, 11]. The northern most occurrence was found in Spitzberg [26, 27]. Finds in Asia were hitherto located only in South China [19, 28]. Devonian sinistrally coiled gyrogonites are known only from today's Alberta, Canada [3] located at that time at the western rim of the Euramerican landmass. During the following Carboniferous period, sinistral forms occupied continental waters in Canada and the Unites States [14] and expanded also over North China [19]. One notes the discrete appearance of the five-celled sinistral Porocharaceae. Charophyte finds are rare in the Carboniferous. This was worldwide a period with extensive wetlands. However, the water must have been acid due to the rotting of high amounts of organic material (the basic material for the coal measures) creating unfavourable condition for the growth and preservation of calcareous organisms.

Ancient Landmass C I _ 3

yt

Dextral gyrogonites

Modem Landmass (^[_J)

O

Sinistral gyrogonites

Figure 5. Palaeogeographical distribution of known Devonian Charophyte localities. Base map for palaeogeography and plate positions 390 My ago redrawn and simplified from the PALEOMAP-project [20, 21]

Charophyte Gyrogonites, the Result of Enantioselective Influence 250 Million Years Ago 371 Siberia )

^ \ \

rS\Gfe^\and i/^^Kazalmstan

\,J\J

PANTHALASSIC

i

OCEAN

\.

/-

\x

j ^ i ^ .
PALEO-TETHYS OCEAN

1 1

^

1 South O * ^ ^ Africa 1^ America " I

J/ /

SEAsia

Ancient Landmass l i i j l p

^

Dextral gyrogonites

Modern Landmass (

O

Sinistral gyrogonites

/

\

[^Europe 7^

PANGEA

Sx A/ort/7 ^\Chir)a

North

p ^ QAmerL /

\ V X

/

Figure 6. Palaeogeographical distribution of known Permian Charophyte localities. Base map for palaeogeography and plate positions 255 My ago redrawn and simplifiedfi-omthe PALEOMAP-project [20, 21]

The following Permian period shows a large majority of localities with sinistral gyrogonites (Fig. 6). The main regions of distribution were the northern and central United States, Canada, Russia, Ukraine and Kazakhstan [23, 29]. A recent find extends them to the Himalayas [30]. Single occurrences in Brazil and Paraguay attest to the presence in the southern hemisphere [31, 32]. During the Permian, the five-celled forms, prefiguring the modem type, prevailed already in number over the ancient sinistral morphologies with six or more cells. Dextral gyrogonites persisted in North China. These finds are of prime importance as they revealed the coexistence of very similar dextral and sinistral forms up to the Late Permian [18, 33] (Fig. 4). At the Permian-Triassic boundary, the morphology shifted definitely in favour of the fivecelled sinistral gyrogonites. Triassic finds are then very frequent and provided abundant populations all over northern and eastern Europe as well as Kazakhstan and North China [29, 34-36]. The main area of distribution clearly was northern Eurasia. Several, although rather poor, materials were reported from the United States in Arizona, New Mexico and Oklahoma [23, 37, 38].

4.

The Permian-Triassic Biological Crisis The definite loss of the dextral gyrogonites coincides with a major event in the history of Earth, the Permian-Triassic boundary biological crisis (PTB). The PTB saw the most drastic mass extinction, generally estimated to the die-off of 90 percent of the marine and 70 percent of the terrestrial organisms. The currently admitted age of the PTB is 251± 3.5 Ma [39]. Explaining the process that caused the extinction is still speculative and has been the matter of diverse, sometimes contradictory hypotheses. Both internal, terrestrial, and external, say extraterrestrial, forces have been suggested. Among the frequently speculated causes are: - volcanic eruptions and subsequent drastic climate change leading to a "volcanic winter"

372

Progress in Biological Chirality ^

Siyeria

^

^ S ^ x-^^

Green

O

^

"~~-~-~-.

^ North /7"^***^^>?5>i^China

V}o

a /WT^^^^^PALEO-TETHYS V V S L ^ V ^ ^^ OCEAN j ^|»;|^ \L^Southem ( 1

A^&icay

VlCrt/na

\

\^SEAsia

/

PANTHALASSIC PANGEA ^

\ O C E A N I Soum \ America ^ \

" j h Africa R \

Modem Landmass (^

^

\

^

^

^

y

QOI

Z\yy^ Ancient Landmass Q J

\

Amsfctba

/_y^^ Dextral gyrogonites

-^— O

Sinistral gyrogonites

Figure 7. Palaeogeographical distribution of known Triassic Charophyte localities. Base map for palaeogeography and plate positions 237 My ago redrawn and simplified from the PALEOMAP-project [20, 21]

followed by a "greenhouse effect"; - increased continental drift and split of the super-continent Pangea that changed the balance between continental and oceanic masses and led to important sea level changes and oceanic anoxia [40-42]. Causes of astronomic origin were suggested with a supernova having approached the Earth near enough to produce gamma radiation. A recent hypothesis concerns a "major bolide impact" (asteroid or comet) and subsequent release of sulphur to the atmosphere [43]. Evidence for an impact event was also deduced from the presence of extraterrestrial gases in fuUerenes [44]. The crisis is still a matter of debate and its impact must be weighted depending on the group of organisms studied and the methods of analysis. Although mass extinction largely affected the land plants [45], recent analysis of global Permian and Triassic plant data led to argue against catastrophic events [46]. Concerning particular reptile groups, the impact was also noted to have had a minor effect [47]. In contrast, size-selective extinction and abrupt loss of large individuals was claimed for gastropods [48]. In so far as the Charophytes are concerned, the chronological succession of the dextral and sinistral populations can not obscure the fact that the change was more or less progressive. The sinistral forms expanded already in the Carboniferous and became dominant in the Permian, that is before the PTB. Nevertheless, the final consequence of the crisis on Charophytes was the definite extinction of the dextral forms. The question raises of which one of the causes cited in literature could have acted on the chirality of the gyrogonites. Drastic climate change from the Carboniferous glaciation to the Permo-Triassic hyper-arid conditions and even an abrupt cold-hot event looks unlikely to be able to affect the Charophytes. These plants occur from the northern polar circle to the Kerguelen Islands at latitude 54° S and even the boreal taxa among the living species produce invariantly sinistral fructifications in the form of organic oospores. Thus, temperature dependant coiling, like it was demonstrated for a number of Foraminifera species [49, 50] can be excluded. In contrast

Charophyte Gyrogonites, the Result of Enantioselective Influence 250 Million Years Ago

373

to frequently cited examples of animals with helical structure such as snails whose asymmetry can be subjected to variation [51, 52], the coiling direction of all post-Palaeozoic Charophytes has been perfectly constant. The interest of the Charophytes for the PTB crisis is that they just did not become all extinct but were subjected to strong enantioselective influence on their coiling direction. Detailed study of the Liaoning sections in North China, where similar dextral and sinistral gyrogonites coexisted up to the Late Permian, possibly could provide new insights into the extinction pattern of these ultimate dextral forms.

5.

Conclusion The Charophytes provide an example of biological chirality and show a phylogenetic transition from one direction of asymmetry to the opposite direction: from dextral to sinistral. The change occurred during the Palaeozoic and can be divided into several steps. From the Late Silurian to the Middle Devonian, only dextral forms have been recorded. In the Devonian, one single locality with a sinistral species is known. The Carboniferous period traced a progressive shift toward an increasing number of finds and species with sinistrally coiled gyrogonites. These forms became largely dominant in the Permian. Only a single dextral species remained in the Late Permian. The Permian-Triassic boundary crisis had a major impact on the chirality of the charophyte fructifications in so far it brought a radical and definite loss of the dextrally coiled gyrogonites. The selection of the sinistrally coiled type of asymmetry was then total and has suffered no exception up to Present, a time span of 250 Million years (Fig. 8). The geological history of the Charophytes raises the question of which one of the phenomena currently under discussion was likely to have had enough helical force to cause the loss of only the dextral forms during the PTB event.

PALAEOZOIC L.S.

200 My Gyrogonites counter clockwise

FTi

MESOZOIC 250 My

PERMIAN TRIASSIC 45 My 45 My

Present

Gyrogonites ONLY clockwise

Figure 8. Gyrogonite's helicity through time, from Late Silurian to Present. PTB, Permian-Triassic boundary; L.S., Late Silurian; M.D., Middle Devonian; My, Million years

374

Progress in Biological Chirality

Nothing is still known about the molecular background, as this conspicuous asymmetry has not yet been investigated from the biological point of view. It may be taken for ascertained that such a stable coiling pattern is directed by a genetic background.

6.

References

[1] T.A. Ishchenko and A. A. Ishchenko, New finds ofCharophytesfrom the Upper Silurian ofPodolia: Systematics and Evolution of Fossil Plants of Ukraine. Kiev Nauk. Dumka, publ., Kiev, 1982 (in Russian). [2J T.A. Ishchenko and L.Y. Saydakovskiy, Doklad Akad Nauk SSSR 220/1 (1974) 209-211 (English translation). [3] G.B. Choquette, Jowrw. Pal. 30 (1956) 1371-1374. [4] 1. Soulie-Marsche, Chirality in Charophytes: stability and evolution from 400 million years to present. In: Advances in BioChirality (Eds. G. Palyi, C. Zucchi and L. CagUoti) Elsevien Amsterdam, 1999, chap. 14. [5J Collins Enghsh Dictionary, HarperCollins Publ.. 2000. [6] C. Martin-Closas, Els Charofits del Cretaci inferior de les conquesperiferiques del bloc de VEbre. Unpubl. Drs thesis, Facultat de Geologia, University of Barcelona, 1989. [7] C. Martin-Closas, R.. Bosch and J. Serra-Kiel, Biomechanics and Evolution ofSpiralization in Charophyte fructifications: The evolution of Plant Architecture (Eds. M.H. Kurmann and A.R. Hemsley) Royal Botanic Gardens, Kew, 1999. [8] L. Grambast, C R. Acad Sci. Paris 249 (1959) 557-559. [9] L. Grambast, Taxon 23 (1974) 463-481. [10] A. Karpinsky, Memoires du Comite geologique n.s. 27, 1-86 (in Russian), 87-166 (in German), Eggers publ., St. Petersbourg, 1906. [11] W.N. Croft, Bull. British Mus (Nat. Hist.) Geology 1 (1952) 189-220. [12] E.G. Ulrich, Contrih. Amer. Pal. 1 (1886) 3-35. [13] R.W. Brown. ^cflJ. Sci. Jour. (36/10) (1946) 344-354. [14] R.E. Peck and G.A. Morales, Micropaleontology \ll^ (1966) 303-324. [15] Hui-nan Lu, I. Soulie-Marsche and Qi-fei Wang, Acta Micropal. Sinica 13 (1996) 1-12. [16] L. Grambast, Naturalia Monspeliensia 14 (1 %2) 63-86. [17] Hui-nan Lu. Acta Micropal. Sinica 14 (1997) 391-404 (in English). [18] Zhen Wang, Acta Micropal. Sinica 1 (1984) 49-60 (in Chinese with English summary). [19] Qi-fei Wang, Jing-lin Yang and Hui-nan Lu./lc/aMcro/767/. Sinica 20 {2003) 199-211 (inEngUsh). [20] C.R. Scotese, Atlas of Earth History, vol. 1: Paleogeography, PALEOMAP-Project, ArUngton, Texas, (2001). [21] C.R. Scotese, http://www.scotese.com (PALEOMAP website) (2002). [22] R E . Peck, Journ. Pal. 8 (1934) 83-119. [23] R E . Peck and J.A.Eyer,7oMrA7.Pfl/. 37(1963) 835-844. [24] H.M.A. Rice, Geol. Surv. Canada paper 70-60 (1970) 1-21. [25] J.E. Conkin, B.M. Conkin, G.S. Gregory and A.T. Hotchkiss, Univ. Louisville, Studies in Pal. andStratigr. n.3 (1974) 1-18. [26] R. Wrona,^c/fl Pal. Polon. 22 (1977) 289-296. [27] M. Feist and N. Grambast-Fessard, llOe congr. natl. Soc. savantes Montpellier, Sciences fasc. V (1985) 65-75. [28] Hui-nan Lu and Shang-zhen Zhang, Acta Micropal. Sinica 1 (1990) 9-17. [29] A. Kisielevsky and D.A. Kukhtinov, Vopr. Stratigr. Palaeontologii 3 (1978) 18-30. [30] D E . Chirananda, Current Science 84 (2003) 987-989. [31] F.W. Sommer, Paleontoi do Parana vol. comem. n. 1 (1954) 175-194, Curitiba, Parana, Brasil. [32] R, Herbst, Anais II Congr. latin, amer. Paleonto. (1981) 249-255 Porto Alegre. [33] Hui-nan Lu and Qi-xin Luo, Acta Micropal. Sinica 1 (1984) 155-170. [34] H. HomafRantzien, Opera Bot. 1 (1954) 1-83 Stockholm. [35] H. Kozur, Freiberger Forschungsber. 280 (1974) 1-56. [36] W. BiX^n, Acta Pal. Polon. 28 (1988) 63-161. [37] K. Kietzke, New Mexico Geol. Soc Guidebook, 38th Field Conf Northeastern New Mexico (1987) 119126.

Charophyte Gyrogonites, the Result of Enantioselective Influence 250 Million Years Ago

375

[38] K. Kietzke, New Mexico Geol. Soc. Guidebook, 40th Field Conf. Southeastern Colorado Plateau (1989) 181-190. [39] H. Yin, W.C. Sweet, B.F. Glenister, G. Kotlyar, H. Kozur, N.D. Newell, J. Sheng, Z. Yang, YD. Zakharov, Newsl. Stratigr. 34/2 (1996) 81-108. [40] D.H. Erwin, The Great Paleozoic Crisis. Life and Death in the Permian, Columbia Univ. Press, New York, 1993. [41] A. Hallam and P.B. Wignall, Mass Extinctions and their Aftermath. Columbia Univ. Press, New York, 1997. [42] H. Kozur, Palaeogeogr., Palaeoclimatol, Palaeoecol. 143 (1998) 221-212. [43] K. Kaiho, Y. Kajiwara, T. Nakano, Y. Miura, H. Kawahata, K. Tazaki, M. Ueshima, Z. Chen, G.R. Shi, Geology 29 (2001) 815-818. [44] L. Becker, J.R. Poreda, A.G. Hunt, T.E. Bunch, M. Rampino, Science 291 (2001) 1530-1533. [45] C.V. Looy, W.A. Brugman, D.L. Ditcher, H. Visscher, Proc. Natl. Acad Sci. USA 96 (1999) 1385713862. [46] P.M. Rees, Geology 30/9 (2002) 827-830. [47] S.P. Modesto, R. J. Damiani, J. Neveling and A.M. Yates, Journ. Vertebrate Pal. 23 (2003) 715-719. [48] J. Payne, Geological Society of America, Seattle annual meeting. Paper No. 157-11 (2003) abstract. [49] H. Hilbrecht, Mitt. Geol. Inst. ETHund Univ. Zurich n. F. n. 300 (1996) 1-93. [50] S. Galeoti and R. Coccioni, Palaeogeogr., Palaeoclimatol, Palaeoecol. 178 (2001) 197-210. [51] A.R. Palmer, Proc. Natl. Acad Sci. USA 93 (1996) 14279-14286. [52] M. Lewin, Left-right asymmetry in animal embryogenesis. In: Advances in BioChirality (Eds. G. Palyi, C. Zucchi and L. CagUoti) Elsevier, Amsterdam, 1999, chap. 12.

Progress in Biological Chirality G. Palyi, C. Zucchi and L. Caglioti (Editors) © 2004 Elsevier Ltd. Allrightsreserved.

Chapter 31 Chirality Transfer in the Formation of the Indole Alkaloids Derived from Secologanin Gyula Beke, Laszlo Karolyhazy, Agnes Patthy-Lukats, Benjamin Podanyi, Laszlo F. Szabo* Department of Organic Chemistry, Semmelweis University, Hogyes utca 7, Budapest, H-1092 Hungary szalasz@szerves. sote. hu

1.

Introduction Chirality is one of the characteristic properties of most natural products, which generally are available from biological sources only in one of the two possible enantiomers, i.e. the life is homochiral. One of the most important classes of non-polymeric natural products is that of the monoterpenoid indole alkaloids. This unique group of compounds comprising more than 2000 individual structures [1] (some of them are shown in Figure 1, and subsequent figures and schemes) has a strong unity and a large variety. The source of the unity is shown by the fact that all of them are produced by Nature from two building blocks, i. e. the biogenic amine tryptamine (1) (or the amino acid tryptophane) and the monoterpenoid glucoside secologanin (2). Most of them are available from three plant families only, i. e. Rubiaceae (including Naucleaceae), Loganiaceae (including Strychnaceae) and Apocynaceae, [2] and common elements can be detected in their structure. However, this unity is paired with a high diversity in structures and properties, due to the presence of several stereogenic elements and fiinctional groups which result in a large variety of fragmentations, cyclizations, and rearrangements. [3] In the second part of the 20* century, the common effort of a high number of scientists gave a rich harvest: many interesting indole alkaloids were isolated, their structure elucidation contributed strongly to the development of the ^H and ^^C magnetic resonance spectroscopy and mass spectrometry, ingenious concepts and highly efficient preparative methods were elaborated for their synthesis, and new potent drugs were introduced into the therapy. Also, the biogenesis of these compounds was intensively investigated by using isotopically labelled compounds, isolated enzymes, tissue cultures and methods of biological engineering. [4, 5, 6, 7] As a consequence, it was possible to detect the main lines of their biosynthesis. However, in spite of this huge effort, there are still large white spots in the details of their formations in the plants. The investigation of their biogenesis is rather difficult, because it involves a long series of derivations, where the test samples are hardly available either from natural sources or by synthesis, the structural differences are small, and the separation of the

378

Progress in Biological Chirality

O HsCOOC^s HJCOOC 7 rauniticine pseudoindoxyl, 3 a rauniticine oxtndoie B 3-isorauniticine pseudoindoxyl, 3p

5a (-)-ajfnalicine, 3a,20p,21a 5b rauniticine, 3a,20a,21p 5c isorauniticine, 3p,20a,21p

17 (-)-ibogamine

^ } I

I "H HgCOOd'''

(+)-p8eudotabersonine

I

22 H 3 C ^ (+)-vincamine

--

QV 23 (+)-eburnamonine

Figure 1. Structure of some selected indole alkaloids

tacanrine

Chirality Transfer in the Formation of the Indole Alkaloids Derived from Secologanin

379

intermediates is complicated. The large databases, especially the Dictionary of Natural Products [1] mainly contributed to the collection of the resuhs, and now constitute a rich source for summarizing details and detect common features. The aim of the present work is to follow the change and transfer of some characteristic structural elements in the different chemical structures and throughout the different plant species. Their special properties, as well as the high number of individual compounds, many of them repeatedly isolated from the same or different plant species, encouraged this approach as well. One of the most special characters of these alkaloids is the chirality, which is tightly connected to their genesis.

2.

Preliminary Remarks Interestingly, there are plant species, which produce a large spectrum of many different types of indole alkaloids. From Catharathus roseus (= Vinca rosea) 43, and Rhazya stricta 61 indole alkaloids were isolated which represent nearly all the main types of structures. In contrast, from Strychnos ngouniensis or Stemmadenia species, only a few, but special alkaloids were obtained, which completed fortunately the collection of the structures mentioned previously. Of course, the results of the isolations depended strongly on the applied techniques and the plant species were selected often by chance. Therefore, in many cases, the absence of a compound in a species can not be significant. However, the copresence of certain structures in the same or closely related species may indicate a common source, or a certain biogenetic connection. The same is true for "dimeric" indole alkaloids in which two different types of substructures are connected by a covalent bond (Figure 2). Anyway, in the evalution of such data, one should be cautious, and the common origin should be advantageously confirmed, e.g. by incorporation of labelled compounds. Really, isotopically labelled ajmalicine supported not only the common biogenetic relation, but the appearance of the labelled compounds in time indicated the approximate order of their formation as well. Chemical ("biomimetic") transformations can not prove the real biochemical transformations, but can make them probable or reasonable. The schemes presented in this paper do not intend to show real mechanisms, but to indicate more or less logical or probable structural relations. Our investigations were encouraged by eariier efforts of M. Hesse and his group. [8, 9] However, their approach was based mainly on the number of the component individual rings, which seems to be a rather mechanical aspect. Our aim was to find biogenetic or at least biomimetic correlations, and not to substitute the experimental investigations, but to detect some structural connections which can confirm and complete previous or future results Some conscious limits were put to our work: a) as database, the Dictionary of Natural Products version 11.2 [1] was used, occasionally completed by the Chemical Abstracts and the Beilstein Cross Fire; b) only indole alkaloids containing the tryptamine and the secologanin subunit in 1:1 ratio ("monomer") were considered (i. e. the "sesquimers" having a 2:1, and the "dimers" having a 2:2 ratio were neglected); c) compounds in which the tryptamine subunit is changed or rearranged, were likewise temporarily disregarded. This means that the present paper includes only the monomeric alkaloids having the tryptamine subunit in unrearranged form. Therefore some important groups of tryptamine-based alkaloids (e. g. the ulein and olivacine alkaloids, as well as the Cinchona and Camptotheca

380

Progress in Biological Chirality ^^=\

^^-^

getssospermme (corynan and strychnan) la

MeOOC

capuvosidine (vobasan and isoplumeran) la lip

MeOO(

14,15-didehydrotetrastachy nine (ibogan and plumeran) II

Hip

H3C— tenuicausine (plumeran and eburnan) H 1116 ip

Ilia

Figure 2. Examples of dimer indole alkaloids

alkaloids) were omitted. In the following, the indole alkaloid name will be used according to these temporary limitations. As it was not possible to follow the transfer of all centers of chirality, our efforts were concentrated on those of the basic skeletons.

3.

A Simplified System of Indole Alkaloids Derived from Secologanin For our discussions, a simplified system of these alkaloids was constructed and is shown in Figure 3. [10, 11, 12] The system is based on three fundamental skeletons of type I, II and III, and on the formation or cleavage of further cycles (Figures 4 and 5). The cleavage of the strategic bond C-15-C-16 in type I skeleton provides the possibility of the formation of the type II and III skeletons. In addition, the main sites of transformations are N-1 and N-4, as well as C-2 and C-7 (i. e. a and P positions of the indole ring) in the tryptamine subunit, C17, C-19, C-21 and C-22, as well as C-16 in the secologanin subunit. Alone the cyclizations inside the secologanin subunit and with one of the nitrogen atoms may result in the formation of 25 tetra- and pentacyclic ring systems shown in Figure 6 [14], many of which appear in the structures of the individual natural products. In the cyclizations, mainly the formation and cleavage of the carba- and azacycles (Figure 4) were considered, as the oxacycles are less

Chirality Transfer in the Formation of the Indole Alkaloids Derived from Secologanin

type il sketeton

3 81

type III skeleton

type I skeleton (in secotoganin)

isoibogan (III) 1

2

17

/18

isopiumeran (II p) 14

strychnan (ip) 432

plumeran (III p) 344

isoebuman (II a) 8

aspldospermatan (I p) 26

eburnan (III a) 56

Rgures Indicate approximate numbers of Isolated alkaloids in Dictionary of Natural Products (Chapman & Hall/CRC Press. Versfon 11/2, 2003)

Figure 3. Main types of indole alkaloids derivedfromsecologanin

characteristic in this respect. The skeletons type II and III can be derived from type I by rearrangements of the carbon framework of the secologanin aglucone subunit. The type I skeleton of the alkaloids, which are formed directly from tryptamine and secologanin in a Mannich-type reaction, contains the carbon framework of secologanin in unrearranged form. In the type II and III skeletons, the carbon framework is rearranged during the biogenesis, i. e. formally the C3 subunit formed by cleavage of the bond C-15-C-16 is transposed through its C-17 atom to C-14 in type II skeleton, and to C-20 in type III. The center of the secologanin subunit is C-15, to which C-3 and C-21 are in equal distance. In the

382

Progress in Biological Chirality

azacyclizations

OCH3

N->C-17 N->C-21 N-»C-22

N->C-19 H

oxa(carba)cyclization8 C>21->C-17 0-17->C-21

0-17->C-19 C-18-^C-17

Figure 4. Possible primary cyclizations in indole alkaloids

o

o

decarboxylation

O

O

cleavage of bond C-4->C-5

O

O

deformylation

in the indole alkaloids, bond C-5 ^ C - 4 is equivalent to C-15->C-16 05^06 C-15->C-14 C-5->C-9 C-15^020

Figure 5. Main types of fragmentations in secologanin derivatives

OCH3

Chirality Transfer in the Formation of the Indole Alkaloids Derived from Secologanin

383

Figure 6. Primary cyclizations of the vincosan skeleton

aglucones (i. e. in most alkaloids) these analogous carbon atoms can be attached either to atom C-2 or C-7 (i. e. to a or (3 position of the indole ring, respectively), providing the possibility of a fiirther level of classification. However, in several alkaloids, neither C-3, nor C-21 is connected to the carbon atoms of the indole ring (seco alkaloids). The biogenesis of all indole alkaloids starts with the Mannich-type reaction of secologanin and tryptamin and results in the formation of the type l a strictosidine (3a) and vincoside (3b). Only the formyl group of secologanin can take part in this reactions, because the other potential reaction sites are blocked in glucoside form. Therefore, in type la alkaloids which correspond to the vincosan skeleton, only C-3 may be connected to the a position of the indole ring. All other alkaloids are formed, mainly after deglucosylation, by further cyclizations (some of them are shown in Figure 4) and fragmentations (shown in Figure 5). The alkaloids of the vincosan skeleton constitute the largest group of indole alkaloids.

384

Progress in Biological Chirality

having more than 1000 representatives (indoxyl derivatives are included) isolated from all the three plant families mentioned above. Characteristic alkaloids of this group are ajmalicine (5a) and yohimbine (4). By further cyclizations several subgroups are formed, but 80 % of the individual alkaloids of type l a contain the corynan skeleton (in Scheme 1) which is only one of the possible 25 skeletons shown in Figure 6. [14] In the type Ip skeleton either C-3 or C-21 of the secologanin subunit may be attached to the p position of the indole ring. This skeleton is found in the strychnan and aspidospermatan alkaloids. The strychnan skeleton may be directly derived from the vincosan skeleton by attachment of ligand C-16 to C-2 coupled by simultaneous shifting of the ligand C-3 from C-2 to C-7 position. This skeleton is the basis of the second largest group of indole alkaloids (432 isolated compound; the oxindole derivatives are included). They have a great variety, with strychnine (10 in Figure 1) and akuammicine (29 in Scheme 2) as typical representatives. Most of these alkaloids were isolated from the Loganiaceae/Strychnaceae and partially from the Apocynaceae families. In contrast, the aspidospermatan alkaloids represent a small, well defined group of alkaloids (26 individual compounds) which were mainly isolated from the Apocynaceae family, with condylocarpine (30 in Scheme 2) as a characteristic member. Interconversion of the two subgroups will be discussed later.

1,2-rearrangement

1,2-rearrangement

22^16-17 oxindole 3S :3/? ratio is - 3 :1

indoxyl 3S configuration only

dihydroxy intermediate

^

T

H3CO' O H pleiocarpaman 3S configuration only

akuammmilan 3S configuration only

corynan skeleton 3S :3R ratio Is ~ 3 : 1

\

O taltx>tan 3S configuration only

yohlmban 3S :3R ratio is - 3 :1

sarpagan 3S configuration only

Scheme 1, Main types of cyclizations in type la indole alkaloids

Chirality Transfer in the Fomiation of the Indole Alkaloids Derived from Secologanin

-H^ H3CO 31 O preaiaiammjcine -CH2O

akuammicine

H ^ ^ ^ H3CO O 31a (X=OH) hypothetic stemmadenine derivatives red.

385

precondylocarpine a|-CH20

N^and red.

31b

32b

stemmadenine

1 Sp-H-stemnrradenine

condylocarpine

Scheme 2. Transition from type la into type ip indole alkaloids

The formal difference between the two rearranged skeletons type II and III is the position of the C2 unit, which is attached to a simple cyclic carbon atom (C-14) in the former, and to a bridgehead carbon atom (C-20) in the latter case. For derivation of the two types, it was supposed that after cleavage of the bond C-15-C-16, conjugated double bond systems involving either C-3 or C-21 are developed, and this property gives again the possibility for fiirther classification. C-17 is attached to C-14 in the type II skeletons (ibogan amd isoplumeran alkaloids), whereas to C-20 in the type III skeletons (plumeran and isoibogan alkaloids). The second bond is formed in the ibogan and isoibogan compounds at C-16, in the plumeran and isoplumeran compounds at C-7. It can be seen that in (iso)plumeran skeleton all cycles are fused ("fused system"), while the (iso)ibogan skeleton contains also bridged cycles ("bridged system"). As C-7 represents the P position of the indole ring, the isoplumeran and plumeran skeletons may be indicated as type Iip and IIip, respectively. In these structures a rearrangement may take place by cleavage of bond C-2-C-16 with simultaneous shifting of the ligand C-3 or C-21 from position C-7 to C-2, i. e. to a position of the indole ring, which resuhs in the formation the isoeburnan (type Ila) and ebuman (type Ilia) skeletons, respectively. Formally, this rearrangement may be considered as the retro process of the transformation of the vincosan (type la) into the strychnan (type IP) skeleton.

386

Progress in Biological Chirality

Interestingly, the number of individual alkaloids according to these "rearranged" skeletons is widely variable. The normal skeletons (plumeran, ibogan, ebuman) are found in definitely higher number of alkaloids (plumeran: 344, ibogan: 84, eburnan: 56), than the iso skeletons (isoplumeran: 14, isoibogan: 1, isoeburnan: 8). The plumeran alkaloids (type IIIP) constitute the third largest group of indole alkaloids having several subgroups, with (-)-tabersonine (19) and (+)-aspidospermidine (20) as typical representatives. Further characteristic alkaloids are (-)-coronaridine (16), (-)-ibogamine (17) and (+)-catharathine (15) in the ibogan (type II), (+)vincamine (22) and (-)-ebumamonine (23) in the eburnan (type Ilia), and pseudotabersonine (21) in the isoplumeran (type Iip) group. The small group of isoibogan (type II) and isoeburnan (type Iip) alkaloids contain only a single parent compound, namely 16hydroxyalloibogamine (17) and tacamine (24), respectively (in the latter case with 7 derivatives).

4.

The Source of Chirality in the Indole Alkaloids at C-15 The unique source of chirality in all indole alkaloids is the C-5 center of secologanin, because the chirality of the other two stereocenters of the precursor is lost after deglucosylation, and tryptamine is achiral. Previous studies suggested that secologanin is formedfi-omacetylcoenzyme-A through mevalolactone. However, recently it was proved, that its biosynthesis runs fi-om pyruvic acid through the hypothetical C5 saccharide 1-deoxy-Dxylulose-5-phosphate. [15, 16, 17] In the coupling reaction of tryptamine and secologanin the S chirality of C-5 of secologanin is transferred into the S chirality of the C-15 center of the indole alkaloids. C-15 is rarely involved in subsequent steps of the biosynthesis. Therefore, most type I indole alkaloids are homochiral at C-15. However, there are two important exceptions. Introduction of unsaturation at C-15, and cleavage of the C-15-C-16 bond, i. e. the key step in the formation of the rearranged skeletons cause the destruction of the chirality at C-15. Nevertheless, even in most type II and III alkaloids, some preference of chirality may be observed, probably as a consequence of the chiral intermediates or transition states (see later).

5.

The Source of Chirality of C-3 in Simple (Fused Polycyclic) Type I Alkaloids It was established by Battersby et coworkers, that, under achiral conditions, the coupling reaction of tryptamine (1) and secologanin (2) afforded both strictosidine (3*S) (3a) and vincoside (3R) (3b) in about equal ratio (in the original paper [18] the configuration of the new center of chirality was given to be opposite, but later it was corrected [19]). The reaction is catalyzed by strictosidine synthase in favour of strictosidine over vincoside [20]. We showed that the stereoselectivity of the enzyme is complete, and formation of vincoside even in traces could not be demonstrated. [21] However, the R configuration at C-3 according to vincoside does appear in about 20% of simpler vincosan alkaloids having no bridged ring. Zenk and coworkers stated that their ". ..resuhs demonstrate unequivocally that strictosidine is the common biosynthetic precursor of alkaloids with 3a as well as 3P heteroyohimbine alkaloids." [22] Unfortunately, the mechanism of the epimerization and the source of the chirality were not established. In addition, we proved that strictosidine synthase does not catalyze the Mannich reaction with Nb-substituted tryptamines. In absence of the enzyme, ligands at Nb of tryptamine increase the selectivity in favour of the 3R epimer [23]. In an ana-

Chirality Transfer in the Formation of the Indole Alkaloids Derived from Secologanin

387

logous reaction of secologanin with dopamine, the enzyme deacetylipecoside synthase was isolated, which favours likewise the R configuration at the analogous center of chirality [24].

6.

Chirality Transfer in Bridged Polycyclic Type l a Alkaloids In the large group of alkaloids corresponding to the corynan skeleton, further cyclizations may take place between centers of the tryptamine and secologanin subunits; some of them are shown in Scheme 1. In the most important ring closures, a covalent bond is formed between C-16 and N-1, C-5 or C-7 corresponding to the pleiocarpaman, akuammilan and sarpagan skeletons, respectively. All alkaloids of these types are homochiral at C-3 and C-15, i. e. both H-3 and H-15 have cis a orientation in the usual representation, which corresponds to S configurations in strictosidine. This stereochemistry is evidently governed by the stable S configuration of C-15, because the appropriate six-membered bridged ringsystem may be formed only in cis position of the two H atoms. Moreover, this fact indicates, that the additional cyclization follows (and does not preceed) the cyclization to N-4. In 3R alkaloids the bridged rings can not be formed. In addition, 3S stereoselectivity was observed also in talbotine (6 in Figure 1) and its derivatives, ahhough the newly formed ring is fused, rather than bridged. However, N-4, having no ligand and being a stronger nucleofile than N-1, is ready for cyclization. These facts suggest that the talbotan type compounds are formed through the pleiocarpaman derivatives, i.e. the first, reversible cyclization takes place at N-4, followed by cyclization at N-1 and finished by re-opening of the previously formed ring at N-4. In this case, cyclization to N-1 provides a temporary bridged system, which can be formed only, if C-3 has S configuration. It should be remarked that the pleiocarpaman and the talbotan derivatives were isolated from closely relative species (e. g. Pleiocarpa picnantha and Pleiocarpa talbotina, respectively). Unfortunately, no experimental data are known concerning the exact chemical or biochemical mechanism of these cyclizations.

7.

Chirality Transfer in Type l a Alkaloids Having a Connection to N-1 In the biogenesis of most indole alkaloids, after the coupling Mannich reaction and deglucosylation, the next step is a further cyclization at N-4. However, there is a special, moderately sized group in which the cyclization takes place at N-1 rather than N-4. This group contains, in addition to the pleiocarpaman and talbotan alkaloids mentioned above, some simpler alkaloids represented by correantoside and kribine (11 and 12 in Figure 1). In most cases, C-17, C-19, C-21 or C-22 (over C-16) of the secologanin subunit takes part in this cyclization. Because of the higher nucleophilicity of N-4 compared to N-1, this cyclization may be expected only, if some factors hinder the ring closure at N-4. Indeed, in these alkaloids (except talbotine and its derivatives which were discussed previously) N-4 is methylated or involved in an unsaturated (aromatic) system. In order to be effective, the methylation (and even the introduction of the unsaturation) should preceed the coupling reaction. The configuration at C-3 is *S' in kribine and its derivatives from Stiychnos elaeocarpa (Strychnaceae), and R in correantoside and its derivatives from Cephaelis corraea (Rubiaceae). This latter species is a close relative of Cephaelis ipecacunha (Rubiaceae), from which, as mentioned previously, an enzyme catalyzing the Mannich reaction in favour ofR

388

Progress in Biological Chirality

configuration at the analogous center of chirality was isolated. Unfortunately, in the present case nothing is known about the source of the chirality. Nevertheless, as likewise mentioned previously, strictosidine synthase does not catalyze the coupling reaction of Nbmethyltryptamine. 8.

Chirality Transfer in Structural Patterns Common in Fundamental Skeletons There are some structural patterns which appear in more than one of the fundamental skeletons, such as the oxindole and indoxyl substructures shown in Scheme 1. Simple representatives of the oxindole alkaloids can be prepared by the coupling reaction of secologanin with 2-oxo-2,3-dihydrotryptamine. [25, 26] However, it seems that in the formation of these alkaloids, the biogenetic route passes through 2-hydroxy, 7-hydroxy and/or 2,7-dihydroxy derivatives, which can be prepared from the type la compounds with hydrogen peroxide in acetic acid and by tuning the reaction conditions appropriately. [27, 28] Both the oxindole and indoxyl systems are formed by 1,2-rearrangement of the C-3 ligand from the a to the P position of the indole ring or in the opposite direction, respectively. Typical representatives are rauniticine oxindole B and rauniticine pseudoindoxyl (7 and 8 in Figure 1). Formally, most of the oxindole and related alkaloids (212 compounds) may be considered as derivatives of the 2,16-5^costrychnan (type IP) skeleton, with exception of 3 compounds which have the type II skeleton (ibogan or 14,17-5ecoibogane skeleton: crassanine, kisantine, tabemoxidine) in their structure. However, in the latter compounds C-3 is not chiral. In the type I oxindole alkaloids C-3 is chiral, and the original configuration should be retained according to the well-known fact that the migrating ligand keeps its configuration in the 1,2rearrangement. Indeed, in this group the 3S/3R ratio is close to that observed in corynan alkaloids. However, in the oxindole alkaloids the original configuration may be modified subsequently by a ring-chain tautomerism. In the group of the indoxyl derivatives comprising nearly 30 alkaloids, representatives of both the vincosan (type la) and the ibogan (type II) skeletons were isolated from plants. In several cases, also the 7-hydroxy-indolenine or 2,7-dihydro-2,7-dihydroxy derivatives as supposed precursors of the 1,2-rearrangement could be isolated from the same species or from close relatives of the same plant family. The configuration at C-3 is retained in all cases according to the mechanism. In this ring system subsequent epimerization at C-3 is not possible. From Uncaria elliptica both the 35* and 3R epimers of rauniticine pseudoindoxyl (8) were isolated, together with their possible educts, the 35 and 3/^ epimers of rauniticine (5a and 5b). In addition, also the 3S epimer of rauniticine oxindole B (7) was isolated from the same species. The last remark lets also to suppose that the indoxyl and the oxindol alkaloids may have a common precursor (e. g. a 7-hydroxyindolenine derivative). 9.

Chirality Transfer from the Type la to the Type ip Skeleton As mentioned previously, the coupling reaction of secologanine with 2,3-dihydro-2oxotryptamine afforded type ip oxindole alkaloids, which could be transformed into type ip strychnan alkaloids themselves (Scheme 1). [6] However, it was demonstrated that in plants, among others, ajmalicine (5a) (type la) was incorporated into akuammicine (29 in Scheme 2) (type IP). Both types of alkaloids (type la geissoschizine and type ip bharginhine) were isolated from the same plant species Rhazya stricta. In addition, several "dimers", e.g.

Chirality Transfer in the Formation of the Indole Alkaloids Derived from Secologanin

3 89

geissospermine (25 in Figure 2), having both a vincosan (type la) and a strychnan (type ip) subunit, were obtained from Geissospermum vellosii. These facts suggest the direct formation of the alkaloids of the strychnan subgroup from the type I vincosan structures, (Scheme 2) without an intermediate oxindole structure. Ajmalicine (5a) (isolated from Vinca rosea and many Ranwolfla species) may be served as an educt of the biosynthesis, and ajmalicine hydroxyindolenine (9) (isolated from Catharanthus roseus tissue culture) as an intermediate. This latter compound can provide the appropriate type ip product by 1,2-rearrangement. In spite of the mixed configuration of C-3 in the type ip oxindole alkaloids (see above), all strychnan alkaloids have S configuration at C-3, due again to the stereochemistry of the rearrangement. One dubious case was later declaired. [29] According to the conformational analysis, the steric requirement of the rearrangement is again the cis orientation of H-3 and H15, which, taking into account the stable S configuration of C-15, is fiilfilled only by the S configuration of C-3 in the type la educt. 10. Chirality Transfer from the Type ip Strychnan to the Type ip Aspidospermatan Alkaloids The aspidospermatan compounds, i.e. a subgroup of ip indole alkaloids, are stereochemical counterparts of some alkaloids of the strychnan subgroup, having a similar chemical structure but opposite orientation of the ligands at C-3 and C-21 (Scheme 2). As mentioned above, typical representative of the strychnans is akuammicin (29) (from several Vinca species) and that of the aspidospermatans is condylocarpine (30) (from Vallesia antillana and Diplorhynchus condylocarpon). Also, the possible precursors preakuammicine (31) (from Vinca rosea) and precondylocarpine (32) (from Vallesia dichotoma) were isolated. According to the proposition of A. I. Scott [6], the intermediates between the two subgroups should be compounds in which the bond C-3~C-7 or C-7-C-21 is cleaved by intramolecular interaction between N-1 and N-4. (31a) and (32a) may be considered as 3,75ecostrychnan or 7,21-5^coaspidospermatan derivatives. Although these hypothetical and probably unstable intermediates not, but their reduced derivative stemmadenine (31b=13) and recently [30] its C-15 epimer (32b) were isolated from biological sources. 14P-hydroxy-(29), (30) and (31b) were obtained from Diplorhynchus condylocarpon, (32b) from Tahernaemontana heyneana. The mechanism of the intercorversion of (31a, X=OH) and (32a) can be easily interpreted by turning of the secologanin subunit around bonds C-5~C-6 and C-15-C-16. By this rotation, of course, the configuration of C-15 is retained, although its orientation in the usual representation reversed. As the annellation of the reclosed central ring system is as tight as that of the starting one, H-3 (in strychnans) and H-21 (in aspidospermatans) have to be in cis relation to H-15. At this point it should be mentioned, that in the Stemmadenia, Vallesia, Diplorhynchus, Rhazia and Vinca species the type ip alkaloids were found together with type la as well as type II and IIIp alkaloids. This fact suggests that the group of the Ip alkaloids (involving also the stemmadenine-type seco-compounds) is the branching regio toward the alkaloids of the skeletons II and III.

11. Transition from Type ip into Type H and Type m p Skeletons In contrast with the type I alkaloids, the type II and III alkaloids were obtained, with a

390

Progress in Biological Chirality

single exception, only from the Apocynaceae family. From stmctural point of view, the essence of the changes providing the rearranged fundamental skeletons is the transposition of the C3 fragment of the secologanin subunit from C-15 to C-14 in type II, and to C-20 in type III skeleton (Scheme 3). The structure of the alkaloids reveals that the site of attachment in the C3 fragment should be changed from C-16 to C-17 as well. Of course, the rearrangement involves the cleavage of the bond C-15-C-16 of type I skeleton by appropriate fragmentation and remodelling of the unsaturated system. Unfortunately, in spite of the huge experimental work and sharp theoretical discussions [5], until now no firm facts are known concerning the exact mechanism of these transformations.

22 E ^^ ibogan (10 84 alkaloids C-17 in p orientation w ith a single exception

^

33B

20->17 1g-»9»

E isoibogan (III) a single representative is know n only

22E plumeran (III p) 344 alkabids acyclic ligands of C-20 and C-21 are in cis a or cis p orientatk>n

22E isoplumeran (lip) 14alkak>ids hydrogens of C-3 and C-14 are in cis p orientatnn

J

L Only one of the enantk>niers is show n

E = COOCH3 isoeburnan (H a) 8 alkak>kls hydrogens of C-3 and C-14 acyclic Hgands of C-20 and in cis p orientatk>n C-21incisaorcisp orientation

Scheme 3. Transitionfromtype ip into type II and III indole alkaloids

Chirality Transfer in the Formation of the Indole Alkaloids Derived from Secologanin

391

However, according to A. I. Scott [6], the changes can be interpreted by Diels-Alder reactions in secodine-type intermediates (33). In Scheme 3, it can be clearly seen that to each of the fundamental skeletons to be formed (ibogan and isoibogan, plumeran and isoplumeran) a hypothetic secodine precursor (33A, 33B, 33C and 33D, respectively) is paired. (33A) and (33C), as well as (33B) and (33D) can be mutually transformed by rotation along the N-4-C-5 bond. However, the mutual transformation of (33A) and (33D) as well as (33B) and (33C) requires proton-catalyzed isomerization of the conjugated double bonds in the partially hydrogenated pyridine ring, and it needs certainly more activation energy than the rotation along the C-N single bond. According to A. I. Scott, the starting point of the transformations may be preakuammicine (29 in Scheme 2), which could be fragmented to the hypothetic stemmadenine intermediate (31a, X=OH). Its acetylated derivative (31a, X=0C0CH3 in Scheme 3) could be more easily fragmented further to the hypothetic secodine derivative (33) which after reduction and migration of one of the double bonds would afford the first immediate secodine intermediate (33A). As (33A) may be readily transformed by rotation into (33C), the ibogan and plumeran alkaloids can be formed easily by a Diels-Alder cyclization. In precondylocarpine, the double bonds of the secologanin subunit are not in appropriate position for an easy fragmentation, and it can not provide a precursor (33B) or (33D) for Diels-Alder cyclization. However, the 33A and 33C can be isomerized more or less easily into 33D and 33B, respectively. These fine structural details may explain the fact that the number of the normal (plumeran and ibogan) alkaloids is much higher than that of the iso (isoplumeran and isoibogan) ones. The main experimental observations concerning this transition are summarized in the following statements: a) Although the hypothetic secodine intermediates themselves not, but 16,17-dihydrosecodin17-ol (14 in Fig. 1) and related compounds were isolated from Rhazya orientalis and relative species. b) Also, several "dimers" having both type II and III subunits, e. g. 14,15didehydrotetrastachynine from Stemmadenia grandiflora (27 in Fig. 2) were isolated. This fact suggests the common origin of type II and III alkaloids. Another dimer alkaloid, capuvosidine (26) (from Capuronetta elegans) having la and Iip subunit, confirms the biosynthetic connection of type I and II alkaloids. c) The cleavage of the strategic bond C-15-C-16 in vincoside derivatives (or even of the bond C-4-C-5 in simple secologanin derivatives) was experimentally proved by us. [31] d) As shown in Scheme 3, in the formation of the rearranged skeletons the supposed DielsAlder reactions would require the conservation of the original double bond C-2-C-7 (in ibogan and isoibogan alkaloids), C-14-C-15 (in plumeran alkaloids), and C-15-C-20 (in the isoplumeran alkaloids), as well as the appearance of a new double bond C-2-C-16 (in the plumeran and isoplumeran alkaloids), C-14-C-15 (in the isoibogan alkaloids), and C15-C-20 (in the ibogan alkaloids). With the exception of the isoibogan skeleton, several representatives of all other skeletons were isolated from natural sources, which fulfill these expectations for unsaturation. Of course, during the subsequent steps of the biosynthesis, partial saturation could take place, so the expected double bonds did not appear in all alkaloids. In the isoibogan group, only a single compound was isolated from a single species, and this is not sufficient for evaluation of chirality transfer. e) In alkaloids which have a fused ring system involving C-3 and C-14 (isoplumeran series)

392 Progress in Biological Chirality or C-20 and C-21 (plumeran series) as common atoms, the ligands (H-3 and H-14 or C-19 and H-21, respectively) are in cis orientation. The Diels-Alder reaction of secodine derivatives would favour this oriention, which is thermodynamically less favourable, f) The application of the Diels-Alder reaction proved to be usefiil in the total synthesis of type II and III indole alkaloids. Several successful approaches are described in the literature, ahhough the educts, the intermediates and the reaction conditions were strongly different from those of the supposed biogenesis. [32, 33, 34] It should be emphasised that these remarks can neither prove the real process of the biosynthesis, nor substitute the necessary experimental investigations about it. However, they indicate that the supposition of such mechanism would not be irrational or irreal. As a consequence of the fragmentation, the chirality of C-15 is lost, and the supposed intermediate secodine derivative (33) has formally no chirality. However, in the new ring systems, several new centers of chirality have been formed with high stereoselectivity, and this fact is reflected in the widely variable number of the isolated alkaloids, too. In the bridged ring system of the ibogan-isoibogan group, the three centers of chirality (C14, C-16 and C-21 in the ibogan and C-3, C-16 and C-20 in the isoibogan group) are tightly interrelated and determine only two stereoseries, having C-17 in a or in (3 orientation. In the ibogan group, all alkaloids with appropriate structure belong to the C-17p stereoseries, with a single exception which is catharanthine belonging to the C-17a stereoseries. Exceptionnally, in the isoibogan group only 1 compound was isolated from a single species, Strychnos ngomiiensis, i.e. a species of the Strychnaceae and not of the Apocynaceae family. Therefore and because of its uncertain stereostructure, the chirality transfer could not be investigated. It should be mentioned that from the same species 1 aspidospermatan and 2 strychnan alkaloids were obtained, too. All these products may have a remote common precursor at the level of stemmadenine. Nothing is known about the cause of the chemotaxonomic exception. In the fused ring system of the plumeran-isoplumeran group there are likewise three centers of chirality in the primary products (C-7, C-20 and C-21 in the plumeran series and C3, C-7 and C-14 in the isoplumeran series) again in the a or P stereoseries. In all alkaloids of this group, the ligands of the common atoms C-20 and C-21 (in the plumeran series) or C-3 and C-14 (in the isoplumeran series) are in cis orientation. However, in the isoplumeran series these ligand (H-3 and H-14) have P orientation, whilst in the plumeran series (C-19 and H-21) are either in a or P orientation, approximatively in 1:1 ratio. These observations do suggest a definite preference of the centers of chirality, which is complete in the isoplumeran series, nearly complete in the ibogan-isoibogan series and poor in the plumeran series. 12. Back Transition from IIip into Ilia Skeleton As mentioned above, in the majority of the plumeran and isoplumeran alkaloids there is a covalent bond between C-7 (i. e. the P position of the indole ring) and C-3 (in the isoplumeran compounds) or C-21 (in the plumeran compounds) which therefore may be considered as Iip and llip alkaloids of type II and III skeletons. This fact gives the possibility of the formation of a last class of indole alkaloids in which the ligand C-3 or C-21 is shifted back from C-7 to C-2, i.e. from the P to the a position of the indole ring (Scheme 3). The shift results in the formation of isoeburnan (type Ila) and ebuman (type Ilia) alkaloids. In both types of

Chirality Transfer in the Fonnation of the Indole Alkaloids Derived from Secologanin

393

peracict

COOCH3 vincadlfformjne (III p)

HO^ H3COOC

I

1

Z

H" H3COOC'

vincamlne (III a ) , a-OH isovlncamlne (III a), p-OH

Scheme 4. Transitionfromtype IIIp into Ilia indole alkaloids

and C-21 (in the eburnans) are in cis position. In the isoeburnan group only tacamine and its 7 derivatives were isolated from a single species {Tabernaemontana eglandulosd) and all of them have the ligands H-3 and H-14 in cis (3 orientation. However, the small number of compounds is not sufficient for evaluation of the chirality transfer. In the eburnan group, the cis a orientation is slightly preferred over cis P orientation of the ligands C-19 and H-21. Also in this case, the exact mechanism of the biosynthesis is not known in sufficient details. However, the fact that the cis a and cis P orientations, observed in the plumeran and isoplumeran compounds, were kept in the eburnan and isoeburnan compounds, suggests a 1,2-rearrangement with retention of configurations. A simplified version of the mechanism proposed by Poisson and coworkers [35] is shown in Scheme 4. Such a rearrangement was successfully applied, under mild ("biomimetic") conditions, in the prepartion of the vincamine and related compounds from tabersonine and its analogues. [36] The appropriate plumeran-eburnan pairs [(-)-vincadifformine and (+)-vincamine, (-)tabersonine and 14,15-didehydro-(+)-vincamine] were isolated from the same or relative species. The biogenetic connection of the plumeran and eburnan skeletons is confirmed also by the dimer alkaloid tenuicausine (28 in Figure 2) having both IIip and Ilia subunit from Melodinus tenuicaudatus. A further common feature of the stereochemistry of plumeran and eburnan alkaloids was also observed. In alkaloids in which the methoxycarbonyl group (C-22) is retained, the ligands H-21 and C-19 are in cis a orientation; in alkaloids in which C-22 is lost, the same ligands are in cis P orientation. The causal connection between these factors is not clear yet. It is also remarkable that several alkaloids of the plumeran (e. g. minovincine and minovincinine), and eburnan groups (e. g. eburnamenine, eburnamonine, 16-epieburnamine and vincamine), were isolated in both enantiomeric forms from the same or from different species. It was very rarely experienced in the other groups of indole alkaloids.

394

Progress in Biological Chirality

13. Conclusions The results can be summarized as follows: In type I indole alkaloids the source of chirality of C-15 is the chirality of C-5 in secologanin. The S configuration of C-3 is completely determined by the enzyme strictosidine synthase and effectively transferred into those alkaloids of type la and 1(3 in which C-15 is involved in a bridged ring system. The source of R configuration of C-3 is not clear. It may come from a second enzyme, or by epimerization of the 3S center. However the mechanism of it is not known. The R chirality of C-3 can not be transferred into the bridged and rearranged systems. Formally, in the type II and III alkaloids the chirality is lost. However, a definite chiral selectivity may be observed in the four larger groups of alkaloids (plumeran, ibogan, eburnan and isoplumeran) and suspectedly in the two small groups (isoeburnan and isoibogan)as well. In the bridged ibogan-isoibogan group the p orientation of C-17 is nearly complete (with a single exception); in the fused systems the cis orientation of the acyclic ligands attached to the common atoms (C-3, C-14 or C-20, C-21) is likewise close to completion. In the isoplumeran group this orientation is exclusively (3; in the plumeran and eburnan groups is a, if the methoxycarbonyl group is retained, and P if it is lost. This chiral selectivity is interpreted by the chirality of the transition states or intermediates. In the plumeran and eburnan groups some alkaloids are available in two enantiomeric forms. Our analysis shows, that the chirality of C-5 in secologanin defines more or less rigorously the stereostructure of most indole alkaloids, and this chiral unity in the molecular multiplicity is a source of beauty. 14.

References

[1] Dictionary of Natural Products, Version 11/2, Chapman and Hall/CRC, New York, London, 2003. [2] In order to be in congruence with ref. [9], in this paper the same botanical classification system is used, as in Ref [3], proposed by A. J. M. Leeuwenberg. According to it, Strychnaceae are involved in Loganiaceae and Naucleaceae in Rubiaceae. [3] L.F. Szabo, Diversity and selectivity in molecular evolution. In: Fundamentals of Life (Eds. G. Palyi, C. Zucchi and L. Caghoti) Elsevier and Accademia Nazionale di Scienze, Lettere ed Arti (Modena), Paris, 2002, pp. 437-451. [4] G.A. CordeU, The Biosynthesis of Indole Alkaloids, Lloydia, 37 (1974) 219-298. [5] Atta-Ur-Rahman and Anwer Basha, Biosynthesis ofIndole Alkaloids, Clarendon Press, Oxford, 1983. [6] A.L Scott, Bioorganic Chemistry 3 (1974) 398-429. [7] A.I. Scott, PureAppl Chem. 65 (1993) 1305-1308. [8] I. Kompis, M. Hesse and H. Schmid, Llodya, 34 (1971) 269-291. [9] M.V. Kisakiirek, A.J. Leeuwenberg and M. Hesse, A chemotaxonomical investigation of the plant families of Apocynaceae, Loganiaceae, and Rubiaceae by their indole alkaloid content. In: Alkaloid: Chemical and Biological Perspectives, Vol. 1 (Ed. S.W. Pelletier) John Wiley and Sons, New York, 1983, p. 211-376. [10] In order to be in accord with ref [9], it was intended to be as close as possible in naming of the basic skeletons. However, renaming could not be avoided in some cases. The name vincosan was used as a general name for all type la alkaloids. In ref [9], the alkaloids of the isoplumeran and isoeburnan groups are involved into the groups of the plumeran and eburnan skeletons, respectively. No compounds of the isoibogan skeleton were mentioned yet. [11] The number of the individual compounds should be considered approximative, partly because it changes continuously by isolation of new compounds, partly because of the registration system of the Dictonary of Natural Products. [12] In this paper, the biogenetic numbering of [ 13 ] is used. [13] J. Le Men and W.I. Taylor, Experientia 21 (1965) 508-510. [14] L. Karolyhazy, A. Patthy-Lukats and L.F. Szabo, J. Phys. Org. Chem. 11 (1998) 622-631.

Chirality Transfer in the Formation of the Indole Alkaloids Derived from Secologanin

3 95

[15] M. Rohmer, M. Seemann, S. Horbach, S. Bringer-Meyer and H. Sahm, /. Amer. Chem. Soc. 118 (1996) 2564-2566. [16] A. Contin, R. van der Heijden, A.W.M. Lefeber and R. Veipoorte, FEES Letters 434 (1998) 413-416. [17] V.S. Dubey, R. Bhalla and R. Luthia, J. Biosci. 28 (2003) 637-646. [18] A.R. Battersby, A.R. Burnett and P.G. Parsons, J. Chem. Soc. (C) (1969) 1193-1200. [19] A. Patthy-Lukats, L. Karolyhazy, L.F. Szabo and B. Podanyi, J. Nat. Prod. 60 (1997) 69-75. [20] J.F. Treimer and M.H. Zenk, Eur J. Biochem. 101 (1979) 225-233. [21] L.F. Szabo, Some aspects of the chemistry of secologanin. In: Studies in Natural Products Chemistry, Vol. 26. Bioactive Natural Products, Part G. (Ed. Atta-Ur-Rahman) Elsevier Science B. V. (2002) 95-148. [22] M. Rueffer, N. Nagakura and M.H. Zenk, Tetrehedron Letters (1978) 1593-1596. [23] G. Beke, A. Patthy-Lukats, B. Podanyi, L.F. Szabo, ChiraUty 13 (2001) 483-487. [24] W. De-Eknamkul, N. Suttipanta and T.M. Kutchan, Phytochemistry 55 (2000) 177-181. [25] R.T. Brown, C.L. Chappie and R. Piatt, Tetrahedron Lett. (1976) 1401-1402. [26] A. Patthy-Lukats, G. Beke, L.F. Szabo andB. Podanyi, J. Nat. Prod 64 (2001) 1032-1039. [27] E.J. Shellard and RJ. Houghton, PlantaMed. 21 (1972) 16-21. [28] R. Stahl and H-J. Borschberg, Helv. Chim. Acta 11 (1994) 1331-1345. [29] J. Bonjoch, D. Sole, S. Garcia-Rubio and J. Bosch, /. Am. Chem. Soc. 119 (1997) 7230-7240. [30] R.K. Grover, S. Srivastva, D.K. Kulshreshtha andR. Roy, Magn. Reson. Chem. 40 (2002) 474-476. [31] L. Karolyhazy, A. Patthy-Lukats, L.F. Szabo and B. Podanyi, Tetrahedron Lett. 41 (2000) 1575-1578. [32] R.M. Williams, Chem. Pharm. Bull. 50 (2002) 711-740. [33] S.A. Kozmin, T. Iwama, Y. Huang and V.H. Rawal, J. Am. Chem. Soc. 124 (2002) 4628-4641. [34] I. Vago, G. Kalaus, I. Greiner. M. Kajtar-Peredy, J. Brlik, L. Szabo and C. Sz^lay, Heterocvcles 55 (2002) 873-880 [35] G. Croquelois, N. Kunesch and J. Poisson, Tetrahedron Lett. (1971) 999-1002. [36] G. Hugel, J. Levy and J. Le Men, C R. Hehd Seances Acad Sci., Ser. C 274 (1972) 1350-1352.

Progress in Biological Chirality G. Palyi, C. Zucchi and L. Caglioti (Editors) © 2004 Elsevier Ltd. Allrightsreserved.

Chapter 32 Origin of Biological Chirality Tetsuyuki Yukawa Coordination Center for Research and Education, The Graduate University for Advanced Studies, Hayama, Miura, Kanagawa 240-0193, JAPAN yukawa@soken. ac.jp

1.

Introduction Asymmetries appear at various stages in the biological evolution. Among them the most fundamental asymmetry is the molecular asymmetry known as the chirality in proteins, RNA's, and DNA's. Proteins of biological origin are known to be formed by left-handed Lamino acids, except glycine which is chirality neutral, while sugars in RNA's and DNA's are selectively right-handed D-ribose. At present it is a complete mystery when, where, and how this asymmetry arose, but I believe it can be understood scientifically as a result of the natural law. Under the circumstance that we do not yet have any reliable scenario of the origin of life, we inevitably give some plausible assumptions at various stages of evolution, especially to the initial conditions. We assume that life began on the Earth within its 4.6 billion years history, and at an early stage of the history when the surface temperature was still very high, chemical evolution started to synthesize amino acids. Origin of the biological chiral asymmetry has been searched for either in natural environments or in the elementary process. I prefer to choose the cause of asymmetry in the elementary process, which I mean the parity violation of weak neutral current, mainly because of its universality. There have been many arguments for and against the parity violation theory as a possible explanation of the asymmetry. The energy difference between parity doublets due to the parity violating interaction (PVED) has been calculated [1], which has given us some hope to explain the preference of L-amino acid and D-ribose because of their energetical stability, although energy differences have been so tiny. It has then been claimed that the self-catalyzing process would enhance small difference of the initial population exponentially large [2, 3]. However, we will see in the following that the parity violation alone is not enough for producing the asymmetry and existence of the self-catalysts does not mean the homo-chirality. Another problem, which has not been taken in consideration so seriously in the asymmetry production, is the second law of thermodynamics, i.e. the entropy increase. Although the selfcatalytic process enhances the asymmetry, the racemization process tends to draw the system back to the state of maximum entropy, namely the thermal equilibrium. It looks true that in living organisms the entropy decreases by expelling high entropy wastes out of the system,

398

Progress in Biological Chirality

and we tend to ignore the entropy increase in biological processes. However, if we are dealing with the chemical evolution prior to the biological evolution, we have to respect the second law of thermodynamics. I would like to give a resolution to this problem by employing the scenario which the particle physics has developed in explaining particle dominance of the universe. In the followings I shall firstly discuss the problem of the parity violation scenario in the origin of chiral-asymmetry, and the resolution on the bases of quantum mechanical description of the racemization process. It is then extended to a system in the thermal environment. Secondly, I shall introduce three necessary conditions for the chiral asymmetry of bio-molecules on the Earth, by rewriting the conditions which have been given originally by Sakharov [4] in order to explain particle asymmetry of the universe. Finally, I shall present a possible scenario of the homo-chirality of proteins.

2.

Parity Violation Revisited

2.1 Energy difference due to the parity violatingforce Let us review here the quantum mechanical energies of an amino acid for states with different chirality due to the symmetry violating force. We write {| 1 >, | d >} being chiral pair states related through the parity operation, P, as \l>P\d>,

\d>P\l>,

(2.1)

where P^ = 1. A Hamiltonian can be split into two parts, H = Hi + H2 , where Hi commutes and H2 anti-commutes with P, respectively; [ P , H i ] = 0,

{P,H2}=0.

(2.2)

The second term, H2 , is responsible to the symmetry breaking. They can be written explicitly as,

H,^-{H^PHP),

H^=-iH-PHP)

.

Then, the energy expectation values for chiral pair states are given by e, = < 11 Hi 11 > + < 11 H211 > = ei + e2, ed = < d | H i | d > + < d | H 2 | d > = < l | PHiP 11 > + < 11 PH2P 11 > = < l | H , | l > - < l | H 2 | l > = ei-e2.

(2.3)

The energy differences between chiral pairs have been calculated for various amino acids [1], which turned out to be about Ae=ei-ed~-10-^^eV

(2.4)

with the correct sign in favour of L-amino acids energetically, but much too small to expect

Origin of Biological Chirality

399

any physical relevance. One may hope that the amplifying mechanism of self-catalytic process will save this difficulty through exponential growth of the population difference [2, 3], AN(t)-[N,(0)-Nd(0)]e^^ . (2.5) However, this process creates both types of chiral molecules exponentially, and as time goes by the larger component, Ni (t), stops growing, because of the shortage of L-amino acids. If both types of amino acids are initially created equally, Nd (t) will catch up eventually. Even though the initial population of L-amino acids happen to be larger due to certain natural environments, the racemization process will take over to diminish the asymmetry and bring the system into the equilibrium distribution. In fact the energy is only meaningful in the equilibrium to give the Boltzmann factors: ^2p&e

N,/N,^e

This reminds us that the direction to homo-chirality is opposite to the thermal equilibrium. 2.2 Quantum motion of an amino acid From the considerations in the last sub-section we find what we should study is not the equilibrium state, but the non-equilibrium process where the asymmetry increases against the racemization. It requires us to describe the system dynamically. To begin with let us consider the motion of an amino acid quantum mechanically. The chiral pair states make the racemization transition by tunnelling, and thus a general state of an amino acid is a linear superposition of the chiral pair states; (2.6)

|^(0>c,(0|/>+c,(0l^>

Motion of the wave fiinction is governed by the Schrodinger equation (with the convention .dc ,^ /—-he dt

(2.7)

where c,{t)

(2.8)

h=

K

Kdj

The Hamiltonian matrix elements are written as hii = ei + e2 , hid = vi + iv2 ,

hdi == vi - iv2 , hdd = ei - e2 ,

(2.9)

with defining vi = < 1 | Hi | d > and iv2 = < 1 | H2 | d > . The off-diagonal components are responsible to the racemization. The equation can be solved analytically, where the explicit form of the solution is given in Appendix 1. In order to relate a solution to the observation process the density matrix is often introduced by P^it) = c^{t)c,*{t) , (2.10) A solution of the Schrodinger equation can be expressed by p(t)^K{t)p{())K*{t) (2.11)

400

Progress in Biological Chirality

where K (t) is the unitary evolution matrix whose components are also found in Appendix 1. If we give the initial distribution symmetric, i.e. equal amount of L- and D-amino acids, by ifi o^ (2.12) P(0)-, ,0 1, the density matrix does not change at anytime, because of unitary nature of the time evolution matrix: K (t) K*(t) = 1. This shows that the parity violating interaction does not induce asymmetry as far as the system starts symmetrically, and evolves unitarily, i.e. conserving the total probability. 2.3 Source of the asymmetry We shall look for the origin of asymmetry in the wave function of a peptide. A peptide is a linear molecule of n-amino acids {ai, a2, .., an}, formed by the dehydration. In general the amino acid ai of the i-th position can have either the L- or the D- chirality except glycine. We shall write the peptide wave function whose amino acid at the i-th position happen to be an Lamino acid li by, |ai*a2*....l,*....an*> | L > .

(2.13)

Here we put * to each amino acid to indicate that it is the residue after dehydration. We also write its chiral conjugate by | D > = Pi| L > , or more explicitly by | D > = |ai*a2*....di*....a„*>.

(2.14)

Now we assume that the overlap integral of states | L > and | D > does not vanish; ^,=- .

(2.15)

There are at least two reasons to support this assumption: the wave ftinction of an amino residue in a peptide is extended, and it is not necessarily the parity eigen-state. When the overlap integral does not vanish there should be a slight modification to the equation of the motion. Writing the overlap matrix as n=\

,

(2.16)

{e \) where we keep eyes only on the racemization of the i-th residue, and have suppressed the idependence of the overlap integral. In general racemization takes place at any position at the same time. The Schrodinger equation now becomes ; 4 [ « " ' c ( 0 ] = //[/7"^c(0], (2.17) at where the Hamiltonian matrix also suffers a modification; H = n "^^^ h n •^^^. This equation can also be solved analytically (see Appendix 1), and the solution in the density matrix form is given by p(t) - n'''^K{t)Yn'''^p{^)n''-\K * {ty'"^ . (2.18) It is clear that even if we choose the initial density matrix symmetric, it evolves in time. When we define the asymmetry function by A(t) - p^^it) - p^jit), we have

Origin of Biological Chirality 401 -^(sm2dtsm7] + 2-^sin dtcosrf) , d d

A(t) = -eM-

(2.19)

where we have defined d ^ = e2^ + vi^ + W , and cosrj = vJ{vl-\-vl)'''^. It has the nonvanishing time average:

A=-sA\~

(2.20)

-y cos 77

It shows that the asymmetry remains finite at least when 8e2COS77 does not vanish. It is interesting to notice that the asymmetry is proportional to the parity violating energy difference e2, and thus the L-amino acid is favoured if the overlap integral ^cos;; is positive. Since the transition is considered to take place in the high temperature environment of the early Earth, we need to make sure that the asymmetry will not be washed out by the heat noise. In Appendix 2 we give the quantum statistical equation under the Gaussian white noise for the case where COST; = 1 (v2=0). The asymmetry fimction averaged over the ensemble is obtained as -.2

< A{t) >= -s

{\-e V, V,'-^X

^

^(cosvj/-—sinvj/)}

(2.21)

with X being the coupling strength to the heat bath. It shows that the asymmetric function converges to an asymptotic value.

<.4(0>-

-^-s-

(2.22)

V, v.'+je

I should notice that the asymptotic form (2.22) is valid only in the case 62 is negligible comparing to A as is discussed in Appendix 2. In Figure 1 I show assuming

Figure 1. Time behaviour of the asymmetric fimction. 3.

Scenario of the Biological Asymmetry

3. J Cosmological scenario of the particle asymmetry In the last section we find that the parity violating force together with the asymmetry of wave functions creates non-vanishing chiral asymmetry of peptides. However, the asymmetry is very small as we can expectfi-omeq.(2.22) and eventually the distribution will reach to the equilibrium. We need something very special to happen so that the asymmetry becomes more significant, while the system tends to the maximum entropy state. It is a common belief of particle physicists such an extraordinary event has happened at the

402

Progress in Biological Chirality

period of creation of the universe. As far as we can observe there are no sign of stars made of anti-matter, although the standard particle theory predicts creation of the equal amounts of particles and anti-particles at and just after the big bang. It was in 1967 Sakharov [4] proposed three conditions necessary to create the cosmological asymmetry; namely, (i) baryon number (here we mean net number of protons and neutrons) should not be conserved. It looks trivial since we find no anti-particle, which is assigned negative baryon number, in nature. The second condition is (ii) the asymmetry of transitions between particle and anti-particle. If they are the same, two transitions will compensate each other and result no net creation of asymmetry. Thirdly, (iii) the universe should be in the non-equilibrium state. Otherwise, the asymmetry is of the same order as the symmetry breaking energy which is negligible. It has been neglected for some time till the C (the charge conjugation) and the CP breaking interaction has been discovered. Together with the baryon number breaking of the grand unified theory (GUT) and the non-equilibrium state of expanding universe, there are possible scenarios for the cosmological asymmetry [5]. Let us illustrate the cosmological scenario briefly leaving the details to literatures which depend much on several uncertain facts in the evolution of the universe. About 13.7 billion years ago just after the Big Bang the universe was filled with high energy photons within a compact space. During expansion of the space photons with energy high enough to create baryon anti-baryon pairs were in the equilibrium with matter and anti-matter. As the universe continued to expand adiabatically, the temperature dropped down, and when high energy photons disappeared, the equilibrium condition was not maintained any more, and the matter and anti-matter distributions have decoupled and temporally fi^ozen. After this period, whenever a baryon met its partner anti-baryon, they annihilate each other to photons, which were not active enough to revive to matter and anti-mater anymore, and became to so-called relic photons in the cosmic background radiation. If there exists the baryon number asymmetry, there should be excess matter which was leftover of the annihilation process. They form stars in this universe. 3.2 Biological scenario for the chiral asymmetry By modifying the cosmological scenario of the matter asymmetric universe we will try to fit for the asymmetry of bio-molecules. The three necessary conditions seem to be readily ftilfilled: (i) the racemization allows the change of numbers of L- and D-molecules; (ii) the parity violation with overlapping wave fiinctions creates asymmetric transition between Land D-chirality states, while (iii) the thermal history of the Earth satisfies the non-equilibrium condition. Let us illustrate a possible scenario of the biological asymmetry under these conditions. At the beginning of the Earth it was covered by the magma ocean. During cooling down by radiation emission the chemical evolution synthesize amino acids either in the super critical atmosphere. The creation occurred through the dominantly parity conserving electro-magnetic force, and the products of L- and D-amino acids are almost the same amount (Figure 2(a)). As temperature went down the rain belt covering the hot atmosphere lowered, and finally reached to Earth's surface. Then the concentration of amino acids got high, and they started interacting each other. At high temperature both polymerization and racemization proceeded quickly, and all the possible racemic mixtures were almost in the equilibrium distribution. As the temperature continued to get lower the racemic transition became very slow because it occurred through tunnelling process, and the distribution was temporaryfi*ozen(Figure 2(b)).

Origin of Biological Chirality

403

tzm

Figure 2. Evolution of L- and D-amino residues in pure L-peptides (L), pure D-peptides (D) and racemic compounds (R) at three stages: (a) the creation of amino acids, (b) the frozen distribution, and (c) the LD araiihilation and excess L

The number distribution of peptides with 1 L-amino residues and d D-amino residues in the equilibrium is given by A^(/,^)c

(l + d)\ l\d\

df-d

(3.1)

Here K is the rate of polymerization and Q is the suppression factor relating to the asymmetry parameter given by eq. (2.22) through

c-

1 - < yl(QO) >

1+ < ^(oo) >

(3.2)

which is obtained by requiring that the numbers of excess amino residues is equal to the asymmetry parameter times total amino residues. It is a common belief that racemates, which I mean a polymer with any mixture of L- and D-amino residues, are inactive in polymerization [6], and only pure L- or D-peptides will enter to play important roles in the biological evolution. In this way the dominant part of entropy is exhausted by racemates. During peptides evolution whenever a L-peptide and a Dpeptide ftised to a racemic compound it became inactive and disappear from the scene of the biological evolution. We make a simplified scenario that an L-peptide of the length n, L(n), and a D-peptide D(n) makes the chemical evolution, L(n) + L(m) «-> L(n+m) D(n) + D(m) <-> D(n+m) L(n) + D(m) ~> R(n+m) , where R(n+m) represents a racemic compound with n+m amino residues. A simple numerical simulation shows the tendency to the mono-chirality as we expect (Figure 3). The excess Lpeptides are of course due to the asymmetry of racemic transition. This concludes the scenario of the origin of biological asymmetry (Figure 2(c)).

404

Progress in Biological Chiralit\

time Figure 3. Numbers of L(D)-amino residues in L(D)-peptides

4.

Discussion We have shown a possibility of explaining the biological asymmetry of chiral-molecules by making use of the cosmological scenario of the matter asymmetric universe. One of the important points which have to be clarified is the appropriateness of parameters to be used for the theory to apply as the scenario on the Earth. The overlap integral 6-of eq.(2.15) is one of them. Another important problem is to make sure the role played by racemic compounds as the entropy carrier. We emphasize that these quantities and properties are basically approachable either numerically or experimentally.

5.

Acknowledgement The author thanks Dr. Mitsuzawa for giving him many useful information and comments. He also thanks to the organizers of the international meeting on the Biological Chirality for giving him a chance to present his work.

6.

References

[1] (a) D. Rein, ATOMKIKozl. Suppl. 16 (1974) 185-193. (b) S.F. Mason, Nature (London) 311 (1984) 19-23. (c) S.F. Mason and G.E. Tranter, Mot. Phys. 53 (1984) 1091-1111. (d) V.I. Goldanskii and V.V. Kuz'min, Sov. Phys. UspeJchi 32 (1989) 1. (e) W. A. Bonner, Orig. Life Evol. Biosphere 21 (1991) 59-111. (f) J.L. Bada, Nature 374 (1995) 594-595. (g) A. MacDermott, Orig. Life Evol. Biosphere 25 (1995) 191-199. (h) P. Lazzeretti and R. Zanasi, Chem. Phys. Lett 279 (1997) 349-354. (i) O. Kikuchi, / Mot. Struct. (Theochem) 589-590 (2002) 183-193.^ [2] F.C. Frank, Biochem. Biophys. Acta 11 (1953) 459-463. [3] (a) D.K. Kondepudi and G.W. Nelson, Nature 314 (1985) 438-441. (b) A.J. Salam,A/o/. Evol. 33 (1991) 105-113. [4] A.D. Sakharov, Zh. Eksp. Teor. Fiz.Pis'ma 5 (1%7) 32. [5J (a) Ya. Zeldovich, Zh. Eksp Teor. Phys. 67 (1974) 2357. (b) M. Yoshimura, Phys. Rev. Lett. 41 (1978) 281-284. (c) S. Dimopoulos and L. Susskind, Phys. Rev. D18 (1978) 4500-4509; (d) idem, Phys. Lett. 81B (1979) 416. (e) D. Toussaint, S.B. Treiman, F. Wilcek and A. Zee, Phys. Rev. D19 (1979) 1036-1045. (f) S. Weinberg, Phys. Rev. Lett. 42 (1979) 850-853. (g) J. Ellis. M.K. Gaillard and D.V. Nanopoulos, Phys. Lett. 80B (1979) 360. [6] J. Sarfati. TJ 12 (1998) 281-284.

Origin of Biological Chirality

405

Appendix 1 to Chapter 32 Quantum Motion of an Amino Acid We begin with giving solutions of eq.(2.7). General solutions of eq.(2.17) can be derived from them by replacing c(0->«^''c(0 ,

and h-^H.

(Al.l)

The solution can be formally written by the unitary operator K(t) as c(t) = ^(0^(0) ,

with Kit) = e-'""' .

(Al.2)

The matrix which makes the Hamiltonian h diagonal is given by U^

^a

-fk''^

(A1.3)

by which the Hamiltonian is transformed to UhU* = E = with d = e2 + vi + V2 ,

(e,+d 0

COST;

0 ^ e^-d J

= v^ /(v, + v^)^^^, and where ^<'-^'

Since K(t) = U*e"*^*U, it is written explicitly as \ QO^dt-z^sindt

/ J l - — I e"^ sindt

d

K(t) = e-''^' /Jl - ^

e^'"^ sin dt

V

U

(A1.4)

cosdt + / ^ s i n dt

Since the density matrix, whose matrix elements have been defined by eq.(2.10), is written as p(t)^K(t)piO)K*(t), (A1.5) the general solution of eq.(2.17) is given by ^1 1 i _1 p(t) - n ^K{t)n^p{0)n^K * {t)n ^ , (Al.6) applying the rule eq.(Al.l), with slight modifications of parameters in eq.(A1.3-4) due to the replacement /? -> ^ , as

VT-

VT

(A1.7)

406

Progress in Biological Chirality

Appendix 2 to Chapter 32 Quantum Statistical Motion of a Peptide We consider the motion of a peptide in a thermal bath which is governed by the Hamiltonian h(t) = h + v(t), (A2.1) where v(t) is a matrix for the thermal perturbation,

which is assumed to be the Gaussian white noise characterized by the correlation function =c^SJ{t-f) (A2.2) Time independent part of the Hamihonian, h, is given by eq's (2.8-9) with vi = 0 for simplicity. We solve the Schrodinger equation by the time-dependent perturbation theory: c(t) = e"'''W(t)c{0) ,

(A2.3)

where W(t) follows the equation of motion i—W{t) = V(t)W(t) , (A2.4) dt with V{t) = e'^^v{t)e~'^'. In order to solve eq.(A2.4) we discritize the time as t = nA where at the end A will be taken to be zero, and at the same time n is chosen to keep /(= nA) fixed, to obtain the continuous Hmit. We shall write a time dependent function as f(t) ->fn in the discrete time scheme. Then eq.(A2.4) is written as

2 "

which is correct up to the second orders in A. Equation of the motion for the density matrix is then given as P.,,=e-'-'w„p„w„*e-''

(A2.6)

M'„=

(A2.7)

with 1 ^ .

2 " The correlation (A2.2) for the descritized scheme is written as < w. A n,>=c 5. - J _ , a,n

b,m

a ab

A

A

nm '

(A2.8) ^

''

Origin of Biological Chirality

with which we evaluate the ensemble averaged density matrix; < p„,, >« ^(A) < M^„p„K^„* > ^ * (A) ,

407

(A2.9)

where K(A) has been given by eq.(A1.3). Regarding that the density matrix at the time n depends on the random fields at time m (^«*>^-Ac2<

PlUn

Pdd,n

Pdl,n ~ Pld,n

Pld,n

Pdl,n

Pdd,n ^ Pll,n

> ,

within the first order in X. Combining these results we obtain for the density matrix elements K = < {Pll,n - PddJ > , iyn ^< {Pld,n " Pdl,n) > , m& Z„ ^< (/?;^^„ + P,,J

> aS

^«+i = ^« - 2 A 7 ^ ' - ^ 2 > ' „ - 2^2Ax-„

J^n.1 - >^« + 2A^d'^-elx„ + 2e,Az„ - 2c,Ay„

(A2.10)

while the trace p^^ „ + p^^ „ is kept constant(=l) . Taking the limit A -> 0 have d (""'] V = -2 dt

UJ

f (A2.11) V 0

e^

0 y

where we set assuming V2 = 0, and write C2 = >^. This can be solved analytically, but not compact, and we assume Q2 is small so that z ~ constant (= e). Then the solution is given approximately as

x(t)^-a ^ ^ ^ { e " " ' '"^' - e-^(cosv,/ - - s i n v/)}

(A2.12)

v,A'+vf

A A + V,

A

since e2 is of the order of lO'^^eV the damping factor of the first terms in the RHS of eq. (A2.12) will be neglected.

Index A ab /w/Y/o electron densities 211 acetaldimine MgporphinCO, adduct 17 acetaldimine, reaction with CO on Mgporphin 17 coordinated to MgporphinCO 22 acetaldimine-CO complex 22 2-acetamido-2,3-dideoxy-3-sulfo-hexose 183 2-acetamido-2,6-dideoxy-6-sulfo-hexose 183 acetic acid 345 Acetobacter 341; 345 acetophenone, oxime, reduction of 228 reduction of, by chromium (II) 224 achiral space 264 achirality deficiency measure 214 acid racemization, within teeth 67 actinolite 147 l,2-di-0-acyl-3-0-(6-deoxy-6-sulfo-a-Dglucopyranosyl)-L-glycerols 181 2-acylamino-2,6-dideoxy-D-glucopyranose-6sulfonates 184 acylglycerols 181 Additive Fuzzy Density Fragmentation (AFDF) 211 adenosine thiophosphorbenzimidazolide, polymerization by uranyl ion 300 adenosine thiophosphorimidazolide, polymerization by uranyl ion 300 Adjustable Density Matrix Assembler (ADMA) method 211 affinity biosensors 289 age determination, at time of death 66 from fossil bones 65 from teeth 66 of fossil bones 70 age, biological 66 calendar 66 chronological 66 dental 66 skeletal 66 alanine, D-to L-transition in crystals of 312 enantioselective synthesis of 231 rate of racemization of 69 albite 142

albumin, as molecular chaperon 279 coupling to host, CD study of 280 structure of. 275 algae 181 alkah feldspar 137 alkylcobalt carbonyl phosphine complexes 30 alkylchromium, intermediate in oxime reduction 230 alkylcobah carbonyls 29 amino acid, biosynthesis, co-evolution of routes of 90 biosynthesis routes 87 distribution, polarity effect on 100 ligand, coordination of 230 molecules, prebiotic synthesis of 207 oxidase 351 quantum motion of 399 racemization, in dentin 67 sequences, in enzymes 3 side chain volume 89 types, and biosynthesis routes 86 a-amino acids 52; 173 a-amino acids, derivatization of 68 endogenous 52 exogenous 52 formation of 49 in meteorites 49 amino acids, and dinucleotides, hydropathy correlation of... 98 as bidentate ligands 223; 226 as sources of chirality 223 as tridentate ligands 223; 226 biological 321 20 canonical 121 carbamoylation of 56 chiral analysis of 350 composition, in wine production 343 D/L ratios in fossil bones 71 epimerization, half life of 71 essential 339 free 340 genetic code for 121 hydropathy order of 97 in age determination 65 in archaeometry 65 in meteorites 128 innutrition 339 in ocean 203 micelle formation by 223

410

Progress in Biological Chirality

natural 334 natural, as ligands 222 near hydrothermal vents 203 nucleic acid-binding 102 quantum description of 405 racemization of 65; 203 racemization, age determination by 67 rate of polymerization of. 403 termal racemization of 206 uneven racemization of 206 violation of chiral identity of enantiomers 334 with apolar side chains 223 amino alcohols 50 1,2-amino alcohols 222 a-amino amides 51; 52 a-amino dinitriles 50 a-amino nitriles 50 2-amino-2,3,(^)-diedoxy-3-D-hexopyranose-3sulfonic acid 184 2-amino-2,3-dideoxy-3-C-sulfo-D-glucopyranose 189 2-amino-2,6-dideoxy-6-sulfo-D-glucose 183 2-amino-2,6-dideoxy-D-glucopyranose-6-sulfonic acid 184 aminoacyl-tRNA synthetases 84; 88 ammonia 13 ammonia or amines, in prebiotic chemistry 50 ammonium dihydrogenphosphate 322 ammonium phosphate 321 amphibian, upper limbs of 157 amphibole 147; 149 anticodon, fishing of 110 initiation of 110 anti-inflammatory strategy 185 archaeological samples 65 archaeological samples, chromatographic analysis of 68 arginine 339 ascorbic acid, in prebiotic molecular evolution 27 asparagine, probability density function of 316 aspartic acid, as indicator of individual age 67 rate of racemization of 69 Aspergillus 276 asymmetric induction, theory of 14 asymmetric microenvironment 264 asymmetric physical forces, global (on Earth) 129 universal (Universe) 129 asymmetric spaces 261 asymmetric specialization.

cerebral specialization of evolutionary basis of ontogenetic basis of. asymmetric state, evolution of spontaneous generation of asymmetric steady states, bifurcation of. asymmetric synthesis, absolute asymmetry, during chemical evolution (molecular), source of autocatalysis autocatalysis, in symmetry breaking of chiral species autocatalytic systems, asymmetry in aziridine, from charge-transfer complexes L-amino acids from

B(0H)4", in prebiotic synthesis bacteria bacteria, D-amino acids in barley malt baryon/anti-baryon pairs Bayer-Villiger microbial oxidations biomolecules, origin of. beer beer, ale Berliner weisse black lager Pilsener behavioural response, reflexive to sensory information benzaldehyde benzylcobalt carbonyl derivatives berberine berberine, interaction with serum albumin Bifibobacterium bifidum bifurcation, general theory of of racemic state Big Bang bilateral movements

155 155 155 167 164 163 13 132 400 99; 174 336 164 130 17 17

49 339 341 344 402 255 125 341; 344 344 344 345 344 344 158 158 276 34 291 291 341 163 54 402 155

Index bilirubin bimolecular reactions, encapsulated binary decision system, evolutionary origin of biochemical processes, time-unidirectional biochirality, origin of origin of in comet origin of in terrestrial rock problem of biological enantioselection biological asymmetry biological asymmetry, and particle asymmetry cosmological origin of scenario of biological chirality, macroscopic models of. origin of biological evolution biological evolution, asymmetry in biological homochirality biological (homo)chirality, origin of biological molecules biological objects, functional symmetry of micro symmetry of structural symmetry of biological signal transduction biological sulfur cycle biological symmetries biology, theoretical biomacromolecular homochirality, origin of biomacromolecules biomimetic synthesis, amino acids by amino acids in enantioselective biomolecular studies, mathematical methods in biomolecules, homochirality of origin of versatility of biorecognition biosensors, for determination of enantiomers withHPLC, combination of

282 264 157 122 79 156 156 80 29 404 401 402 401 366 29 257; 397 398 397 233 29; 173 13 122 122 122 29 182 123 35 3 4 221 221 221 209 125 127 215 289 3 50 351

biosyntheses biosynthesis families, precursors to biosynthetic processes, in aqueous medium.... biotechnological industry bisisoxazolines bivalve, D-amino acids in bones borates, in prebiotic chemistry boric esters boson Z^ Bradyrhizobiumjaponicum Brain of vertebrate, bifurcated bread Brevibacterium brewers yeast 2-bromo-3-methyl butyric acid (±)-2-bromo-3 -methylbutyric acid (±)-2-bromobutyric acid (±)-2-bromovaleric acid Bucherer-Bergs reaction R,S-2-butanol R,S-2-butyl chloride

411 221 90 221 221 222 321 65 50 51 207 181 155 348 341 344 269 268 268 268 52 266 266

C

C=N double bond, enantioselective reduction of reduction of C=0 double bond, reduction of cabbage, fermented juice calcite calcite crystal surfaces calcite scalenohedron calcite, as biominerals chiral surfaces of crystal morphology of hexagonal unit cell of Miller indices of scalenohedron of surface topology of Cahfornia Mastitis Test capillary electrophoresis capillary electrophoresis, for chiral separation capsules, from achiral precursors

228 228 224 347 347 137; 145; 149 145 150 145 145 146 145 145 145 146 341 350 351 261

412

Progress in Biological Chirality

memory of 263 synthesis of. 261 carbamates 50 carbamic acids 50 carbohydrate-derived sultones 194 carbohydrates 173 carbohydrates, from carbon monoxide and hydrogen 23 carbon dioxide, in prebiotic chemistry 50 in prebiotic synthesis 49 carbon monoxide 13 carbon monoxide clusters 13 carbon monoxide. charge transfer complex with porphin 16 in formation of D-sugars 13 in formation of L-amino acids 13 in prebiotic evolution 13 reaction to D-2'-deoxyril)ose 23 reaction to D-ribose 23 reaction withacetaldimine onMgporphin 17 reaction withMgporphin 21 reactions of 18 sugarfrom,on Mgporphin 23 carbonyl compounds, in prebiotic chemistry 50 3-carboxy-4-methyl-5-propyl-2-iiiran propanoic acid 277 carotene, role of in vision 238 carotenoids 278 casein, proteolysis of 342 catalysis, enantioselective 177 cation coordination 150 centrosymmetric crystals 265 ceramides 181 cerebral hemisphere, dominant (left) 158 non-dominant (right) 158 cerebral laterahzation, Polya model for 155 charged sugar oligomers 185 Charophyte Gyrogonites 365 Charophytes, asymmetry of 373 classification of 368 climate change effect on 372 from Alberta (Canada) 368; 370 from Arizona (USA) 371 from Brazil 371 from Indiana (USA) 367 from Kazakhstan 371 from Kerguelen Islands 372 from Liaoning Province (China) 369

from New Mexico (USA) from Oklahoma (USA) from Paraguay from Ukraine from Ural region geological history of in Devonian period in Early Carboniferous in Late Permian in Mesozoic Era in Palaeozoic Era in Permian in Permian period in Permian-Triassic boundary in post-Palaeozoic Era in Triassic Era cheese cheese, GranaPadano Parmigiano-Reggiano Pecorino ripened chemical energetics, co-emergence of co-evolution of chemical evolution chemical evolution, prebiotic chemoselectivity CHFClBr, infrared spectra of enantiomers chiral amplification chiral asymmetry, biological scenario for general theory of spontaneous generation of chiral autocatalysis chiral biomolecules chiral crown, ether stationary phase chiral crystal growth chiral crystal surface chiral crystalline surfaces chiral electron density clouds chiral enhancement chiral environments, on mineral surfaces chiral features, local chiral geochemical environments chiral guests chiral identity. violation of, macroscopic violation of, microscopic chiral impurity chiral induction

371 371 371 367; 368 367 373 365; 370 368 368 369 365; 370 369 365 366; 371 373 370 341; 342 342 342 342 342 49 49 336; 398 203 221 131 139 402 159 159 222 209 350 142 138 149 210 329 137 139 137 264 335 335 132 35

Index chiral information chiral information, transfer of chiral interaction chiral iridium complex, Mossbauer effect of chiral iron complex, Mossbauer effect of chiral memory, of capsule molecules chiral molecular discrimination chiral molecules, as diastereoisomers formation in centrifugal fields formation in electric fields formation in gravitational fields formation in magnetic fields formation of in living orgaiusms NMR parameters of requisite of life chiral mutations chiral organic ligands chiral spaces chiral species, autocatalysis of chiral surface charge distribution chiral symmetry breaking transitions, sensitivity of chiral synmietry breaking, equations for structural chiraUty induction chirality induction, achiral parts of electron density and entropy balance by selection mechanisms qualification of chirality reduction chirality reduction, and entropy balance chirality, at different structural levels autocatalytic amplification of biological models of choice of helical induced levels of molecular no computer code for ofDNA of proteins ofRNA peristatic propagation of

225 225 167 130 131 263 137 311 311 311 311 311 173 153; 233 311 233 3 222 221; 264 164 150 166 166 166 209; 215 210 210 210 212 209; 215 210 170 29; 233 29 257 30 272 170 272 160 397 397 397 264 171

413

symmetry breaking transition 167 transfer by amino acid ligands 221 chiral-symmetry-breaking transition 169 Chlorella pyranoidosa 181 chlorophyll, present day 19 chromanone, reduction of, optimization of 227 chronuum(II) aminopolycarboxylate complexes, reagents in synthesis 222 chromium(II), CD study of 229 chiral perturbation of 230 UV-VIS study of 229 cider 345 circular dichroism 331 circular dichroism, induced 279 circular dichroism spectra, of albumin 280 of chromium (H) 229 ofdiazoketo-rodopsin 241 of ketone reduction 227 of(llZ)-lockedretinals 245 of oxime reduction 229 of retinal 245 of rhodopsin 241 of serum albumin 280 oftryptophanase 321; 324 of tyrosine 333 circularly polarized sunlight, photochemical effect of 29 circularly polarized light 6; 311 circularly polarized light, of neutron stars 132 clinoamphibole 147 clinopyroxene 137; 142; 143 clockwork mechanism 32 clockwork, chiral 29 molecular 29 clockwork-principle 33 coacervate, evolution of 154 cobalt, essential for some orgaiusms 300 cocoa beans 348 codon, initiation of 110 codon-anticodon, conflicts between 109 minihelices 84 co-emergence, dynamics 49 coencapsulation 264 coencapsulation complexes 265

414

Progress in Biological ChiraHty

coffee coffee beans coffee. D-amino acids in roasting of cognitive binding comets, delivery of chiral material by complex molecular systems complexed metal ions complexity, level of of life forms composition of food, chemical concentrations, fluctuations of conceptual space coniigurational chirality conformational chirality, exciton coupling in conformational potential function conformations, chirahty developing in the crystalline phase enantiomeric preferences conglomerates, aqueous solutions of crystallization of in meteorites nonracemic partially resolved coordinated hgands coordination catalysis, intermediates of copper, chiral surfaces of essential for living organisms Coriolis force cosmological asymmetry coumarin, complex with serum albumin natural coupled conrotation, in molecular systems cow's milk crackers, D-amino acids in crayfish crayfish, D-amino acids in crocetin crocetin, complex with serum albumin Crocus sativus

348 348 348 348 99 80 210 221 4 216 339 259 158 33 282 32 33 31 30 30 80 39 79 80 80 29 29 137 300 311 402 281 278 31 341 348 343 321 286 286 286

crystal formation, kinetics 329 crystal growth surfaces 139; 149 crystal morphology 137 crystal nucleation, rate of 334 crystallization, as cooperative process 329 from supersaturated solutions of enantiomers.330 of conglomerates 39 ofdiastereomers 39 ofenantiomers 39 crystallographic axes 138 crystals, achiral 138 chiral 138 euhedral 137 cubic close-packed (CCP) metals, chiral surfaces of 137 curcumin 283 curcumin, complex with serum albumin 284 cyanate 52 cyanide anion 50 cyanobacteria 181 cyclopentadienylrhodium dimeric complexes, diastereomers of 41 cytochrome, D-amino acid in 69 D D-2'-deoxyribose nucleoside, formation of 26 D-3 -mercapto-2-methylpropionic acid 350 D-alanine 204; 341 D-alanine, as a molecular marker 343 in age determination 70 inarchaea 321 ineubacteria 321 in multicellular organisms 321 racemization of 204 D-allo-isoleucine 65; 68 D-amino acids 129; 203; 311; 340; 343 D-amino acids, analytic techniques 321 determination of 349 evolution of 402 in Bacteria 341 in bone samples 65 in crackers 348 in crustaceans 343 indict 352 in fermented foods 352 in foods 340; 352

Index in fossil shell 65 in fruits 349 in gelatines 349 in honey 348 in juices of fruits 351 in juices of vegetables 351 in juices sauerkraut 348 in milk products 348 in orange juice 351 in sauces 348 in tomato 349 in tomato products 348 in tooth samples 65 in vegetables 349 in vinegars 348 macroscopic identity by 329 MO description of 400 D-arabinose 24 dark energy 126 dark matter 126 Darwinian evolution 177 D-aspartic acid 341 D-aspartic acid, in age determination 70 in age estimation of teeth 72 inarchaea 321 ineubacteria 321 in multicellular organisms 321 death, as chirality reduction 216 density functional theory 210 1 -deoxy-1 -C-methylene-p-D-glucopyranose sulphonic acid 193 6-deoxy-6-sulfo-D-glucopyranose 181 6-deoxy-6-sulfo-D-glucose 182 3 -0-(6-deoxy-6-sulfo-a-D-glucopyranosyl)-1,2-di-Ohexadecanoyl-L-glycerol 181 3-0-(6-deoxy-6-sulfo-a-D-glucopyranosyl)-Lglycerol 181 6-deoxy-D-xy/o-hex-5-enopyranosides 182 (+)-deoxynojirimycin 192 deoxyribonucleoprotein 101 deoxyribose, from carbon monoxide 27 production of 23 reactions leading to 27 derivatization reagents 350 destruction, mutual 176 D-gluconolactone derivatives 186 D-glutamic acid 341 D-glutamic acid, in age determination 70 in age estimation of teeth 72 D-glyceraldehyde, formation of. 18

415

D-histidine, in age determination 70 diammoniumhydrogen phosphate 322 diastereomer, co-crystallization 46 diastereomeric complexes 263; 264 diastereomeric isoindole derivatives 350 diastereomers, co-crystallize 46 crystallization of 39; 46 energy content of 39 formation of 32 molecular shapes of 40 racemate crystals of 41 separated by chromatography 40 separation by fractional crystallization 40 solid solutions of 41 diastereoselection, dipole effects in 269 magnetic effects in 269 steric effects in 269 diastereoselectivity 221 diazoketo-ret6 241 diet, D-amino acids in 352 differential equations 177 dinucleotides and amino acids, hydropathy correlation of 98 diopside 143; 144 dipeptide formation, enantioselectivity in 58 D-isoleucine 341 dissipative structure, far-from-equilibrium 164 dissymmetry, amplification of 54; 56 divaline, formation of 58 D-leucine 341 D-lysine 341 D-mannonoIactone derivatives 186 D-methionine 341 D-molecules, adsorption energies of 137 DNA 153 DNA, helix, five-symmetry of 119 linear sequence in 160 segments of 83 synthesis, possible mechanism of 298 D-omitine 341 D-phenylalanine 341 D-phenylalanine, in age determination 70 D-proline 341 D-quartz crystals 130

416

Progress in Biological Chirality

D-ribo-P-furanosyl 3'-5' phosphodiester linkage297 drug-drug interactions 271 drugs, anti-inflammatory 185 anti-metastatic 185 co-administration of 271 D-serine 341 D-serine. inarchaea 321 ineubacteria 321 in multicellular organisms 321 D-serine, N-stearoyl, micelles of octamers of 335 D-sugars 125; 355 D-sugars, carbon monoxide in, formation of 13 from carbon monoxide on porphin template.... 17 homochirality of 161 inDNA 311 inRNA 311 related to D-glyceraldehyde 13 D-threonine 341 D-tryptophan 321 D-tyrosine 129; 330; 341 D-valine 341 D-xylose 24 dynamical system, bifurcation in 258 evolution of 259 multistable 258 dynamics, Darwinian-type 5 E early Earth, hydrosphere of 13 surface of. 81 early metabolism, origin of homochirahty in 326 Earth. appearance of organic molecules on 127 atmosphere of 127 biomasson 80 chiral molecules by comets to 80 evolution of 126 primitive, energy on 62 primitive, non-racemic a-methyl-amino acids. 54 primitive, a-amino acids on 62 Earth's crust 137 Eggenberger 153 Eigen 4 electron density modelling 210 electroweak force 207 embryophytes 366 enantiomer transition state,

Gibbs energy of. enantiomeric atomic surfaces enantiomeric capsules enantiomeric enrichment, in nonracemic conglomerates enantiomeric excess enantiomeric excess, limits to enantiomeric purity enantiomeric purity, by crystallization enantiomers, biological effects of. crystallization of energy content of energy difference of interconversion of magnetic gradient difference of NMR spectra of. relaxation time difference of scalar properties of spin-spin relaxation time of vector properties of enantiopure biomaterial, accumulations of enantiopure compounds, melting point of. enantioselective catalytic synthesis, by transition metal complexes enantioselective syntheses, by molecular catalysis enantioselectivity enantioselectivity, by amino acid ligands in ketone reduction enantioseparation enatiomeric excess, amplification of encapsulation complexes, chiral space in energy, chemical terrestrial Enterococccus entropy increase entropy, carried by racemates enzyme stereospecificity enzyme stereospecificity, flexibility of enzymes. Early Eochara wickenendi, Choquette

168 149 261; 262 79 173 177 173 80 340 39 39 317 32 317 312 312 39 315 39 81 265 222 29 221; 261 225 225 350 173 261 61 62 347 397 404 321 321 321 367; 368

Eochara,

Choquette genus

368

Index epimerization, half life of, of amino acids in geochronology escapement/gear connection, slipping malfunction of E-selectins ester group, autosolvation by T] 3-coordination of evolution evolution, Darwinian type mutation jumps in on Earth, biological on Earth, chemical pre-biological stochastic process in evolutionary system, fitness landscape of exciton coupling, effect in conformational chiraUty

71 65 35 185 33 33 209 7 11 5 5 5; 7 5 7 282

F fatty acids 286 feldspar 142; 144; 149 feldspar, alkali 142 plagioclase 142 femtosecond photochemistry, in vision 239 fermentation 341 fermentation methods 254 fermented, fish 343 foods, D-amino acids 352 milk, pasteurised 341 meat 343 ferrocenyl-ketyl radicals 222 Fibonacci numbers 119; 122 fitness landscape, hierarchy of 8 five-symmetry 119; 122 five-symmetry, ofDNAheHx 119 flavonoids, in food 277 stmcturesof 278 flavouring compound analysis 253 flow reactor, simulating a hydrothermal environment 204 fluctuations, initial stage of 154 (+)-/(-)-l-(9-fluorenyl)ethyl chloroformate 350 food aroma, biosynthesis in 253

biotransformations in volatile compounds in food, biotechnology chirality in taste of volatile components of Foraminifera foremilk formaldehyde fossil bones, amino acids D/L ratios in D-amino acid in fossils in Hungary, age determination of, amino acids in Frank free energy, sources of french pressed cheese frog's skin, D-amino acids in fruits, D-amino acids in

417 253 253 341 340 340 275 370 341 13 71 69 66 163 173 342 321 349

G gas chromatograhy gear slippage gelatines, D-amino acids in Gemmichara Gemmichara sinensis, Z.Wang Gemmichara, Wang Zhen genus gene-modified microrganisms, as biocatalysts genetic code triplets, correspondence to amino acids genetic code, algebraic analysis of aRSl in formation of chemistry in evolution of complexity in formation of formed by co-evolution functional functional model of GPS in formation of Ho sector in formation of in Archaea inEukaryotes matrix representation of palindromic pairs in paUndromic pairs in, formation of palindromic triplets in quaternary number system in

350 32 349 369 369 368 254 84 121 104 112 105 87 97 106 103 103 95 95 120 95 103 94 121

418

Progress in Biological Chiralit>

root of selection in evolution of self-correction of self-organization in evolution of stages of formation of structural model of symmetry in thermodynamic stability of triplet pairs in variants of geochronology, epimerization in glucose glucose, probability density function of glutamic acid glutamic acid, crystallization of rate of racemization of glyceraldehyde, formation from CO glycerol, formation from CO glycine glycolipids glycomimetics glycosaminoglycans gold. chiral surfaces of golden section golden section. self-similarity coefficient in grains Grana Padano cheese grape must gypsum gypsum, chiral faces of euhedral crystals of gyrogonites, asynunetry of. helicityof orientation of

121 113 257 112 103; 105 105 120; 121 94 94; 96 91 65 344 316 60 80 69 21 21 331 181 185 185 137 119; 122 119 339 342 343; 345 147: 149 148 148 365; 369 365 366

H Hadean continent haemorrhage, risk of Haldane half-sandwich complexes Halley comet, biomolecules in Halococcus species Hamiltonian, off-diagonal elements and

52 271 125 42 128 183; 184 racemization

399

splitting symmetry breaking handedness, impossibility to code linear sequence Hartree-Fock frequencies Hartree-Fock frequencies, corrected experimental theoretical Helicobacter pylori Helicobacter pylori, carbohydrate-binding specificity of hematite hierarchical fitness landscape hierarchical landscapes high performance liquid chromatography high performance liquid chromatography, coupled with enzyme reactors precolumn derivatization hippuricacid histidine histidine, rate ofracemizationof historical anthropology historical anthropology, age in historical populations, study of HN-RT, inhibition of Hohenberg-Kohn theorem holographic electron density theorem homochiral life, emergence of macromolecules in homochiral macromolecules, selection of homochiral template homochirality in biomolecules homochirahty, asymmetric physical force in origin of bifiircation origin of biochemical biomacromolecular by asymmetric adsorption by bifurcation by circularly polarized light by electric and magnetic field by gravitational fields by magnetic moments of enantiomers chance origin of condition of life evolution of from space indispensable condition of life levels of

398 398 160 17 17 17 17 187 187 141 11 11 350 351 350 277 339 69 66 77 66 182 210 211 81 149 11 3 215 128 129 137 4 129 54 129 129 129 129 128 3 54 132 79 169

Index origin by fluctuations 129 origin of biological 3 origin of problem 7 origins of 125; 128; 321 preceding life 126 quantum effects in origin of 133 rise of 54 second order phase transition 133 theory of hierarchical 159 homogeneous catalysts 176 honey 348 honey, D-amino acids in 348 white fir 348 human age, by anthropological methods 74 by D/L amino acid contents 74 human brain 155; 157 human brain, D-amino acids in 321 hypocampal formation 158 postnatal maturation of 155 human eye, D-amino acids in 321 human health 352 human teeth, D/L-aspartic acid content 76 D/L-glutamic acid content 76 D-amino acids and age 73 L-aspartic acid content 76 L-glutamic acid content 76 hydantoins 52 hydrogen 13 hydrogen bonding, intermolecular 266 hydrogen bonds 221; 261-265; 269; 329 hydrogen cyanide 13 hydrogen cyanide, in prebiotic chemistry 50 in prebiotic synthesis 49 hydrolysis, competitive 54 competitive, dissymmetry of 54 hydrothermal environments 203 a-hydroxy acids 52 a-hydroxy amides 51; 52 3-hydroxybutyric acid 268 a-hydroxynitriles 50 hypercube, for sequence description 4 hypercycle, enantiopure 258 hyphenated techniques, of analysis 350

419

/ iminiumions 50 imino-diacids 52 ImpA (a-anomeric), polymerization by uranyl ion 303 indole-3-acetic acid 27^ indoxyl sulfate 277 information, genetic and systemic HI informational molecules 122 informational molecules, monochirality of 122 intermolecular forces 261 interstellar grains 132 ion-exchange column 350 iridium complex (chiral), Mossbauer effect of 130 iridium complexes, diastereomers of 46 iron complex (chiral), Mossbauer effect of 131 iron, essential for living organisms 300 isocyanates 52 isodensity contour, molecular 214 surface, of molecules 213 isoleucine 68; 339 isoleucine, epimerization of 70 in age determination 70 isopropyl chloride 266 R(-)-4-(3 -isothiocyanatopyrrolidin-1 -yl)-7-(N,Ndimethylaminosulfonyl)-2,1,3 -benzoxadiazoles3 50 S(+)-4-(3 -isothiocyanatopyrrolidin-1 -yl)-7-(N,Ndimethylaminosulfonyl)-2,1,3 -benzoxadiazoles3 50

juice processmg.

.349

K a-ketocarboxylic acids, enantioselective reduction of ketone reduction, CD study of cyclic, by chromium (II) enantioselectivity in ketyl radical intermediate in mechanism of. open chain by chromium (H) UV-VIS study of ketones, cyclic asymmetric

230 227 224 225 227 227 224 227 255

420

Progress in Biological Chirality

flavour in cheese ketyl radical intermediate in ketone reduction kinetic modelling, of evolution Kohlrausch-Williams-Watts law Kondepudi catastrophic mechanism kuruma prawn

276 227 62 11 311 343

lactic acid 342 Lactobacillus 254; 341; 348 Lactobacillus brevis 347 Lactobacillus bulgahcus 342 Lactobacillus plantarum 347 lactones, flavouring 255 L-alanine 204 L-alanine. formation of. 18 in bone analysis 68 racemization of 204 L-amino acids 125; 169; 203; 340; 355; 398 L-amino acids, carbon monoxide in formation of 13 evolution of 402 exclusive selection of 321 favoured by theory 401 formation of. 22 from aziridine 17 homochirahty of 161 in proteins 311 macroscopic identity by 329 MO description of 400 racemization 65 related to D-glyceraldehyde 13 taste of. 340 L-aspartic acid, in bone analysis 68 L-DOPA 254 lead ion. catalyst for oligoRNA 304 catalyst in nucleoside oligomerization 299 leafy green vegetables 339 lecithin retinol acyltransferase (LRAT) 240 legumes 339 Lenz'slaw 21 Leonardos! a jinxiensis. Z.Wang 369 Leonardosia, Sommer genus 368 leucine 339 Leuconostoc mesenteroides 347 L-fuconolactone derivatives 186 L-glutamic acid.

in bone analysis L-histidine, in bone analysis life, as chirality induction as negentropy current homochirahty as condition of mirror-antipode origin of origin of on Earth preceded by homochirahty lightning, as free energy source limestone lipid bilayer, orientation of rhodopsin in L-isoleucine living beings, fundamental constituents of living organisms, characteristic of components of. entropy of living systems, specific synunetry of L-molecules, adsorption energies of lobsters local chiral environments local electron density long-term memories, stability of L-phenylalanine, in bone analysis L-quartz crystals L-selectins L-serine, N-stearoyl, micelles of octamers of L-sugars L-threonine, in bone analysis L-tryptophan L-tryptophan, in bone analysis interaction with serum albumin L-tyrosine L-valine, in bone analysis Lyapunov parameters lysine lysozyme, D-amino acid in

68 68 216 257 79 6 397 125 126 173 145 241 65 83 355 173 216 119 137 343 149 210 158 68 130 185 335 129; 311 68 321 68 288 130; 330 68 259 339 69

Index M macrophytes 365 magnesium ions, in chlorophyll 19 magnesium, chelated by porphin 19 magnesiumporphin complex, HOMO of 20; 21 LOMOof 20 resonant structures of 20; 21 magnesiumporphin, acetaldimine and CO 22 adduct with acetaldimine and CO 17 coordinated to CO 22 reaction with CO 21 magnesiumporphinCO complex 20 magnetic field, effect on spin diffusion 314 effect on spin-spin relaxation time 314 Maillard reaction 340; 344 malolactic bacteria 343 maltose 344 Mammals 339 Mammals, placement of organs in 160 (S)-mandelic acid 266 (RVmandelonitrile 268 manganese, essential for living organisms 300 marble 145 marine minerals 145 Mars 137 marsh crab 343 mastitis 341 mercury ion, coordination property of 300 messenger RNA 83 metabolic cycles, self-sustaining 154 metabolic stability 99 metal-ligand contacts 261 metal-ligand interactions 264 meteorites 52 meteorites, amino acids in 128 conglomerates in 79 enantiomeric excess in 132; 133 L-amino acids in 132 methionine 339 methyl 2-deoxy-2-C-sulfono-a-D-mannopyranoside 189 methyl 2-deoxy-2-C-sulfo-p-D-glucopyranose.. 189 methyl 2-thiotrityl-p-D-glucopyranoside surrogate 189

421

methyl 4-S-acetyl-2,3,6-tri-0-benzyl-4-thio-a-Dgalactopyranoside 190 methyl 4-S-acetyl-2,3,6-tri-0-benzyl-4-thio-a-Dglucopyranoside 190 methyl 6-deoxy-2,3 -0-isopropylidene-4-S-acetyl-4thio-a-L-mannopyranoside 192 methyl 6-deoxy-2,3-0-isopropylidene-4-S-acetyl-4thio-a-L-talopyranoside 192 methyl 6-deoxy-6-C-sulfo-a-D-galactopyranoside 192 methyl ketones, oxidation of 255 a-methyl-amino acids 54 micelles, chiral organization of 335 Micrococcus 341 Micrococcus luteus 91 microtubules, of proteins milk products, D-amino acids in Miller indices Miller's experiments minerals, ferromagnesian miniaturization, molecular level of Miocene evaporite deposit mirror movements, in infants Moellerina greenii molecular capsules, bimolecular reactions in used for enantioselection molecular chaperon, albumin molecular clockwork, chiral molecular engine molecular evolution molecular populations, in early life molecular recognition molecular recognition motif molecular switches molecule guests molecule hosts molecules, complexity of MoUer-Plesset perturbation theory molybdenum, essential for living organisms monomer strands, racemic doublet of monosodium glutamate Moon

171 348 138 52 142 29 149 156 367; 368 264 265 279 29 60; 61; 62 215 154 42 47 29 261 261 128 17 300 154 348 137

422

Progress in Biological Chirality

morphogenesis, human morphological asymmetry morphological asymmetry. and molecular asymmetry in Mammals origins of movements, contralateral involuntary mirror of infants of young children mRNA, considered as genes muramic acids Murchison meteorite Murchison meteorite, nonracemic material in Murray Kentucky meteorite, nonracemic material in muscone mutation mutation jumps mutation jumps, hierarchical level of mutational explosion

160 161; 171 171 159 161; 171 155 155 155 155 155 83 185 79 79 79 255 177 8 10 5

N N-(0)-trifluoroacetyl alkyl esters

351

N2O3,

reactions of

57

N2O4,

reactions of 57 N-acetylcysteine 350 N-acetylpenicillamine 350 N-alkyloxycarbonyl-trimethylsilyl derivatives... 351 natural amino acids. symmetry violation by 329 nature, achiral inorganic 257 chiral organic 257 NaHCOs, in prebiotic synthesis 49 N-carbamoyl amino acids 52; 56 N-carbamoyl peptides 53 N-carbamoyl valine 56 N-carbamoylamino acids 52 negentropy. and life 257 nested ultrametric spheres, hierarchy of 9 neuro-transmitter substances, chemical 158 neutrino.

polarized flows of neutron star, polarized gamma-radiation by rotating NH3, in prebiotic synthesis nickel, essential for some organisms N-isobutyryl-L(or D)-cysteine nitrogenous metabolites nitrous acid NMR studies, solvent effect in NO, in prebiotic synthesis reactions of NO/O2, mixture (+)-nojirimycin Na Nozaki-Hiyama reaction N-perfluoroacyl alkyl esters nuclear weak force nucleation centers, of crystals nucleoprotein nucleoside oligomerization, catalyzed by lead ions nucleoside phosphorimidazolide, prebiotic synthesis of nucleoside sultons nucleosides. from carbon monoxide reactions leading to nucleotide analogs nucleotide bases, hydropathy of nucleotide sequences, inRNA nucleotides, from carbon monoxide reactions leading to

257 6 207 49 300 350 339 52 312 49 57 52; 57; 62 192 52 222 351 335 329 99 299 298 196 27 27 197 85 3 27 27

O O2.

in prebiotic synthesis reactions of ocean, primitive sea floor in ochratoxin, interaction with serum albumin odorants, binding to protein volatile components of food

49 57 53; 203 207 276 275 276

Index oligoadenylate formation, uranyl ion-catalyzed oligoadenylate, fonnation on a poly(U) template template synthesis of. oligoribonucleotides, by template-directed chain elongation oligouridylate, mechanism of formation of template synthesis of. olivine ontogenesis, in age estimation Oparin Oparin's theory early life opsin, reaction with ll-c/>retinal reaction with 13 -desmethylretinal transient dark activity in optical isomers organic chemistry, synthetic organic compounds, abiotic synthesis of organic molecules, adsorption of appearance of on Earth organs, asymmetric placement of origin(s)oflife origin(s)oflife, molecular Orion Nebula, circularly polarized light of orthoclase ortho-phthaldialdehyde osmium complexes, diastereomers of oxazaborolidines oxide minerals, chiral surfaces of oxime reduction, CD study of mechanism of UV-VIS study of oximes, reduction of oximinocarboxylic acid, enantioselective reduction of

300 307 305 308 307 306; 307 149 66 125 154 248 247 249 340 35 153 141 127 160 35 49 132 142 350 46 222 137 229 229 229 229 232

P parity violating, electro-weak interactions energy energy difference

166 336 130; 317

423

energy difference between enantiomers 311 interaction 397 weak interaction 129; 132; 133 p-decay 207 p-radiation 166 parity violation 398 Parmesan cheese 254 Parmesan cheese, ageing markers in 255 epimerization of amino acids in 255 free amino acids in 255 free fatty acids in 255 migration process in 255 volatile components of 255 Parmigiano-Reggiano cheese 342 Pasteur 39; 125; 311; 340 Pecorino cheese 342 Pediococcus 347 Pennicillum 276 R,S-2-pentanol 266 pentoses, formation of 23 peptides, catalytic effects of 49 co-emergence of 49 co-evolution of 49 emergence of, the catalytic activities of 62 evolution of 54 formation rate of 58 hetero 54 homo 54 homochiral 62 hydrolysis of 54 quantum description of 406 synthesis of 58 peptidoglycan 341 Permian-Triassic boundary, biological crisis 371 crisis, enantioselective influence of 373 perturbation theory, time-dependent 15 phagocytosis, in eye 238 pharmaceuticals, chiral piuification of 137 phase transition, second order and homochirality 133 phenylacetylcobalt tricarbonyl triphenylphosphine derivatives 34 phenylalanine 329; 339 phenylalanine, enantioselective synthesis of 231 rateofracemizationof. 69 phenylbutazone 271 phenylcyclopropane 266 (R)-l-phenylethanol 268

424

Progress in Biological Chirality

1-phenylethanol, reductive synthesis of. 1 -phenylethyl amine, reductive synthesis of. phenylglycine, enantioselective synthesis of para-hydroxy philosophy, Democritian trend of Platonian trend of. phospholipids photochemistry, with circularly polarised light phyllotaxis, spiral phylogenetic tree phylogenetic tree, hierarchy of phylogenetic trees pianostool, inverted inverted structure of Ru complexes (+)-pinanediol plant polysaccharides plants platinum, chiral surfaces of polarized light, effect on porphin effect on tetrahydroporphin Polya Polya experiments polycondensation, kinetic measurements of polyglycines polymers polypeptide, chirality levels in polypeptides polyribonucleic acids, chirality of. essential tasks of helical strands of replication by translation by polysaccharides ponasterone A population, degradation of Porocharaceae, Grambast family of porphin, ab initio treatment of. fi^ee electron treatment of LCAO treatment of polarized light effect on

225 229 231; 232 331 125; 133 125; 133 181 13 366 8 8 88 47 45 261 185 339 137 15 15 153 154 57 58 185 169 173 79 79 79 79 79 173 283 10 368 14 14 14 15

template for CO porphyrins, charge-transfer complexes with nitrogen compounds in geological deposit magneto-optical rotatory dispersion of postnatal, brain maturation development Praesycidium siluricum prawns pre-biological, enantioselection evolution prebiotic chemistry, essential molecules of primary pump in reducing medium in prebiotic chiral selection prebiotic environments, calcite in prebiotic evolution, hierarchical relationships in random factors in prebiotic molecules, enantiopurity of prebiotic reaction, cooperative premotor cortex, of left-handed subject of right-handed subject prenols primary pump primary pump, in prebiotic pump on tidal beaches primeval atmosphere primitive synthetase activities Primochara calvata principal dinucleotide, symmetries in matrix of products, pharmaceutical Propionibacterium protein conformation, and odorant binding complexity of evolution of protein residue hydropathy protein stability, and punctuation systems rules of protein synthesis, initiation of termination of without translation

19

13 13 13 155 156 368 343 29 5 50 56 13 137 145 7 7 79 54 155 155 181 53; 60 56 62 13 92 368 86 340 341 276 107 107 89 100 99 108 110 93

Index proteins, amino acid distribution in complete dietary hydrolysis of incomplete transmembrane a-helical segment of proteolysis of casein P-selectin receptor inhibition P-selectins Pseudomoellerina, Wang Zhen genus pseudotetrasaccharide psychrotropic bacteria purge and trap analysis pyridazinyl ring, inversion of pyridoxal 5'-phosphate pyridoxal phosphate pyrite, catalyst for biomolecules pyroxenes

100 339 339 68 339 185 169 342 182 185 368 187 341 253 264 322 254 128 142; 149

Q quantum chemistry calculations 210 quartz 137; 138; 144; 149 quartz, crystal faces of 140 crystals 130 helices in 140 left handed 140 right handed 140 quercetin 278 R R,C0R2, in prebiotic synthesis R3NH2, in prebiotic synthesis racemates, melting points of racemic compounds, as entropy carrier racemic mixture racemic world racemization, and Hamiltonian elements apparent rates of. racemized amino acids, determined with HPLC radioactivity, as free energy source radiocarbon method

49 49 265 404 173 153 399 207 70 173 65

425

random choice, memory of 4 regioselectivity 221 repHcation cycle, life time of. 154 retinal analogs, for conformation study 249 retinal pigment epithelium (RPE) 240 retinal, W-cis238 11-CW-, chromophore, in rhodopsin 246 1 \-cis-, NMR study of 246 CD study of 245 conformation of 245 (1 lZ)-3-diazo-4-oxo[15-^Hi]241 [15-'H]-3-diazo-4-oxo-10,13-ethano-l l-c/5-. 241 (1 lZ)-locked cyclopropyl245 retmoids, adamantyl allenic 243 1 l-c/5-retinol 240 trans-XQi\no\ 240 rhodium complexes, diastereomers, molecular recognition of 42 diastereomers, NMR study of 41 diastereomers of 41; 46 rhodium, chiral center on 41 rhodopsin visual cycle 239 rhodopsin, batho239 CD spectra of 241 chromophore conformation in 247 chromophore of 237 diazoketo241 diazoketo-, bleaching intermediates of 243 diazoketo-, CD spectra of. 241 diazoketo-, UV/VIS spectra of 241 from bovine eye 240 in human eye 237 interaction with water molecules 241 lumi240 meta-I240 meta-Il240 orientation of in lipid bilayer 241 photo 239 photo-affinity studies of 240 photocycle of. 243 UVA^IS spectra of. 241 X-ray structure of 240 ribonucleoprotein 101 ribose ketene, formation of 25 ribose phosphate, formation of 25 ribose, from carbon monoxide 27

426

Progress in Biological Chirabt>

probability density function of reactions leading to ribozymes RNA RNA (oligo), lead ion catalyst for ligation on RNA template mixed sequence synthesis of template synthesis of RNA. a-anomerof P-anomerof as information carrier as primordial genes binding by proteins considered as genes hetero-chiral hetero-linkage in homo-chiral homo-linkage in metal ion catalyst for prebiotic synthesis model of pyranose type selection of isomers of synthesis non-enzymatic template for rephcation world rock lobster rock-forming minerals rocks, feldspars in igneous metamorphic sedimentar>' rotational freedom, reduction of rumen bacteria Rumer transformation ruthenium complexes, diastereomers of molecular recognition of pianostool structure of

316 27 92 49: 153 304 307 307 308 309 309 297 93 101 83 309 309 308 308 297 297 297 308 298 297 92 49 343 139 142 142 142 142 29 341 120 43; 46 43 43

S Saccharomyces caHsbergensis Saccharomyces cerevisiae Saccharomyces exiguous Salam-process sauces, D-amino acids in sauerkraut sauerkraut juices, D-amino acids in Schrodinger equation

344 344 348 133 348 341; 347 348 399

sea floor, in ocean seafood quahty selectin-carbohydrate interaction selectivity, chemodiastereoenantioregioself-catalytic process, amplifying mechanism of self-complementary molecular subunits self-organization, intramolecular self-replication, by assembling monomers sequences space serum albumin serum albumin, biosensor response to CD study of chirahtyof complex difiusion coefficients of. complex relaxation rates of. complex studied by DOSY complex studied by NOESY complex with bilirubin complex with coumarin complex with crocetin complex with curcumin equilibria of ICD study of interaction with antioxidants interaction with berberine interaction with flavours interaction with food components interaction with L-tryptophan interaction with ochratoxin interaction with salycilate interaction with toxins interaction with vanillin ligand binding equihbria of ligand binding kinetics multiple binding sites of NMR studied by PQCI study of QCM study of SPR study of shore crab silicate minerals, chiral surfaces of Silurian, late silver ion, coordination property of silver,

207 343 185 221 221 221 221 399 261 31 154 4 271 290 280 279 287 287 287 288 283 281 286 284 281 283 275 291 275 275; 291 288 276 287 275 289 273 274 272 286 291 289 289 343 137 365 300

Index chiral surfaces of snails snow crab sodium ammonium tartrate Solar system, evolution of solid phase microextraction soluble drugs, concentration of sour cabbage sourdough, fermentation of space-time symmetry, of Universe Spallanzani spin-spin relaxation time, measurement of ofenantiomers. spin-spin relaxation, magnetic field effect on spirits sponge spontaneous chiral symmetry breaking, general theory of........ spontaneous symmetry breaking spontaneous symmetry breaking, general features of 7C-7C stacking starfish starter cultures .„. steady state , stereoisomerism, new forms of stochastic population growth, with feed back stochastic process, in evolution stoneworts Strecker reaction Streptococcus Streptococcus thermophiius (R)-styrene oxide sugar C-sulfonic acids, secondary sugar esters sugar phosphates sugar sulfates sugar sulfonic acids, anomeric sugar sulfoulosonic acids , sugar sulphonic acid amide...... sugar, C-sulfonic acids 3 '-C-sulfonated lactose derivatives Sulfoquinovose sulfoquinovosyl monoacylglycerols

137 370 343 39 126 253 271 347 348 ,123 ...,.125 315 .315 314 345 181 161; 164 ...171 162 221 181 341 177 264 153 5 365 51 254; 341 ,342 266 189 185 185 185 192 185 197 181 187 182 181

sulfoquinovosyldiacylglycerols sultones, fused to sugar ring sunlight, as free energy source super-continent Pangea symmetry breaking, kinetic model of symmetry deficiency measure symmetry deficiency measure, external internal symmetiy, P, spontaneous breaking of synthetase classes synthetases, coupled to tRNA system, farfromequilibrium Szent-Gyorgyi, Albert

427 181 196 173 372 56 214 213 212

329 257 86; 88 104 173 216

T tartaric acid, crystals of taste, bitter sour sweet teeth temperature, as free energy source template effects of host molecules tetrahydroporphin, ab /w/Y/o treatment of free electron treatment of LCAO treatment of polarized light effect on thermodynamics , thermodynamics, second law of thiol compound, chiral..... threonine threonine, crystallization of rateof racemizationof tidal beaches, primary pump on tomato products, D-amino acids in .,..,. topaz..,., transition metal catalysts , transition metal complexes, chiral ligands in

340 340 340 340 65 173 261 14 14 14 15 397 397 350 339 80 69 62 348; 349 ..149 222 222

428

Progress in Biological Chirality

transition metal ion, - amino acid complexes, equilibria catalysts forprebiotic syntheses equilibria tremolite triammonium phosphate tris(l,2- ethanediamine) cobalt(III) tris(l,2-ethanediamine) iridium(III) tRNA derivation tRNA dimers, thermal stability of tRNA pairs, in protein synthesis successive addition of tRNA, coupled to synthetases tryptophan tryptophan, probability density function of. rate of racemization of tryptophanase tryptophanase, ammonium phosphate effect on CD study of conformation and activity of conformational changes of effect of hydrophobic amino acids on fluorescence spectroscopy study of folding and activity of kinetic constants of kinetic study of Michaelis constant of subtle conformational changes temperature effect on turmeric, yellow dye of tyrosine D,L-tyrosine tyrosine, CD study of crystallization of interaction with ortho-v/ater solubility of in water supersaturated solutions of, tyrosine-phenol liase

223 300 223 147 322 132 132 92 93 95 95 104 254; 339 315 69 321 321 321; 324 326 324 326 324 326 323 323 323 324 326 283 254; 329 329 333 330; 334 336 329 329 254

U udder inflammation ulfoqumovosyl diacylglycerols ulosonic acids unit cell Universe, evolution of particle asymmetry of uranyl ion,

341 182 185 138 125; 126 398

as a catalyst for oligonucleotide synthesis coordination property of. uremic toxins

300 300 277

V valine 60; 339 valine, in age determination 70 rate of racemization of 69 van der Waal's forces 261 vanillin 276 vanillin, interaction with serum albumin 289 vegetables, D-amino acids in 349 Vester-Ulbricht process 129; 130 vinegar 341; 345 vinegar, "aceto balsamico tradizionale di Modena" 346 balsamic 346 D-amino acids in 348 from cider 346 from red wine 346 from sherry 346 from white wine 346 vision, age-related macular degeneration (AMD) of. 238 black and white 237 p-carotene m 238 colors in 237 femtosecond photochemistry in 239 neural signal of. 238 of cow 237 of dog 237 of frog 237 of goldfish 237 of horse 237 of human 237 of owl 237 of quid 239 of rat 237 of salmon 238 of sheep 237 of tadpole 239 scopotic 237 vitamin A in 238 visual transduction cycle 239 visual transduction path 244 volatile aroma compounds, in wine 343

warfarin water

271; 278 13

Index water, brackish plants in fresh plants in in prebiotic chemistry in prebiotic synthesis orthoor^/io-interaction with tyrosine parawavefunctions, perturbed by magnetic field weak force weak nuclear forces, asymmetry of wine wine fermentation wine, fortified red

365 365 50 49 336 336 .....336 15 311

rose white woolly-handed crab

429 343 343 343

F yeast yeast extracts yeast, of baker of brewer of wine

343 348 348 348 348

29 341; 343; 345 343 344 343

zmc, essential for living organisms.

.300