Advances in Supramolecular Chemistry. Volume 7

",4

---t -<

I'H

-1-

C r">

!"1"1

~n O~

>" Z

C

This Page Intentionally Left Blank

ADVANCES IN SUPRAMOL Editor: GEORGE W. GOKEL Department of Molecular Biology and Pharmacology Washington University School of Medicine St. Louis, Missouri VOLUME 7

JAI PRESS INC.

Stamford, Connecticut

Copyright 9 2000 by JAI PRESS INC. 100 Prospect Street Stamford, Connecticut 06904-0811 All rights reserved. No part of this publication may be reproduced, stored on a retrieval system, or transmitted in any way, or by any means, electronic, mechanical, photocopying, filming, recording, or otherwise without prior permission in writing from the publisher. ISBN: 0-7623-0678-5 Manufactured in the United States of America

CONTENTS

LIST OF CONTRIBUTORS

vii

PREFACE George W. Goke/

xi

SUPRAMOLECULAR ASSEMBLIES IN NATURAL AND ARTIFICIAL ION CHANNELS Gregory J. Kirkovits and C Dennis Hall ION RECOGNITION AND TRANSPORT BY POLYo(R)-3-HYDROXYBUTAN OATES AND INORGANIC POLYPHOSPHATES

Rosetta N. Reusch

49

FUNCTIONALIZED MACROCYCLIC LIGANDS AS SENSORY MOLECULES FOR METAL IONS

Guoping Xue, Paul B. Savage, Jerald S. Bradshaw, Xian X. Zhang, and Reed M. Izatt

99

CHIRALITY IN CALIXARENES AND CALIXARENE ASSEMBLIES

Myroslav Vysotsky, Christian Schmidt, and Volker B6hmer

FROM MOLECULES TO CRYSTALS: SUPRAMOLECULAR SYNTHESIS OF SOLIDS

Brian Moulton and Michael J. Zaworotko

INDEX

139

235 285

This Page Intentionally Left Blank

LIST OF CONTRIBUTORS

Volker B6hmer

Institute fur Organische Chemie Johannes Gutenberg Universit~itMainz Mainz, Germany

Jerald S. Bradshaw

Department of Chemistry and Biochemistry Brigham Young University Provo, Utah

C. Dennis Hall

Department of Chemistry King's College, University of London London, England

Reed M. Izatt

Department of Chemistry and Biochemistry Brigham Young University Provo, Utah

Gregory J. Kirkovits

Department of Chemistry King's College, University of London London, England

Brian/Vloulton

Department of Chemistry University of South Florida Tampa, Florida

Rosetta N. Reusch

Department of Microbiology Michigan State University East Lansing, Michigan

Paul B. Savage

Department of Chemistry and Biochemistry Brigham Young University Provo, Utah

Christian 5chmidt

Institut fur Organische Chemie Johannes Gutenberg Universit~t Mainz Mainz, Germany vii

viii

LIST OF CONTRIBUTORS

Myroslav Vysotsky

Institut fur Organische Chemie Johannes Gutenberg Universit~at Mainz Mainz, Germany

Guoping Xue

Department of Chemistry and Biochemistry Brigham Young University Provo, Utah

Michael J. Zaworotko

Department of Chemistry University of South Florida Tampa, Florida

Xian X. Zhang

Department of Chemistry and Biochemistry Brigham Young University Provo, Utah

PREFACE

As the field called "supramolecular chemistry" advances and enlarges, this series of monographs becomes ever broader in scope. The goal of Advances in Supramolecular Chemistry to present a variety of articles that encompass the current scope of supramolecular science. This volume continues to reflect that broad range but also reflects some personal bias. The latter is apparent in the two chapters dealing with channel compounds. Each of these chapters, though, is broad in its own way. Reusch's presentation is very close to the biology interface and describes poly-3-hydroxybutanoate channels. These have been studies by using electrophysiological techniques but their structure and activity have been confirmed in the classical way by total synthesis. Hall and Kirkovits' description of synthetic channels describes some of their recent work to incorporate an electrochemical switch within the channel and thence within the phospholipid bilayer. Hall's molecules might be referred to as macrocyclic compounds that are sensors for metal ions. This topic is the subject of the chapter written by Bradshaw and coworkers. The compounds in question are not channels at all but were developed from what might be called the "organic-analytical" perspective. Calixarenes, a receptor system of incredibly broad current application, is described from the stereochemical perspective by a pioneer in this important field. Current and authoritative as this chapter is, an important message is that much remains to be undertaken and understood. ix

x

PREFACE

In the final chapter of this volume, Moulton and Zaworotko present a broad description of the emerging discipline of "crystal engineering." The explosion in this area into a field in recent years could not be adequately described by any single chapter. Still, this report introduces and tantalizes the reader with a range of exciting results. This is clearly a field where more remains to be discovered than has been unearthed to date. It is hoped that this broad view of supramolecular chemistry and the many directions it leads will be of interest to those already in the field. As in previous volumes, it is hoped that those outside the field may see extensions of their own work that will bring them into it. George W. Gokel Editor

SUPRAMOLECULAR ASSEMBLIES IN NATURAL AND ARTIFICIAL ION CHANNELS

Gregory J. Kirkovits and C. Dennis Hall

1. 2.

3.

4. 5. 6.

Introduction: Ion Channels Across Cell Membranes . . . . . . . . . . . . . . . Natural Ion Channel Proteins . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1. Potassium Selectivity . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2. Gramicidin: A Model for Ion Channel Conduction . . . . . . . . . . . . Single Channel Measurements . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1. The Planar Lipid Bilayer Method . . . . . . . . . . . . . . . . . . . . . . 3.2. Patch-Clamping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Synthetic Peptide Ion Channel Models . . . . . . . . . . . . . . . . . . . . . . Engineered Natural Ion Channels . . . . . . . . . . . . . . . . . . . . . . . . . Classification of Artificial (Non-Peptide) Ion Channel Models . . . . ...... 6.1. Early Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.2. Proton Channels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.3. Macrocycle Based Mimics . . . . . . . . . . . . . . . . . . . . . . . . . 6.4. Macrocyclic Peptides . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.5. Synthetic Channels via Aggregate Formation . . . . . . . . . . . . . . . 6.6. Ion Channel Forming Polymers . . . . . . . . . . . . . . . . . . . . . .

Advances in Supramolecular Chemistry Volume 7, pages 1--47. Copyright 9 2000 by JAI Press Inc. All rights of reproduction in any form reserved. ISBN: 0-7623-0678-5

2 3 4 8 8 9 8 10 11 13 15 15 18 33 36 41

2

GREGORY J. KIRKOVITS and C. DENNIS HALL 6.7. Modeling Potassium Channel Selectivity . . . . . . . . . . . . . . . . . 7. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9

42 44 44

INTRODUCTION: ION CHANNELS ACROSS CELL MEMBRANES

Conduction of ions across cell membranes is required to modulate the ionic composition of the intracellular medium within all living cells. The phospholipid membrane bilayer, however, constitutes a hydrophobicbarrier that isolates a living cell from its surrounding environment. Consequently, the membrane is normally impermeable to small ions because a considerable amount of energy is required to transfer a charged species, such as Na § from a high dielectric medium (e.g. water where it is heavily solvated) through the low dielectric medium of a nonpolar membrane and hence into (or out of) the cell. The problem of how this barrier is overcome has intrigued biologists, biochemists, and chemists for many years since understanding of this phenomenon is probably the essential key to understanding cellular function. Rapid ion flow across membranes is known to occur via carrier and channel molecules (Figure 1) and the primary function of both mechanisms is to provide a low-energy pathway for ions to traverse the cell membrane. Traditionally, the carrier molecule was viewed as a "ferryboat," binding the ion at one interface and transporting it through the membrane as a complex, l The modern interpretation of the carrier mechanism however, is somewhat different. The molecule actually spans the membrane and appears to exist in two conformations. A reorientation step then leaves the binding sites alternately exposed to the intra- and extracellular media. Channel molecules, on the other hand, provide a contiguous pore, i.e. a tunnel-like path for ion transport. More specifically, channel proteins form distorted cylinders with hydrophobic surfaces that interact with the lipid and hydrophilic surfaces lining the inner pore. The pore provides an aqueous (or partially hydrated) environment lined with polar residues that can interact

8'

|

JO lO

@

(9

|, 1

IO ....

@ |

~'

u,

8"i

......

'. . . .

:

.~0

,o 10

I0

Figure 1. Schematic illustrations of the mechanisms of facilitated ion transport through a membrane bilayer: carrier (left) and channel (right). M + and L represent cation and ligand respectively. Reproduced with permission from G. Gokel, Acc. Chem. Res. 1996, 29, 425-432.

Natural and Artificial Ion Channels

3

favorably with ions to facilitate their transport while the protein shields the ion from the low dielectric medium of the membrane lipid. Rates of transport differ significantly for the two mechanisms. Channels are capable of permitting ion transport at rates exceeding 107 ions s-1, while the maximum velocity of carriers, limited by the rate of the conformational changes they undergo is ca. 104 ions s -1. A great deal of information has accumulated in recent years about natural ion channel proteins due to advances in techniques such as patch-clamping, ion channel reconstitution, site-directed mutagenesis, and crystallography. However, a clear understanding of the chemical and mechanistic nature of channel function has yet to be obtained. In most cases it is not known with certainty which protein segments span the membrane, how those segments are anchored in the membrane, which residues align to form the pore, which groups serve as the relay elements and what determines selectivity. The size and complexity of the protein aggregates and the inherent difficulty of isolation hinder solutions to these problems. Since ion channels are reputed to be involved in several diseases 2 intense efforts have been devoted to the design and study of model systems that may provide an insight as to the function of natural ion channel proteins and the subject falls clearly into the arena of supramolecular chemistry.

2. NATURAL ION CHANNEL PROTEINS When open, the channel provides a water-filled pore that extends across the entire length of the bilayer. The pore is much wider than an ion over most of its length, but narrows in a small stretch known as the selectivity filter (vide infra), where ionic selectivity is established (Figure 2). However, the flow of ions across cell membranes is not continual but a precisely regulated biological process, a direct result of the ion channels opening or closing in response to a specific stimulus. The response is referred to as gating. Gating requires a conformational change of the pore that moves a "gate" into and out of an occluding position. A specific biological stimulus is sensed by a channel which controls the probability of the pore opening and closing. Ion channels are categorized into three types in terms of the biological signals that give rise to the gating. These are the voltage-dependent channels which respond to a change in the electrical potential of the cell membrane, the ligand-gated (or receptor-operated) channels which are activated by the binding of a ligand to an associated cell receptor (e.g. the acetylcholine receptor channel), 3 and the mechano-sensitive channels that are controlled by extension of the cell membranes. Ion channels are ubiquitous and undoubtedly occur in the lipid bilayer of all cell membranes. They have many functions and some are critical to cellular activity. These include the generation of cellular energy, the production of electric signals in the nervous system, and a variety of signal-transduction processes. The inorganic ions involved in this signaling include sodium, potassium, and calcium. Each

4

GREGORY J. KIRKOVITS and C. DENNIS HALL iOrl fhlx

~ l selectivity filter

voltage sensor

r

|

/ gate

b

membrane pore

Figure 2. Schematic diagram of a voltage-gated channel. Reproduced with permission from S. Futaki, Biopolymers 1998, p. 76, Fig. 1. channel is designed for selective ion permeability, and can be distinguished further by the stimulus that causes it to open or close. Sodium ion transport is fundamental to the generation of electric signals in nerve cells and muscles and an important class of sodium channel is the voltage-activated sodium channel (VASC) which is found in the membranes of nerve and muscle cells. As the name suggests it is reversibly opened or closed by changes in transmembrane potential. It is a single polypeptide chain, approximately 260 kDa in molecular mass, built of four repeating units each one composed of six membrane-spanning m-helical stretches (Figure 3). One stretch in each of the four units has a five- to sevenfold repeat of the sequence, (-Arg-X-X-) where Arg (or sometimes Lys) is the positively charged amino acid arginine (or lysine). It is thought that this stretch senses the change in voltage of the cell membrane which gates the sodium channel. 3 Remarkably, the amino acid sequence of the VASC is similar to that of two other (Ca 2§ and K § voltage-gated channels.

2.1. Potassium Selectivity One of the long held aims of ion channel research has been to understand ionic selectivity. This has been exemplified by the study of potassium channels. The functions of potassium channels are extensive and include brain cell communication and the maintenance of cell size and shape. They are also possibly the most highly regulated of all ion channels, and exhibit remarkable selectivity--a factor of 10a-fold higher for K + over Na +. Structurally, they fall into two categories: the voltage-gated potassium channels (K v) and inwardly rectifying potassium channels

(Kir).

Natural and Artificial Ion Channels

5

Figure 3. Possible model of a voltage-gated K§ channel (i) showing the six transmembrane spanning helices. Included is the helical stretch that contains the positively charged amino acids that are thought to sense the transmembrane voltage and hence gate the channel. The pore loop, which contains the selectivity filter is shown to the right of the model. Both the N- and C- termini are intracellular. (ii) Viewed from above, the functional channel is composed of four identical subunits. Reprinted with permission from Nicholls, D. G., Proteins, Transmitters and Synapses, 1994, p. 52, Blackwell Science, Oxford. Data were also taken with permission from Trends in Neuroscience, 1990, 13, 201-206, Guy, H. R. and Conti, F., "Pursuing the structure of voltage-gated channels," and from Trends in Pharmacol. Sci. 1992, 13, 359-365, Pongs, O., "Structural basis of voltage-gated K§ channel pharmacology," with permission from Elsevier Science. The K v channels make up a large molecular family, the members of which are identical in essential aspects of structure and mechanism. Very little structural information was available until the successful cloning and sequencing of the Shaker gene of Drosophila, which codes for a voltage-gated K § channel. 4 As the name implies, the I~ channels "gate" in response to changes in the membrane potential of the cell. Structurally, they consist of four subunits that assemble around the channel pore (cf. Figure 3). Each subunit is composed of five hydrophobic transmembrane segments and one transmembrane segment that contains a positive charge on every third residue and plays the role of the voltage sensor. The pore region is rather hydrophobic and contains a number of highly conserved aromatic residues. This discovery was followed by that of the inward-rectifier channels (Kir) which, remarkably, only allow the flow ofK § ions in the inward direction against the natural concentration gradient. Structurally, they are smaller than the K v channels, consisting of only two transmembrane domains per subunit which flank the pore region. However, the pore sequence was found to be highly homologous to that of the K v channels 5 and, like the K~, channels, the functional channel protein is a tetramer typically composed of four identical subunits. Sequence alignment of all the cloned K § channels revealed that, without exception, a critical homologous amino acid

6

GREGORY J. KIRKOVITS and C. DENNIS HALL

sequence exists in the pore region. 5 The signature sequence is Thr-Xxx-Thr-ThrXxx-Gly-Tyr-Gly and in particular consists of a highly conserved Gly-Tyr-Gly (G-Y-G) sequence which appears to be essential for ion selectivity. It was found that mutation of this sequence disrupted the ability of the channel to discriminate between K § and Na § This raised important questions as to what enabled the channels to differentiate so effectively between K § and Na § The implication was that strong energetic interactions existed between K § ions and the pore which led to the proposal that perhaps cation-re interactions were responsible for establishing selectivity. 6 One model in particular suggested that tyrosine (Tyr) formed a tetrameric cage with the corresponding tyrosines on the other subunits serving as the narrowest region of the pore and being fundamental to ion selectivity. In order to investigate this hypothesis Dougherty undertook theoretical studies of the interaction of small cations with benzene both in the gas phase and in water. 7 He showed that in water the affinity of a cation-x site follows the order K § > Rb § >> Na § Li § which is qualitatively the same sequence as observed for K § channels. The major breakthrough in defining selectivity came, however, in 1998 with the determination of the crystal structure of a K § channel from the bacterium Streptomyces lividans (referred to as the KcsA K§ channel), a member of the Kir family. 8 The structure shows that the two transmembrane spanning sequences are essentially o~-helical. A total of eight helices (two from each subunit) wrap around the central pore and open up slightly toward the extracellular surface of the channel, rather like a flower. The structure reveals that the peptide backbone carbonyl oxygen atoms of the critical G-Y-G sequence actually form the selectivity filter. The pore region forms a loop involving a small "pore" helix followed by a turn that contains the G-Y-G sequence (Figure 4). In this conformation, however, the Tyr side chain is unable to contribute any cation-r~ interactions to ion binding because itpoints away from the channel. Three consecutive carbonyl atoms from each subunit, however, point inward, providing a ring of oxygen atoms that participate in cation binding. The analogy to crown ethers is obvious but in addition the amide carbonyl oxygen atoms provide a more formidable negative electrostatic potential. By incorporating the more electron dense Rb § ions, visualization of ions within the pore region was achieved. In order to explain the selectivity it was proposed that the G-Y-G carbonyl groups are held in precisely the correct geometry for binding K § by the rest of the protein structure which thus prevents the ionophore from adjusting to provide favorable interactions with Na § The highly conserved Tyr residue also plays a fundamental role in defining this molecular framework by hydrogen bonding to a tryptophan (Trp) residue in the pore helix, and additionally forming favorable van der Waals interactions with another Trp in the pore helix of an adjacent subunit. Thus a framework of aromatic amino acids is produced that provides a "layer of springs" to hold the carbonyl oxygen atoms in precisely the correct position. In so d~ing, a K § ion is permitted to enter the selectivity filter where evidence suggests that this ion dehydrates (almost completely). However, the energetic cost of the dehydration

Natural and Artificial Ion Channels

,

7

ooo

pore helix

~c c

Figure 4. Schematic representation of the KcsA K§ channel structure. Two of the four subunits are shown; the other two lie above and below the plain of the page. Each subunit is composed of the N- and C-terminated 0~-helices, the pore helix, and the loop containing the conserved G-Y-G sequence. The selectivity filter resides near the extracellular face of the membrane. Three positive ions are shown as they appear approximately in the crystal structure. Reproduced with author's permission from Dougherty, D. A.; Lester, H. A. Angew. Chem., Int. ed. Engl. 1998, 37, 2330. is compensated by the carbonyl oxygen atoms, which come into very close contact with the K § ions. On the other hand, the precisely regulated geometry impedes the selectivity filter from accommodating a Na § ion since the carbonyl oxygen atoms cannot approach close enough to compensate for the energetic cost of dehydration of the smaller Na § ion. Although a member of the Kir family, the KscA potassium cation channel pore region resembles that of a K v sequence especially in view of the crucial Trp residue which is conserved in all members of the K v family. In Kir channels the Trp residue is actually replaced by leucine and phenylalanine and hence the Trp-Tyr hydrogen bonds that provide the framework for K § selectivity, are not applicable to Kir channels where a different sort of contact is required. Immediately below the selectivity filter, which comprises the narrowest part of the ionophore, the channel opens up into a cavity (10/~, in diameter) that holds a "pool" of water to stabilize and shield an ion from what would otherwise be the most hydrophobic part of the membrane. Additional stabilization is provided by the C-termini end of the four pore helices which have substantial negative electrostatic potential due to the helix dipole and point right at the pool. 9 The hydrophobic lining of the pool is also probably conducive to the high flux of ions by providing a lowresistance pathway.

8

GREGORY J. KIRKOVITS and C. DENNIS HALL

The selectivity filter contains more than one ion (Figure 4) and this appears to be fundamental to the high flux rate. The proposed model is that the binding of a second ion at an adjacent site follows the binding of the first ion. The binding of the second ion destabilizes the first electrostatically and consequently pushes the first out of the filter into the cavity. The second ion takes the place of the first ion and a new ion moves in. The process is analogous to a "billiard-ball" effect, driven by the high concentrations of K § ions in physiological solutions (vide infra for an analogous example in an artificial system).

2.2. Gramicidin" A Model for Ion Channel Conduction To date, the most extensively studied natural ionophore is gramicidin, 1~ a polypeptide antibiotic isolated from the bacterium Bacillus brevis. Indeed, the idea of a channel-like structure for ion transport was inferred from its study. However, unlike channel proteins, which are exclusively composed of L-amino acids, each alternate amino acid in gramicidin has D-stereochemistry. It is composed of 16 residues, 15 of which are amino acids. The structure is summarized below, in which Xxx has the following identities" gramicidin A (gA), Trp; gB, Phe; gC, Tyr; gD, a mixture of gA, gB, g C , - 8 0 " 5 915. HCONH-L-VaI-D-Gly-L-Ala-D-Leu-L-AIa-D-Val-L-Val-D-ValL-Trp-D-Leu-L-Xxx-D-Leu-L-Trp-D-Leu-L-Trp-CONHCH2CH2OH Cation transport is mediated by two peptide chains of gramicidin which transiently hydrogen-bond between their formyl end groups to form tail-to-tail dimers 11 about 26/~ in length that span the membrane. The alternating D- and L-configuration results in a [363-helix in which the polypeptide backbone forms the interior of a hydrophilic water-filled pore, and the hydrophobic side chains interact with the lipid bilayer. The resultant coil-like structure is stabilized by -NH..-O- hydrogen bonds that extend parallel to the pore axis from one turn of the helix to the next. The internal diameter of ca. 4 ]k permits the passive diffusion of monovalent cations aided by the carbonyl groups at rates in excess of 107 ions s-l. It should be noted, however, that this low molecular weight experimentally amenable polypeptide is not necessarily a model for natural tx-helical channel proteins because the latter form transmembrane pores by aggregation.

3. SINGLE CHANNEL MEASUREMENTS The development of biophysical techniques to make possible measurements of single ion channels in membrane bilayers has been fundamental to the advances in the understanding of how natural ionophores, particularly ion channel proteins, operate at the molecular level. The two principal techniques are the planar lipid

Natural and Artificial Ion Channels

9

bilayer method and patch-clamping and it is pertinent to provide a brief description of each in order to understand the subsequent sections.

3.1. The Planar Lipid Bilayer Method This technique essentially comprises two aqueous compartments containing electrolyte connected by a pinhole across which a bilayer of phospholipid molecules is spread. An electrical potential is applied across the bilayer and in the presence of an ionophore the resultant current flowing between the two chambers is measured as a function of time. 12 The method is often used to assess the activity of synthetic peptides, and has been used to measure the channel activity of several non-peptide channels. Of particular significance is its use in allowing the measurement of the single channel properties of isolated channel-forming proteins that have been reintroduced back into planar bilayer membranes. 13 The benefits of using the planar bilayer method are threefold: (i) the conducting ions and osmotic conditions of the bathing solutions can be varied and in so doing the ion selectivity sequences of the channel former can be determined; (ii) during the course of the experiment, additions, and perfusions can be made to the bathing solutions and hence the introduction of activating ligands can be investigated; and (iii) the lipid composition of the bilayer can be altered allowing the effects of membrane surface changes on membrane behavior to be tested.

Analysis of Single Channel Data As discussed above, channels fluctuate between opened and closed states. By monitoring the current flowing across the bilayer as a function of time it is thus possible to distinguish between these states. Transitions between the open and closed states of a channel molecule appear as a series of discrete and integral square-wave pulses of current since the actual current relaxation between 0 ions s -1 (channel closed) and >10 6 ions s-l (channel open) is too fast to measure. Ideally, the data should exhibit a series of rectangular pulses of constant amplitude and should be of random duration. Each integral transition is representative of a single channel opening. When channels are open, ions go through the membrane according to the difference in the voltage and the ion concentration on both sides of the membrane. The ions passing through the channel are observed as a channel current. It is therefore possible to calculate the ion flux (conductance, normally in picoSiemens) for the channel molecule by dividing the observed peak height (current) by the applied voltage, i.e. conductance = current/voltage. Conductance is dependent on many factors including the pore size and the pore length.

10

GREGORY J. KIRKOVITS and C. DENNIS HALL

Determining Ion Selectivity Many channels discriminate between ions and transfer biological signals across cell membranes by only allowing the passage of specific ions. This selectivity is dependent on various factors including pore size, electrostatic interaction between the ions and the pore-forming amino acids, and the dehydration energy of the ions involved. 14Such selectivity can vary from broad discrimination between anions and cations or more specifically cations of the same charge, e.g. Na + and K +. As noted above, one of the advantages of the planar bilayer technique is the ability to introduce ionic concentration gradients across the two chambers. Consequently, this permits the determination of channel permeability for one particular ion over another by measurement of the reversal potential (Vrev). This is simply the potential at which there is no net ionic current in the presence of the transmembrane concentration gradient. For example, if the only electrolyte present is KC1 at a concentration of Ccis on one side of the bilayer and of concentration Ct~ans on the other then the reversal potential, Vrev of a channel in a bilayer can be expressed by the Goldman-Hodgkin-Katz equation 1 as:

RT [PKCtrans+Pc]Ccisl

Vrev= 7

In PKCcis +

PcICuans

(1)

where PK and Pcl are the relative permeabilities of K + and C1- ions, respectively. Thus the permeability ratio for the two ions can be calculated by: PK _ I0~Ctrans - Ccisl PcI

Ctrans- Ccis

where c~ = exp(FVrev/RT ). Alternatively, Equation 1 can be used to determine the permeability ratio of ions of like charge when their concentrations are equal on either side of the bilayer. In this manner, ionic selectivity sequences can be established for a channel former.

3.2. Patch-Clamping The method of patch-clamping permits monitoring of the activity of single channels in cell membranes. 3 Experimentally, a glass micropipette (ca. 1 gtm in diameter) is applied to the surface of a cell, and by the use of slight suction, a piece of the cell membrane can be drawn into the pipette. Alternatively, this patch can be detached from the membrane. In so doing, the isolated piece of membrane may contain only one or at most a few channels. Consequently, the patch, seals off the end of the micropipette. The pipette can then be immersed in bathing solutions of different composition and different solutions may be introduced into the body of the pipette to permit control of the aqueous environments. The setup is completed by applying an electrical potential across the membrane patch. In order for the

Natural and Artificial Ion Channels

(i)

On-cell

11

(ii) Whole-cell

suck

~pull (iii) i n s i d e - o u ~

~ pull (iv)

Outside-ou~

Figure 5. The four-patch clamp recording methods. In each example, the cell is shown to the left, the micropipette to the right. For a discussion of the methods see text. Reproduced with permission from Hille, B., "Ionic channels of excitable membranes" 1st ed., Sinauer, Sunderland, MA (1984) p. 217.

method to be successful, the patch must form an extremely tight (high resistance) barrier across the aperture of the pipette. This is achieved by fire-polishing the pipette. Records of channel activity can then be monitored like the planar bilayer method, as current versus time traces. There are four protocols for performing the patch-clamp experiment (Figure 5). These are listed as follows: (i) On-cell in which the pipette is pressed against the membrane and left in place, thereby allowing the conductance of a tiny patch to be measured; (ii) Whole-cell where suction breaks the membrane thus enabling measurements of the entire cell; (iii) Inside-out and (iv) Outside-out: in both these cases, the pipette is pulled out of the membrane thus isolating that patch.

4. SYNTHETIC PEPTIDE ION CHANNEL MODELS As discussed above, the majority of known natural channel-forming structures involve large protein aggregates that contain a number of helical transmembrane segments acting in concert to control transmembrane ion and potential gradients. Naturally, the size and complexity of these molecules has meant that elucidation of the mechanisms governing their function has been difficult. One approach to this problem has been to study much smaller synthetic peptides, and by using those amino acids implicated in channel function, effectively "recreate" the channel pore structure with membrane spanning amphiphilic and/or hydrophobic o~-helical

12

GREGORY J. KIRKOVITS and C. DENNIS HALL

motifs. Of particular relevance has been the development of methods to control the assembly of peptides to form a more defined structure. Mutter and Montal synthesized template-assembled synthetic proteins (TASP) to control transmembrane assembly of helical peptides. 14 The TASP were designed to adopt a globular, four-helix bundle channel pore structure which produced conductance events characteristic of channel-forming peptides, while the shortest analogue elicited only erratic current fluctuations. The helical segments of natural proteins usually consist of multiple nonidentical helices, displaying sophisticated channel functions. To this end, Futaki has used the template approach to organize nonidentical helical peptide segments to form channel structures. 15 DeGrado et al. 16a'17prepared 21-residue amphiphilic peptides composed of only leucine (L) and serine (S) residues. Leucine was chosen to provide the apolar face of the helix because of its hydrophobicity and helix-forming propensity, while serine was chosen to provide the polar, yet uncharged face of the helix. The length of the helix corresponded to that of the hydrophobic portion of a membrane bilayer and the two sequences, (LSSLLSL)3 and (LSLLLSL) 3 gave different ion channel activity. The former peptide was found to produce cation selectivity similar to that of the acetylcholine receptor, while the latter was found to be proton-selective. Energy-minimization modeling suggested that this was due to the formation of parallel hexamer- and tetramer-helix bundles, respectively, in the membrane and attempts were made to stabilize the four-helix bundle structure by using a porphyrin template. 16b The subsequent results of planar bilayer studies showed that the template exerted a major influence on open lifetime and voltage dependence and that the channels were unimolecular. The transmembrane m-helices formed by Ac-(LSSLLSL) 3 (Ac = acetyl; Cterminus = CONH2), although formally neutral, have partial charges of +0.5 and -0.5 near their N- and C-termini, respectively, due to alignment of their amide groups. 18 This electrostatic asymmetry can account not only for the preferred transmembrane orientation of the peptide but also the nonlinear single-channel current-voltage curves (rectification) observed for Ac-(LSSLLSL) 3. DeGrado et al. extended this model by the addition of formal charged residues to just one end of the helix. In so doing, they were able to demonstrate how the electrostatic environment at the mouth of the channel can have a large effect on conductance and cation selectivity. By introducing a crown ether unit at the C-terminal region of hydrophobic helical peptides, Otoda et al. 19 were able to demonstrate increased stabilization of the peptide aggregate in the membrane by the formation of sandwich-type complexes with large cations. Ion channel activity was also increased due to the ability of the crown peptide to bind ions to the terminal portion of the hydrophobic helix bundle at the water-lipid interface. Ueda et al. 2~considered the problem of insolubility of hydrophobic peptides which restricts the distribution of peptides in water to a phospholipid bilayer membrane. In consequence they constructed a hydrophobic helix bundle shielded by hydrophilic peptides that acts rather like an umbrella.

Natural and Artificial Ion Channels

13

5. ENGINEERED NATURAL ION CHANNELS Another approach to the study of the mechanisms of transport has been to alter the performance of some of the more simplistic natural ion channels by structural changes. By adding a positively charged group to the C-terminal end of a gramicidin monomer near the mouth of the channel as exemplified by 1, Woolley et al. 2l reasoned that cation conductance could be modulated depending on the proximity of the charge to the channel entrance. They observed, not only single-channel currents of 1 that were of reduced amplitude relative to gramicidin, but more importantly, different current levels during the lifetime of the gramicidin-ethylenediamine dimer of 1. These different levels were due to the charged-NH~ group whose effect varied with cis-trans isomerization of the carbamate functionalities. In fact carbamate conformation was found to govern the proximity of the positive charge to the channel mouth, thereby inducing partial blocking of the entrance to the channel. By incorporating the photoisomerizable azobenzene unit in 2 they achieved a further degree of control over the current levels via photochemical trans cis isomerization of the azo linkage. 22 Schreiber et al. 23a connected two gramicidin units using tartaric acid derivatives (dioxolane) to form a covalently linked hybrid, thereby eliminating channel disruption as seen in the native channel. However, the ion channel conducting properties of these molecules were observed to be dependent on the stereochemistry of the tartaric acid linker. In particular, the (R,R)-dioxolane linker (a stereochemical mismatch with the helix geometry of gramicidin A) exhibited rapid interruptions in current (flickers). However, by introducing sterically demanding substituents onto the central linker, they were able to eliminate this behavior thus demonstrating that the observed gating phenomenon was due to flipping of the dioxolane ring into and out of the channel (Figure 6). 23b Sansom and co-workers 24 attached redox-active ferrocene units to the C-terminal end of alamethicin to explore the effect of oxidation on channel formation. Briefly, alamethicin is an antibiotic peptide, composed of 20 amino acids, which forms ion channels by self-association in lipid bilayers, in a voltage-dependent manner. Evidence supports the "barrel-stave" model of channel formation, in which czhelical monomers of the peptide associate to form an aggregate transmembrane pore. Different sized aggregates are observed to give multiple conductance states of alamethicin. The single-channel properties and voltage sensitivity of 3 and 4 were similar to alamethicin, although higher cis-positive potentials were required

o

o

9

1

NH~

2 (trans-meta-azobenzene)

14

GREGORY J. KIRKOVITS and C. DENNIS HALL

Figure 6. Schematic representation of the intramolecular linked gramicidin dimer showing flipping of the (R,R)-dioxolane ring into and out of the conduction pore. Reproduced with permission from Stankovic, C. J.; Heinemann, S. H.; Schreiber, S. L., J. Am. Chem. Soc. 1990, 112, 3703.

to elicit activity. This was thought to be due to steric and/or conformational demands imposed on the peptides by the bulky substituents. In addition, 4 exhibited a transition with unusually protracted open lifetimes. The effect of oxidation on channel formation was investigated by treatment with the oxidizing agent ceric ammonium nitrate (CAN). The in situ addition of CAN to 3 caused a time-dependent decrease in channel formation at constant potential, consistent with the presence of positively charged head groups decreasing the propensity of alamethicin peptides to form channels. The long lifetimes of 4 were selectivity eliminated, whereas alamethicin itself was unaffected by CAN. Premixing of the peptides with CAN produced similar results. Indeed, approximately twice the concentration of peptide was required to induce channel activity as compared to peptide in the absence of CAN. Conductance histograms showed that the oxidized peptides spent a greater percentage of their time in the closed state when compared to the reduced forms.

3

4

Natural and Artificial Ion Channels

15

6. CLASSIFICATION OF ARTIFICIAL (NON-PEPTIDE) ION CHANNEL MODELS Within the last 20 years the curiosity of the organic chemist has spawned an interest in the study of non-peptide channel models. These have been primarily designed to incorporate the structural elements of natural ion channels required for channel activity. The aim has been to contribute to the mechanistic understanding of the significantly more complex natural channel proteins. Perhaps a more realistic target however, is their application to the development of new pharmaceuticals and drug-delivery systems. 25Synthetic peptide models of ion channels function mainly as well-defined aggregates which create an assembled pore for ion transport. 17 Non-peptide channel models, however, can be loosely categorized into two functional modes 26'27 as follows: (i) unimolecular membrane-spanning tunnels, which provide "relay" elements for cation transport (commonly macrocyclic crown ethers), and (ii) transport through self-assembled aggregates. Usually amphiphilic in nature, the latter form the transmembrane tunnel by recognition and organization within the membrane bilayer. Both systems require the application of principles evolved in studies of supramolecular chemistry.

6.1. Early Examples The earliest example of a non-peptidic channel model was prepared by Tabushi et al. 28 It consisted of a ~-cyclodextrin, attached to four hydrophobic tails designed to afford a "half-channel?' The transport of copper and cobalt was assessed in artificial liposomes (kco(u) = 4.5 x 10-4s-Z), the rate being much faster than in the absence of the half-channel. In the same year, Lehn 29 reported a solid-state model of a molecular tunnel consisting of stacked macrocyclic polyethers, with K § ions located alternatively inside and on top of successive macrocycles. Following the theme of the macrocyclic tunnel model, Nolte 3~ prepared an oligomer, 5, of isocyanide with benzo-18-crown-6 side chains which possessed a rigid helical backbone with four repeating units per helical turn, thereby generating four tunnel-like tubes composed of the macrocyclic tings aligned on top of one another and spaced ca. 4 ,/k apart. The system displayed features akin to natural

N=C

'~

-

""c~f'c- 'R

.N II

R'

v

-0

L

0

o.J

16

GREGORY J. KIRKOVITS and C. DENNIS HALL

ionophores comprising (i) pores with a polar interior and apolar exterior; (ii) a hydrophilic top and bottom to face the aqueous medium inside and outside the membrane, and (iii) a chain length of ca. 40/~ to bridge a typical membrane bilayer. Vesicles of dihexadecyl phosphate were prepared with a UV-active dye trapped in the inner aqueous volume. The rate of Co 2§ transport into the vesicles was determined by monitoring the increase in absorbance of the C02§ complex using UV-vis spectrophotometry (kco(u)= 10-4s-I). No Co 2§ transport was detected in the absence of 5. 6.2.

Proton Channels

In an attempt to prepare flux-promoting compounds Menger 31 synthesized phosphatidylcholine derivatives incorporating a polyether side chain. These proved incapable of promoting ion movement across bilayers, but intermediates in their synthesis bearing the general structure RO(CH2CH20),,R' proved to be active. Proton transport across phospholipid bilayers was assessed using a pH-fluorescence spectroscopy technique by incorporating acid-responsive pyranine dye trapped within the vesicles. By varying the three sections of the general structure, Menger was able to maximize proton transport, which was achieved with CHa(CH2)10 COO(CH2CH20)sCH2Ph, 6. The three sections comprised a hydrophobic polymethylene chain, a relatively hydrophilic ester plus ethyleneoxy groups and a benzyl head group, presumably to serve as a membrane anchor. It was recognized that a single molecule of 6 was of insufficient length to span the membrane and that a minimum of two molecules would need to align in order to traverse the entire lipid bilayer and promote proton passage. The transport ability of 6 exceeded that of gramicidin and the simple cartier molecule 18-crown-6 proved inactive under identical conditions. Dubowchik 32 synthesized a linear bola-amphiphilic octamine 7, in which each nitrogen was attached to an adamantylmethyl group terminated at either end by a propylsulfonate headgroup. Fully extended, 7 was designed to be 48 A in length, thus capable of spanning a typical membrane bilayer. Nonaggregated and essentially non-protonated 7 was intended to dissipate a proton gradient with proton transfer occurring along the length of the chain and with the backbone effectively acting as a proton wire. Phosphatidylcholine vesicles were prepared, the pH of the

7

Natural and Artificial Ion Channels

17

external aqueous solution was raised to 7.5-8.0 and the change in pH over a period of time was monitored inthe presence of 7. The pH dropped steadily over two hours almost immediately upon addition of 7. Control experiments showed there was only a very slow leakage of protons most probably due to lysis of a small fraction of defective liposomes. The results were consistent with the transporter acting as designed, but membrane defects caused by other kinds of interaction with 7 were also considered possible. Using the principle of ion pair formation between ammonium cations and the phosphate anions of lipids, Matile et al. 33 prepared 8, an amphiphilic polyamine dendrimer. Rather than acting as a membrane channel, 8 was expected to form reversible membrane defects in the lipid bilayer. The steroid moiety was expected to act as the hydrophobic anchor for bilayer orientation and steric bulk was expected to prevent the polyamine penetrating the bilayer. Proton transport was assessed in unilamellar vesicles using the pH-fluorescence technique in which the external pH was increased to 7.8 relative to the internal pH at 7.4. The results demonstrated that 8 was almost as active as gramicidin, and maximal flux was achieved in ca. 20 s.

HaN~NH.

fi

/

§

J 9 "J H3N

"'q 8"WA"

NH

8

Menger had previously shown that the ion flux across synthetic phospholipids could be promoted by attaching alkyl groups of various sizes to the hydrophobic portion of the lipid, presumably by creating reversible defects in the bilayer. 34 The rigid molecule, 9 was designed to resemble the polyene and polyol segments of amphotericin B (AmB) (Figure 7), but with the charged terminal anchoring group absent. 35 Briefly, AmB is a low molecular weight polyene antibiotic that forms channels by the self-assembly of 8-10 monomers. The model is that of the "barrel-stave", in which the hydroxylated portion of each molecule faces into the channel pore and extends across the bilayer. Oligo(p-phenylene) units were chosen for the backbone because of their hydrophobicity, conductivity, conformational flexibility, and luminescence properties. Small unilamellar vesicles containing. entrapped HPTS (8-hydroxypyrene-l,3,6-trisulfonic acid) were used to monitor proton transport again using pH-fluorescence spectroscopy. In the presence of a pH

18

GREGORY J. KIRKOVITS and C. DENNIS HALL

H r, 3"

~

~

~

~

~

~

~

.

H3C

H.J

/.. OH

OH

-

OH

NH2

OH

amphoteriein O

0

OH HO

OHHO

O0

OH HO

OHHO

O0

OH HO

OHHO

O0

OH HO

OH

0

Figure 7. Structure of amphotericin B (top) and oligo(phenylene) 9 (bottom). gradient, changes in internal pH facilitated by 9 showed that the length of the rigid rod must match that of the bilayer for optimal activity. 35a Compound 9 also demonstrated significant H § > K § selectivity. Without 9 and in the presence of the selective potassium cartier, valinomycin ([K§ = [K§ only minor increases in HPTS emission were seen. However, in the presence of 9 the flux rate was 16-fold higher implying that under these conditions 9 mediates H§ § exchange with H § > K § selectivity. Transport was explained by the hydrogen-bonded chain (HBC) mechanism. This is a two-step "hop-and-turn" process in which proton passage occurs via successive hypothetical "hops" followed by sequential "turns" in order for the new O - H bond to return to the initial configuration and acquire another proton. Fluorescence quenching experiments using spin-labeled lipids confirmed the transmembrane orientation of octamer 9. 35b Overall, this presents a picture of a proton-selective membrane-spanning and presumably unimolecular channel model. Matile reported poor bilayer solubility of rigid-rod 9 so tackled this problem by incorporating lateral propylene side chains. 35r This was expected to increase lipophilicity and also to improve the flexibility of the hydroxyl groups thus facilitating proton transport by the HBC mechanism. The structural modifications did not appear to have a detrimental effect on the active structure and increased ion flux rates were seen for the propylene-substituted oligo(phenylenes).

6.3. Macrocycle Based Mimics The key to Fyles original design 36 for an artificial ion channel (Figure 8) was based on a central rigidifying macro-ring using polycarboxylate 18-crown-6 de-

Natural and Artificial Ion Channels

19

rived from tartaric acid [COOH-(CHOH)2-COOH ]. Fyles exploited the availability of tartaric acid in its optically active and meso forms to direct "wall" units from the core. These units incorporated either or both polar and nonpolar functionalities to provide structural control (i.e. alignment with the bilayer phospholipids) and a tunnel-like path for transport within the bilayer. The crown ether was composed of one, two, or three tartaric acid units such that the structure consisted of either two, four, or six wall units. To each wall unit was attached a hydrophilic head (H) group (glucose, 2-mercaptoacetic acid or 3-mercaptopropanol) to serve as an anchor and to assist in establishing the transmembrane configuration. Cation flux was studied using the pH-stat method for a variety of compounds with systematic structural variations. Subunits that were well represented at the lower end of activity included mercaptopropanol (head group), the long wall unit (dodecanediol chains) and the lipophilic wall group {-(OCHECH2)30-chains }. In a group of otherwise identical structures, changes in the head group from glucose through mercaptopropanol to ethanol resulted in an increase in proton flux. Four wall units appeared to be most conducive to activity, but somewhat surprisingly the sixth wall unit was found predominantly at the lower end of the scale. The best wall unit was a hydrophilic and lipophilic balance, composed of both octyl and -(OCHECH2)30- chains (Figure 8). This provided the extended molecule with a length appropriate to span the bilayer. Interestingly, a combination of kinetic and inhibition studies showed that several of the molecules studied actually functioned as carders. Based on similar ideas, Lehn et al. 37 constructed an artificial ion channel utilizing a central tetracarboxy-18-crown-6 unit for labile alkali-metal complexation. Substitution of two pairs of axially oriented lipophilic dendrimer-like arrays with terminal carboxyl groups for anchoring at water-membrane interfaces, was expected to provide a so-called "bouquet" molecule with an extended transmembrane orientation. The central macrocyclic nucleus was also replaced with ~cyclodextrin to provide a larger rigid hole of internal diameter, ca. 6/~ (Figure 9). Alkali metal ion NMR spectroscopy was employed to assess cation transport across the phospholipid bilayer of liposomes. Briefly, the presence of alkali metal ions in the outer aqueous solution can be distinguished from internal ions by the presence of a shift reagent. Two resonance signals were therefore seen which corresponded to internal (unshifted signal) and external (shifted signal) metal ions. Vesicles with external shift reagent, were created with opposing gradients in Na § (outside) and Li § (inside) concentration. Transport of the cations down their concentration gradients due to the presence of the bouquet in the membrane was therefore followed directly using 23Na and 7Li NMR spectroscopy as a function of time. Ion transport activities were seen to be approximately the same for the two systems, and it was concluded that ion transfer proceeded by a one-for-one Na § ion for Li § ion exchange (cation-cation antiport). Rates of transport for the ~cyclodextrin derivative through gel-state membranes were similar to those observed in thefluid state, a result consistent with channel function.

20

GREGORY J. KIRKOVITS and C. DENNIS HALL 140

head

H

OI4

"-

+

=

o4-%o

H

oo

~

~o

c-o

-'-~o.]=~ core ~~176

o~~--o, o - ~ -,-

_~o

;o,c~o

%o.J

o

~

o~~

o-~o~,

-4H

H

Figure & Generalized structure and example of the channel mimics prepared by Fyles and co-workers. In 1990 Gokel and co-workers published evidence of a membrane-insertable, sodium cation-conducting channel active in large unilamellar vesicles (LUVs) using a synthetic compound based upon diaza-18-crown-6. 38a Several design principles were applied to the preparation of tris(macrocycle) 10, which contained a central macrocycle similar to the model systems developed by Lehn and Fyles but designed to be flexible (Figure 10). First, the channel former had to correspond to the length of a typical phospholipid bilayer (ca. 30/~). Dodecyl alkyl chains were chosen for this purpose and also to serve as connective units. Second, head groups were required to provide a point of cation entry and exit and to anchor the structure in an extended conformation in the bilayer. Work undertaken in the same laboratory demonstrated that amphiphilic crown ethers could aggregate into stable vesicles 39 and that bola-amphiphilic crowns (crown-hydrocarbon chain crown) could form monolayer lipid membranes. 4~ Hence crown ethers were selected to employ an elementary principle of cation selectivity; in addition, the choice of diaza-18crown-6 permitted derivatization at nitrogen while retaining conformational flexi-

Natural and Artificial Ion Channels

21

RO~O '~OR

OR

O

o~

o

~

OR

o

~

o

o__,y-o---~oLT -~o o~ o "~,,"-" o

F o ~F ~

J

0"

f

oF/

~

~--~,,~ "'~ ~

beta'cycl~

. j o--.-J/)"- ,H

o)

o/

O

?

OR

o

L.. OCH ~" - - 7

,._.,

O~o RO

o

5

o

o~

OR

o

o

Figure 9. Structures of the bouquet molecules incorporating an 18-crown-6 center (left) and a 13-cyclodextrin center (right).

bility. Third, a central unit to capture and relay the cation through the bilayer was required. For synthetic simplicity a diazacrown was also chosen such that the overall structure would afford a hypothetical "tunnel." By this approach the ether oxygen atoms of the three macro-rings would play a central role as relay elements. Cation flux rates were assessed for 10 directly in phospholipid vesicles using a dynamic 23Na NMR method. In this procedure, addition of a Dy 3+ shift reagent to an aqueous vesicle solution prepared in the presence of Na § results in an external Na § signal shifted with respect to internal Na § With the incorporation of a channel former, the line width broadens, reflecting dynamic exchange between internal and external Na + through the membrane channel. Consequently, the linewidth varies directly with cation flux and a plot of [channel] versus linewidth permits determi-

22

GREGORY J. KIRKOVITS and C. DENNIS HALL

/"'X

C~ Y

/--"x

0- 3

C~

N-(CH2)~3-N

0- 3

C~

N--(CHz)~N

CoM.../0-2 Cok...../09 10, Y=N(CH2)llCH 3 14, Y - N" y / ~ A

A

11, Y=NH

NO 2

Y

CoX_.._/0-2

12, Y=O

15, Y - N" y /

0- 3

~

13, Y=NCH2C6H 5

OCHaI6'

Y--N

H

Figure 10. Tris(macrocycles) prepared by Gokel and co-workers.

nation of the rate constant from the slope of the graph. The rate for 10 was found to be 13.5 s-l, ca. 40-fold greater than the simple carrier molecule, N,N'-di-dodecyl4,13-diaza-18-crown-6, but ca. 100-fold poorer than gramicidin at the same concentration. The initial findings were followed by a comprehensive study in which a series of bis- and tris-(macrocyclic) compounds with structural variations were prepared. 38b Initially, proton transport was monitored across phospholipid bilayers by using a fluorescence technique. In this method a pH-sensitive pyranine dye is trapped within vesicles. If the external pH is rapidly lowered, in the presence of a proton conductor, protons will flow into the vesicles. The pH of the internal solution is thus lowered and this is reflected by a decay in the corresponding fluorescence signal of the dye. When the covalent spacer units were simply changed from alkyl -(CH2)12- in 10, to oxyethylene -(CH2CH20)3CH2CH 2- proton flux showed a marked fall. This result was most unexpected as it had been anticipated that a larger number of donor atoms would increase the cation flux by lowering the transport energy needed in passage. It was deduced that more donor atoms lead to stronger binding whereas the dynamics of channel activity require weak complexation. Cation flux rates of the tris(macrocycles) were determined by 23Na NMR spectroscopy as described above and were standardized relative to gramicidin which was given an arbitrary value of 100. The original design hypothesis envisaged the three macro-rings lying parallel to one another, thus providing a tunnel-like path. However, when the central diaza-18-crown-6 ring was replaced by diaza-15crown-5 and subsequently by open-chained O(CH2CH20)3 , transport rates were respectively 28, 25, and 14. The results suggested that although the central ring is important, the cation does not necessarily pass through it. It was thus postulated that the central ring of 10 lies parallel to the phospholipids thereby lowering the polarity in the bilayer midplane and reducing the distance a cation is required to travel unaided by donor atoms. Removing the dodecyl side chains (as illustrated by

Natural and Artificial Ion Channels

23

11) did not affect the rate but when the head group diazacrowns were replaced with aza-18-crown-6 (e.g. 12), activity was lost (<2). Altering the side chains from dodecyl to benzyl (as in 13) increased the respective transport rates from 28 to 39. The explanations for these observations were based on membrane stabilization. It was expected that the tris-macrocycles would be protonated at their terminal rings. Therefore, the lack of activity for the aza-18-crown-6 derivative suggested that protonation of the only available nitrogen atom disrupted the transmembrane configuration, whereas protonation of the diazacrown derivative(s) occurred at a position which favored channel function. The increased activity on incorporation of the benzyl side chain was ascribed to the arene residues acting as anchors. Interaction with the phospholipid polar head groups would enhance transmembrane stabilization and thus increase the flux rate. It is interesting to note that channel fragments of 10 were unable to conduct cations, presumably because they were not of sufficient length to span the membrane. Subsequent studies 38c'e confirmed the ability of the compounds to partition completely from water into a hydrophobic solvent, thereby excluding the possibility that differences in cation flux reflected different solubilities in the bilayer vesicles. It was anticipated that if the terminal rings served as openings for the passage of Na § ions, altering the electron density on the distal benzyl groups would affect the rate of cation transport. Indeed, substitution of the benzyl group in the 4-position of 13 with a nitro (14) or a methoxy group (15) was observed to cause the relative rates of 23Na cation transport to be slower (30) and faster (43), respectively. An indole-terminated tris-macrocycle, 16 was prepared 38't which was expected to either stabilize the tris-macrocycle with respect to the bilayer or hydrogen bond internally from the indole N - H to a proximate oxygen or nitrogen atom, thus effectively blocking the channel opening. Indeed, no detectable cation flux was detected in 23Na NMR studies. The result was substantiated by calculations and models that showed hydrogen bond formation did indeed hold the indole ring over the opening, thereby 'closing' the channel. The presence of an intramolecular hydrogen bond in the IR spectra of the tris(macrocycle) at varying concentrations was consistent with this finding. The planar bilayer method was also used to assess the ability of channels 13-15 to conduct Na § cations in a phospholipid bilayer. 38e Typical traces were observed consisting of rectangular pulses of constant amplitude between open and closed states, strongly implying a channel mechanism for the tris-macrocycles. The cation fluxes (conductance) were expected to correlate with the substituent-dependent fluxes observed in the 23Na NMR studies, but in practice very little difference was seen. For an applied potential range of +60 to +140 mV, compounds 13-15 obeyed linear current-voltage relationships, and the calculated conductance was ca. 13 pS. This is presumably because the channel formers create similarly sized, water-filled pores in the bilayer. However, preliminary experiments did suggest that the open times of the channels increased in the order p-NO 2 (14) < p-H (13) < p-OCH 3 (15). Indeed, in a separate communication, TM the methoxybenzyl channel 15 was re-

24

GREGORY J. KIRKOVITS and C. DENNIS HALL

ported to exhibit single-channels events open for periods of more than 30 s. A series of fluorescence experiments with dansyl- and N-methylindolyl-terminated tris(macrocycles) indicated that the terminal rings were located in the headgroup region of the phospholipid bilayer and that the compounds functioned as monomers. 38g Based on the accumulated experimental evidence, a postulate for the channel conformation of the tris(macrocycle) was proposed in which the terminal rings serve as head groups and anchor the transmembrane structure thus stabilizing a chain of water molecules and cations in transit by using dipolar and hydrogen bonding interactions (Figure 11). The role of the terminal rings in Na + transport was also investigated by preparing two macrocycles (17 and 18) which differed in their terminal ring size. 3sh The rates determined in vesicles for Na+-transport using 23Na NMR methods were 16 and 26 for 17 and 18, respectively, relative to gramicidin (=100) suggesting that the size difference in the terminal tings may restrict the cation release rate. In contrast, the calculated conductance of both channels was essentially identical at ca. 13 pS. The activity of each channel, however, was markedly different. Typically, traces for 17 were of no more than 5 s in duration with single-channel transitions dominant, whereas for 18 numerous multiple-channel openings were observed, frequently of

/--o o-~ F/

,T

-N

/--o o-~

N-(CH2)EN

[-o

N--(CI'L~)EN

o-~ N

~f ' /

."1

17, n= O; 18,n = I

,

{

0

o

0 o

}

L . ~ o ~ o

o

<

o---,,~

:J 0

~

[

0

.J

Figure 11. Postulated channel conformation of 13 in a phospholipid bilayer.

0

}

Natural and Artificial Ion Channels

25

>5 s duration, implying that the increased ring size of 18 had an observable effect on cation flux. Similar observations were made when a naphthylmethyl-substituted tris-macrocycle was compared with benzyl-substituted 13. 38i Assessment by planar bilayer methods revealed discrete, integral opening and closing behavior, indicative of channel formation, and near identical conductances, ca. 13 pS. However, the naphthyl-channel gave rise to more open events with a greater number of multiple transitions at lower concentration. The aim was to evaluate the effect of an increase in aromatic surface area on Na § cation flux, but although an effect was observable it was not quantifiable. The length of the hydrocarbon chain linking each of the macro-rings was altered to assess the extended conformation within a bilayer environment. 38j In this exercise the chain length of 13 was varied from 8 carbons through to 16 carbons. Assessment of Na § flux using the 23Na NMR method showed, most interestingly, a complete loss of activity with the 8-carbon spacer chain. This implied that the shorter chain tris-macrocycle was simply not long enough to span the hydrocarbon portion of the phospholipid bilayer which, at the same time, ruled out the possibility of a carrier mechanism for cation transport by these compounds. Compounds 19 and 20 with a calix[4]arene as the central unit in a 1,3-alternate conformation were examined for their ability to conduct Na § ions using the planar bilayer method. 41 Unlike the tris-macrocycles which exhibited discrete conductance steps, the calixarenes displayed erratic "bursts" of up to 300 pS. This was thought to be due to the terminal tings on either side of the calix rapidly exchanging positions, i.e. sweeping outwards, thereby making the opening in the membrane much larger and thus increasing conductance. Using the principles outlined in Gokel's pioneering work, a serious attempt has now been made to introduce redox-active centers in the form of ferrocene or cobaltocenium units into potential artificial ion channels capable of spanning the lipid bilayer. Thus, a series of compounds (21-32) has been synthesized, fully characterized and in some cases tested for cation transport by both patch clamp and planar bilayer techniques. 42 Table 1 shows the first nine of these compounds containing four 18- or 21-dia~ crown tings each with an overall length designed to span a lipid bilayer of approximately 30/~. The first six compounds contain either ferrocene (21-23) or cobaltocenium (24-26) as the central linking units whereas the last three (27-29) contain glutaryl units designed to mimic the metallocenes for size but lacking, of course, the central redox-active unit. The last three compounds (30-32) all comprise three macrocyclic tings with cobaltocenium units attached to a terminal ring and in one case (32) with ferrocene at the other end. Apart from the redox-active units, the design strategy was devised to test the efficacy of 18 versus 21 crowns for ion transport and the value of decyl chains or anthrylmethyl units as "anchors" for the prospective channels. The anthryl units also offered an opportunity to detect cation complexation and establish host-guost stoichiometry by the extremely sensitive technique of fluorescence spectroscopy.

26

GREGORY J. KIRKOVITS and C. DENNIS HALL

0

0

t.o

("

0

<- o.9

/ 19, R=H 20, R =tBu

In the event all the channels containing metallocene units showed electrochemical activity but (21-23) showed an irreversible wave at ca. 720 mV (vs. Ag+/Ag). Addition of increasing amounts of cations (Na § K § Mg 2§ and Ca 2§ showed the expected anodic shifts 43 and at excess concentrations of cations a return to "pseudoreversibility" was observed in some cases. The host-guest stoichiometry appeared to be ca. 1:3 (or 4) consistent with three or four rings accommodating the cations and this was confirmed for 22 by 13C NMR and by fluorescence spectroscopy. A speculative explanation of the irreversibility of the CV wave is that the ferricenium form of the uncomplexed host is stabilized by interaction with the tertiary amino functions and that reversibility is reestablished on complexation of the tertiary amine nitrogens with the cations. Similar observations have been made before 44-47 and the explanation receives additional support from the observation that (24-26) and (30-32) all show reversible CV behavior. Clearly the oxidation potential of the cobaltocene system is not high enough to induce interaction with the tertiary amino functions. With (24-26) addition of alkali and alkaline-earth cations again generally induced modest anodic shifts with the exception of 26 and Mg 2§ which gave

27

Natural and Artificial Ion Channels Table 1.

Synthetic Tetrakis(Macrocyclic) Compounds Prepared for Study

Central Unit (L)

Ring Size (n)

Sidearm (R)

No.

Notation

ferrocene ferrocene ferrocene cobaltocenium cobaltocenium cobaltocenium glutaryl glutaryl glutaryl

1 1 2 1 1 2 1 1 2

decyl 9-anthrylmethyl 9-anthrylmethyl decyl 9-anthrylmethyl 9-anthrylmethyl decyl 9-anthrylmethyl 9-anthrylmethyl

21 22 23 24 25 26 27 28 29

Fc-N 18N-Cl0 Fc-N 18N-Ant Fc-N21 N-Ant Co-N 18N-C10 Co-N 18N-Ant Co-N21N-Ant G-N 18N-C10 G-N 18N-Ant G-N21 N-Ant

"

L__

~"

._...]

oG

a

,(C H2),o

"3 ,o ~,,~--Vo- -o-~ n

o ~-jo- -o--~ L_. --J n ~_~,C-o __o - ~

Note:

,(ell2),0

*Generalizedstructureshowingthe variableelementsof the tetrakis(macrocyclic)compoundsshownin Table 1.

small, but significant, cathodic shifts on addition of excess cation. In the latter case, the electron density on the metallocene unit must increase on complexation and this was ascribed to conformational changes associated with simultaneous binding between the small Mg 2+ cation and both the amide carbonyl oxygen and the tertiary nitrogen atoms. The conformational change may have caused sufficient twisting of the carbonyl groups out of the plane of the cp rings to inhibit electron-withdrawing /---k

/--"x

(-o

0- 3

(~--o

o

Co o

30, R '= C(O)Co, R 2= Boc; 31, R ~= C(O)Co, R 2= H; 32, R'= C(O)Co, R l= C(O)Fc.

/---k

0-3 o

Co o

0- 3 o

28

GREGORY J. KIRKOVITS and C. DENNIS HALL

resonance interaction between the metallocene and the complexed carbonyl functions. Unfortunately, no X-ray crystallographic data is available to confirm this hypothesis. With 32, the CV waves showed classical reversibility throughout and the cathodic shifts, although modest (maxima -45 to -75 mV dependent on the nature and the amount of the cation) were evident for all the cations studied (Na § Ca 2+, and Mg 2+) at both the redox-active centers. This was again ascribed to inhibition of carbonyl resonance on complexation and the reversibility of the ferrocene system indicates no interaction with the tertiary nitrogen atoms in the molecule is possible. Artificial channel activity was tested by the patch clamp technique with 21 and by the planar lipid bilayer method with 21 and several other channels in the series. An "inside-out" patch of cellular membrane was exposed to a solution of 21 and then tested for K + transport across the membrane in symmetrical K + conditions, at holding potentials ranging from +80 to -80 mV. The results are shown in Figure 12. In control recordings patches held at +60 mV for example, showed no activity (Figure 12i). These were then treated with solutions of 21 (10 IxM in DMSO) when channel events (ca. 2.5 pA) were immediately observed (Figure 12ii) often with multiple openings (n = 4 patches). The activity declined after persistent washout

Figure 12. Ion channel activity in neuronal membrane at (i) -60 mV, control; (ii) -60 mV after exposure to 21 (10 I~M); (iii)0 mV; (iv) +60 mV in symmetrical K§ conditions. The short horizontal bars at the left of the traces indicate the current level of the closed state.

Natural and Artificial Ion Channels

29

for 10 min but was observed at both negative and positive potentials (Figure 12iv) with qualitatively different opening and closing kinetics and eliminated at 0 mV (Figure 12iii) as expected in the symmetrical conditions used. The data shows clear evidence of ion transport promoted by 21 at both negative and positive potentials. In addition, changing the potential across the patch from negative to positive resulted in the reduction of the transmembrane current. For example, at -80 mV the average single channel current amplitude was -4.0 pA (equivalent to a conductance of 50 pS) whereas at +80 mV the same average current was +2.4 pA (30 pS) implying a modest degree of rectification. The results suggested the incorporation of 21 in the membrane bilayer but were necessarily viewed with some caution since it was conceivable that 21 was in some way activating endogenous channels within the biological membrane. The planar bilayer method was therefore adopted to check this possibility. The ability of compounds 21, 22, 23, 25, and 28-32 to conduct cations was examined in planar bilayers composed of phosphatidylethanolamine (PE) painted across a 200 ~tm diameter aperture in a septum between two aqueous compartments filled with 600 mM KC1 (10 mM HEPES; pH 7.2) (Figure 13). Bilayer quiescence was confirmed at +148 mV and bilayer stability was observed to be unaffected by the addition of up to 5 ~tL of DMSO. Stock solutions of the channel compounds (10 I.tM) were prepared immediately prior to use, and 0.1-5 laL (i.e. 1-50 pmol of substance) in DMSO was added to the stirred solution in the cis chamber. Typically, channel insertion into the bilayer occurred between 30 s and 30 min. Single-channel recordings, for all the tetrakis(macrocycles) studied, were dominated by bursts of irregular openings. This was characterized by very brief flickers (of millisecond duration), and an apparent continuum of conductances (from a few pS up to hundreds of pS in magnitude). The activity typically began with occasional single events, which increased in frequency and magnitude. Often activity dimin-

Figure 13. Planar bilayer traces for (A) 1 pmol 31 at +50 mV and (B) at-50 mV using symmetrical KCI conditions (600 mM, 10 mM HEPES, pH 7.2). The short horizontal bars at the left of the traces indicate the current level of the closed state.

30

GREGORY J. KIRKOVITS and C. DENNIS HALL

ished temporarily, but was also liable to increase to such an extent that the bilayer sometimes became leaky and eventually ruptured. The erratic behavior was seen at a range of potentials (+ 10 to + 148 mV) but occurred with a higher frequency and magnitude at larger potentials. The irregular behavior was recorded for all the tetrakis(macrocycles) and similar behavior has been observed for a number of synthetic peptides, 15b'16a'2~and synthetic channel formers. 41'48,50d'57b'62 Channels are normally observed as discrete integral square-wave pulses of constant amplitude and the disparity between this type of observation and the erratic behavior observed suggests fundamental differences in the mechanism of activity. It is possible that the fluctuations in current are due to transient disruptions to the lipid bilayer resulting from poor packing between the phospholipids and the macrocyclic ligands rather than the formation of a tunnel-like path. This was certainly considered a possibility in the oligo-THF peptide model system (vide infra) 48 and the fact that membrane rupture often occurred is compelling evidence in support of the postulate. Perhaps the macrocycles can form aggregate structures within the bilayer and in so doing they disturb the lipid structure resulting in "leak" currents at the interface between the lipids and macrocyclic aggregates. These are seen as rapid changes in flux (high conductance spikes) as molecules move into and out of the bilayer. The aggregates may also be stable for a short time giving rise to a continuum of conductances, after which they either break up (the corresponding activity diminishes) or become sufficiently large to induce membrane rupture. There was, however, some evidence for channel formation in which the crown tings aligned in a transmembrane configuration to provide a tunnel-like path for conduction. For example, although 21 displayed erratic behavior it also gave step-like fluctuations in current with a single open level of ca. 8 pS and relatively protracted lifetimes of up to 10 s. Compound 22 with the anthryl "anchor" exhibited ca. 25 min of sustained single open-level transitions varying in lifetime from 50 ms to nearly 20 s. This was recorded after a period of quiescence, which was preceded by a period of irregular high activity. Beyond the range o f - 4 0 to +80 mV, erratic spiking began to occur, which eventually overwhelmed the discrete open levels. The current-voltage relationship for 22 (Figure 14) shows a mean conductance of ca. 61 pS. The second compound to exhibit single-channel activity was G-N21N-Ant (29) which also exhibited a period of regular openings after a period of no activity. Open lifetimes of up to ca. 10 s were observed and the single-channel activity again came to a conclusion with the appearance of the burst-like activity. The current-voltage relationship for 29 (Figure 14) is Ohmic and shows a mean conductance of ca. 78 pS. Compound 31 with a terminal cobaltocenium unit was studied only once and after a period of irregular brief flickers, exhibited integral single-level openings of constant amplitude over a potential range of +70 mV with lifetimes up to ca. 10 s. Examples of the step-like transitions observed are shown in Figure 13 and the current-voltage relationship (Figure 14) shows a mean conductance of ca. 47 pS.

Natural and Artificial Ion Channels

31 5

5

(i)

(ii)

4

3 2

~

r. . . . . . . . , -70 -JO

2

n I0

9

30

lO

70

-3

4

gs =61pS

-5

-6

g,--78ps

hoklngmmial (mY)

(iii)

5

4 3

""

n ~

~

9

-3

4

~ I0

30

50

g~= 47pS

-5 hadins mea~-t (mY)

Figure 14. Current-voltage relationships for (i) 10 pM 22, (ii) 10 pM 29, and (iii) 1 pM 31 using symmetrical KCI conditions (600 raM, 10 mM HEPES, pH 7.2).

The average conductances obtained for 22, 29, and 31 were, respectively, 61, 78, and 47 pS, significantly higher than observed for 21 but in good agreement with the patch-clamp results. The reasons for this are unclear and may suggest a third, as yet unknown conformation for cation passage. Certainly, the higher conductances appear to conflict with the argument for macrocycle-controlled transport when compared to the flux rates of the Gokel channels. On the other hand, a higher conductance was observed for the 21-crown system than the 18-crown systems, implying that the larger ring-size mediates less-hindered passage of K § ions. It should be emphasized that any inference for active transmembrane conformations of the compounds is somewhat speculative in view of the paucity of data. The ultimate aim of the project was to see if rectification (asymmetric currentvoltage responses) or gating could be achieved by use of the redox-active metallocene moieties. Not only did channel-like function prove difficult to establish but the potentials required to oxidize the Fe(II) center for example (ca. 720 mV), were beyond the limits of the experimental setup. Indeed, membrane stability at very high voltages (>200 mV) is questionable. 41 With the small amounts of data obtained (Figure 14) it is difficult to substantiate whether any rectification is realized.

32

GREGORY J. KIRKOVITS and C. DENNIS HALL

However, two observations should be noted. First, it is clear from Figure 14(ii) that as expected, the current/voltage response for 29 is linear, i.e. Ohmic. In the case of 22 however [Figure 14(i)], there is an indication of rectification since the slope on the anodic side is less than that on the cathodic side. These data are qualitatively similar to those observed for 21 using the patch-clamp approach. This conclusion however must be treated with a degree of caution in view of the very limited amount of data available. Second, a marked voltage dependence on the erratic mode of activity when studying 22 was observed. Switching from large negative to large positive potentials (- 100 to + 100 mV or - 150 to +150 mV) caused activity to cease which suggests that it is channel formation rather than bilayer perturbation that elicits the spiking events. The qualitatively similar mean conductances calculated for the compounds that display single-level open transitions (22, 29, and 31) suggest that the channel-active conformation of the molecule in the bilayer is common to all three compounds. The protracted open lifetimes were similar in length to the Gokel channel formers, TM and taken together with the observed ring size-conduction relationship, would appear to be consistent with a fixed conduction pathway arising by alignment of the macrocycles, in a stable transmembrane conformation. It is interesting to note that the single-channel events recorded were only obtained within the range -70 to +70 mV, beyond which irregular activity predominated. This observation is similar to the activity of the macrocyclic peptides of Voyer et al. 5~ which only displayed more regular transitions at lower applied potentials. Dougherty has commented on the possible role of aromatic side chains in the function of naturally occurring proteins 7 and Gokel has shown qualitatively that an increase in re-surface area of substituted tris(macrocycles) results in more channel openings and multiple transitions. 38i It is therefore interesting to note that the anthryl-containing compounds seemed to incorporate into the lipid bilayers more readily than the others, suggesting that the anthracene unit may provide increased bilayer solubilization via membrane anchoring. Boheim and co-workers 48 used oligo(tetrahydrofuran)s as a building block for a polyether helical peptide, 33, which consisted of five helical turns, to potentially span the lipophilic part of a lipid bilayer. Although the compound is not strictly macrocyclic, the tetrahydrofurans do provide a coil-like array of oxygen donor atoms. The stereogenic centers bearing the methyl groups also serve to stabilize the helical structure in the region of the peptide bond. Planar-bilayer studies of N-deprotected compound 33 in symmetrical KC1

R, --N

H

'H H

H 33

o J9

R2

Natural and Artificial Ion Channels

33

conditions exhibited irregular spiking events rather than discrete step-like fluctuations in current. These events were of millisecond duration and varied in intensity (up to 100 pA). The possibility that 33 could give rise to transient permeation structures that disturbed the lipid structure was offered as an alternative explanation to channel formation.

6.4. Macrocyclic Peptides Although containing amino acids, the following examples do not function via self-assembly of membrane-spanning ct-helical peptides to form a central pore. Instead, ion transport is mediated through tunnel-like structures and as a consequence they may be classified as macrocyclic in nature. Using a segment condensation strategy, Voyer and co-workers 49prepared a 21-amino acid peptide composed of 15 L-leucines and six 21-crown-7 L-phenylalanines (34). These amino acids were chosen for their high propensity to form o~-helices. The compound was designed to form an artificial ion channel of ca. 30/~ in length by stacking of the 21-crown-7 rings when adopting an tx-helical conformation in a membrane bilayer. Cation flux was assessed in unilamellar vesicles using a pH-stat method. 5~ Essentially, transport across a bilayer was followed by monitoring the volume of titrant needed to maintain a set pH in response to proton-cation antiport as a function of time. Vesicles with an internal pH of 6.6 were diluted with an external solution, adjusted to pH 7.6 to generate the proton gradient. A proton carrier [FCCP = (4-(trifluoromethoxy)phenyl)-hydrazone] was used to ensure rapid proton transport and various alkali metal cations were then added. In the absence of transporter or metal cation, no proton transport was observed. Addition of 34 in the presence of ions resulted in proton efflux, and the transport ability was assessed by the release of protons which were continuously neutralized to maintain the gradient. Rapid release Of protons was achieved with saturation being reached in ca. 2 min.

o•H

0 ~N

o

'

o

..9.

H

o-

3 0

0

0

0

0

0

02

%o.j

0

%oj 34

o,)

34

GREGORY J. KIRKOVITS and C. DENNIS HALL

Transporter 34 exhibited the same activity for the series, Li t , Na t, K t, Rb t, and Cs t. Interestingly, the heptapeptide analogue was too short to span the bilayer but functioned as a carrier. Subsequently, the ion channel activity of 34 was studied using planar bilayer methods. 5~ Peptide 34 was either introduced to the bulk KC1 solution or mixed with the lipid sample prior to bilayer formation. In the first method, incorporation proved difficult, possibly due to the high tendency of 34 to form aggregates. In the second method the stability of the bilayer was perturbed and single-channels were not obtained. It was postulated that this was due to the adsorption of the peptide onto the surface of the membrane due to electrostatic interactions of the crown ethers with the polar head groups. However as soon as the peptide was incorporated successfully into the bilayer, i.e. peptide parallel to the lipid hydrocarbons, singlechannel events were recorded indicative of 34 functioning in a unimolecular fashion. The difficulty experienced in the incorporation of the artificial channel was demonstrated with a fully protected analogue of 34 in which the L-leucine residues were replaced with L-alanines. 5~ Planar bilayer experiments using symmetrical NaCI conditions exhibited irregular-spiking activity under large potentials (>_+60 mV). This was ascribed to transient adsorption of the peptide on the membrane. Partially deprotected analogues behaved similarly but displayed a more regular activity at lower potentials (-40 to +40 mV). Single channel measurements of the fully deprotected peptide, however, were recorded at +_20 mV (amplitude range, 1-2 pA) indicating that the polar head groups anchored this analogue more strongly in the bilayer. Partial blockage of the ionic currents were observed in the presence of Cs + and guanidinium ions (bound by ligands of the 21-crown-7 type) suggesting that Na t ions traveled through the crown rings but the possibility of helical peptide aggregates forming channels could not be ruled out. Ghadiri 51 prepared an artificial ion channel based on a self-assembled 13-sheet peptide architecture. The eight residue peptide, cyclo-[-(Trp-D-Leu)aGln-D-Leu], 35, was composed of alternating L-tryptophan and D-leucine residues and a L-glutamine residue in order to attain a flat ring conformation thus providing an internal diameter of ca. 7.5 ,/~. The stacking of eight to ten subunits of the cyclic peptide, via a network of inter-subunit hydrogen bonds provided a tubular structure (a peptide nanotube) long enough to span a typical lipid membrane. Two factors were seen to induce the self-assembly of the channel structure: (i) the enthalpic contribution of a large number of hydrogen-bonds and (ii) an increase in the lipid chain entropy arising from the interactions of the hydrophobic side chains with the lipids in the non-polar environment. The hydrogen bonded network was confirmed by Fourier transform-infrared spectroscopy with the observation of a N - H stretching band at 3272 cm -1. Proton transport was monitored via pH-fluorescence spectroscopy using vesicles having internal pH 6.5 and external pH 5.5 again in the presence of an entrapped pH-sensitive dye. The ion transport ability of 35 was found to be slightly higher

Natural and Artificial Ion Channels

35

H2N

A

~

NH

35, cyclo[-(Trp-D-Leu),GIn-D-Leu] than that of gramicidin A and amphotericin B. The two control peptides cyclo[(Gln-D-Leu)4-],-which lacks the necessary hydrophobic side-chains for partitioning into the bilayer, and cyclo[-(MeN-D-Ala-Phe)4-] 51d-of sufficient hydrophobicity but predisposed to forming dimeric units, were both unable to promote proton transport under similar conditions. Thus the experiments showed the requirement of hydrophobic surface characteristics and extended hydrogen bonding as a means to provide channel structures long enough to span the lipid bilayer. Planar bilayer experiments showed brief step-like fluctuations in current, with 35 exhibiting single-channel conductances of 55 pS in 500 mM NaC1 and 65 pS in 500 mM KCI at concentrations in the order of 10-7 M in the subphase; the lack of selectivity was attributed to the large internal cavity. Open channel lifetimes showed a marked dependence on peptide concentration. For example, raising the added peptide concentration to 2.0 x 10-5 M produced open channel lifetimes of >30 s. Interestingly, by increasing the ring size of the cyclic peptide to cyclo[Gln-(D-LeuTrP)4-D-Leu] (pore diameter 10/~), Ghadiri s2 was able to claim the first example of an artificial transmembrane structure with the ability to transport glucose across unilamellar lipid vesicles. As an alternative to the cyclic D,L-O~-peptides, Ghadiri used cyclic tetramers of ct-unsubstituted-13-chiral-I$-amino acids (13a-amino acids) to form structurally analogous channel assemblies (36-38). 53 However, owing to the extra carbon atom in each residue, the cyclic 133-peptide subunits were expected to give a unidirectional arrangement of the polar amide backbone functionalities, with NH and carbonyl groups lying on opposite faces of the peptide ring. Stacking of the peptide subunits in a parallel manner was therefore anticipated to provide the tubular structure with a macrodipole moment similar to that observed in or-helices. Consequently, it was expected that the macrodipole would influence conductance through voltage gating

36

GREGORY J. KIRKOVITS and C. DENNIS HALL

36 R'= R~= R~

H

F

NH 2 J ' Hk 37 R'= i [ ' - " ~ 'R'=

R.,J =~

o~N

H"R

0

H

38R'=R~= ~ - ~

and rectification. The cyclic tetramer was chosen to provide conformational rigidity and a high degree of preorganization for self-assembly with an internal pore diameter of 2.6-2.7 A. Previous work by the group 51indicated that tryptophan-rich cyclic peptides partitioned efficiently into an apolar membrane and coupled with the implied role in membrane anchoring, led to a study of combinations of 36-38. Proton transport experiments, as described above showed that the activities of 36 and 37 were similar to the cyclic D,L-0t-peptides. However, 38 exhibited no activity, possibly due to precipitation of the peptide upon addition to the liposome suspension. In planar lipid bilayer experiments, compounds 36 and 37 displayed single channel opening and closing events as well as multiple transitions of individual channels. The observed conductance for 36 was 56 pS with 37 exhibiting similar activity. Channel gating was thought to be due to ring opening and collapse, as well as rapid assembly-disassembly of the peptide nanotube.

6.5. SyntheticChannels via Aggregate Formation Kobuke et al. 54 demonstrated channel activity of an ion pair composed of an oligoether-carboxylate anion combined with a dioctadecyldimethylammonium cation, 39, which aggregates as shown in Figure 15. The relatively hydrophilic oligoether chain was expected to provide a polar inner surface for cation transport surrounded by two hydrophobic long alkyl chains of the ammonium counterpart which would constitute the outer wall. In an extended conformation, the chain lengths were approximately similar to a lipid monolayer and therefore alignment and self-assembly of individual monomers of 39 within the lipid bilayer would be required to form the active transmembrane-spanning aggregate. Single-channel measurements of 39 in planar bilayers revealed three types of independent but constant conductance levels. These comprised (i) high conductance levels (700-1400 pS) which stayed open almost indefinitely; (ii) an intermediate level (70-100 pS) with millisecond opening and closing kinetics, and (iii) a low conductance level (10.85 + 1.58 pS) which showed relatively slow gating. The

37

Natural and Artificial Ion Channels

Figure 15. (i)Ion-pair channel former 39. (ii) Hypothetical half-channel structure, which is connected to another similar half-channel in the other half of the monolayer. The black and white columns represent the oligoether-carboxylate anion and the two-tailed hydrophobic ammonium counter ion, respectively. Adapted version of a figure reprinted from Kohuke, Y.; Morita, K., Inorg. Chim. Acta 283, 173 with permission from Elsevier Science. permeability ratio between K § and CI- was found to be ca. 5 which is consistent with the oligoether chains acting as the cation-conducting pore. The channels did not exhibit cation selectivity since conductances in 0.5 M NaCI and 0.5 M KCI were almost identical. Interestingly, the single-alkyl chain analogues of 39 were unable to form stable ionic channels. By replacing the carboxylate group of 39 (electrically neutral) with phosphate (40), an overall negative charge was introduced to the head group. 55 Consequently, it was proposed that the resultant dipole of 40 would give rise to voltage-gated activity. On incorporation into planar lipid bilayers, characteristic single channels were observed but more interestingly, the open probabilities showed a marked dependence on the applied voltage, which varied with each experiment. This suggested dipole reorientation in response to the applied potential as the origin of the voltage-gated response.

411

38

GREGORY J. KIRKOVITS and C. DENNIS HALL 0

-o

MeaN

0 41

o,

o.,~ ~

.o.p:_

O

O

oO

'

"

MeaN

~

0

42 O~_/O P O-" \0

Me3~'~k~ 0 " ~ ' ] ~ O~ 43 Ion pairs in which the ammonium cation was substituted with the oligoether chain, coupled with either steric acid, 41, L-tx-phosphatidic acid, 42 or dioctadecyl phosphate, 43, also showed activity in planar bilayers consistent with the formation of ion channels. 56 Observations with 41 were rather limited suggesting that a single alkyl chain is insufficient to cover the oligoether-pore conducting element. Channel currents with 42, however, were very stable with multiple openings being frequently observed at conductivities varying in the range 4-25 pS. The results suggested relative ease of incorporation of ion-pair 42 as well as the coexistence of nonidentical channel aggregates. The channels exhibited cation over anion selectivity, indicative of transport mediated by an inner pore composed of the o

O

o

o

RS o

~

o

~ 1 7 6 1 7 6

44, R = HO,C" / ~

S~f"

o

o

o

o

O

O

O ~ O O-v- G" ~ G -

45, R -- U~N~

vO

CO,H

o OH

46, R = H O o ~ O H o

HOzC" ~

s

O~O~o~0

o~X'ls

$~, O

47, G = CI-I~,r/= 1

O

O

o

O~G~G~O

48, G = CH~,n = 2

s O

CO,H

Natural and Artificial Ion Channels

39

The chains are attached to positively charged ammonium cations and, despite the reverse ionic character of the head groups that form the pore (compare 41-43 with Figure 15), the observed cation selectivity suggests that the inner membrane structure and not the charged head group determines ion selectivity. Fyles 25 prepared a series of bis-macrocyclic bola-amphiphiles (44-46), which were designed to be lipophilic, approximately columnar in shape and capable of stabilizing an ion in transmembrane passage via the formation of discretely sized aggregates. These compounds appear to have developed a reincarnation of an earlier channel model (see Figure 9). Transport of a series of alkali metals was assessed using the pH-stat method described above. The apparent kinetic order for all the compounds studied was approximately two indicating an aggregate as the active species. Multiple-steps of uniform conductance, with durations of a second or more, were most frequently observed in planar bilayer experiments, suggesting multiple copies of a single aggregate. Specific conductances were broadly in the range 7-35 pS, depending on the cation and lipid preparation used and the compounds exhibited a modest cation over anion selectivity. The results support a model of cation transport through an aqueous channel between two or more membrane spanning bola-amphiphiles. Modifications to the hydrophobic chains (compounds 47 and 48) induced a unique current-time response in planar bilayers. 57a Compound 49 also displayed very interesting channel activity when studied using planar bilayer methods. 57b It was reasoned that the inherent molecular dipole of 49 would inhibit its transfer through the bilayer, thus imposing an asymmetric distribution in which the more polar end (succinate) would reside at the side of the bilayer to which it was added and hence that voltage-gated activity would be observed. When 49 was added to the cis chamber and a voltage-ramp potential was applied (- 100 to + 100 mV), significant cation flux was recorded only at negative potentials. Over a period of time, however, the current-voltage asymmetry disappeared and current was seen at both negative and positive potentials. It was reasoned that at the early stages, orientation of 49 was indeed asymmetrical but that over a period of time the succinate headgroups began to penetrate the bilayer, thus producing both transmembrane orientations. Indeed, when 49 was added to the trans side, current was initially observed only at positive potentials. Pre-mixing of 49 with the lipid prior to bilayer formation gave a linear current-voltage response. Under "normal" conditions (i.e. at constant applied potential) 49 showed very brief bursts of irregular openings. However, when the pH of the electrolyte solution was lowered from 6.4 to 5.9, an open state of constant amplitude (-8.65 + 0.5 pA at -120 mV, o o

H02C~ s

~ o

o

o

1

~ ~/~176

7

6

S"~ o

s

0~0~/~0~0

o

o

49

o

40

GREGORY J. KIRKOVITS and C. DENNIS HALL

-72 + 4 pS) with a lifetime of minutes was observed. It was suggested that at lower pH a greater percentage of the succinate head groups were protonated, thus reducing electrostatic repulsions between the dianionic succinate head groups in an orientated aggregate and therefore increasing random transmembrane distribution of 49. Introduction of a low concentration of barium salts also induced discrete single channels, presumably by reducing head group repulsions within an aggregate by masking the anionic charge through association with the divalent cation. Regen 58a used amphotericin B as a design basis for the structural requirements of an ion channel. Composed of a rigid, hydrophobic sterol nucleus and flexible hydrophilic chains, 50 was expected to function via self-assembly in a membrane bilayer. Cation flux was assessed using Lehn's NMR method and comparisons were made with the bouquet molecules. Ionophoric activity was similar to the more sophisticated bouquet molecules (maximum entry achieved in ca. 60 h), but Na § transport was significantly less than recorded with amphotericin B. The rate of transport showed a strong dependence on the concentration of 50, indicating the formation of active aggregates. 58b Detailed studies of the rate of transport versus ionophore concentration revealed two distinct processes attributed to a monomer-active carrier species at "low" concentration and an active aggregate at "high" concentration. Transport rates for an analogue of 50 were thus assessed in both fluid- and gel-phase membranes. At low concentration (1 mol% sterol), no Na § transport was detected in gel-phase membranes but transport was seen in the fluid phase. In contrast, at high concentration (2.5 mol% sterol) transport was effected in both fluid- and gel-phase membranes. This clearly demonstrated that at low concentrations a carrier mechanism dominates but that at high concentration a channel mechanism operates. Regen 59 also prepared a molecule that mimics the structure and properties of the naturally occurring sterol, squalamine. It was shown that 51 was membraneselective in that transport of protons and/or anions via the polyamine chain was favored in negatively charged bilayers (egg phosphatidylglycerol) over bilayers that were electrically neutral [egg phosphatidylcholine (egg PC)], a feature which was ascribed to surface charge interactions. Within a monolayer leaflet, 51 was thought to fold on itself by forming a salt bridge, thus exposing a zwitterionic head group at the membrane surface. Since this zwitterion would

50, R = H or CH~

Natural and Artificial Ion Channels

41

(ii)

(i) H

H

_NH~X//x~ N~

$

O

N

0

51 Figure 16. (i) Squalamine mimic sterol 51. (ii) Hypothetical arrangement of aggregated 51 in a monolayer leaflet. Alignment with another aggregate in the other monolayer is required to give the transmembrane channel. Reproduced with permission from Merritt, M.; Lanier, M.; Deng, G.; Regen, S. L., J. Am. Chem. Soc. 1998, 120, 8496.

favor other zwitterionic head groups as nearest neighbors (e.g. those of egg PC), formation of stable internalized aggregates of sterol 51 might be preferable in an environment of negatively charged phospholipids as represented in Figure 16ii. In the absence of the ionizable primary amine (through N-acetylation) or alternatively, with the polyamine chain replaced by an ether linkage, negligible activity was observed in pH-discharge experiments.

6.6. Ion Channel Forming Polymers Seebach 6~studied the single-channel properties of oligomers of poly[(R)-3-hydroxybutanoate], P(3-HB), 52, a stereoregular polymer produced as a storage material by prokaryotic microorganisms. The unprotected derivatives of chain length > 16 monomer units exhibited typical single-channel behavior in the presence ofRb § and Ba 2§ ions at concentrations of 0.1-5% of lipid. A linear current-voltage relationship was seen, but measurements on more than one membrane for a particular oligomer led to different mean values of conductance and the values did not correlate with the different concentrations used. In a subsequent study Seebach et al. 61 synthesized a 128-mer of hydroxybutanoic acid complexed with inorganic calcium polyphosphate and showed that it could form calcium-selective channels that were indistinguishable from poly(3-hydroxybutyrate)/calcium polyphosphate voltage-gated channels isolated from Escherichia coli plasma membranes. Similarly, a synthetic polyelectrolyte, poly(2-ethylacrylic acid) (PEAA) was shown to be active in artificial membranes and pH-sensitive. 62 Modes of activity n n

52

, 6 3 2 6 4 96

42

GREGORY I. KIRKOVITS and C. DENNIS HALL

included discrete conductance steps, but irregular open-channel events and fastflickering spikes were more commonly observed. Analysis of the discrete activity in 200 mM NaCI solution demonstrated multiple independent conductance states induced by PEAA, in the range 90--400 pS. The observed PEAA channels were seen to be cation-selective but with variable permeability ratios of Na + over C1- in the range, 2-11. Formation of ion channels was also shown to be pH-dependent. For example, exposure to a PEAA solution at pH 7.6 produced no activity but on switching the bathing solution to pH 6.0, activity was observed.

6.7. Modeling Potassium Channel Selectivity The near perfect selectivity of potassium channels has been the focus of much recent attention. Due to the presence of tyrosine residues in the selectivity filter (vide supra), it was suggested that cation-n interactions 6 might play a role in achieving selectivity. The postulate promoted interest in the area 63 and calixarenes were obvious candidates to provide suitable models to establish experimental evidence for the concept. To this end, Beer 64 reported the crystal structure of a calix[4]tube which displayed potassium selectivity (Figure 15i). Similarly, Cragg et al. 65 described the crystal structure of a 2 : 1, Na:oxacalix[3]arene complex (Figure 17ii), which provided insight into the so-called "billiard ball" effect. 8 Preliminary electrophysiological experiments suggested that the calix inhibited K § flux but allowed Na § transport, thus performing the function of a selectivity filter. Kobuke et al. 66a reported the preparation and properties of a macrocyclic amphiphile based on resorcinol which showed K § selectivity. The basis for the selectivity was the rigid tetraphenylene unit, which was expected to present a physical barrier to the selection of cations on the basis of size (Figure 18). On incorporation into a planar lipid bilayer, characteristics of a single ion channel were recorded as evidenced by integral transitions between open and closed states (in the order of milliseconds to seconds) and the observation of a stable and constant conductance (6.1 + 0.8 pS). Only the longer chain analogue 53, however, produced stable channel currents showing that prolonged activity could only be observed

tl~utF~u

(i)

~Bu

tB%,.,

(ii)

~

Figure 17. (i) Beer's potassium selective calix[4]tube. (ii)Cragg's oxacalix[3]arene.

43

Natural and Artificial Ion Channels

C HO~OH

0~0

HO H R - ~ ~ HO

~

I

H

H

R OH

(

-OH

53, R = (CH2),,CH, 54, R = (CH2),oCH,

C

)

~~

H

~-o.

Figure 18. The resorcinol based half-channel and its hypothetical structure in a lipid bilayer. Reproduced with author's permission from Tanaka, Y.; Kobuke, Y.; Sokabe, M., Angew. Chem. Int. Ed. Engl. 1995, 34, 693.

when the amphiphile had a length capable of spanning half the membrane. It was proposed that 53 recognized the positioning of another molecule of 53 in the other half of the bilayer, thereby forming a tail-to-tail dimer which provided an overall tunnel for cation passage (Figure 18). This proposal was supported by the observation that when 53 was added to only one side of the membrane, channel activity was not observed. Under salt gradient conditions, 53 also revealed selectivity for potassium over sodium (permeability ratio, PK§ § = 3) and the passage of K § ions was blocked by a counter Rb § gradient. Attractive x-cation interactions between K § and the aryl system were thought to contribute to the lowering of the potential barrier for K § transport, the pore being large enough for K § passage (r = 1.33/~) but not for the Rb § cation (r = 1.47/~). Interestingly, Kobuke 66b has recently reported the ability of 53 to self-assemble into stable vesicles. Matile 67 developed the first ligand-gated K+-selective channel by exploiting n-cation interactions. The rigid-rod oligo(p-phenylene)system described earlier was modified to 55 by incorporating an external binding ligand. The iminodiacetate (IDA) link of 55 was then used to bind polyhistidine (pHis) through Cu 2§ (Figure 19ii) and self-assembly of 55 was expected to produce a membrane-spanning slide for cations. Cation transport was assessed across vesicle bilayer membranes and the rate of K§ for complexed 55 was significantly greater than for Na § However, the observed selectivity was much lower for the free ligand which implied ligand-in-

44

GREGORY J. KIRKOVITS and C. DENNIS HALL (i)

n slide

oN

ligand bindingsite

55 o

(ii)

!

o

.0

~ .0

-o--~,::..oO--~,::.. Figure 19, (i) Structure of rigid-rod "x-slide" 55. (ii) The ligand binding site showing binding of polyhistidine (pills) to iminodiacetate (IDA) through Cu 2+.

duced organization of the ~ slides with complexed 55. The hypothesis was supported by the observed inhibition of transport for 55 in the presence of external tetraethylammonium cation (TEA+), a common K + channel blocker which is thought to act by binding to the cyclic arene tetrad in the selectivity filter. 8 Consequently cation selectivity may be due to the ligand-assembled arene arrays.

7. CONCLUSION A diverse number of approaches to the design of artificial ion channels and the study of their transport has been described. The ability of these systems to be effective as ion channels varies considerably, but taken together they show that the activity of structurally "simple" ionophores can mimic those of their more illustrious natural counterparts. The challenge of controlling selectivity and the phenomenon of gating or rectification remains to be overcome in artificial systems but the expansion of knowledge in the field of supramolecular chemistry promises to resolve many of the outstanding questions in the near future.

REFERENCES 1. 2. 3. 4. 5. 6. 7.

Hille, B. Ionic Channel of Excitable Membranes, 1st ed.; Sinauer: Sunderland, MA, 1984. Laniado,M.; Lalani, E.-L. Chemistry & Industry 1999, 100-104 and references cited therein. Stein,W. D. Channels, Carriers, and Pumps; AcademicPress: New York, 1990. Miller,C. Science 1991, 252, 1092-1096. Dougherty,D. A.; I,ester, H. A. Angew. Che~, Int. Ed. Engl. 1998, 37, 2329-2331. Ma, J. C.; Dougherty,D. A. Chent Rev. 1997, 97, 1303-1324 and references cited therein. Dougherty,D. A. Science 1996, 271, 163-168 and references cited therein.

Natural and Artificial Ion Channels

45

8. Doyle, D. A.; Cabral, J. M.; Pfuetzner, R. A.; Kuo, A.; Gulbis, J. M.; Cohen, S. L.; Chait, B. T.; MacKinnon, R. Science 1998, 280, 69-77. 9. Roux, B.; MacKinnon, R. Science 1999, 285, 100-102. 10. (a) Urry, D. W. Proc. Natl. Acad. Sci. USA 1971, 68, 672-676. (b) Urry, D. W.; GoodaU, M. C.; Glickson, J. D.; Mayers, D. F. Proc. Nat. Acad. Sci. USA 1971, 68, 1907-1911. 11. (a) Wallace, B. A.; Ravikumar, K. Science 1988, 241, 182-187. (b) Langs, D. A. Science 1988, 241, 188-191. 12. Manning, S. D. Ph.D. Thesis, University of London, 1989. 13. Williams, A. J. In Microelectrode Techniques--The Plymouth Workshop Handbook, 2nd ed.; Ogden, D., Ed.; The Company of Biologists Limited: Cambridge, 1994, pp. 79-101 and references cited therein. 14. (a) Mutter, M.; Tuchscherer, G. G.; Miller, C.; Altmann, K.-H.; Carey, R. I.; Wyss, D. F.; Labhardt, A. M.; Rivier, J. E. J. Am. Chem. Soc. 1992, 114, 1463-1470. (b) Grove, A.; Mutter, M.; Rivier, J. E.; Montal, M. J. Am. Chem. Soc. 1993, 115, 5919-5924. 15. Futaki, S. Biopolyrners 1998, 47, 75-81. 16. (a) r.~ar, J. D.; Wasserman, Z. R.; DeGrado, W. F. Science 1988, 240, 1177-1181. (b) ]~kerfeldt, K. S.; Kim, R. M.; Camac, D.; Groves, J. T.; Lear, J. D.; DeGrado, W. F. J. Am. Chem- Soc. 1992, 114, 9656-9657. 17. Akerfeldt, K. S.; Lear, J. D.; Wasserman, Z. R.; Chung, L. A.; I)eGrado, W. F. Acc. Chem. Res. 1993, 26, 191-197. 18. Lear, J. D.; Schneider, J. P.; Kienker, P. K.; DeGrado, W. E J. Am- Chem- Soc. 1997, 119, 3212-3217 and references cited therein. 19. Otoda, K.; Kimura, S.; Imanishi, Y. J. Chem- Soc., Perkin Trans 1 1993, 30, 11-30. 20. Ueda, H.; Kimura, S.; Imanishi, Y. Chem Commun. 1998, 363-364. 21. Woolley, G. A.; Jaikaran, A. S. I.; Zhang, Z.; Peng, S.J. Am. Chem- Soc. 1995, 117, 4448-4454. 22. Lien, L.; Jaikaran, D. C. J.; Zhang, Z.; Woolley, G. A. J. Am. Chem- Soc. 1996,118, 12222-12223. 23. (a) Stankovic, C. J.; Heinemann, S. H.; Delfino, J. M.; Sigworth, E J.; Schreiber, S. L. Science 1989, 244, 813-817. (b) Stankovic, C. J.; Heinemann, S. H.; Schreiber, S. L. J. Ant Chem- Soc. 1990, 112, 3702-3704. 24. Schmitt, J. D.; Sansom, M. S. E; Kerr, I. D.; Lunt, G. G.; Eisenthal, R. Biochemistry 1997, 36, 1115-1122. 25. Fyles, T. M.; Loock, D.; van Straaten-Nijenhuis, W. E; Zhou, X. J. Org. Chem- 1996, 61, 8866-8874 and references cited therein. 26. Gokel, G. W.; Murillo, O. Acc. Chem- Res. 1996, 29, 425-432. 27. Fyles, T. M. Current Opin. Chem. Biol. 1997, I, 497-505. 28. Tabushi, I.; Kuroda, Y.; Yokota, K. Tetrahedron Lett. 1982, 23, 4601-4604. 29. Behr, J.-E, Lehn, J.-M.; Dock, A.-C.; Moras, D. Nature 1982, 295, 526-527. 30. (a) Neeval, J. G.; Nolte, R. J. M. Tetrahedron Lett. 1984, 25, 2263-2266. (b) Kragten, U. E; Roks, M. E M.; Nolte, R. J. M. J. Chem- Soc., Chem. Commun. 1985, 1275-1276. 31. Menger, E M.; Davis, D. S.; Persichetti, R. A.; Lee, J.-J. J. Am. Chem- Soc. 1990,112, 2451-2452. 32. Dubowchik, G. M.; Firestone, R. A. Tetrahedron Lett. 1996, 37, 6465-6468. 33. Sakai, N.; Matile, S. Tetrahedron Lett. 1997, 38, 2613-2616. 34. Menger, E M.; Aikens, E Angew. Chem., Int. Ed. Engl. 1992, 31, 898-900. 35. (a) Sakai, N.; Brennan, K. C.; Weiss, L. A.; Matile, S. J. Am- Chem- Soc. 1997, 119, 8726-8727. (b) Weiss, L. A.; Sakai, N.; Ghebremariam, B.; Ni, C.; Matile, S. J. Am- Chem- Soc. 1997, 119, 12412-12419. (c) Ni, C.; Matile, S. J. Chem- Soc., Chem. Commun. 1998, 755-756. 36. (a) Cannichael, V. E.; Dutton, E J.; Fyles, T. M.; James, T. D.; Swan, J. A.; Zojaji, M. J. Am. Chem. Soc. 1989, 111, 767-769. (b) Fyles, T. M.; James, T. D.; Kaye, K. C. Can. J. Chem- 1990, 68, 976-978. (c) Fyles, T. M.; Kaye, K. C.; James, T. D.; Smiley, D. W. M. Tetrahedron Lett. 1990, 31, 1233-1236. (d) Fyles, T. M.; James, T. D.; Pryhitka, A.; Zojaji, M. J. Org. Chem. 1993, 58, 7456-7468. (e) Fyles, T. M.; James, T. D.; Kaye, K. C.J. Am. Chem- Soc. 1993,115,12315-12321.

46

GREGORY I. KIRKOVITS and C. DENNIS HALL

37. (a) Jullien, L.; Lehn, J.-M. Tetrahedron Lett. 1988, 29, 3803-3806. (b) Canceill, J.; Juillien, L.; Lacombe, L.; Lehn, J.-M. Helv. Chim. Acta 1992, 75, 791-812. (c) Pregel, M. J.; Jullien, L.; Lehn, J.-M. Angew. Chem., Int. Ed. Engl. 1992, 31, 1637-1640. (d) Jullien, L.; Lazrak, T.; Canceill, J.; Lacombe, L.; Lelm, J.-M. J. Chem. Soc., Perkin Trans. 2 1993, 1011-1020. (e) Pregel, M. J.; Juillien, L.; Canceill, J.; Lacombe, L.; Lehn, J.-M. J. Chem. Soc., Perkin Trans. 2 1995, 417--426. 38. (a) Nakano, A.; Xie, Q.; MaUen, J. V.; Echegoyen, L.; Gokel, G. W. J. Ant Chem. Soc. 1990, 112, 1287-1289. (b) Murillo, O.; Watanabe, S.; Nakano, A.; Gokel, G. W. J. Am. Chem. Soc. 1995, 117, 7665-7679 and references cited therein. (c) Murillo, O.; Suzuki, I.; Abel, E.; Gokel, G. W. J. Am. Chem. Soc. 1996,118, 7628-7629. (d) Murillo, O.; Abel, E.; Maguire, G. E. M.; Gokel, G. W.; Gokel, G. W. J. Chem. Soc., Chem. Commun. 1996, 2147-2148. (e) Murillo, O.; Suzuki, I.; Abel, E.; Murray, C. L.; Meadows, E. S.; Jin, T.; Gokel, G. W. J. Am. Chem. Soc. 1997, 119, 5540-5549. (f) Abel, E.; Meadows, E. S.; Suzuki, I.; Jin, T.; Gokel, G. W. J. Chem. Soc., Chem. Commun. 1997, 1145-1146. (g) Abel, E.; Maguire, G. E. M.; Meadows, E. S.; Murillo, O.; Jin, T.; Gokel, G. W. J. Am. Chem. Soc. 1997, 119, 9061-9062. (h) Murray, C. L.; Meadows, E. S.; Murillo, O.; Gokel, G. W. J. Am. Cheat Soc. 1997, 119, 7887-7888. (i) Maguire, G. E. M.; Meadows, E. S.; Murray, C. L.; Gokel, G. W. Tetrahedron Lett. 1997, 38, 6339-6342. (j) Murray, C. L.; Gokel, G. W. J. Chem. Soc., Chem. Commun. 1998, 2477-2478. 39. (a) Echegoyen, L. E.; Hernandez, J. C.; Kaifer, A. E.; Gokel, G. W.; Echegoyen, L. J. Chem. Soc., Chem. Commun. 1988, 836-837. (b) De Wall, S. L.; Wang, K.; Berger, D. R.; Watanabe, S.; Hemandez, J. C.; Gokel, G. W. J. Org. Chem. 1997, 62, 6784-6791. 40. (a) Mufioz, S.; Mall6n, J. V.; Nakano, A.; Chen, Z.; Gay, I.; Echegoyen, L.; Gokel, G. W. J. Chem. Soc., Chem. Commun. 1992, 520-522. (b) Mufioz, S.; Mall6n, J. V.; Nakano, A.; Chen, Z.; Gay, I.; Echegoyen, L.; Gokel, G. W. J. Am. Chem. Soc. 1993, 115, 1705-1711. 41. de Mendoza, J.; Cuevas, E; Prados, P.; Meadows, E. S.; Gokel, G. W. Angew. Chem., Int. Ed. Engl. 1998, 37, 1534-1536. 42. (a) Hall, C. D.; Kirkovits, G. J.; Hall, A. C. J. Chem. Soc., Chem. Commun. 1999, 1897. (b) Kirkovits, G. J. Ph.D. Thesis, University of London, 2000. 43. (a) Medina, J. C.; Goodnow, T. T.; Bott, S.; Atwood, J. L.; Kaifer, A. E.; Gokel, G. J. Chem. Soc., Chem. Commun. 1991, 290-292. (b) Hall, C. D.; Sharpe, N. W.; Danks, I. P.; Sang, Y. P. J. Chem. Soc., Chem. Commun. 1989, 419-421. (c) Plenio, H.; Yang, J.; Diodone, R.; Heinze, J. Inorg. Chem. 1994, 33, 4098-4104. (d) Plenio, H.; Diodone, R. lnorg. Chem. 1995, 34, 3964-3972. (e) Plenio, H.; Aberle, C. Organometallics 1997, 16, 5950-5957. (f) Medina, J. C.; Goodnow, T. T.; Rojas, M. T.; Atwood, J. L.; Lynn, B. C.; Kaifer, A. E. (g) Hall, C. D.; Tucker, J. H. R.; Sharpe, N. W. Organometallics 1991, 10, 1727-1731. 44. Beer, P. D. Chem. Soc. Rev. 1989, 18, 409-450. 45. Beer, P. D.; Chen, Z.; Ogden, M. I. J. Chem. Soc. Faraday Trans. 1995, 91(2), 295-302. 46. Beer, P. D.; Chen, Z.; Drew, M. G. B.; Kingston, J.; Ogden, M.; Spencer, P. J. Chem. Soc., Chem. Commun. 1993, 1046-1 048. 47. Plenio, H.; Diodone, R. Inorg. Chem. 1995, 34, 3964-3972. 48. Wagner, H.; Harms, K.; Koert, U.; Meder, S.; Boheim, G. Angew. Chem., Int. Ed. Engl. 1996, 35, 2643-2646. 49. Voyer, N.J. Am. Chem. Soc. 1991, 113, 1818-1821. 50. (a) Voyer, N.; Robitaille, M. J. Am. Chem. Soc. 1995, 117, 6599-6600. (b) Voyer, N.; Lamothe, J. Tetmhedron 1995, 51, 9241-9284. (c) Meillon, J.-C.; Voyer, N.Angew. Chem., Int. Ed. Engl. 1997, 36, 967-969. (d) Voyer, N.; Potvin, L.; Rousseau, E. J. Chem. Soc., Perkin Trans. 2 1997, 1469-1471. 51. (a) Ghadiri, M. R.; Granja, J. R.; Milligan, R. A.; McRee, D. E.; Khanzanovich, N. Nature 1993, 366, 324-327. (b) Ghadiri, M. R.; Granja, J. R.; Buehler, L. K. Nature 1994, 369, 301-304. (c) Khanzanovich, N.; Granja, J. R.; McRee, D. E.; Milligan, R. A.; Ghadiri, M. R. J. Am. Chem. Soc. 1994, 116, 6011-6012. (d) Ghadiri, M. R.; Kobayashi, K.; Granja, J. R.; Chadha, R. K.; McRee, D. E. Angew. Chem., Int. Ed. Engl. 1995, 34, 93-95. (e) Kobayashi, K.; Granja, J. R.; Ghadiri, M.

Natural and Artificial Ion Channels

52. 53. 54. 55. 56. 57. 58.

59. 60. 61. 62. 63. 64. 65. 66. 67.

47

R. Angew. Chent, Int. Ed. Engl. 1995, 34, 95-98. (f) Engels, M.; Bashford, D.; Ghadiri, M. R. J. Ant Chent Soc. 1995, 117, 9151-9158. (g) Hartgerink, J. D.; Granja, J. R.; Milligan, R. A.; Ghadiri, M. R. J. Ant Chent Soc. 1996, 118, 43-50. (h) Kim, H. S.; Hartgerink, J. D.; Ghadiri, M. R. J. Ant Chem. Soc. 1998, 120, 4417-4424. (i) Hartgerink, J. D.; Clark, T. D.; Ghadiri, M. R. Chem. Fur. J. 1998, 4, 1367-1372. Granja, J. R.; Ghadiri, M. R. J. Ant Chent Soc. 1994, 116, 10785-10786. Clark, T. D.; Buehler, L. K.; Ghadiri, M. R. J. Ant Chent Soc. 1998, 120, 651-656. Kobuke, Y.; Ueda, K.; Sokabe, M. J. Ant Chem. Soc. 1992, 114, 7618-7622. Kobuke, Y.; Uoda, K.; Sokabe, M. Chent Lett. 1995, 435-436. Kobuke, Y.; Morita, K. lnorg. Chint Acta 1998, 283, 167-174. (a) Fyles, T. M.; Loock, D.; Zhou, X. Can. J. Chent 1998, 76, 1015-1026. (b) Fyles, T. M.; Loock, D.; Zhou, X. J. Ant Chent Soc. 1998, 120, 2997-3003. (a) Stadler, E.; Dedek, P.; Yamashita, K.; Regen, S. L. J. Am. Chent Soc. 1994, 116, 6677-6682. (b) Deng, G.; Merritt, M." Yamashita, K." Janout, V.; Sadownik, A." Regen, S. L. J. Ant Chent Soc. 1996, 118, 3307-3308. (a) Deng, G.; Dewa, T.; Regen, S. L. J. Ant Chent Soc. 1996, 118, 8975-8976. (b) Merritt, M.; Lanier, M.; Deng, G.; Regen, S. L. J. Ant Chent Soc. 1998, 120, 8494-8501. Seebach, D.; Brunner, A.; Biirger, H. M.; Reusch, R. N.; Bramble, L. L. Helv. Chint Acta 1996, 79, 507-517. Das, S.; Lengweiler, U. D.; Seebach, D.; Reusch, R. N. Proc. Natl. Acad. Sci. USA 1997, 94, 9075-9079. Chung, J. C.; Gross, D. J.; Thomas, J. L.; Tirrell, D. A.; Opsahl-Ong, L. R. Macmmolecules 1996, 29, 4636--4641. De Wall, S. D.; Meadows, E. S.; Barbour, L. J.; Gokel, G. W. J. Ant Chent Soc. 1999, 121, 5613-5614. Schmitt, P.; Beer, P. D.; Drew, M. G. B.; Sheen, P. D. Angew. Chent, Int. Ed. Engl. 1997, 36, 1840-1842. Cragg, P. J.; Allen, M. C.; Steed, J. W. J. Chent Soc., Chent Commun. 1999, 553-554. (a) Tanaka, Y.; Kobuke, Y.; Sokabe, M.Angew. Chent, Int. Ed. Engl. 1995, 34, 693-694. (b) Tanaka, Y.; Miyachi, M.; Kobuke, Y. Angew. Chent, Int. Ed. Engl. 1999, 38, 504-506. Tedesco, M. M.; Ghebremariam, B.; Sakai, N.; Matile, S. Angew. Chent, Int. Ed. Engl. 1999, 38, 540-543.

This Page Intentionally Left Blank

ION RECOGNITION AND TRANSPORT BY POLY-(R)-3-HYD ROXYB UTYRATES AND INORGANIC POkYPHOSPHATES

Rosetta N. Reusch

1. 2.

3.

4.

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Inorganic Polyphosphates (PolyPs) . . . . . . . . . . . . . . . . . . . . . . . . 2.1. Occurrence and Synthesis . . . . . . . . . . . . . . . . . . . . . . . . . 2.2. Ion Selection and Transport . . . . . . . . . . . . . . . . . . . . . . . . . Poly-(R)-3-hydroxybutyrates . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1. Occurrence and Synthesis . . . . . . . . . . . . . . . . . . . . . . . . . 3.2. Amphiphilic and Solvent Properties . . . . . . . . . . . . . . . . . . . . 3.3. Channels in Planar Lipid Bilayers . . . . . . . . . . . . . . . . . . . . . Complexes of PHB and PolyP . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.1. PHB/PolyP Complexes in Bacterial Membranes . . . . . . . . . . . . . . 4.2. E. coli PHB/PolyP Complexes as Ion Channels . . . . . . . . . . . . . . 4.3. 9Synthetic Ion Channels from PHB 128 and PolyPs . . . . . . . . . . . . . 4.4. Synthetic Ion Channels from OHB19/23 and PolyPs . . . . . . . . . . . . 4.5. Selectivity of PHB/PolyP Ion Channels . . . . . . . . . . . . . . . . . . 4.6. Block of PHB/PolyP Ion Channels . . . . . . . . . . . . . . . . . . . . . 4.7. Gating of PHB/PolyP Ion Channels . . . . . . . . . . . . . . . . . . . .

Advances in Supramolecular Chemistry Volume 7, pages 49-98. Copyright 9 2000 by JAI Press Inc. All rights of reproduction in any form reserved. ISBN: 0-7623-0678-5 49

50 50 50 52 53 53 55 58 63 63 66 68 70 72 76 76

50

ROSETTAN. REUSCH

4.8. Voltage-Dependenceof PHB/PolyP Ion Channels . . . . . . . . . . . . 5. Structureand Mechanism of Ion Transport . . . . . . . . . . . . . . . . . . . 6. SupramolecularComplexes of Protein, PHB, and PolyP . . . . . . . . . . . . 6.1. PHB/PolyPComplexes in a CaATPase Pump . . . . . . . . . . . . . . . 6.2. PHB and PolyP in a Potassium Channel . . . . . . . . . . . . . . . . . 7. PHB/PolyPComplexes as DNA Channels? . . . . . . . . . . . . . . . . . . . 8. EvolutionaryAspects and Concluding Remarks . . . . . . . . . . . . . . . . . Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

79 79 83 83 85 90 93 94 95

1. INTRODUCTION Poly-(R)-3-hydroxybutyrates (PHBs) and inorganic polyphosphates (polyPs) are simple linear homopolymers, derived from acetate and phosphate, respectively. Both polymers are widely distributed in biological organisms, from the most primitive bacteria to the most highly evolved plants and animals. 1-21 Each has singular molecular characteristics that contribute to a capacity for ion recognition and transport. Together, they have all the requisite attributes of physiological ion channels. Accordingly, complexes of PHBs and polyPs, isolated from the plasma membranes of bacteria 22-25 or prepared from the synthetic polymers, 26 form voltage-activated calcium-selective channels across planar lipid bilayers that display many of the characteristics of protein calcium channels. 27'28 There is increasing evidence that the roles of PHB and polyP in ion transfer have been conserved. The two polymers were recently found to be components of two transport proteins, namely the CaATPase pump of human erythrocytes and the potassium channel of Streptomyces lividans, 29"3~testifying to their relevance and versatility in ion regulation. In this review, we first discuss the individual properties of polyPs and PHBs related to their function in ion transport, and then consider how these two structurally dissimilar polymers may act in synergy to form ion-selective transmembrane channels. Finally, we examine protein ion transporters that contain PHBs and polyPs, and consider how these three macromolecules may come together in supramolecular channels to refine ion recognition and regulate ion transport.

2.

INORGANIC POLYPHOSPHATES (POLYPS) 2.1. Occurrence and Synthesis

Inorganic polyphosphates (polyPs), also called condensed phosphates, are linear polymers of tetrahedral orthophosphate units linked through common oxygen atoms by phosphoanhydride bonds (Figure 1A). PolyPs can be formed in the laboratory or in nature by temperature- or pressure-induced condensation of acid orthophosphate salts; accordingly, they are prepared synthetically by heating the

PHB and Polyp Ion Transport

~~

P

51

~o~ /o~ ~o~ /o~ /o~ /o~ ~o~ /o~_ 7 P

_

[K(PO3)ln

P

_

P

o

[Ca0'O3)2ln

P

_

P

o_

.

J'o. .J-"o. or

[Na(eO3)ln

[KBa2&O3)sln

Figure 1. A. Chain of polyphosphate (polyP) residues. B. Chain configurations of polyP salts. Drawings show a few of many possible conformations taken from crystal structures. 33'35 Individual phosphate residues, each with a monovalent negative charge, are represented by a tetrahedron. The rotational flexibility of the P-O-P bonds joining the tetrahedra allows the chains to twist into a large variety of conformations, depending on the associated cations. 27 appropriate dihydrogen phosphate salts and they may be found in the environment in volcanic residues and deep-sea ocean vents. 31'32 nMHEPO 4 ~ (MP03) ~ + nI-~O Phosphate anhydrides are the principle stores of biochemical energy. The phosphoanhydride bond of polyP has a standard free energy of hydrolysis of approximately -9.7 kcal, within the range of that of other cellular phosphorylating agents such as ATP (-7.3 kcal) and phosphocreatine (-10.3 kcal). 33 As a consequence of its intermediate position in the thermodynamic scale, polyPs can both transfer and accept phosphates from other phosphate anhydrides, and this is apparently the

52

ROSETTA N. REUSCH

means used to synthesize and degrade polyPs in biological cells. The enzyme polyphosphate kinase (PPK), discovered in E. coli by Komberg et al., 34 and since reported in many other organisms, 35-37 catalyzes the reversible transfer of the y phosphate of ATP to the end of a polyphosphate chain. ATP + (polyP),, ~ ADP + (polyP)n+l The same enzyme can also catalyze the transfer of terminal phosphate units from polyP to GDP, UDP, and CDP. 38 Enzymes that catalyze the transfer of phosphate from polyP to AMP or to NAD or to glucose have also been described in bacteria. 7's Other phosphate anhydrides with high standard free energy of hydrolysis can donate high-energy phosphates to polyP; e.g. an enzyme that catalyzes reversible transfer of phosphates from 1,3-diphosphoglycerate to polyP has been identified in Neisseria crassa. 2'39 Homologous polyP-synthesizing enzymes have not been identified in eukaryotes, but proteins that are reversibly phosphorylated by ATP are potential catalysts for polyP synthesis (see 6.1). Enzymes that hydrolyze polyphosphates, endopolyphosphatases, and exopolyphosphatases have been detected in diverse prokaryotic and eukaryotic cells. 7'8 Many biological functions have been proposed for polyPs including: phosphate reserve, ATP substitute and energy source, chelator, buffer, regulator of enzymes and gene activity, cell wall component, antiviral agent, transport of ions, and transport of DNA. 2-s

2.2. Ion Selection and Transport Orthophosphates are especially adapted to form polyelectrolytes because of their trivalent negative charge. Since the first pK of phosphates is --2 and successive ionization constants differ by factors of 105, phosphate units retain a monovalent negative charge at neutral pH, even after they are linked together. This negative charge protects polyPs from rapid hydrolytic attack by water, making them kinetically stable under physiological conditions in spite of their pronounced thermodynamic instability, so that they can persist in aqueous environments. 4~ PolyPs have remarkable ion exchange and ion transport properties. The phosphate group is an optimal coordination partner for metal cations in that it exhibits a flexible tetrahedral geometry with variable bond lengths and angles. 43 Unlike more rigid chelating agents, whose structures are specifically suited to the coordination requirements of particular cations, phosphates adjust their geometry to conform to the binding cations. PolyPs also exhibit a low-energy barrier to rotation about the P - O - P bond bridging the tetrahedra, so that polyP chains can twist into a number of different conformations (Figure 1B). Depending largely on the coordination preferences of the associated cations, bond angles vary from about 120 ~ to 180 ~ and the orientations of the linked tetrahedra range from eclipsed to staggered conformations. 31'43'44An important consequence of the unusual flexibility of polyPs is that unitary negative charges on adjacent tetrahedra can approach within 4/~, a distance suitable for binding divalent cations. 45 The stronger electro-

PHB and PolyP Ion Transport

53

static interaction of divalent cations leads to their preferential selection over monovalent cations. Binding energy rather than ion size or coordination geometry is the rationale for the ability of polyPs to sequester trace amounts of divalent cations from solutions containing monovalent salts. This natural capacity, which makes them effective agents in water softening, 31 is an important attribute for effecting selective permeation through ion channels. However, when divalent cations are not available, polyPs will efficiently bind monovalent cations. Ion exchange relationships in polyPs follow the same trend as the hydration energies and abide by the sequence: M +2 > Li + > Na + > K+.43 In summary, polyPs are flexible polyelectrolytes with the capacity to attract and bind diverse cations. They select among cations primarily by charge, and to a lesser extent by size, but not by coordination geometry. Their backbones form ladders of closely spaced, identical binding sites for transport of cations in response to concentration or voltage gradients. These remarkable polyfunctional molecules also have the capacity to grow or diminish in size by accepting or donating high-energy phosphate units. However, their high charge density creates a strong electrostatic barrier to penetrating a lipid bilayer and they do not discriminate well among cations of the same charge. In order to form physiologically effective ion channels, polyPs must associate with macromolecules that enable them to traverse the bilayer and enhance their ability to distinguish cations. The simplest and perhaps most primitive macromolecule to associate with polyP to form selective ion channels is poly (R)-3-hydroxybutyrate (PHB).

3. POLY-(R)-3-HYDROXYBUTYRATES 3.1. Occurrence and Synthesis Poly(R)-3-hydroxybutyrates (PHBs) are linear homopolymers of (R)-3-hydroxybutyric acid (Figure 2A). PHBs of high molecular weight (10s- 106Da) were discovered by Lemoigne in 19259 in cytoplasmic granules (or inclusion bodies) of Bacillus megaterium. Subsequently, granules composed of nearly pure (97%), high molecular weight polyester have been found in a wide variety of archaebacteria and eubacteria.l~ The osmotically neutral PHB granules are accumulated by these prokaryotic organisms as storage material and can account for up to 80% of the cell dry weight. More recently, low molecular weight PHBs (<13,000 Da) were found in a wide variety of prokaryotic and eukaryotic cells. Low molecular weight PHBs were first detected in membranes of granule-forming bacteria, 22'23 and later in bacteria that do not produce PHB granules, in fungi, and in diverse plant and animal tissues. 13'1721,46 This low molecular weight form of the polyester is always found complexed to other macromolecules; hence it is sometimes referred to as complexed PHB or cPHB. The majority of cPHB is complexed to proteins, cPHB-conjugated proteins are found in membranes, organelles, and cytoplasms. Apparently, cPHBs "dissolve" in lipid bilayers and also in hydrophobic pockets of proteins.

54

ROSETTA N. REUSCH

/• CH3 ~SCoA+ CH3/I~SCoA O

CoA-SH o

II

o

I)

CH3/C~cH(C~SCoA

OH

I

O

II

CH,/Ir'CH~"C'SCoA H CoA-SH~ CH3

,c,2

0

---['-~l CH(" J n |

Figure2. A. Chain of (R)-3-hydroxybutyrate residues (PHB). B. Common pathway for synthesis of PHBs in bacteria. 11 High molecular weight PHB granules are synthesized in bacteria from acetylCoA when their growth is limited by the absence of nutrients other than carbon (N, O, p)10-12 (Figure 2B). Under these conditions, acetyl-CoA is diverted from synthesis of proteins and nucleotides to synthesis of PHB using a three-step pathway: condensation of acetyl-CoA to acetoacetyl-CoA, reduction by NADPH to form (R)-3-hydroxybutyryl-CoA, and polymerization with release of CoA to yield PHB. II Frequently, the inclusion granules contain not only PHB but also its homologues, collectively known as polyhydroxyalkanoates (PHAs). 15'16PHAs have assumed considerable commercial importance as biodegradable plastics because of

PHB and PolyP Ion Transport

55

their polypropylene-like material properties. 15'47 The enzymes responsible for synthesis and degradation of PHA granules have been identified and cloned, 48-5~but these enzymes do not regulate the metabolism of cPHB. Synthases for cPHB have not yet been identified, but it has recently been found that some RNases hydrolyze proteinconjugated cPHB (unpublished result, Bryant and Reusch). Whereas PHB granules a r e found only in certain prokaryotes, low molecular weight cPHB, like polyP, is a universal cell constituent. It is cPHB this is the subject of this report.

3.2. Amphiphilic and Solvent Properties In the solid state, PHBs form left-handed, 21 helices with a 5.96/~ pitch 51-55 that fold in lamellar crystallites of--50 ]k thickness; 54'56'57 in solution, they retain a flexible helical form. 58'59 PHBs are amphiphilic molecules; the repeating unit contains both hydrophobic methyl groups and hydrophilic ester carbonyl oxygens. The shape of the molecule can be described by four backbone dihedral angles (assuming constant bond angles). Computer modeling studies (Insight II, Biosyms) suggest that the amphiphilic PHB molecule may assume a large number of exolipophilic endopolarophilic forms with ester groups in the preferred antiperiplanar orientation, allowing it to act as an intermediary between polyP and the bilayer 25'27 (Figure 3). The conformation of the molecule is highly sensitive to small changes in one or more of the less constrained backbone dihedrals. A change of only 1~ in

Figure 3. Some exolipophilic-endopolarophilic helical conformations of PHB. Computer-generated models (Insight II, Biosyms) of the amphiphilic PHB molecule in helical conformations with the methyl groups facing outward and the ester groups facing inward. All three conformers are 140 residues in length and have the ester groups in the preferred antiperiplanar arrangement (see Table 1). Carbon-black: hydrogen-blue; oxygen-red. A. Ester groups are aligned at each turn of the helix; B. ester groups alternate; C. ester groups are helically skewed. 27

56

ROSETTA N. REUSCH

a single dihedral angle can effect a change in helical rise of as much as 2.5 /~k27 (Table 1). The argument is not that PHB itself retains such conformations, rather that the flexibility of the molecule enables it to wrap around other macromolecules and conform to their contours. An equally important property of PHB is its capacity to "dissolve" salts. Indeed, PHB may be unique among biological polymers in having the structural features, shared by a small group of synthetic polymers, that are necessary for forming ion-conducting salt complexes known as polymer electrolytes. 6~ We can sum up the important characteristics of this class of polymers as: (1) flexible backbones with low barriers to bond rotation to ease segmental motions of the polymer chain; (2) heteroatoms that have sufficient electron donor power to form coordinate bonds with cations; and (3) a suitable distance between heteroatoms to permit the formation of multiple intrapolymer coordinate bonds to cations. The stability of these polymer salt complexes is enhanced, as compared to complexes formed by small molecules, by the entropic advantage known as the "polymer effect", attributable to the cooperative effect of binding to neighboring ligands attached to a common backbone. The most well-known member of this class is the polyether, polyethylene oxide, whose complexes with lithium perchlorate have been used commercially in lithium batteries. 6~ The good solvating power of polyethylene oxide is attributed to an optimal spacing of the electron-donating ether oxygens along a flexible backbone that allows multiple contacts between the polymer backbone and cations. When this distance is decreased, as in polymethylene oxide, chain flexibility is greatly reduced; when it is increased, as in 1,3-polypropylene oxide, the distance between

Table 1. Backbone Dihedral Angles and Helix Parameters for Some Exolipophilic-Endopolarophilic Conformations of PHBa'b

ii

iH

PHB140 Structures Residues/Turn

14 - 176 85.8

12.5 -1.0

-155.5

C l -C2-C3 -01 Rise A

- 104.1 4.5

2.4

16.4

16.1

Notes:

B -174

01-C1-C2-C3

Internal Diam (]k)

i'

Conformational Variants of A (change in dihedral angle)

A

C2-C3-Ol-Cl C 3 - O l - C l - C2

i'

0

1.0

91.7

C 9 176 73.9

-1.0

1.0

-153.7 -142.5

7.0

5.5

3.0

6.7

2.7

5.0

5.0

16.8

15.9

17.0

16.8

16.1

14.8

11.0

-1.0

1.0 -112.7 -142.5

~Ref.27.

bBondangles are the same in all conformers: O 2 - C 3 - O

1-

125~ O 2 - C 3 - C

2-

125~ C2-C3-O i = 110~

PHB and PolyP Ion Transport

57

oxygens is too great to allow the polymer to assume the low-energy conformations that maximize polymer-cation coordinations. 6~ In both cases, the ability to solvate cations is lost or greatly diminished. Substituents pendant to the backbone, as in 1,2-polypropylene oxide (1,2 PPO), do not abolish the capacity of the polymer to solvate cations but may attenuate it by increasing steric hindrance. The carbonyl oxygens of esters are weaker electron donors than ether oxygens; nevertheless, polyesters with suitable spacing between ester groups also form ion-conducting salt complexes. Watanabe et al. 63 have shown that poly-13-propiolactone, also called poly-3-hydroxypropionic acid, forms ion conducting salts with lithium perchlorate. Poly-13-propiolactone bears the same relationship to PHB as does polyethylene oxide to 1,3-polypropylene oxide (Figure 4), thus it should not be surprising that PHB forms ion-conducting complexes with lithium perchlorate, 64 albeit the conductivity of these complexes is greatly improved by mixing PHB with other PHAs to reduce crystallinity. It is also important to consider the types of salts that form polymer electrolyte complexes. As an aprotic polymer, PHB does not have hydrogen-bond donating groups needed to solvate anions. Accordingly, it "dissolves" salts of large anions that have a diffused charge requiring little solvation. 6~ The capacity of PHBs to solvate cations is also limited. Ester carbonyl oxygens are weak Lewis bases of low polarity and can form coordinate bonds only with hard cations that have large solvation energies (Table 2). Salts of polyanions are preferred due to the cooperative effect. In summary, one would expect that PHBs are best able to dissolve salts composed of large diffuse polyanions and hard cations, and partition them into hydrophobic regions of bilayers and proteins.

4c_c_,c,_o

c_c_o

o

PEO

PPL

C

1 c-c-c-o

O

C

.

4,c_ c

PHB

1,2-PPO

Figure 4. Backbone structures of salt-solvating polymers. The figure shows the similarity of backbone structure, with optimal spacing between electron-donating oxygens, of polymers that form ion-conducting salt complexes. PPL-poly-3-propiolactone; PEO: polyethylene oxide; PPO: 1,2- polypropylene oxide. 9

v

-

9

1 8

58

ROSETTA N. REUSCH Table 2. Ionic Radii, Hydration Numbers, Free Energies of Solvation, Surface Charge Densities, and Coordination Geometry of Alkali and Alkaline Earth Cations of Interest

Cation

Ionic Radius* (A )

Hydration No. a

-AG ~ (kcab'mole)

Coordination Geometry

Na §

0.98

6

98.5

octahedral

K§

1.33

6

80.5

octahedral

Mg 2+

0.72

6

454

regular octahedral

Ca 2+

1.06

7,8

379

irregular cubic

Sr2+

1.27

7,8

340

irregular cubic

Ba 2+

1.43

7,8

314

irregular cubic

Note: "Ref.98.

Polymer electrolytes conduct cations by segmental motions of the polymer backbone that carry the cations from one complexation site to the next. 6~ Since this requires significant fluidity, the polymers are conductive only in the amorphous state, i.e. above the crystalline to gel transition temperature. For the bulk polymer of PHB, this temperature is in the range of 0 ~ to 10 ~ Accordingly, single molecules of PHB dissolved in fluid lipid bilayers should be capable of considerable segmental motions at physiological temperatures. There has been substantial experimental verification of the capacity of long-chain polymers (>50 units), short-chain oligomers, and cyclic oligomers of (R)-3-hydroxybutyrate to solvate salts. Seebach et al. 65'66 prepared crown ester complexes from cyclic trimers (triolides) of (R)-3-hydroxybutyrate and sodium thiocyanate; Biirger and Seebach 67 demonstrated that cyclic oligolides and oligomers of (R)-3hydroxybutyrate (OHBs) transport alkali and alkaline earth picrates across methylene chloride layers in U-tubes; Reusch and Reusch 64 prepared conducting complexes from PHB and its homologue, poly-(R)-3-hydroxyvalerate, with lithium perchlorate; and Fritz et al. 68 showed that OHBs can transport Ca 2§ into liposomes in the presence of a Ca 2§gradient. Finally, as discussed below, PHBs make bilayers permeant to ions.

3.3. Channels in Planar Lipid Bilayers The ability of PHBs and OHBs to conduct salts across phospholipids bilayers was examined in a planar bilayer voltage-clamp setup. 69'7~In this system, a bilayer is formed between two aqueous solutions by "painting" a decane solution of phospholipids across a small aperture (ca. 0.2 mm in diameter) in a partition separating two chambers containing aqueous salt solutions. The hydrocarbon drains away and the phospholipids spontaneously arrange themselves into a bilayer (black lipid membrane or BLM). The aqueous solution on one side (cis side) of the bilayer

PHB and PolyP Ion Transport

59

Figure 5. Planar bilayer setup. The system consists of two aqueous solutions, labeled cis and trans, separated by a planar bilayer. External voltage commands are applied to the cis side, with the trans side maintained at ~round (defined as zero voltage).

OSC--oscilloscope; VCR--recording tape system.

represents the cell cytoplasm and the solution on the other side (trans) represents the aqueous environment outside the cell. The trans side is maintained as ground and external voltage steps are applied to the cis side (Figure 5). Seebach et al. 71'72 prepared monodisperse oligomers (MwlM n = 1.0) of (R)-3-hydroxybutyrate using an exponential fragment-coupling strategy, and incorporated them individually into planar lipid bilayers of synthetic 1,palmitoyl-2,oleoyl-phosphatidyl choline (16:0,18:1 PC) between symmetric solutions of 60 mM RbC1, 5 mM MgC12, 10 mM Hepes CsOH, pH 7.2. 73 Channel-forming activity was observed for OHBs of 16, 32, 64, and 96 monomer units when concentrations of these oligomers were 0.1-1.0% of bilayer lipids (w/w) (Figure 6) and a voltage _>60 mV was applied to the cis side. The stepwise fluctuations were also observed under these conditions when natural short-chain PHBs (-60 units), obtained by acid-catalyzed degradation of PHB granules, were incorporated into bilayers of the same composition at concentrations of 5% (w/w). Stereoregularity was not essential; channel activity was observed with low molecular weight polymers containing 39% isotactic diads (Mw ~ 4 x 103, M n -- 6 x 102, Mw/M n = 6.7) at concentrations of 5% (w/w). In all cases, the channel conductances were highly variable. It was possible to obtain linear current-voltage relationships for a given single channel in the bilayer, but measurements of single channels formed with the same sample in a fresh bilayer yielded quite different conductances. Many of the single channels exhibited high open probabilities, indicating they were essentially open pores. As expected, they did not discriminate well among cations. The current records were similar to those observed with ionic and nonionic detergents, such as Triton X- 100, octyl glucoside, or sodium dodecyl sulfate, in planar bilayers. 74'75

60

ROSETTA N. REUSCH

I

t

~

H

o

80 mV OH pO =0.51 96

200~------] 3pAl /

N.620

o

|.l

17

26

3s [psi

Figure 6. Single-channel currents of synthetic PHB96. Left:. Representative current fluctuations obtained when the given voltage was applied at a planar bilayer made from 1-palmitoyl, 2-oleoyl, phosphatidylcholine (16:0,18:1, PC) containing 0.1 to 1% of 96met of PHB between symmetric solutions of 60 mM RbCI, 5 mM MgCI2, 10 mM Hepes CsOH, pH 7.2. The solid horizontal bar in each record indicates the current level with the channel closed, pO is the probability that the channel is in the open state. Right:. Corresponding conductivi~ histograms. N indicates the total number of observations that have been analyzed. '~ In addition to high concentration, oligomer size and end group structure proved to be important factors in channel formation. High molecular weight PHBs from natural sources (>3000 units) and very small synthetic oligomers of <8 units did not form channels, even at concentrations of 5% (w/w). Oligomers of 16, 32, 64 and 96 units in which the end groups had been derivatized (COOH group by tert-butyl and OH group by PhCH2) displayed no single-channel activity at similar high concentrations. The arrangement of the oligomers in the bilayer is uncertain, but if one assumes that PHB in the bilayer preserves the 21 helicity, 5.96/~ pitch, determined from X-ray diffraction of crystalline PHB, 51 - 55 then the length of molecules containing 16 monomer units corresponds well to the width of the 16:0,18:1 PC bilayer, estimated as 48/~ from electrical capacitance measurements. 76 This bilayer width also corresponds to the average 50 ,/k thickness of lamellar crystallites formed by synthetic PHBs >16 units. 72'77 It is presumed that each PHB molecule crosses the hydrophobic region and is stabilized at each end by the formation of hydrogen bonds from the terminal hydroxyl and carboxyl groups to the ester groups of the phospholipids (Figure 7). These data imply that chains >16 units fold back on themselves. Considering the high concentrations of the polyesters required to form channels, it seems probable that the pores are formed by aggregates of several molecules of appropriate size. Seebach et al. 71 propose that the oligomers form islands of lamellar crystallites in the bilayer, and that ion permeability results from areas of mismatch at the interfacial regions between phospholipids and the oligomers. When the oligomers are of mixed lengths, higher concentrations are required, suggesting that there is selection for oligomers of approprirte size.

PHB and Polyp Ion Transport

61

Figure 7. Schematic representation of (R)-3-hydroxY7butyrateoligomers (OHBs) of 32 units incorporated into planar phospholipid bilayer. 3

Fritz et al.68 examined the channel-forming activity of synthetic oligomers of 8, 16, 32, and 64 units in bilayers of liposomes, which encapsulated the fluorescent dye, Quin-2. Ca 2§ transport into the vesicles was followed by observing the Ca2§ Quin-2 absorption at 264 nm. In the presence of a concentration gradient, the 32mers and 64mers, but not the 8mers or 16mers, transported Ca 2§ into vesicles at a rate comparable to the ionophore, calcimycin. 78 The OHB-mediated Ca 2§ transport was inhibited by the presence of La 3§ A carrier-like mechanism was suggested, with Ca 2§ ions hopping between complexing carbonyl groups. Transport of other cations was not examined; hence specificity for Ca 2§ was not established. More insights into the relationship between polyester size and bilayer width were provided by Das et al., 79 who examined the channel-forming ability of synthetic oligomers of (R)-3-hydroxybutyrate of M n 1670, Mw/Mn 1.2, and isotacticity 94%. Since these oligomers have an average of 19 residues by M n measurements and 23 residues by M w measurements, they are referred to as OHBI9t23. The oligomers were prepared synthetically by Jedlinsky et al. 8~ via regioselective ring-opening polymerization of (S)-13-butyrolactone, catalyzed by a supramolecular complex of a crown ether and sodium (R)-3-hydroxybutyrate. This process is less laborious than the exponential fragment-coupling process above. Although polymers or oligomers formed by this method do not have the advantage of identical length, the size range for a given preparation is relatively narrow and the end groups (OH and COOH groups) are the same as those present in natural PHB, as shown by NMR and ESI-MS spectroscopy, sl OHBI9~ did not form channels in bilayers of 16:0 18:1 PC and cholesterol (5:1 w/w) at concentrations up to 2.5%. 79 This failure was attributed to a mismatch between oligomer length and bilayer width. Solid-state measurements based on the above 2 l helix conformation indicate that the length of oligomers with 19 units and 23 units are ca. 57/~ and 69/~, respectively. Geometric considerations indicate that at least six oligomers must assemble to form a pore large enough to conduct ions. 6a For OHB 19t23, the probability is high that most of these oligomers will be too short to fold properly and too long to remain fully extended in a bilayer of 48/~. To test this hypothesis, the complexes were incorporated into bilayers of 1,2 dieicosenoyl phosphatidyl choline (di20:l PC), cholesterol (5:1 w/w), and separately into bilayers of 1,2 dierucoyl

62

ROSE-IrA N. REUSCH AO

16000

15 O t-.

w :3 r

O

LI.

OJ

6

"26o ~o"s6o Time, ms

0

86o

......

| w, I I

8

12

1e

32000

15 O

Q.

r-

g

tO

4

Amplitude, pA

BO

<:

8000

L.

t

16000-1

i1

0J

O""

0| ............

"200" " "4()0" " "600" " "8()0

4

Time, ms

C.

8

12

16

Amplitude, pA

8 oO~176176176176 ~176176176176 ~176176

~176176176176 .~176176176176176176 2

...'"'~.'"'""

c-

-1

0

O

Potential, mV

Figure 8. Profiles of single channel currents of biomimetic OHB19/23 in planar lipid bilayers. A. OHB19/23 in bilayers of 1,2-dieicosenoyl phosphatidylcholine (di20 PC). B. OHB19/23 in bilayers of 1,2-dierucoyl phosphatidylcholine (di22 PC). OHB19/23was incorporated into planar lipid bilayers, composed of synthetic phosphatidylcholines and cholesterol (5:1 w/w), between symmetric bathing solutions of 200 mM CaCI2, 5 mM MgCI2, 10 mM Tris-Hepes, pH 7.4 at 22 ~ Data was filtered at 1 kHz. All points amplitude histograms are shown to the right of each trace. C. Nonselectivity of OHB19/23. Single-channel current-voltage relationships of OHB19/23 incorporated in planar bilayer membranes composed of synthetic di22:1 PC/cholesterol (5:1 w/w), between asymmetric bathing solutions of 65 mM CaCI2, 10 mM NaCI, 5 mM MgCI2, 10 mM Tris-Hepes, pH 7.4 (cis side) and 200 mM NaCI, 0.1 mM CaCI2, 5 mM MgCI2, 10 mM Tris-Hepes, pH 7.4 (trans side) at 22 ~ The Nernst equilibrium potentials (calculated from concentrations) were Eca =-82 mV, EO = +9 mY, and ENa = +76 mV. Error bars were smaller than the symbol sizes in most cases. The dotted lines indicate 95% confidence limits.

PHB and PolyP Ion Transport

63

phosphatidyl choline (di22:1 PC), cholesterol (5:1 w/w). Since each additional methylene group adds about 1.27/~ to the bilayer,82 the widths of di20:l PC and di22:1 PC bilayers are estimated as 53 and 58/~, respectively. At concentrations >1% ofphospholipids (w/w), OHB 19t23displayed activity in both bilayers; however, channel formation was observed much more frequently in di22:1 PC. Representative current records of the dominant channels observed in each bilayer are shown in Figure 8A. The channels in both cases were essentially open pores of fluctuating conductance, similar to those formed by the synthetic oligomers above (Figure 6). As shown by the all points amplitude histograms 83 (Figure 8B), channels in di22:1 PC had a lower average conductance as well as a much narrower amplitude distribution than in di20:1 PC (61 + 4 vs. 109 + 13 pS), suggesting that more organized channels were formed in the wider bilayer. Over long periods of recording (>10 min), full closures were brief (order of ms) and extremely rare (<1 per min), indicating a very high open probability (>0.99). The channels formed by OHBI9/23 displayed no cation selectivity. This was tested by determining the reversal potential (zero-current potential) when the channels were incorporated into bilayers formed between aqueous solutions of unequal ion composition; Ca2§ dominant on one side and Na§ dominant on the opposite side. The Nernst equilibrium potentials (calculated from concentrations) 84 are Eca= -82 mV, Ecl= +9 mV, and ENa = +76 mV so that preference for Ca2§ would be demonstrated by a negative reversal potential and a preference for Na§ as a positive reversal potential. As shown in Figure 8C, the reversal potential, estimated graphically from the single-channel current-voltage relationships, was zero indicating no preference for divalent or monovalent cations. In summary, OHBs form nonselective ion channels in planar lipid bilayers, provided that the ratio of OHBs to lipid is high and there is a rough correspondence between oligomer length and bilayer width. The data suggest that OHBs longer than bilayer width can bend back in hairpin turns to fit within the confines of the hydrophobic region, but the folded segments should each be of approximately bilayer length. It is postulated that short-chain OHBs aggregate in clusters, with molecules arranged perpendicular to the bilayer and parallel to fatty acyl chains. The results demonstrate the ability of oligomers and polymers of (R)-3-hydroxybutyrate to make bilayers permeable to ions; however, the OHB "channels" cannot select among cations by charge, size, or coordination geometry. Like polyP, PHB has some of the essential properties of physiological ion channels but not others. Auspiciously, the two polymers complement each other, and together they possess all the requisite characteristics of well-regulated ion-selective channels.

4. COMPLEXES OF PHB A N D POLYP 4.1. PHB/polyPComplexes in Bacterial Membranes PHB/polyP complexes were fn-st discovered in the plasma membranes of bacteria by Reusch and Sadoff22-25'85during spectrofluorometric studies of membrane structure, using the hydrophobic probe, N-phenyl- 1-naphthylamine (NPN). When NPN is added to cell suspensions, it partitions into the hydrocarbon region of the cell membranes and

64

ROSETTA N. REUSCH

responds to changes in the viscosity or polarity of the bilayer by a change in fluorescence intensity. This procedure, developed by Overath and Traiible, s6 results in minimal disturbance of cell processes, and reports lipid transitions specifically and with reasonable agreement with transitions determined by light-scattering and X-ray diffraction. The major transition in membranes of log-phase cells is the broad gel to liquid-crystalline transition of the phospholipids that begins below 0 ~ and ends at 20-24 ~ While observing the thermotropic transitions in bacterial membranes at stages of cell

>Ior) z Lu I-z ,..,,,

uJ (/) z uJ

(b

O0 ILl nO :::) .J IJ. UJ ~>

b -

a

i

I..J IJJ n-

0

I

10

1

20

I

30

1

40

|

50

i

60

T E M P E R A T U R E (~

Figure 9. A. Thermotropic fluorescence spectra of E. coil DH1 cells using the hydrophobic probe, N-phenyl-l-naphthylamine (NPN). (a) Mid-log phase cells; (b) stationary lphase cells; (c) cells made genetically transformable by the method of Hanahan. 46 NPN was added to 4 mL of cell culture to a final concentration of 1 ~M and the thermotropic fluorescence spectra were recorded. 24 Measurements were made at increasing temperature (ca. 2 ~ per min). Excitation: 360 nm; emission: 410 nm. Measurements were made at increasing temperature (ca. 2 ~ per rain). B. Effects of physical treatments on the thermotropic transitions in genetically competent E. coil DH1. (a) Thermotropic transitions at descending temperature; (b) cells pelleted at low speed and suspended in supernatant; (c) as in b but suspended in equal volume of distilled water; (d) as in (b) but suspended in 10 mM phosphate buffer, pH 7.4. Excitation: 360 nm; emission: 410 nm. Fluorescent probe was NPN. Measurement (a) was made at decreasing temperature and (b), (c), (d) at increasing temperatures (ca. 2 ~ per min).

PHB and PolyP Ion Transport

65

l,m (1t z w tzI lal v) z u tn lal 0

w

.I w

L

I0

|

20

I

30

TEMPERATURE

l

l

40

50

_

1

60

('C)

Figure 9. Continued. differentiation, a new relatively sharp transition at --56 ~ was noted. The fluorescence peak at --56 ~ was weak or imperceptible in cells during mid-log phase growth, became more evident in stationary-phase cells, and was most intense in genetically transformable cells, i.e. cells that showed high competence to take up exogenous DNA 23-25 (Figure 9A). The transition was irreversible (Figure 9Ba). When the cells were collected and resuspended in the original growth medium, in water, or in buffer, the sharp transition diminished in intensity or disappeared and was replaced with broad fluorescence at lower temperatures (Figure 9Bb,c,d). This suggested that the structure responsible for the transition is labile and composed of lipid-soluble and aqueous-soluble components. The lipidic component was identified as PHB by chemical analysis of the isolated plasma membranes. 22-25 The MW was estimated as --13,000 Da (-150 units) by viscosity measurements 22 and --12,000 Da (-140 units) by nonaqueous size-exclusion chromatography. 19 The water-soluble component was identified as polyP by chemical analysis, 22-2s and its size was estimated as 55-65 residues (MW ca. 5000 Da) by acrylamide gel electrophoresis. 87 The neutralizing cations for polyP were determined to be Ca 2§ by graphite furnace atomic absorption spectrometry. 25 It was suggested that the sharp rise in fluorescence is caused by the dissociation of PHB/Ca(polyP) complexes. As the polymers separate, the polar polyP molecules move to the aqueous interface and the PHB chains are released into the bilayer, effecting an increase in

66

ROSETTA N. REUSCH

bilayer viscosity and a consequent increase in NPN fluorescence. Since the polymers cannot reassociate when temperatures are lowered, the viscosity of the bilayer (and fluorescence intensity) increases at decreasing temperatures until the liquidcrystalline to gel transitions of the phospholipids begin. Freeze-fracture electron microscopy studies of the membranes of E. coli and A. vinelandii by Reusch et al. 24 provide evidence of structural changes that support the fluorescence data (Figure 10). Freeze-fracture micrographs of log-phase cells show a typical mosaic of particles and pits on both concave and convex surfaces of the plasma membranes. However, as complexed PHB was increasingly incorporated into the membranes, as determined by analysis of the purified membranes and evidenced by the intensity of the thermotropic transition at - 56 ~ the micrographs revealed the formation of small semi-regular plaques in the plasma membranes (arrows) that possess shallow particles. The plaques grew in size and frequency as the concentration of membrane PHB and intensity of the PHB/polyP transition increased.

4.2. E. coli PHB/polyP Complexes as Ion Channels The ability of E. coli PHB/polyP complexes to form calcium-selective channels in planar bilayers was investigated in the planar bilayer system described above (Figure 5). E. coli DH5~ cells were made genetically competent to increase the concentration of PHB/polyP in the membranes. Then vesicles were prepared from the cell envelopes, and added to the cis side of a planar bilayer formed by synthetic 16:0, 18:1 PC between symmetric bathing solutions of 250 mM CaCI 2, 5 mM MgC12, 10 mM Tris Hepes, pH 7.3. The complexes were allowed to insert spontaneously into the bilayer. 27 No activity was observed in the absence of an applied

Figure 10. Representative freeze-fracture electron micrograph of competent E. coli DH1. The micrograph shows the typical appearance of small semi-regular plaques (arrows) in the plasma membranes of E. coli DH1 cells after treatment to make them genetically transformable by the method of Hanahan. 146 These cells have sharp thermotropic transitions at -56 ~ when examined as in Figure 9A. 24

PHB and Polyp Ion Transport CHCI3

80 mV

PM

120 m V

PM

_0_.

~

$0 mV ..II_.

~ . j ' ~ - ~ ]~:

Tr-I I I* 'T I'q _+ ~

-_--,:_ -

--

. . . . . . . . . . .

67

~

p

~

.

l

~

r

-,

,~. . . . . . . . .

r

..-

|~

Figure 11. Profiles of calcium channels from E. coiL Single-channel currents observed in membrane vesicles and extracts of E. coli competent cells when incorporated into planar bilayers of 16:0,18:1, PC between symmetric aqueous bathing solutions of 250 mM CaCI2, 5 mM MgCI2 in 10 mM Iris Hepes, pH 7.3. 27 Line 1. Membrane vesicles of whole cells (WC) were added to the cis bathing solution and the solution was gently stirred to allow spontaneous incorporation of the vesicles into the bilayer. Single-channel currents were activated by a voltage step >60 mV. A representative current record at 100 mV is shown. Lines 2-3. Plasma membrane vesicles (PM) were incorporated into the bilayer and single-channel currents were activated as above. Representative records at 80 mV and 120 mV are shown. Line 4. A chloroform extract of genetically competent cells was added to a decane solution of 16:0, 18:1, PC. The chloroform was removed by evaporation, and the remaining lipid mixture was used to form the bilayer. Single-channel currents were activated as above. A representative current record at 80 mV is shown.

voltage, but when a potential of>60 mV was held for several minutes, single-channel currents were observed signifying the presence of Ca2+-permeant channels (Figure 11, line 1). The envelopes were then separated on sucrose gradients into plasma and outer membranes. Each was tested in a similar manner, but channel activity was observed only with the plasma membrane vesicles (Figure 11, lines 2,3). This is consistent with earlier findings, by chemical analysis and freeze fracture electron microscopy, that PHB/polyP complexes are confined to the plasma membranes. 23,24 The calcium channel activity observed in plasma membrane vesicles of competent E. coli was essentially the same as that for PHB/polyP complexes extracted from plasma membranes into chloroform solution (Figure 1 l, line 4). The molecu-

68

ROSETTA N. REUSCH

lar weight of the extracted PHB/polyP complexes was estimated as 17,000 + 4000 Da by size-exclusion chromatography. 27 This measurement, together with the molecular weights of the component polymers, indicates that the complexes are formed from one strand of PHB (~ 12,000 Da) and one strand of Ca(polyP) (~5000 Da). To establish the composition of the channels still further, the PHB/Ca(polyP) complexes were reconstituted from PHB recovered from E. coli and Ca(polyP) prepared from commercial sodium polyphosphate and calcium chloride. This solution was then premixed with phospholipid and used to form a bilayer. Finally, the complexes were formed in situ. PHB was mixed with phospholipids before painting the bilayer; Ca(polyP) was then added to the aqueous bathing solutions, and a potential was applied to induce the formation of the complexes within the bilayer. For all preparations, the concentration of PHB in the bilayer was restricted to one-hundredth or less of the amount required to form PHB channels. Singlechannel currents similar to those described above were obtained with each of these procedures. 27

4.3. Synthetic Ion Channels from PHB128 and PolyPs To resolve remaining doubts as to whether the channel activity was indeed effected by PHB/polyP complexes and not by trace protein contaminants, Das et al. 2s performed a total synthesis of the channel complex from (R)-3-hydroxybutanoic acid, sodium polyphosphate, and calcium chloride. Lengweiler et al. 26 prepared a 128mer of (R)-3-hydroxybutyrate (PHB 12s)by an exponential fragmentcoupling strategy. This synthetic polymer has a MW of 11.4 kDa, which is close to that of E. coli PHB (~ 12 kDa). Calcium polyphosphate was prepared from sodium polyphosphate glass (av residue number 65) and calcium chloride. PHB128/polyP complexes were formed by adding a chloroform solution of PHBI28 to an excess of dry Ca(polyP) (av 65 units). After evaporation of the chloroform, the dry mixture of polymers was heated briefly in a microwave oven, and then suspended in chloroform and gently mixed in a bath ultrasonicator. PolyPs are highly insoluble in chloroform; hence, only polyP complexed with PHB12s is found in solution. Uncomplexed PHBI28 will also be present, but the concentration of the complexes in the bilayer was maintained at concentrations too low (< 0.01% of phospholipid w/w) for the formation of PHBI2 s channels. The size of complexed polyP was determined by acrylamide gel electrophoresis to be in the same range (50-70 residues) as that in the E. coli complexes. 27 The current records of channels formed in planar bilayers by the synthetic complexes were indistinguishable from those of PHB/polyP complexes extracted from E. coli (Figure 12), and the conductances of the synthetic and E. coli channels were equivalent; 101 + 6 pS and 104 + 12 pS (Figure 13), respectively. These results show clearly that protein is not essential to the observed channel activity.

PHB and PolyP Ion Transport

69

Figure 12. Representative single-channel current fluctuations of synthetic and E. coli PHB/POlyP complexes at various clamping potentials. Left: Synthetic PHB128/polyP; Right: Channels extracted from competent cells of E. coil DH5a. Complexes were incorporated into planar lipid bilayers composed of 16:0, 18:1, PC and cholesterol (5:1; w/w) between aqueous bathing solutions of 200 mM CaCI2, 5 mM MgCl2, 10 mM Tris Hepes, pH 7.4 at 22 ~ The bars at the side of each profile indicate the fully closed state of the channel. Clamping potentials with respect to ground are indicated at the left side.28

70

ROSETTA N. REUSCH

g

i

....

, .......

-120 -80

A~ 6

AO 0

0 . _0 . . . . . .

40

g 0_5

40

80

120

-10 -15

Holding Potential, mV Figure 13. Current-voltage relations for synthetic PHB128/polyP (E3) and E. coli PHB/polyP channel complexes (A). The conductance of the channel for Ca 2+ in symmetric solutions, under the experimental conditions described in Figure 12, is 101 + 6 pS for the synthetic channels and 104 + 12 pS for the E. coli channels. The data points represent mean values of 10 observations. 28

4.4. Synthetic Ion Channels from OHB19/23 and PolyPs Ion channels were also prepared from synthetic oligomers and polyPs. Das et al. 79 formed complexes from synthetic OHB19/23 with Ca(polyP) (av 65 units) following the procedure used to prepare the synthetic complexes of PHBl28 and Ca(polyP) above. The channel activity of the complexes was examined in bilayers of di22:1 PC and cholesterol (5:1 w/w) between symmetric bathing solutions of 200 mM CaC12, 5 mM MgC12, 10 mM Tris-Hepes, and pH 7.4 at 22 ~ The current records of OHB19/2a/polyE shown in Figure 14A, display high conductance and complex channel activity. All points amplitude histograms for single channel records at different clamping potentials indicate a major fully open state with a conductance of 260 _+8 pS and a minor open state with conductance of 153 _ 3 pS (Figure 14B). These conductances were substantially higher than the -61 pS conductance of OHB 19/23in the same bilayer, or the ~ 100 pS conductance reported for the major open state of the biological or synthetic PHB/polyP complexes. The channel structure is unknown but it is proposed that polyP stretches across the bilayer, encircled and solvated by an indeterminate number of OHBs. The polyP polyanion, with its high negative charge and conformational polymorphism, would reasonably be responsive to voltage change. The space between the two polymers, lined with ester carbonyl oxygens on one side and phosphoryl oxygens on the other, can accommodate multiple conductive pathways for cations. The presence of multiple lanes for ion transport may explain

PHB and PolyP Ion Transport

71

Figure 14. Characteristics of PHBIgI23/POlyP complexes in di22:1 PC/cholesterol bilayers. A. Profiles of single channel currents. PHB19123/polyP complexes were incorporated into planar lipid bilayers, composed of synthetic di22:1 PC/cholesterol (5:1 w/w), between symmetric bathing solutions of 200 mM CaCI2, 5 mM MgCI2, I 0 mM Tris-Hepes, pH 7.4 at 22 ~ Data was filtered at I kHz. B. Current-voltage relations for PHB19123/polyPchannel complexes. All points amplitude histograms were constructed for single channel records at each indicated potential. 81 Data was filtered at 2 kHz. The points show the mean peak position of Gaussian distributions, fit by a simplex least-square procedure at respective clamping potentials. The data shown here are mean amplitudes of the major open state from several experiments. Best fit obtained by linear regression yields a single-channel conductance of 260 + 8 pS. Each symbol represents an independent experiment.

72

ROSETTAN. REUSCH 260pS

BO

30.

o,,~

20.

<

10. / o

,= e.-

0

.1 6

r

_ "01 1

So

,9' o

1o

Potential, mV Figure 14. Continued.

the high conductance of OHB 19/23/polyP channels as compared to the singlelane OHB19/23 channels. However, OHBl9/23/polyP channels also show higher conductance than the multi-lane channels formed by E. coli PHB/polyP and PHB 128/polyP complexes. This suggests that channels formed by polyP with the oligomers are more disordered than those formed by polyP with single long molecules of PHB.

4.5. Selectivity of PHB/PolyP Ion Channels Protein Ca 2§ channels are highly permeant to Ca 2§ Sr2§ and Ba 2§ In the presence of these divalent cations, they select strongly against monovalent cations, such as K § and Na+S4; however, the channels become highly permeable to these monovalent cations when all divalent cations are removed. These characteristics are shared by the Ca 2§ channels formed in bilayers by E. coli or synthetic PHB12s/polyP complexes. The channels are selective for divalent over monovalent cations, permeant to Ca 2§ Sr 2§ and Ba 2§ and become permeable to monovalent cations when divalent cations are absent. 27'28 The preference of PHB/polyP for divalent cations was established by determining the reversal potential when the channels were incorporated into bilayers formed between aqueous solutions of unequal ion composition. As noted above (3.3), the reversal potential was estimated graphically from the single-channel current-voltage relationships, and this value was compared with the equilibrium potential for each cation in a perfectly selective channel, as calculated from the Nernst equation. 84

PHB and PolyP Ion Transport

73

A

0.8

<~ 0.6 c

-

0.4

o

i 0.2 ~ o

I

-60

"

I'"

-40

9 'I

A ~ v.`"

-20

I

0

9

I

"

20

I

40

9

I

60

"

i

80

"

I00

Voltoge (mV) B

.

4

3 2.

w=,==

i .)

0 r./3

-120

-80 1 - 4 0

.11"

" s'0"' 1 o

-2

Potential, mV Figure 15. A. Selectivity of E. coli PHB/POlyP complexes for Sr2+ over Na+. Singlechannel current-voltage relations of PHB/polyP complexes, extracted from E. coil and purified by HPLC size-exclusion chromatography, and incorporated into bilayers of 16:0, 18:1, PC between unequal solutions; cis 200 mM SrCI2, 10 mM Tris Hepes, pH 7.3, and trans (ground) 288 mM NaCI, 8 mM SrCI2, 10 mM Tris Hepes, pH 7.3. 27 The equilibrium potentials, calculated from concentrations, were Eca= -40 mV, ECl --- +7 mY, and ENais nominally + infinity. B. Selectivity of synthetic PHB12~polyP complexes for Ca2§ over Na§ Single-channel current-voltage relations of PHB128/polyP complexes, incorporated in planar lipid bilayers composed of 16:0, 18:1, PC and cholesterol (5:1; w/w) between unequal solutions; cis 65 mM CaCI2, 10 mM NaCI, 5 mM MgCI2, 10 mM Tris Hepes, pH 7.4, and trans (ground) 200 mM NaCI, 0.1 mM CaCi2, 5 mM MgCI2, 10 mM Tris Hepes, pH 7.4.28

74

ROSETTA N. REUSCH 8

C

6

o~176 ~176

4

<=

..-'"ii.-" .... "

2

r

"~~

"

40

80

120

:t

Potential, mV

Figure 15. Continued. The equilibrium potentials (calculated from concentration) were ECa = - 8 2 mV, ECl = +9 mV, and ENa is +76 inV. Error bars indicate standard deviation of the mean (n = 3). C. Selectivity of synthetic PHB19/23/polyP complexes for Ca 2+ over Na +. Single-channel current-voltage relationships of PHB19/23/polyP complexes incorporated in planar bilayer membranes composed of synthetic di22:1 PC/cholesterol (5:1 w/w), between asymmetric bathing solutions of 65 mM CaCI2, 10 mM NaCI, 5 mM MgCI2, 10 mM Tris-Hepes, pH 7.4 (cis side) and 200 mM NaCI, 0.1 mM CaCI2, 5 mM MgCI2, 10 mM Tris-Hepes, pH 7.4 (trans side) at 22 ~ The straight line indicates the best fit obtained by linear regression, yielding a permeability ratio (Ca2+: Na+) of-4:1. The dotted lines indicate 95% confidence limits. Selectivity for divalent over monovalent cations by E. coli PHB/polyP complexes was demonstrated in bilayers of 16:0, 18:1 PC between unequal solutions of Sr2§ and Na § at pH 7.3 (Figure 15A). The reversal potential was -38 mV, close to the equilibrium potential for Sr2§ o f - 4 0 mV, whereas the Nernst equilibrium potential for Na § is nominally plus infinity and that for C1- is +7 mV. Preferential selection of divalent cations by synthetic PHB 12s/polyP complexes was demonstrated in 16:0, 18:1 PC bilayers and cholesterol (5:1; w/w) between unequal solutions of Ca 2§ and Na § at pH 7.4 (Figure 15B). The reversal potential was -67 mV, close to the Nernst equilibrium potential for Ca 2§ o f - 8 2 mV, whereas the equilibrium potentials for Na § and CI- are +76 mV and +9 mV, respectively. Synthetic OHBl9a3/polyP complexes demonstrated weak but explicit selectivity for divalent over monovalent ions 79 when incorporated in planar bilayers composed of synthetic di22:1 PC, cholesterol (5:1 w/w) between unequal solutions of Ca 2§ and Na § at pH 7.4 (Figure 15C). The reversal potential was -20 mV; the Nernst equilibrium potentials were the same as for the synthetic complexes above. The data indicate selectivity for Ca 2§ over Na § of about 4:1, a significant improvement over channels formed by the oligomers alone (see Figure 8C), but still much poorer discrimination than the >90:1 selectivity demonstrated by the natural and synthetic

75

PHB and Polyp Ion Transport

A a~

B

60 m

200.~

80

m 2 pA

c3

o

40

20

5 b i 6 6 i 6" [La3+], ~ M Figure 16. A. Block of E. coil PHB/POlyP channels by transition metal cation, La3§ Representative single-channel current steps at 120 mV for PHB/polyP complexes, extracted from E. coil, purified by size-exclusion chromatography, and incorporated into bilayers of 16:0, 18:1, PC between symmetric solutions of 100 mM CaCI2,2 mM MgCI2,10 mM Tris Hepes, pH 7.3, and with LaCI3 added to the cis side as stated.27 (a) No LaCI3; (b) 0.20 mM LaCI3; (c) 0.55 mM LaCI3. B. Block of synthetic PHB128/POlyP channels by transition metal cation, La3+. The bilayer composed of 16:0, 18:1, PC and cholesterol (5:1; w/w) was formed between aqueous bathing solutions of 200 mM CaCI2, 5 mM MgCI2, 10 mM Tris Hepes, pH 7.4. After incorporation of the channel, activities were recorded for 5 min at a clamping potential of-80 mV. Then, LaCI3 was added to the trans compartment to achieve the indicated concentrations. The bath was stirred and activities were recorded after 1 min of addition of La+3. Data points represent mean values of the amplitude histograms; error bars show the standard deviation from the mean. 28

76

ROSETTA N. REUSCH

PHB/polyP complexes. Though superior to uncomplexed OHB19r23 in many respects, OHB19rJpolyP channels apparently do not have the degree of structural organization required for stringent discrimination of cations. This demands binding cavities with precise ligand geometries, more easily structured from single long PHB molecules than from multiple short OHBs.

4.6. Block of PHB/PolyP Ion Channels Another characteristic of protein calcium channels, shared by PHB/polyP channel complexes, is blockage by low concentrations (10 to 20 mM) of transition metal cations. 84 Lanthanum ion is a particularly potent calcium channel blocker. The sequence of blocking effectiveness is in general La 3§ > Co 2§ > Cd 2§ > Mg 2§ The E. coli-derived PHB/polyP complexes 27 were completely blocked by 0.6 mM La 3§ (Figure 16A), 1.5 mM Co 2§ or 8 mM Cd 2§ Nearly complete block of singlechannel currents was observed for the PHB 12a/polyP complexes 28 at concentrations >0.1 mM La 3§ (Figure 16B). Mg 2§ also had an affect on the currents. At low concentrations (2 mM), Mg 2§ appeared to increase the stability of E. coli PHB/polyP channel complexes and acted as a mild blocker, making it easier to observe channel openings and closures. 27 For this reason, low concentrations of Mg 2§ are customarily included in the buffers. Raising Mg 2§ concentrations to higher levels significantly diminished single-current magnitudes, and nearly complete block of Ca 2§ currents was observed in the E. coli channels when Mg 2§ concentrations reached 20 mM.

4.7. Gating of PHB/PolyP Ion Channels PHB/polyP complexes, isolated from E. coli or synthetic PHB128/polyP complexes, exhibit two distinct gating modes (modes 1 and 2) in planar lipid bilayers. 28'88In Figure 12, mode 1 is shown at positive potentials and mode 2 at negative potentials, but the channels gate in both modes at positive and at negative potentials. In mode 1, the channels display long openings of the order of several seconds with infrequent and brief closures to the fully closed state; in mode 2, the channel activity is characterized by long bursts with flickering between the fully open state of 87 + 3 pS and a major subconductance state of 56 + 2 pS, interrupted occasionally by closures to the fully closed state (<0.5% of total time) (Figure 17A,B). The channel exhibits a number of other subconductance states, but with much lower frequency. In both modes, the channel occasionally enters a long closed state of several seconds duration. The two gating modes are kinetically connected as switching between modes is observed, albeit infrequently. The switching can take place in either directionwfrom mode 1 to mode 2 or the reverse (Figure 17C). The factors that determine preference for a specific mode or affect switching between modes are presently unknown, but the complex gating kinetics of this channel may be attributable to the multitude of conductive conformations that can be adopted by these multi-lane, multi-binding site, amorphous complexes in

PHB and PolyP Ion Transport

77

Figure 17. Gating modes of PHB/polyP channels from E. coli. Recordings show ion channel activity of the extracted complexes in two kinetically distinct modes of gating: (A) mode 1 and (B) mode 2. 88 Complexes were extracted from competent E. coli DH50t and incorporated into planar lipid bilayers composed of 16:0,18:1 PC and cholesterol (5:1 w/w) between symmetric bathing solutions of 200 mM CaCI2, 5 mM MgCI2, 10 mM Tris-Hepes, pH 7.4 at 22 ~ Positions of fully closed state, fully open state and the major subconductance state of mode 2 are indicated as C, O, and S, respectively, on the left side of the tracings. Note that the vertical bar indicating current scale is 15 pA for A and 10 pA for B and C. All point amplitude histograms for selected traces in each mode are shown at the side of respective traces with corresponding mean conductances. In C, transitions from mode 2 to mode 1 and back are indicated by arrows. Clamping potential with respect to ground (trans) was -120 mV for A and + 100 mV for B and C. Data shown was filtered at 1.5 kHz for A and B and at 1 kHz for C.

response to potential differences. It is possible to envision many mechanisms by which current amplitudes or gating profiles could be modulated. For example, movement of a large slowly permeant cation through one lane could distort local PHB and/or polyP conformations and affect Ca 2+ flow in neighboring lanes.

78

ROSETTA N. REUSCH

,o

AO

I

0"6t

m o

It.

0.4 0.2.

-Ito" -~o" -~o" o " 4'0 " 8'o " ~ o

Potential, mV

Bo

0.50.4. 0.3. O

0.2 9 9 -~6-~o

0.1 ~o

o"" "4'o" " s'o" " "~o

Potential, mV Figure 18. Voltage dependence of open probability of E. coli PHB/polyP and OHB19/2.v'polyP complexes. A. The data shows the probability of opening of E. coli PHB/polyP channels to the fully open state from the subconductance state at indicated clamping potentials. 88 The linear voltage dependence indicates an increasing probability of detecting the channel in the fully open state (o) at more positive clamping potentials. The experimental conditions were the same as in Figure 17. In all cases, the total time of recordings analyzed exceeded 150 s. The time spent by the channel in the fully closed state was lessthan 0.5%. The data was filtered at 2 kHz and acquired at 20 kHz. Fit by nonlinear regression (Boltzman) is shown by the solid line. B. The data show the open probability of OHB19/23/polyP complexes, incorporated into planar lipid bilayers composed of 1,2-erucoyl phosphatidylcholine (di22:1 PC) and cholesterol (5:1 w/w) between symmetric bathing solutions of 200 mM CaCI2, 5 mM MgCI2, 10 mM Tris-Hepes, pH 7.4 at 22 ~ at the indicated clamping potentials. 81 Data was filtered at 1 kHz.

PHB and PolyP Ion Transport

79

Nonpermeant cations may block one or more lanes while permeant ions continue to move through remaining open lanes. In this regard, concurrent block of adjoining lanes may have a different effect than block of lanes at opposite sides of the channel.

4.8. Voltage-Dependence of PHB/PolyP Ion Channels PHB/polyP complexes, like protein calcium channels, display voltage dependence. 84 Several kinetic properties of E. coli PHB/polyP channels were voltage-dependent, 88 including the residence time of the channels in fully open and major subconductance states, and the probability of full openings from the subconductance state (Figure 18A). 88 A given single channel displayed one of two voltagedependent responses that were mirror images of each other. This indicates an asymmetry in the channel structure and suggests that a given single-channel may adopt one of two opposite orientations in the bilayer. Since polyP molecules are identical at both ends, it is surmised that the asymmetry is due to the different end residues of the PHB molecule. PHB has a carboxyl group (ionized at physiological pH) at the head and a hydroxyl group at the tail, creating a structural asymmetry that permits individual molecules to incorporate into the bilayer in head-to tail or tail-to head directions. The importance of end residue structure to channel activity was demonstrated by the inability of the synthetic oligomers to form channels when the hydroxyl or carboxyl end groups were derivatized. 73 Synthetic tris(macrocycle) cation channels have also shown remarkable sensitivity to structural changes at the distal ends. 89 In the cell, PHB synthesis takes place in the cytoplasm, thus it is assumed that all PHB/polyP channels in the bacterial plasma membrane are oriented in the same direction and display the same pattern of voltage dependence. The open probability of OHBl9a3/polyP channels also showed voltage dependence. 81 Most frequently, open probabilities were higher at more positive potentials (Figure 18B). About 10% of single channels showed the opposite pattern, i.e. higher open probabilities at more negative potentials. Completely symmetrical relationships were never observed. This indicates that OHB19r23/polyP channels, like E. coli PHB/polyP channels, are asymmetric and the structural asymmetry is likely established by the different end groups of the oligomers that allows them to assume one of two opposite orientations in the bilayer. End groups likely play a more important role in oligomer channels since they are more numerous. The entropically favored arrangements of OHBs in these channels may be those in which oligomer alignment is random.

5. STRUCTUREAND MECHANISM OF ION TRANSPORT All peptide and protein ion channels, as well as synthetic ion channels, are amphiphilic structures with an outer coat of nonpolar residues and a lining of polar and charged residues. 89-97 These attributes are provided in a cooperative fashion by

80

ROSETTA N. REUSCH

the two structurally dissimilar homopolymers, polyP and PHB. The strong selectivity of PHB/polyP complexes for Ca 2+ indicates that the geometry of the binding cavities formed by the coordinating groups of the two polymers must be optimal for distinguishing Ca 2+, i.e. seven or eight oxygen ligands arranged in irregular geometry. 98'99 Such a binding cavity would discriminate well against the other major physiological divalent cation, Mg 2+, which prefers six ligands arranged in regular octahedral geometry. 1~176 Ca 2+ is unique among physiological cations in its ability to tolerate variant bond angles and bond lengths, making it the cation of choice for cross-linking disparate macromolecules. The molecular structure of PHB/CapolyP channel complexes remains uncertain; however, some assumptions can be made concerning the general organization of the complexes from the physical properties and sizes of the polymers, and the low

C I

:c----ac~i2 ."

J / ~,,cH / /~=o j"

.%,.o ~o ~ /o

Figure 19. Ca(polyP) solvated by PHB. A. Drawing depicting the solvation of Ca(polyP) by PHB in the bilayer formin~ a Ca2+-selective channel. 13 B. Drawing of the cross section of PHB/polyP channel. 3 C. Putative coordination geometry of Ca2+ in PHB/POlyP. Ca2+ forms ionic bonds with four phosphoryl oxygens of polyP and ion-dipole bonds with four ester carbonyl oxygens of PHB to form a neutral complex with distorted cubic geometry.25

PHB and Polyp Ion Transport

81

dielectric environment they inhabit. It is clear that the highly polar polyanionic polyP must be shielded from the hydrophobic region of the bilayer by the amphiphilic PHB. As solvating polymers, PHBs "dissolve" polyP salts by encircling them and replacing the water of hydration about the cations with coordinate bonds to its ester carbonyl oxygens (Figure 19A). The calcium ions, held by ionic bonds to the core polyP helix, form weak ion-dipole bonds to ester carbonyl oxygens. This has the effect of crosslinking the turns of the PHB helix. 25 In the idealized model, each Ca 2+ is held in an ionophoretic cage solvated by eight oxygen atoms in distorted cubic geometrymfour phosphoryl oxygens from two adjoining units of polyP (total charge of-2), and four ester carbonyl oxygens (two neighboring PHB units from one turn and two contiguous units from the turn directly above or below) (Figure 19B,C). The interchangeable nature of cations allows PHB/polyP complexes in the bilayer to act as dynamic molecular sieves, selecting Ca 2+ from other physiological cations by charge, coordination geometry, and size. The two models proposed for the channel complexes by Reusch et al. 2s'27'2s and Seebach et al. 66'72 have this general structure. The former model assumes that PHB/polyP is basically a polymer electrolyte complex and proposes that PHB has a coiled conformation such as it displays in solution, 58'59 while the latter suggests that PHB maintains the lamellar helix form of its solid state. 73 The result of both these arrangements is a pliant structure of two discrete polymers bridged together by lanes of cations 25'27'28 (Figure 20A,B). Normal molecular motions of the polymers such as twisting or stretching movements, or sliding or rotation of polyP

Figure 20. Diagrammatic representation of the PHB/polyP channel structure. The central cylinder represents the polyP helix, which contains pairs of closely spaced monovalent negative charges that provide a ladder of binding sites for Ca2+. A. The Ca(polyP) is surrounded and solvated by an exolipophilic-endopolarophilic helix of PHB.ZUB. The Ca(polyP) is surrounded and solvated by a 13-sheet-like structure of PHB similar to the lamellar structure of solid PHB.28

82

ROSETTA N. REUSCH

within the homogeneous environment provided by PHB may alter channel geometry and effect changes in current amplitude and gating. One would expect that some channel conformations are more stable than others and consequently more probable. The major subconductance state of mode 2 (Figure 17) may represent a particularly stable conformation for PHB/polyP in 16:0, 18:1 PC bilayers. It may not be possible to resolve the structure of the complexes with certainty. In the Reusch model, the complexes have the liquidic properties of polymer electrolytes and this suggests a family of conformations rather than a single defined structure. In the Seebach model, several PHB molecules are involved in surrounding polyP. The individual PHB chains are free to adopt various positions in the phospholipid lattice; hence, a well-defined structure is again unlikely. Nonetheless, further studies may help to decide between the two views of the arrangement of PHB molecules in the complexes and delineate the more probable conformations. The mechanism of ion conduction by PHB/polyP channel complexes can be rationalized in terms of the structures and properties of the component polymers. One view of how the channel may operate in the cell membrane or planar bilayer follows. A helix of Ca(polyP), surrounded and solvated by PHB, forms a salt bridge extending from the cytoplasm to the periplasm. Multiple conductive passageways are formed between the two polymers; the outer wall is lined with solvating oxygens, the inner wall is girdled by monovalent, phosphoryl anions. At the external interface, cations are drawn to the mouth of the channel by polyanionic polyP, and divalent cations are preferentially bound. Ca 2+ is preferentially selected by virtue of its well-suited coordination geometry and the relatively rapid rate at which it undergo replacement of water of hydration. Sr2+ and Ba 2§ are not physiological cations, but since they have the same coordination geometry as Ca 2§ they are also permeant in an artificial system 27'2s (Figure 19D). Since cation-binding sites on polyP are identical and spaced at frequent intervals, there is no net potential energy cost to cation movement within the channel. Segmental motions of the PHB backbone and librational movements of ester carbonyl oxygens carry Ca 2+ from site to site in parallel single-file lanes until internal concentrations rise to an appropriate level or the membrane is again polarized. Transition metal cations, particularly trivalent cations like La 3+, bind tightly to polyP at the interface but have difficulty entering because of their unsuitable coordination requirements, and consequently they block the ion flow. The physiological role of PHB/polyP complexes has not been established, but bacterial cells like eukaryotic cells maintain low internal Ca2§ 1~176 and there is increasing evidence pointing to calcium involvement in a number of important cellular functions, such as chemotaxis, cell division, heat shock, pathogenicity, and differentiation. 1~176 Systems for calcium export have been identified. 1~176 Mechanisms for calcium entry are less well known. An L-type channel was reported in Bacillus subtilis, 11~ and Jones et al. and Holland et al. 1~ recently demonstrated the presence of a putative Ca 2§ influx channel in stationary phase E. coli,

PHB and Polyp Ion Transport

83

inhibited by La 3+. The properties of this channel were consistent with the behavior of PHB/polyP calcium channel complexes. They suggested that a second Ca 2+ channel, not inhibited by La 3+, may also be present. In summary, PHB/polyP complexes share many of the notable characteristics reported for proteinaceous Ca 2+ channels: i.e. selectivity for divalent over monovalent cations; permeance to Ca 2+, Sr 2+, and Ba2+; block by transition metal cations; voltage activation; and voltage dependence. Moreover, the structures postulated for the complexes bear a striking resemblance to the single-file, multiple-site Ca 2+ channel structures described by Hagiwara and Byerly, Hess and Tsien, McCleskey and Almers, and Tsien et al. 113-116

6. SUPRAMOLECULAR COMPLEXES OF PROTEIN, PHB, AND POLYP 6.1. PHB/PolyPComplexesin a CaATPasePump The structural features of PHB/polyP channels, coupled with the ability of polyPs to "grow" by accepting phosphoryl groups from ATP and other phosphate anhydrides, 41-45 form the basis for the following hypothetical mechanism for Ca 2§ extrusion. 25 As described above, the Ca(polyP) helix traversing the PHB pore creates a bridge of calcium ions connecting the pools of Ca 2§ in the cytoplasm and external medium. Although there is a considerable Ca 2§ gradient across the membrane, estimated for E. coli as 2 mM outside and 0.1-0.3 ktM inside, 1~176 the exceptionally strong ionic bond between Ca 2§ and polyP prevents an inward flow of Ca 2§ The extreme insolubility of calcium polyphosphates testifies to the strength of this ionic bond. At the same time, Ca 2§ is held to PHB by relatively weak ion-dipole bonds. This organization implies that the CapolyP chain can be secreted through the PHB sheath by extending the polyP chain at the cytoplasmic side of the membrane. As the appended phosphate units move into the PHB channel, Ca 2§ is sequestered from the cytoplasm. At the periplasmic face, polyP may be discharged into the periplasm or degraded by exopolyphosphatases. 117-12~The ultimate effect of these reactions is to expend ATP energy to draw Ca 2§ from the cytoplasm and extrude it into the periplasm. This, in effect, is a CaATPase pump. To test the above hypothesis, Reusch et al. 3~ analyzed the human erythrocyte Ca2+-ATPase pump for polyP and PHB. This protein is the first plasma membrane enzyme demonstrated to work as a Ca 2§ pump, and it is the sole transporter of Ca 2§ in the cell. 121'122 Moreover, the Ca2+-ATPase is ubiquitous in plasma membranes of a variety of cell types in eukaryotic organisms. 123 The Ca2+-ATPase was purified to homogeneity by calmodulin affinity chromatography 124-125and separated by SDS-PAGE. The presence of polyP in the 135 kDa protein was signaled by its reaction to the cationic dye, o-toluidine blue. PolyPs of >5 units cause a shift in the absorption maximum of o-toluidine blue towards shorter wavelengths, i.e. from a maximum at 507 nm (blue) to 530 nm (violet-red). 126 The

84

ROSETTA N. REUSCH

absorption shift is specific for inorganic polyphosphates, inasmuch as RNA and DNA are stained blue. The identity of polyP in the Ca2§ was confirmed, and its quantity determined, by an enzymatic assay in which E. coli polyphosphate kinase is used to convert polyP and laC-ADP to laC-ATp.127 The presence of PHB in the Ca2§ was demonstrated by a positive reaction to anti-PHB IgG on a Western blot (Figure 21A). This comigration of PHB with the Ca2§ indicates a strong association between the polyester and the protein, because PHB is uncharged and does not itself migrate on gels. Furthermore, the PHB was not removed by extraction with chloroform, suggesting that the association was covalent. The PHB in the purified protein was measured by a chemical assay in which the polyester is converted via ~elimination to its unique degradation product, crotonic acid, by heating in concentrated sulfuric acid. 5~ Under the conditions of the assay, no amino acids or homopolymers of amino acids produce significant amounts of crotonic acid. 5~Though only small quantities of the polymers were found (0.5 ktg polyP/mg protein and 1.3 l,tg PHB/mg protein), the amounts are likely to be greatly understated since both PHB and polyP are subject to extensive chemical and enzymatic hydrolysis during the lengthy isolation and purification procedures. The results have importance in that they place these two polymers, with proven capacities to facilitate ion transport, in a human calciumtransporting structure. According to the hypothesis, the polyP chain must be extended at the cytoplasmic side to "pump" Ca 2§ out of the cell. In bacteria, polyP chains can be extended by

Figure 21. A. Reaction of Ca2+-ATPase protein to anti-PHB IgG. Purified Ca2§ protein (8.5 I~g per lane) was separated by electrophoresis on a 10% polyacrylamide gel. 30 Left lane 1: Coomassie blue stain of Ca2+-ATPase protein. Left lane 2: Western blot of Ca2+-ATPase protein probed with rabbit anti-PHB IgG. Second antibody was anti-rabbit IgG conjugated to alkaline phosphatase. Color development was with the alkaline phosphatase substrate kit (Bio-Rad). B. Phosphorylation of the Ca2+-ATPase by [32p]polyP. Purified Ca2+-ATPase protein (2 ~g) was phosphorylated at room temperature by [32p]polyP and separated by electrophoresis on a 10% polyacrylamide gel. 30 Right lane 1: Coomassie blue stain of phosphorylated Ca2+-ATPase. Right lane 2: Autoradiogram of phosphorylated Ca2+-ATPase.

PHB and PolyP Ion Transport

85

polyphosphate kinases, which transfer high energy phosphates between ATP and polyP. 4~ Polyphosphate kinases have not been detected in eukaryotes; however, some of the activities established for polyphosphate kinase (PPK) by Ahn and Kornberg 129 were also demonstrated for the human erythrocyte Ca2§ by Niggli et al. 13~Both enzymes can be phosphorylated by [y-32p]ATP, producing a phosphoenzyme that can transfer the phosphoryl group back to ADP and regenerate ATP.129-131 A histidine residue is phosphorylated in PPK, and an aspartate residue in the CaATPase. [y-32p]ATP + PPK ~-) ADP + PPK.(32p-PO~) [y-32p]ATP + CaATPase ~ ADP + CaATPase.(32p-PO3) Inasmuch as ATP and polyP have similar phosphorylating potentials, the ability of [32p](polyP) to phosphorylate the CaATPase was examined by autoradiography of an SDS-PAGE gel (Figure 21A). [32p](polyP) + CaATPase ~ CaATPase,(32p-PO3) Since aspartyl phosphates have higher phosphorylating potentials than histidine phosphates, it was postulated that the [32p]polyP-phosphorylated Ca2§ could transfer the phosphate to a polyP chain as well as back to ADP. The capacity of the [32p]polyP phosphorylated Ca2§ to carry out these reactions was confirmed (Figure 22A,B), thus demonstrating that the CaATPase has all the enzymatic activities associated with polyphosphate kinases. 129 CaATPaseo(32p-PO3) + ADP ~ [32p]ATP CaATPaseo(32p-PO3) + (polyP) a ~-~ [32p](polyP)n+l The manner in which the PHB and polyP interact with the protein is not known and it remains uncertain whether they play an active role in the export of Ca2+; however, the presence of these highly conserved polymers and enzymatic activities in the erythrocyte Ca2§ is consistent with the molecular mechanism for Ca2§ export suggested above, i.e. the protein may be viewed as a device that transports Ca 2§ out of the cytoplasm by catalyzing the transfer ofphosphoryl groups from ATP to polyP.

6.2. PHB and Polyp in a Potassium Channel The presence of PHB and polyP in a human calcium pump indicates that the role of these polymers in ion transport has been conserved, and further suggests that the polymers are likely constituents of other ion channels and pumps. Recently, the potassium channel of the Gram-positive soil bacterium, Streptomyces lividans (KcsA), attracted great interest as the first ion channel to have its structure analyzed

86

ROSETTA N. REUSCH

Fi~.ure 22. A. Formation of [32p]ATP from [32p]polyP and ADP by Ca2+-ATPase. The

[3 P]polyP, with average chain length >400 residues, was incubated with I mM ADP, 100 mM KCI, 4 mM MgCI2, 0.1 mM CaCI2 in 10 mM Hepes-KOH (pH 7.4) and Ca2+-ATPase for 30 min at 37 ~ ADP and ADP (15 nmol each) were added as carriers, and ADP, ATP, and polyP were resolved by thin layer chromatography on PEI cellulose, visualized by UV, and autoradiography. Brackets indicate regions of the chromatogram occupied by long chain polyP, ATP, and ADP standards after development. 3~ Right lane: Reaction in the presence of Ca2+-ATPase. Over 80% of the radioactive phosphate in the product comigrated with ATP. The remainder, which appears to be short-chain [32p]polyP resulting from incomplete reaction, formed a band between the origin and the ATP band. Left lane: Control. Result in the absence of Ca2+-ATPase. B. Elongation of polyPs by [y-32P]ATP. PolyPs (I 0 l~g) was incubated with 2 ~g Ca2+-ATPase, [y-32p]ATP (30 Ci/mmol), 50 mM Hepes-KOH, pH 7.4, 130 mM KCI, 8 mM MgCI2, 0.15 mM n-dodecyl octaethylene glycol monoether, 20 I~M 16:0, 18:1, PC, I mM EGTA, I mM CaCI2 (30). Electrophoresis was performed on 20% acrylamide, 7 M urea minigels. PolyPs was selected for this assay because increases in chain length are more noticeable on gels for short-chain polyphosphates. Standards were polyPs, [32p]polyP, xylene cyanol and bromphenol blue. The position of the dye markers was calibrated against short-chain polyPs that had been prepared and measured as described by Clark and Wood. 148Visualization was with o-toluidine blue and autoradiography. Lane/'PolyPs and [y-32p]-ATP in the presence of Ca2+-ATPase and CaCI2 as above. Lane 2: Same as lane I except CaCI2 was absent and solution contained I 0 mM EGTA. Lane 3" Same as lane I except Ca2+-ATPase was absent.

PHB and PolyP Ion Transport

87

by X-ray crystallography. 132Although KcsA is a prokaryotic channel, segments of the amino acid sequence show similarity to the sequences of eukaryotic K + channels. 133'134 The correspondence is particularly strong for the residues of the highly conserved signature sequence in the pore region and is also significant for the residues of the inner helix. The existence of these sequences, critical to the ion conduction and gating processes, indicates that KcsA has essentially the same pore structure and mechanism of potassium selectivity as eukaryotic potassium channels. This conservation of structure and function made KcsA a good subject for testing the above hypothesis. KcsA is a homotetramer of 17.6 kDa subunits. It has been cloned, overexpressed, and purified to homogeneity, and its stability and electrophysiological behavior have been examined. 135-142 Its small size, high levels of expression, and stability in detergent solutions make it an excellent candidate for structural and biochemical studies. X-ray analysis of KcsA crystals with data to 3.2/~ resolution, revealed that residues 23 to 119 of four identical subunits create an inverted teepee structure with a 12/~ long narrow pore at the outer end 132 (Figure 23A). Each subunit has two transmembrane or-helices connected by a pore region of roughly 30 amino acids, a common structural motif in K+ channels. The outer helix faces the lipid bilayer and the inner helix faces the central pore. There is a ~ 10/~ cavity in the center of the channel, lined with hydrophobic amino acids. PolyP was detected in KcsA tetramers, but not in monomers, on SDS-PAGE gels by metachromatic reaction to o-toluidine blue stain. A band of free polyP was also visible, suggesting that polyP is released when tetramers dissociate. The identity of polyP was tested by observing the effects of treatment with exopolyphosphatase from S. cerevisiae. This enzyme removes orthophosphate units, in the range of 3 to 1000 units, processively from the end of polyP chains. 45 The free polyP was degraded, but tetramer-associated polyP was not affected, indicating it was inaccessible to the enzyme. PolyP in KcsA was estimated by enzymatic assay, using E. coli polyphosphate kinase to transfer phosphates from polyP to 14C-ADP yielding laC-ATp127 (Figure 23B), as -17 t.tg/mg KcsA or -15 polyP monomer units per tetramer. Loss of polyP during purification of KcsA was likely but was not estimated. PHB was detected in both tetramer and monomer species of KcsA by reaction to anti-PHB IgG on Western blots (Figure 23C), and estimated by chemical assay 5~ as 34 I.tg/mg protein or 28 monomer units PHB per KcsA tetramer. Presuming that the PHB is divided equally between the 4 subunits, this suggests there are -7 PHB units per KcsA subunit. PHB was not removed from the protein by repeated extraction with warm chloroform, suggesting the bonding between them is covalent. While performing electrotransfer of KcsA from gels to nitrocellulose membranes, it was noted that the tetramer transferred nearly completely out of the gel in buffers of pH 8.6, whereas transfer of the subunits required buffers of much higher pH (11.3). This implied a large difference in the isoelectric points (pls) of

88

ROSETTA N. REUSCH

Figure 23. A. Model of KcsA. Ribbon representation of the tetramer as an integral membrane protein. Four identical subunits create an inverted teepee. The inner helix of each subunit faces the central pore while the outer helix faces the phospholipid membrane. There is a relatively wide water-filled cavity (about 10 t~ wide) near the middle of the membrane that narrows into the selectivity filter at the extracellular end (from Doyle et al. 132) B. Formation of 14C-ATP from 14C-ADP and polyp from KcsA. Samples were incubated with 50 mM KHepes, pH 7.2, 40 mM (NH4)2SO4, 4 mM MgCI2, 0.5 mM [14C]ADP, 2000 units polyphosphate kinase, at 37 ~ for 45 rain. 127 ADP and ATP (5 mM each) were added as carriers and ADP and ATP were resolved by thin layer chromatography on PEI cellulose, and visualized by UV and autoradiography. 29 Brackets indicate regions of the chromatogram occupied by ADP and ATP standards after development with 1M LiCI, 1M HCOOH. Lane 1:0.1 l~g polyp (type 45). Lane 2: As in lane 4 but without polyphosphate kinase. Lane 3: Native KcsA. Lane 4: polyP isolated from KcsA. Arrow indicates origin. C. Presence of PHB in KcsA tetramers and monomers. The figure shows (A) 12% SDS-PAGE gel of partially dissociated KcsA visualized with Coomassie Brilliant Blue stain and (B) Western blot of a similar gel probed with anti-PHB IgG. Second antibody was conjugated to alkaline phosphatase. Samples were heated for 1 rain at 70 ~ before loading. 9 D. Analytical isoelectric focusing of KcsA. Focusing was carried out on 0.4 mm gels of 5% acrylamicte:bis (33.7:1), 20 mM n-dodecyl-13-D-maltoside, 5% pH 3-10 ampholytes formed on gel support film. The gel was stained with Coomassie Brilliant Blue R-250, crocein scarlet and destained with 40% methanol, 10% acetic acid. 29 Lane 1: pl standards. Lane 2: Unheated KcsA. Lane 3: KcsA heated for 5 min at 90 ~

PHB and PolyP Ion Transport

89

the subunits and the tetramer. The pI of KcsA subunits could not be determined experimentally because of their extreme insolubility, even at high concentrations of nonionic detergent, but the pI calculated by ExPAsy is 10.3.143-144 The pI of the tetramers was determined experimentally by analytical isoelectrofocusing as 6.57.5 (Figure 23D). The high pI of the subunits is consistent with the excess of positively charged residues (19+, 13-), mainly arginines, which are concentrated at the intracellular end of the protein. The more neutral pI of the tetramers is consistent with its inclusion of polyP. As a polyanion, polyP can attract and bind together the four subunits. At the same time, it converts the net charge of the channel from positive to negative, and provides an attractive entryway and staircase of negative charges for K § transport. By drawing the basic residues inward, it reduces the charge at the protein surface. The neutralizing effect of polyP also helps to explain the remarkable stability of the tetramer, demonstrated by its resistance to dissociation in nonionic detergent for over a month at room temperature and for at least an hour in the ionic detergent, sodium dodecyl sulfate. The tetramer also survives exposure to heat up to --60 ~ for at least 30 min. 136'137'139 Neither PHB nor polyP are discernible in the electron density maps of KcsA recently reported by Doyle et al.; 132 however, only 60% of the channel residues, comprising the most hydrophobic region (residues 23-119), were imaged. Much of the intracellular entryway (residues 126-160) had been excised and most of the remaining intracellular residues were disordered. The excision of 20% of the

Figure 23. Continued.

90

ROSETTAN. REUSCH

protein, including the most highly charged region (17 of 35 residues, 9+, 8-) may have eliminated some or all of the PHB and polyP. Even if still present, both PHB and polyP are polymorphic, hence they are difficult to image. It was postulated that KcsA protein creates an environment in which the natural selectivity of PHB/polyP for Ca 2§ is altered to favor selection for K § PolyP attracts all cations to the intracellular entryway. At physiological pH, polyPs are fully charged and preferentially bind divalent cations, but strategic placement of basic residues of the KcsA protein (primarily arginines) at the intracellular end may attenuate the negative charge density of polyP, and transform its cation preference from divalent to monovalent. Discrimination between K § and Na § is more difficult and requires the cooperation of all three polymers. PHB and polyP provide the oxygen ligands (or most of them) that form the binding cavities, but protein architecture defines the distance between the ligands of PHB and polyP, and between the PHB ligands on neighboring subunits. By attaching a PHB chain to an amino acid residue of each subunit, the carbonyl ester oxygens of neighboring PHB chains are kept too far apart by the protein configuration to effectively replace the hydration shell of Na § Transport of K§ then proceeds by stepwise movement up the polyP ladder. The locations of PHB and polyP in KcsA are presently unknown. Estimates of molecular size, based on van der Waal radii of component atoms, indicate that the width of a PHB chain is --4/~, and that of polyP --3/~. The Pauling radius of K § is 1.33/~. This implies that a cavity of-- 13.5/~ diameter would accommodate a PHB/ K(polyP) complex, though considerably less space may be required if the methyl groups of PHB are nestled within hydrophobic pockets along the protein wall. Allowing for experimental error, the - 10 ,/k cavity in the center of the channel, lined with hydrophobic amino acids, could accommodate a PHB/K(polyP) complex with the polyP chain trailing down to an arginine rich region of the intracellular entryway. It is also possible the polymers are located entirely in the intracellular portion that was not imaged. In this view, the S. lividans K § channel is a supramolecular structure in which selection and transport of K § is effected by cooperative interactions of protein, PHB and polyP. This hypothetical mechanism does not address and may not satisfy the wealth of experimental data acquired for this and other potassium channels. However, the presence in KcsA of these proven facilitators of ion selection and transport should be taken into account when resolving the manner in which the channel performs these tasks.

7. PHB/POLYP COMPLEXES AS DNA CHANNELS? PHB/polyP complexes were first discovered in genetically competent bacteria. 22 Although the complexes are present in the plasma membranes of log-phase cells of diverse organisms, their concentrations are low under optimal growth conditions.

PHB and PolyP Ion Transport

91 ......,

'"9

"'111.Sh

140 "1

IS

~20h

12Q

.-';9'I i lO0

.~ //,'/

o

-~

so

9

eO

i

/// ///

"'\,,,,

,,_ .......... .... ,,,,,//'

10'

| . . . . . T e m o e r ll tur e ~

Figure 24. A. Thermotropic fluorescence spectra in A. vinelandii UW1 cells, using the hydrophobic probe N-phenyl-l-naphthylamine (NPN), during log-phase vegetative growth and at various stages of genetic transformability. Log-phase growth was in Burk nitrogen-free buffer, pH 7.2, plus ammonium acetate (1.1 mg/mL) and glucose (1%). Cells were cultured at 30 ~ with moderate aeration. For development of transformability, cells were transferred to the same medium, except that Fe2+ was omitted from the buffer, and cultured as above.22'149At the indicated times, NPN was added to 4 mL of cell culture to a final concentration of 10-5 M and the thermotropic fluorescence spectra were recorded. Fluorescence intensities are relative. Excitation: 360 nm; emission: 410 nm. Measurements were made at increasing temperature (ca. 2 ~ per min). B. Relationship between the concentration of membrane PHB; and transformation efficiency. Relationship between the concentration of membrane PHB (IHT/ILT) (A) and transformation efficiency (e). Membrane PHB was estimated by the ratio of the fluorescence intensity of the PHB/polyP transition (--56 ~ (IHT) and the phospholipid gel to liquid-crystalline transition (below 24 ~ (ILT). Transformation efficiency was determined by the method of Page and yon Tigerstrom. 149 C. Evidence for a PHB/DNA complex. Chloroform-soluble donor ssDNA extracted from E. coil RR1 cells in early stages of DNA uptake. Cells were made genetically competent by the method of Hanahan, 146 and transformed with 32p-DNA prepared from the same organism. DNA uptake was interrupted after 5 rain. The cell pellet was washed sequentially with methanol, methanol:acetone (1:1 ), acetone and the dry residue was extracted with dry cold chloroform. Solvated DNA was recovered by washing the chloroform extract with 10 mM EDTA, pH 7.5. Electrophoresis was carried out on 15% acrylamide:bis (30:1 ). Lane 1,2: Untreated extract. Lane 3: Digestion with mung bean nuclease. Lane 4: Digestion with lambda exonuclease. Lane 5: Digestion with RNase 1A. 18

92

ROSETTAN. REUSCH

e

z

,,,~"~'~ "',,

X

I-

i-I

4

i,-4

..'2

A

8,000

z -,I m

~

oa

4,o0o

lP m r m

It,~

eo

g m

oS

11

9

i

I

I

I

14

1t

10

M

||

|4

I,I~IR8

Figure 24.

Continued.

Had it not been for the remarkable increase in their numbers when the cells are made genetically transformable and the sensitivity of the fluorescence assay with NPN, it is likely they would have remained undetected for many years to come (Figure 24A). It was the striking correlation between PHB/polyP concentrations and transformation efficiencies in A. vinelandii, B. subtilis, and E. coli that led Reusch et al. 22'23'25 to postulate that the complexes play a role in DNA transmembrane transport (Figure 24B). Here we will briefly consider the experimental and structural evidence in support of this hypothesis, and speculate on a mechanism by which PHB/polyP may be used for DNA transmembrane transport.

PHB and PolyP Ion Transport

93

For the purpose of this discussion, the processes leading to DNA transmembrane transfer can be divided into three stages: (1) development of competence for DNA uptake; (2) DNA binding; and (3) DNA uptake. The details of these procedures vary considerably for different organisms; nevertheless, the effect of the first stage in diverse organisms is a conspicuous increase in the concentration of PHB/polyP complexes in the plasma membranes. In organisms such as A. vinelandii that develop competence physiologically, i.e. by culturing them in minimal medium with a limiting nutrient, one observes a slow increase in the concentration of PHB/polyP membrane complexes (Figure 24A), that corresponds with genetic transformability (Figure 24B). 23 In organisms such as E. coli that develop competence by physicochemical treatment, there is a striking increasing in membrane concentrations of PHB/polyP when log-phase cells are suspended in the competence-inducing buffers 24 (see Figure 9A). Again, there is a strong correspondence between concentration of PHB/polyP membrane complexes and transformability, and when formation of the complexes is prevented by any means, transformation is inhibited. 24'145 The second step of DNA binding requires divalent cations, and only certain ones will do - Mg 2§ Ca 2§ Mn 2§ Sr2§ All can bind of DNA to polyP by cross-bridging phosphate residues from each polymer, and none are strong blockers of PHB/polyP channels. DNA uptake occurs only during the last step, when cells with elevated concentrations of PHB/polyP are returned to log-phase growth in normal growth media. The thermotropic fluorescence spectra at this time show a rapid decrease in the intensity of the 56 ~ fluorescence peak, indicating that the excess complexes are rapidly withdrawn from the membranes. As polyP is retrieved by cytoplasmic enzymes, it may draw the bound DNA molecule into and through the PHB channel. In support of this hypothesis, single-stranded donor DNA was found complexed to PHB, isolated from E. coli RR 1 recipient cells, when DNA uptake was interrupted (Figure 24C). From this viewpoint, the various procedures for competence development and DNA transformation are simply resourceful methods for changing the direction of movement of polyP within the PHB pore from outward (as in a pump) to inward, thus allowing the organic polyphosphate, DNA, to replace the inorganic polyphosphate, polyP, in the PHB pore where it can be drawn into the cell. Q

EVOLUTIONARY ASPECTS AND CONCLUDING REMARKS

Simplicity of structure and ease of formation makes it reasonable to infer that polyP and PHB were components of early cells, possibly preceding RNA. PolyPs are prebiotic molecules, formed by condensation of phosphates in volcanic condensates and thermal vents at the bottom of the ocean. 2'34'~46The synthesis of PHB is more demanding but requires only acetate and a reducing agent (see Figure 2B), both available in the primordial milieu. The conservation of these rudimentary homopoly-

94

ROSETTA N. REUSCH

mers throughout evolution suggests they have fundamental physiological roles. Perhaps the most important is the selective binding and transmembrane transfer of ions. The formation of PHB and its association with polyP may have occurred soon after elemental molecules of the primordial soup became separated from their environment by an enveloping lipid film. The creation of replicating molecules such as RNAs required a system for reducing internal Ca 2§ as Ca 2§ would tend to precipitate essential organic phosphate intermediates. PolyPs have a natural capacity for sequestering divalent cations, but stoichiometric amounts are needed and they also remove Mg 2§ a critical cation for many biosynthetic reactions. PHBs make the lipid layer permeable to ions, but ion flow is indiscriminate and regulated by concentration gradients that would tend to equilibrate inner and outer salt compositions. The solvation of polyPs by PHBs results in the formation of salt cells across the lipid layer, so that a few PHB/polyP molecules can serve as vehicles for export (or import) of multitudes of cations. But PHBs do more than position the polyPs across the lipid barrier--they greatly improve cation discrimination. The binding cavities formed by the two diverse polymers contain multiple coordinating ligands arranged in an irregular geometry that discriminate in favor of Ca 2§ and against Mg 2§ Simpler, more efficient, easier to assemble, Ca2§ channels than PHB/polyP are difficult to imagine. Every atom of both polymers contributes to the task of ion selection and transport. No deletions or substitutions are conceivable that would improve their effectiveness. PHB/polyP complexes are versatile structures and may be used not only to control intracellular Ca 2§ but also to produce streams of ions to supply energy for the biochemical reactions of the primordial cell, carry intracellular signals, and transfer informational molecules such as DNA. The recent discoveries of PHB and polyP in a human calcium pump and bacterial potassium channel suggest that the "naked" PHB/polyP complexes found in bacteria are progenitors of protein ion transporters. The process by which protein channels and pumps may have evolved from PHB/polyP complexes is unknown; however, one may surmise that over time proteins surrounded the complexes to support and regulate their activity. At first, the association may have been noncovalent, but subsequently PHB may have become tethered to the protein by a covalent bond. By this view, many of the channels and pumps of prokaryotes and eukaryotes may be supramolecular structures in which protein, polyP, and PHB join together for efficient regulation of transmembrane ion transport.

ACKNOWLEDGMENTS I thank my colleagues and collaborators Harold Sadoff, William Reusch, Linda Bramble, Ruiping Huang, April Gruhn, Sudipto Das, Yuri Shalin (Michigan State), Arthur Kornberg (Stanford), Dieter Seebach (ETH), Danuta Kosk-Kosicka (John Hopkins), Zbigniew Jedlinsky (Polish Academy of Sciences), for their invaluable contributions to these studies. I am

PHB and Polyp Ion Transport

95

also grateful to the National Science Foundation and National Institutes of Health for financial support.

REFERENCES

,

10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 0.

21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39.

Harold, E M. Bacteriol. Rev. 1967, 30, 772. Kulaev, I. S. In The Biochemistry of Inorganic Polyphosphates; Brookes, R. E, Ed.; John Wiley and Sons: New York, 1979. Kulaev, I. S.; Vagabov, V. M. Adv. in Microb. Phys. 1983, 24, 83. Wood, H. G.; Clark, J. E. Ann. Rev. Biochent 1988, 57, 235. Komberg, A. J. Bacteriol. 1995, 177, 491. Kumble, K. D.; Komberg, A. J. Biol. Chent 1995, 270, 5818. Komberg, A. Prog. Mol. Subcell. Biol. 1999, 23, 1. Schr6der, H. C.; Lorenz, B.; Kurz, L.; Muller, W. E. Prog. Mol. Subcell. Biol. 1999, 23, 45. Lemoigne, M. Ann. Inst. Pasteur 1927, 41, 148. Schlegel, H. G.; Gottschalk, G. Angew. Chem. 1962, 74, 342. Dawes, E. A.; Senior, P. J. Adv. Microb. Physiol. 1973, 10, 135. Holmes, P. A. In Developments in Crystalline Polymers-2; Bassett, D. C., Ed.; Elsevier: New York, 1988, 1. Reusch, R. N. Proc. Soc. Exp. Biol. Med. 1989, 191,377. Doi, Y. Microbial. Polyesters; VCH: New York, 1990. Anderson, A. J.; Dawes, E. A. Microbiol. Rev. 1990, 54, 450. Brand, H.; Gross, R. A.; Lenz, R. W.; Fuller, R. C. Adv. Biochent Eng. Biotech. 1990, 41, 77. Reusch, R. N.; Sparrow, A. W.; Gardiner, J. Biochint Biophys. Acta 1992, 1123, 33. Reusch, R. N. FEMS Microbiol. Rev. 1992, 103, 119. Seebach, D.; Brunner, A.; BUrger, H. M.; Schneider, J.; Reusch, R. N. Eur. J. Biochem. 1994, 224, 317. MUller, H. M.; Seebach, D. Angew. Chem. 1994, 32, 477. Reusch, R. N.; Gruhn, A. G. 1996 Internat. Syrup. on Bact. Polyhydroxyalkanoates; NRC Research Press: Ottawa, 1997, p. 10. Reusch, R. N.; Sadoff, H. L. J. Bacteriol. 1983, 156, 778. Reusch, R. N.; Hiske, T. W.; Sadoff, H. L. J. Bacteriol. 1986, 168, 553. Reusch, R.; Hiske, T.; Sadoff, H.; Harris, R.; Beveridge, T. Can. J. Microbiol. 1987, 33, 435. Reusch, R. N.; Sadoff, H. L. Proc. Natl. Acad. Sci. USA 1988, 85, 4176. Lengweiler, U. D.; Fritz, M. G.; Seebach, D. Helv. Chint Acta 1996, 79, 670. Reusch, R. N.; Huang, R.; Bramble, L. L. Biophys. J. 1995, 69, 754. Das, S.; Lengweiler, U. D.; Seebach, D.; Reusch, R. N. Proc. Natl. Acad. Sci. USA 1997, 94, 9075. Reusch, R. N. Biochemistry 1999, 47, 15666. Reusch, R. N.; Huang, R.; Kosk-Kosicka, D. FEBS Lett. 1997, 412, 592. Corbridge, D. E. C. Stud. Inorg. Chem. 1985, 6, 170. Yamagata, Y.; Watanabe, H.; Saitoh, M.; Namba, T. Nature 1991, 352, 516. Yoshida, A. J. Biochem. (Tokyo) 1955, 42, 163. Komberg, A.; Komberg, S.; Simms, E. Biochi~ Biophys. Acta 1956, 20, 215. Robinson, N. A.; Wood, H. G. J. Biol. Chent 1986, 261, 4481. Tinsley, C. R.; Manjula, B. N.; Gotschlich, E. C. Infect. lmmun. 1993, 61, 3703. Tzeng, C. M.; Komberg, A. Mol. Microbiol. 1998, 29, 381. Kuroda, A.; Komberg, A. Proc. Natl. Acad. Sci. USA 1997, 94, 439. Kulaev, I. S.; Vagabov, V. M. Biokhimiya 1971, 36, 587 (in Russian).

96

ROSETTA N. REUSCH

40. Lippmann, E In Phosphorus Metabolism; McElroy,W. D.; Glass, H. B., Eds.; John Hopkins Press: Baltimore, MD 1951, p. 521. 41. Davis, B. D. Arch. Biochent Biophys. 1958, 78, 497. 42. Westheimer, E H. Science 1987, 235, 1173. 43. Matheja, J.; Degens, E. T. In Structural Molecular Biology of Phosphates; Gustav Fischer Verlag: Stuttgart, 1971, p. 78. 44. Majling, J.; Hanic, E Topics Phos. Chent 1980, 10, 341. 45. Eisenman, G.; Horn, R. J. Memb. Biol. 1983, 76, 197. 46. Huang, R.; Reusch, R. N. J. Biol. Chem. 1996, 271, 22196. 47. Hankermeyer, C. R.; Tjeerdema, R. S. Rev. Environ. Contant Toxicol. 1999, 159, 1. 48. Slater, S. C.; Voige, W. H.; Dennis, D. E. J. Bacteriol. 1988, 170, 4431. 49. Schubert, P.; Steinbuchel, A. J. Bacteriol. 1988, 170, 5837. 50. Peoples, O. P.; Sinskey, A. J. J. Biol. Chem. 1989, 264, 15298. 51. Comibert, J.; Marchessault, R. H. J. Mol. Biol. 1972, 71,735. 52. Yokouchi, M.; Chatani, Y.; Tadokoro, H.; Tani, H.; Teranishi, K. Polymer 1973, 14, 267. 53. Comibert, J.; Marchessault, R. H. Macromol. 1975, 8, 296. 54. Marchessault, R. H.; Coulombe, S.; Morikawa, H.; Okamura, K.; Revol, J. E Can. J. Chem. 1981, 59,38. 55. Br0ckner, S.; Meille, A.V.; Malpezzi, L.; Cesaro, A.; Navarini, L.; Tombolini, R. Macromol. 1988, 21,967. 56. Barham, P. J.; Keller, A.; Otun, E. L.; Holmes, P. A. J. Mater Sci. 1984, 19, 2781. 57. Welland, E. L.; Stejny, J.; Halter, A.; Keller, A. Polynt Commun. 1989, 30, 302. 58. Marchessault, R. H.; Okamura, K.; Su, C. J. Macromol. 1970, 3, 735. 59. Akita, S.; Elaga, Y.; Miyski, Y.; Fujita, H. Macromol. 1976, 9, 774. 60. Armand, M. B. In Polymer Electrolyte Reviews-l; MacCallum, J. R.; Vincent, C. A., Eds.; Elsevier Applied Science: New York, 1987, p. 1. 61. MacCallum, J. R.; Vincent, C. A. In Polymer Electrolyte Reviews 1; MacCallum, J. R.; Vincent, C. A., Eds.; Elsevier Applied Science: New York, 1987, p. 23. 62. Gray, E M. In Solid Polymer Electrolytes; VCH: New York, 1992, p. 1. 63. Watanabe, M.; Ogatu, N. In Polymer Electrolyte Reviews I MacCallum, J. R.; Vincent, C. A., Eds.; Elsevier Applied Science: New York, 1987, p. 39. 64. Reusch, R. N.; Reusch, W. H. 1993, U.S. Patent No. 5,266,422. 65. Seebach, D.; BUrger, M.; Plattner, D. A. Helv. Chim. Acta 1993, 76, 2581. 66. Seebach, D.; Brunner, A.; Bachmann, B.; Hoffmann, T.; K0hnle, E N.; Lengweiler, U. D. Ernst Shering Res. Found. 1996, 28, 1. 67. B0rger, H. M.; Seebach, D. Helv. Chim. Acta 1993, 76, 2570. 68. Fritz, M. G.; Walde, P.; Seebach, D. Macromol. 1999, 32, 574. 69. Miller, C. Physiol. Revs. 1983, 63, 1209. 70. Alvarez, O.; Benos, D.; Latorre, R. J. Electrophysiol. Tech. 1985, 12, 159. 71. Seebach, D.; Brunner, A.; B0rger, H. M.; Reusch, R. N.; Bramble, L. L. Helv. Chint Acta 1996, 79, 507. 72. Plattner, D. A.; Brunner, A.; Dobler, M.; MOiler, H.-M.; Petter, W.; Zbinden, P.; Seebach, D. Helv. Chirn. Acta 1993, 76, 2004. 73. Seebach, D.; B0rger, H. M.; M011er, H.-M.; Lengweiler, U. D.; Beck, A. K. Sykes, K. E.; Barker, P. J.; Barham, P. J. Helv. Chim. Acta 1994, 77, 1099. 74. van Zutphen, H.; Merola, A. J.; Brierley, G. P.; Cornwell, D. G. Arch. Biochem. Biophys. 1972, 152,755. 75. Tanaka, J. C.; Furman, R. E.; Barchi, R. L. In Ion Channel Reconstitution; Miller, C., Ed.; Plenum Press: New York, 1986, p. 277. 76. Fettiplace, R.; Haydon, D. A. Physiol. Rev. 1980, 60, 510.

PHB and PolyP Ion Transport

97

77. Sykes, K. E.; McMaster, T. J.; Miles, M. J.; Barker, E A.; Barham, P. J.; Seebach, D.; MOiler, H.-M.; Lengweiler, U. D. J. Mat. Sci. 1995, 30, 623. 78. Balasubramanian, S. V.; Sidkar, S. K.; Easwaran, K. R. K. Biochent Biophys. Res. Commun. 1992, 189, 1038. 79. Das, S.; Reusch, R. N. Macrornol. 1999, in press. 80. Jedlinsky, Z.; Kurcok, P.; Lenz, R. W. Macmmol. 1998, 31, 6718. 81. Jedlinsky, Z.; Adamus, G.; Kowaiczuk, M.; Schubert, R.; Szewezuk, Z.; Stefanowicz, E Rapid Comm- Mass Spectrom. 1998, 12, 357. 82. Small, D. M. The Physical Chemistry of Lipids; Plenum Press: New York, 1986. 83. Colquhoun, D.; Sigworth, E J. In Single Channel Recording; Sakman, B.; Neher, E., Eds.; Plenum: New York, 1983, p. 191. 84. Hille, B. Ionic Channels of Excitable Membranes; Sinauer Associates: Sunderland, MA, 1992. 85. Reusch, R. N. Prog. Mol. Subcell. Biol. 1999, 23, 151. 86. Overath, E; Tratible, H. Biochent 1973, 12, 2625. 87. Castuma, C.; Huang, R.; Komberg, A.; Reusch, R. N. Z Biol. Chem. 1995, 270, 12980. 88. Das, S.; Reusch, R. N. J. Membrane Biol. 1999, 170, 135. 89. Abel, E.; Meadows, E. S.; Suzuki, I.; Jin, T.; Gokel, G. W.; Murillo, O. Chem- Commun. 1997, 1145. 90. Urry, D. W. Top. Curt Chent 1985, 128, 175. 91. Christensen, B.; Fink, J.; Merrifield, R. B.; Mauzerall, D. Proc. Natl. Acad Sci USA 1988, 85, 5072. 92. Sansom, M. S. E Prog. Biophys. Mol. Biol. 1991, 55, 139. 93. Kobuke, Y.; Ueda, K.; Sokabe, M. J. Am. Chem- Soc. 1992, 114, 7618. 94. Akerfeldt, K. S.; Lear, J. D.; Wasserman, Z. R.; Chung, L. A.; DeGrado, W. E Acc. Chem- Res. 1993, 26, 19 I. 95. Epand, R. M. In The Amphipathic Helix; CRC Press: Ann Arbor, MI, 1993, p. 222. 96. Fyles, T. M.; James, T. D.; Kaye, K. C.J. Am- Chent Soc. 1993, 115, 12315. 97. Gokel, G.; Murillo, O. Acc. Cheat Res. 1996, 29, 425. 98. Lehn, J. M. Struct. Bonding (Berlin) 1973, 16, 1. 99. Simon, W.; Morf, W. E.; Meier, P. C. Struct. Bonding (Berlin) 1973, 16, 113. 100. Martin, R. B. Metal Ions Biol. Syst. 1990, 26, 1. 101. Chang, C. E J. Bacteriol. 1986, 167, 935. 102. Gangola, P.; Rosen, B. P. J. Biol. Chem. 1987, 26, 12570. 103. Jones, H. E.; Holland, I. B.; Baker, H. L.; Campbell, A. K. Cell Calcium 1999, 25, 265. 104. Holland, I. B.; Jones, H. E.; Campbell, A. K.; Jacq, A. Biochimie 1999, 81,901. 105. Straley, S. C.; Plano, G. V.; Skrzypek, E.; Haddix, P. L.; Fields, K. A. Mol. Microbiol. 1993, 6, 1005. 106. Tisa, L. S.; Olivera, B. M.; Adler, J. J. Bacteriol. 1993, 175, 1235. 107. Tisa, L. S.; Adler, J. Proc. Natl. Acad. Sci. USA 1994, 89, 11804. 108. Norris, V.; Grant, S.; Freestone, P.; Canvin, J.; Sheikh, E N.; Toth, I.; Trinei, M.; Modha, K.; Norman, R. I. J. Bacteriol. 1996, 178, 3677. 109. Lynn, A. R.; Rosen, B. E In Ion Transport in Procaryotes; Rosen, B. E" Silver, S., Eds.; Academic Press: New York, 1987, p. 18 I. 110. Ivey,D. M.; Guffanti, A. A.; Zemsky, J.; Pinner, E.; Karpel, R.; Padan, E.; Schuldiner, S.; Krulwich, T. A. J. Biol. Chent 1993, 268, 11296. 111. Matsushita, T.; Hirata, H.; Kusaka, I. Ann. N.Y. Acad. Sci. 1989, 560, 276. 112. Saimi, Y.; Loukin, S. H.; Zhou, X. L.; Martinac, B.; Kung, C. Methods Enzymol. 1999, 294, 507. 113. Hagiwara, S.; Byerly, L. Ann. Rev. Neurosci. 1981, 4, 69. 114. Hess, E; Tsien, R. W. Nature 1984, 309, 453. 115. McCleskey, E. W.; Almers, W. Proc. Natl. Acad. Sci. USA 1985, 82, 7149.

98

ROSETTA N. REUSCH

116. Tsien, R. W.; Hess, R; McCleskey, E. W.; Rosenberg, R. L. Ann. Rev. Biophys. Chem. 1987, 16, 265. 117. Andreeva, N. A.; Okorokov, L. A. Biokhimiya 1990, 55, 2286. 118. Lorenz, B.; Muller, W. E. G.; Kulaev, I. S.; Schroder, H. C. J. Biol. Chem. 1994, 269, 22198. 119. Wurst, H.; Komberg, A. J. Biol. Chem. 1994, 269, 10996. 120. Akiyama, M.; Crooke, E.; Komberg, A. J. Biol. Chem. 1993, 268, 633. 121. Dunham, E. T.; Glynn, I. M. J. Physiol. 1961, 156, 274. 122. Schatzmann, H. J. Experientia 1966, 22, 364. 123. Penniston, J. T. In Calcium and Cell Function; Academic Press: New York, 1983, p. 99. 124. Kosk-Kosicka, D.; Inesi, G. FEBS Lett. 1985, 189, 67. 125. Kosk-Kosicka, D.; Scaillet, S.; Inesi, G. J. Biol. Chem. 1986, 261, 3333. 126. Griffin, J. B.; Davidian, N. M.; Penniall, R. J. Biol. Chem. 1965, 240, 4427. 127. Crooke, E.; Akiyama, M.; Rao, N. N.; Komberg, A. J. Biol. Chem. 1994, 269, 6290. 128. Karr, D. B.; Waters, J. K.; Emerich, D. W. Appl. Environ. Microbiol. 1983, 46, 1339. 129. Ahn, K.; Komberg, A. J. Biol. Chem. 1990, 265, 11734. 130. Niggli, V.; Penniston, J.; Carafoli, E. J. Biol. Chem. 1979, 254, 9955. 131. Chiesi, M.; Zurini, M.; Carafoli, E. Biochemistry 1984, 23, 2595. 132. Doyle, A. D.; Cabral, J. M.; Pfuetzner, R. A.; Kuo, A.; Gulbis, J. M.; Cohen, S. L.; Chait, B. T.; MacKinnon, R. Science 1998, 280, 69. 133. MacKinnon, R. S.; Cohen, L.; Kuo, A.; Lee, A.; Chait, B. T. Science 1998, 280, 106. 134. MacKinnon, R.; Doyle, D. A. Nature Struct. Biol. 1997, 4, 877. 135. Schrempf, H.; Schmidt, O.; Ktimmeden, R.; Sinnah, D.; MUller, M.; Betzler, T.; Steinkamp, ; Wagner, R. EMBO J. 1995, 14, 5170. 136. Heginbotham, L.; Odessey, E.; Miller, C. Biochemistry 1997, 36, 10335. 137. Cordes, D. M.; Perozo, E. Biochemistry 1997, 35, 10343. 138. Li, H. L.; Sui, H. X.; Ghanshani, S.; Lee, S.; Walian, P. J.; Wu, C. L.; Chandy, K. G.; Jap, B. K. J. Mol. Biol. 1998, 282, 211. 139. Heginbotham, L.; Kolmakova-Partensky, L.; Miller, C. J. Gen. Physiol. 1998, 111,741. 140. Cuello, L. G.; Romero, J. G.; Cortes, D. M.; Perozo, E. Biochemistry 1998, 37, 3229. 141. Perozo, E.; Cortes, D. M.; Cuello, L. G. Science 1999, 285, 73. 142. Heginbotham, L.; LeMasurier, M.; Kolmakova-Partensky, L.; Miller, C. J. Gem Physiol. 1999, 114,551. 143. Bjellqvist, B.; Hughes, G. J.; Pasquali, Ch.; Paquet, N.; Ravier, E; Sanchez, J.-Ch.; Frutiger, S.; Hochstrasser, D. E Electrophoresis 1993, 14, 1023. 144. W'dldns, M. R.; Gasteiger, E.; Bairoch, A.; Sanchez, J.-C.; Williams, K. L.; Appel, R. D.; Hochstrasser, D. E 2-D Proteome Analysis Protocols; Link, A. J., Ed.; Humana Press: NJ, 1998. 145. Huang, R.; Reusch, R. N. J. Bacteriol. 1995, 177, 486. 146. Kulaev, I. J. Biol. Physics 1994, 20, 255. 147. Hanahan, D. J. Mol. Biol. 1983, 166, 557. 148. Clark, J.; Wood, H. G. Anal. Biochem. 1987, 161,280. 149. Page, W. J.; von Tigerstrom, M. J. Bacteriol. 1979, 139, 1058.

FUNCTIONALIZED MACROCYCLIC LIGANDS AS SENSORY MOLECULES FOR METAL IONS

Guoping Xue, Paul B. Savage, Jerald S. Bradshaw, Xian X. Zhang, and Reed M. Izatt

1. 2.

3.

4.

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Chemosensors for the Alkali Metal Cations . . . . . . . . . . . . . . . . . . . 2.1. Fluorescent Sensors for the Alkali Metal Cations . . . . . . . . . . . . 2.2. Crowned Spirobenzopyrans as Alkali Metal Cation Sensors . . . . . . . Chemosensors for Alkaline Earth Metal Cations . . . . . . . . . . . . . . . . 3.1. Fluorescent Sensors for Alkaline Earth Metal Cations . . . . . . . . . . 3.2. Cryptand and Bibrachial Lariat Type Crowned Spirobenzopyrans as Alkaline Earth Metal Cation Sensors . . . . . . . . . . . . . . . . . . . 3.3. Azacrown Ethers Functionalized with 8-Hydroxyquinoline and Its Derivatives as Metal-Ion Sensors . . . . . . . . . . . . . . . . . . . . . Chemosensors for Transition Metal Ions . . . . . . . . . . . . . . . . . . . . 4.1. Fluorescent Sensors for Transition Metal Ions . . . . . . . . . . . . . . 4.2. Macrocycles Functionalized with the Dansyl Chromophore as Zinc (II) Chemosensors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Advances in Supramolecular Chemistry Volume 7, pages 99-137. Copyright 9 2000 by JAI Press Inc. All rights of reproduction in any form reserved. ISBN: 0-7623-0678-5 99

100 101 101 108 112 112 114 117 121 121 127

100

XUE, SAVAGE, BRADSHAW, ZHANG, and IZATT

4.3. MacrocyclesFunctionalized with 8-Hydroxyquinoline and Its Derivatives as Transition Metal Ion Sensors . . . . . . . . . . . 5. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

128 132 132

1. I N T R O D U C T I O N The biological importance of many metal ions is well established. 1'2 However, the same and other metal ions can be toxic to life when present at certain concentrations in the environment, water supplies, the food chain, and industrial chemicals and products. Consequently, an intensive effort has been devoted to develop various sensory molecular receptors capable of recognizing, sensing, and selectively transporting these positively charged substrates so that the concentrations of these metal ions in aqueous or nonaqueous media can be monitored quantitatively, metal ions of commercial value can be recovered from waste solutions, and certain toxic transition and posttransition metal ions in the environment can be removed. 1-7 The goal of this research is the design and construction of ion-selective sensors by assembling a specific metal ion-receptor with a subunit capable of signaling the occurrence of receptor-substrate interaction. The efficiency of the sensor is based on: (i) metal ion affinity and selectivity; (ii) the steric orientation and length of the linker arm (spacer) between the receptor and sensor units; (iii) the interaction between the substrate and the signaling unit; and (iv) the sensitivity, measuring ranges, and simplicity of measurement of the displayed signal. Chemosensors employing fiber optic technology could prove useful for real-time and real-space detection and quantification of metal ions provided synthetic fluoroionophores and chromoionophores capable of signaling the complexation of metal ions with useful selectivity are developed. Optical sensors for metal ions have been reported; s-13 however, they have generally lacked ion selectivity. A number of chromoionophoric and fluoroionophoric dyes, including zincon, 13 8-hydroxyquinoline-5-sulfonic acid, l~ calcichrome, 11 and lipophilized PAR 1,12 have been immobilized on solid supports yielding metal ion sensors. These systems responded well with many metal ions but they do not display significant ion selectivity. The critical requirement for the design of an ion-specific chemosensor is the selective binding of the target ion by the chemosensor. The development of supramolecular chemistry has led to a large variety of macrocyclic compounds with

1

FunctionalizedMacrocyclic Ligands

101

Figure 1. Schematicrepresentationof macrocycle-basedchemosensor. high affinities and selectivities for metal ions, 14-17 and some of them have been used to develop metal-ion sensors. 4-6 Functionalization of macrocyclic compounds with additional ligating units is an effective way to increase metal-ion-complexing ability and selectivity. 18'19 Furthermore, the additional ligating units with responsive functions may also act as the signaling units to provide measurable response to ion binding. The two different processes occurring during metal-ion detection by macrocycle-based chemosensors, i.e. molecular recognition and signal transduction, are illustrated in Figure 1. Among the different chemosensors, those which are fluorescence-based present many advantages such as high sensitivity (single molecule detection is possible), low cost, ease of performance, and versatility. The fluorescent-based sensors offer subnanometer spatial resolution with submicron visualization and submillisecond temporal resolution. 2~ Thus, numerous fluorescence-based chemosensors have been prepared and investigated. 4-6 On the other hand, chromoionophores are also distinct and worthwhile subjects in their own right. 27-3~ In this chapter, we will concentrate on the development of macrocycle-based fluorescent chemosensors for metal ions. Some examples from our laboratory will be discussed in detail. We have also inserted a few paragraphs dealing with the crowned spirobenzopyran ion-sensing systems since this new type of chemosensor has been developed only recently.

2. CHEMOSENSORS FOR THE ALKALI METAL CATIONS 2.1. FluorescentSensors for the Alkali Metal Cations The great impact in metal-ion coordination chemistry caused by the discovery of the macrocyclic polyethers 3l has led to new designs of chemosensors for the alkali cations. Application of a supramolecular receptor to fluorescence sensing was first

102

XUE, SAVAGE, BRADSHAW, ZHANG, and IZATT

described by Sousa in 1977 when naphthalene was outfitted with crown ether receptors as in compounds 2 and 3. 32 The observed fluorescence changes for these ligands upon binding alkali metal cations in ethanol were attributed to a heavy atom effect (for Cs § and Rb § and a complexation-induced change in triplet energy relative to ground- and singlet-state energies. Examination of dibenzo- 18-crown-6 (4) also showed small but significant alkali metal cation induced fluorescence and phosphorescence modulations. 33'34Phosphorescence modulation was subsequently addressed in another time-resolved study on 5. 35 Two important results to emerge from these studies were the significance of spin-orbit coupling effects with the heavier alkali metal cations and coordination-induced ligand rigidification. The latter effect was of particular importance in bifluorophoric systems such as 4. 33 More highly functionalized macrocyclic s e n s o r s 6, 36,37 7, 38 and 8 39,40 w e r e also developed. These pioneering cases showed small responses to the alkali cations, but a valuable principle was thereby established: cation-induced spectral wavelength shifts in either direction are a natural consequence of ionic charge coupling with excited-state dipoles. 38-40 Macrocyclic crown ether receptors continue to be important in the growth of integrated fluorophore receptors 41 with ICT (internal charge transfer) excited states. 4 Coumarin fluorophores are the sensing agents in 9-12. 42-47 Compound 1143 is a sensor for Li § whereas 12 senses extracellular K§ 45 Lariat crown ether 13 has several interesting features such as cation-controlled absorption spectra and photocontrollable cation binding. 47 However, the photoisomerization process responsible for the latter property also causes reduced fluorescence quantum yields. Compound 14 is structurally related to thermotropic liquid crystal materials. 48'49This similarity may be responsible for some of its spectral characteristics even in homogeneous solution where K § causes a substantial blue-shifted fluorescence. Long communication wavelengths are achieved by 15 50 and 16, 51 but small ion-induced spectral effects remain a problem.

oC-O o

o .3

2

3

o

k.o.J 4

o

%_,o..; 5

103

FunctionalizedMacrocyclic Ligands o

"~0"~

N~

Me

o~., ,.,...~o I

k,,oJ

o 6

k,,oJ Me2N~OTO

',~7%,-, 1 (

o__~o,~o

o

,~

0

0

~,o,J 10

o~

~o~o~ ~~o~o~

,~C02Me

o

"

12

CCk"

A high point in this line of research is Tsien's sensor 17 52 which selectively responded to physiological levels of sodium ions. 41'52 Sensor 17 for Na § is based on an integrated fluorophore-receptor system 41 and allows the Na § -induced deconjugation of fluorophore segments by a conformational change to allow additional t r a n s diaxial ligation of Na § held in the diazacrown ether moiety. Cryptand 18, 53 which is structurally more elaborate, also achieves selective measurement of Na § in physiological media. The K § sensory coumarocryptand 19 54 also possesses a dialkoxy coumarin fluorophore with an ICT excited state and an overlapping [2.2.2] cryptand moiety. However, the similarity with the previous cases stops here. The

104

XUE, SAVAGE, BRADSHAW, ZHANG, and IZATT

C~~~

o~xo

o

~~~7~ o~

kso;

rx

o_~ c o o_~ o

o

Q~

13

-

O___~cN

14

(o~_,oC~~.o~o~ o~o.~

o

~o~o

o

15

i~o'-'o--~ o_,o.9~

§ N-M~

16

17

F

F

o

o

k...,....; N-~XC -

o; ~ 18

105

Functionalized Macrocyclic Ligands

nitrogen atoms in cryptand 19 are not conjugated with the fluorophore and can act as electron donors in a PET (photoinduced electron transfer) system. 21 The experimental results of K§ fluorescence enhancement by a factor of about 3 in 19 without any significant wavelength shift are in line with PET sensory behavior. de Silva et al. 55 have incorporated the core receptor of Tsien's sensor 17 into a "fluorophore-spacer-receptor" system 41 so that Na§ fluorescence switching can be arranged by suppression of PET. The fluorescence emission spectrum of 20 displays large Na§ increases in intensity and shows good discrimination against K § while several other ions cause almost no effect. Protons, naturally, give rise to large fluorescence enhancements 56 but the effects are negligible above pH 6. This result is very significant since these two ions (Na § and K § are principal signal carders during information processing within the nervous system. Cryptand 2157 combines PET design principles with the idea of complexation-induced conformational changes previously used in this context by Tsien. 58 However, the conformational changes are expected to be of a smaller magnitude due to the restrictions caused by the macrobicyclic framework. Nevertheless, the effect proves to be sufficient to cause ion-induced redox potential changes in the receptor module which gives rise to order-of-magnitude enhancement of fluorescence with K § and Rb § ions. In contrast with the dibenzo-18-crown-6 integrated fluorophore-receptor system 4, 33,34the first use of PET systems with benzocrown ether receptors arrived a decade later in the form of 22. 59 Receptor 22 shows fluorescence switching "on" by Na § with excellent selectivity against protons. Earlier PET systems employed azacrown

N"~O~/~O"~,N

oJ

C~e

MeQ

,~ ~ C N

Me

0-

20

19

O/---NO

N~O~/~O~N

CN

106

XUE, SAVAGE, BRADSHAW, ZHANG, and IZATT

ether receptors 236~ and 24. 62 Receptor 23 with its diazacrown ether receptor and an anthracene-containing alkane strap has found use as a constrained triple exciplex system63 and as a receptor in structural studies, 64'65in addition to its pioneering role as a fluorescent PET sensor for alkali metal cations. Ion-induced fluorescence enhancement is influenced by large conformational alterations caused by the hydrogen bonding ability of the solvent and by the nature of the cation itself. Structurally simpler receptor 24 allows extensive solvation of its electron transfer state and leads to low cation-free fluorescence quantum yields which in turn sets the stage for large cation-induced fluorescence enhancements. 66 The magnitude of these enhancements can be related to the electronic field strength of the crucial lone electron pair on the nitrogen atom, to the charge density of the guest ion, and to the center-to-center distance between the ion and the nitrogen atom. Another notable result 59'62 is that the ion-binding constants measured fluorimetrically with these "fluorophore-spacer-receptor" systems are virtually identical to the values found for the parent receptors by ground state techniques. 14'15 Among alkali metal cations, there has been much interest in Li § and Li § ionophores, 67 due to their actual and potential applications in science, medicine, and technology. 68-7~However, little is known about designing an ionophore showing high selectivity for Li § not only because Li § has a small ionic diameter and is strongly hydrated in water, but also because Na § ions are much more abundant in nature. In this regard, it has been shown that small aza-cages with a macrobicyclic arrangement such as 25-34 are able to selectively encapsulate Li § in aqueous solution, but none shows an appreciable interaction with Na+. 71-79

CN

C'o

24 n-O, 1

23

Functionalized Macrocyclic Ligands

107

Me 2 5 X = NH, n = 1

26 X=NH, n=2 28 X=N-Me, n=2 30 X=S, n=2

2 7 X = N-Me, n = 1 29 X=O,

n.-2

31

X = Cell4, n = 1

33

X = CH=. n = 2

32

X - N-(CH2)11CH 3, n - 2

3 4 X = N-Bz, n - 1

Ciampolini et al. s~ have prepared three anthracene-functionalized aza-cages 35-37 as selective Li § receptors and chromoionophores. These three ligands present a small three- dimensional cavity able to selectively encapsulate Li + and the aromatic sensor is situated in an external position outside the macrocyclic cavity, but in such a way that its optical properties could be perturbed by metal complexation. The experimental results indicate that all three ligands are able to selectively bind Li § in alcoholic solution while the other alkali metal cations are not complexed. Such remarkable selectivity is due to the small dimensions of the macrobicyclic cavity in which alkali cations larger than Li § cannot be encapsulated. As expected, 36 forms a binuclear Li § complex with each ion-lodged in a single-cage subunit. The photochemical properties of the ligands are strongly influenced by protonation. Moreover, upon complexation with Li § an increase in the fluorescence emission of the sensor, relative to that of the free ligand, is observed in methanol. Unfortu-

~,.,.N,,J N

Me.

N

Me

N Me

..

N~

3s

Me

37 Me 36

NL,~

108

XUE, SAVAGE, BRADSHAW, ZHANG, and IZATT

nately, the presence of the anthracene moiety renders the ligands insoluble in aqueous solution. Kubo and co-workers prepared and studied N,N'-bis(1-naphthylmethyl)-4,13diaza-18-crown-6 (38), 81 N,N'-bis(9-anthrylmethyl)-4,13-diaza-18-crown-6 (39), 82 and N,N'-bis(1-pyrenylmethyl)-4,13-diaza- 18-crown-6 (40) 83 for use as fluorescent cation sensors. The observed cation binding strengths and corresponding fluorescence changes were generally modest. This is surprising for such widely used and strongly fluorescent residues. In a study by Gokel and co-workers, 84-86 intimate C-H...O hydrogen bonding between the aromatic sidearms and the proximate macroring oxygen atoms were observed. It is these intramolecular interactions that appear to define both conformations and binding affinities of compounds 38-40, and in turn, diminish their effectiveness as sensors. These results will be very helpful for the design of functionalized macrocyclic ligands as chemosensors. 2.2.

Crowned Spirobenzopyrans as Alkali Metal Cation Sensors

Spirobenzopyran derivatives are typical photochromic compounds that undergo reversible isomerization to the corresponding colored zwitterionic merocyanine isomers under UV and visible irradiation or heat. 87 Extensive studies have been devoted to their applications to photochemical control of the physical properties of ligating agents in solution. The examples are in photocontrol of membrane transport, 88'89 membrane potentials, 9~ and polymer rheology. 94'95 Incorporation of a crown ether moiety, which is able to complex metal ions, into the spirobenzopyran skeleton is expected to provide the parent spirobenzopyran with important new properties. Inouye and co-workers have developed various types of spirobenzopyrans bearing a monoazacrown ring as a recognition site such as 41-46. 96-98 These crowned spirobenzopyrans exhibited no absorption bands above 450 nm in nonpolar solvents (e.g. CH3CN, CHCI 3, acetone) indicating the closed spiropyran forms. However, addition of alkali metal iodides to these solutions causes distinct CH2

38 /---o R-N

o N-R

oJ

39

R-

x__../

I 40

R-

~ c~

Functionalized Macrocyclic Ligancls

109 Me. Me

Me.

Me

~NO 0

2

o

02?

41 n = l 42 n = 2 43 n=3 44 n,,4

R

45

R-

46

..

o-

changes in their absorption spectra. When a fivefold molar quantity of LiI was added to the acetonitrile solutions of 41 and 42, new absorption bands appeared (41: ~. max = 530 nm, e = 4700; 42: ~, max = 530 nm, e = 10,000). Only negligible changes were observed upon addition of the other alkali metal iodides. While 43 provided a small but significant selective coloration with NaI, cation-induced hypsochromic band shifts caused by Li § Na § K § Rb § and Cs § ions, which decreased in that order, were a result of electrostatic interactions between the complexed cations and the p-nitrophenolate dipole of the merocyanine. An NMR study revealed that this color change can be attributed to isomerization from the closed spiropyran form to the merocyanine structure (Scheme 1). The position of the complexed cation in the molecule was found to be important for isomerization. Spirobenzopyran 4798 (Scheme 2) was designed to recognize cations in which the complexed cations could interact with the other oxygen of the closed spiropyran and not the phenolate oxygen of the opened merocyanine. Indeed, isomerization of 47 to the open-colored Me.

Me

NO2 MI

Scheme 1.

110

XUE, SAVAGE, BRADSHAW, ZHANG, and IZATT

NO2

Me. Me

NO2 Me

0 MI

0

/__o oJ

0

o3

47'

47

Scheme 2.

merocyanine was most strongly suppressed by the presence of K +, which was expected to be strongly recognized by the crown ring. The length of the spacer chain connecting the spirobenzopyran unit and the crown ether unit also affects the coloration efficiency. Color selectivities of spirobenzopyrans 48-51 in which the spirobenzopyran units were much further separated from the crown ether units by alkyl chains than in 41-44 were examined. 98 Although the behavior of Li + is still unknown and remains to be clarified, 48-50 provided the highest coloration for Na +, K +, and Cs +, respectively. Each of Na +, K +, and Cs + is larger than the size of the cavity of the crown ether of the respective spirobenzopyrans. This finding suggests that in the opened merocyanines, alkali metal cations are favorably located in the space made by the spirobenzopyran unit and the crown ether unit as depicted in Figure 2. In spirobenzopyrans 45, 46, and 48-51, the molar absorptivities in the presence of the alkali cations are considerably smaller as compared with those of 41-44 and 47, and in particular, 51 was scarcely affected upon addition of any alkali metal cation. The low coloring efficiency might result from an entropic disadvantage: reduced probability of the existence of the complexed cations in the neighborhood of the phenolate oxygen of the merocyanines. All of these results showed that isomerization of the crowned spirobenzopyrans to the open-colored meroMe. Me NO2

48-51

n= 1-4

Functionalized Macrocyclic Ligands

111 NO2

Figure 2. Schematic representation of the colored complex between the spirobenzopyrans 48-50 and an alkali cation. cyamines was induced by recognition of alkali metal cations and the selectivity of the coloration was governed by several factors: (1) the size of the crown ether ring; (2) the position of recognition; (3) electronic properties of both the complexed cations and the merocyanine dipoles; and (4) the length of the spacer chains connecting the spirobenzopyran units and the crown ether units. Spirobenzopyran derivatives 52-54 bearing a monoazacrown moiety, such as 12-crown-4, 15-crown-5 and 18-crown-6 at the 8-position, are additional examples independently developed by Kimura and co-workers. 99-1~ Crowned spirobenzopyrans 52 and 53 exhibit a similar coloration selectivity to that of 41 and 42. It has been pointed out that the coloration behavior does not correlate well with the metal ion recognition feature of the crown ether moiety as estimated from the ring size. In the cation-induced spectral change for 54, complexation with Li § or Na § promoted isomerization to the corresponding merocyanine form more strongly than did complexation with K § This is not in accord with the generally accepted view that 18-crown-6 derivatives interact more strongly with K § than with Na § or Li § This also demonstrated that isomerization of the spirobenzopyran to the merocyanines was governed not only by recognition of alkali metal cations by the crown

Me_ Me

NO.z

o

52-54

n - 1-3

o

112

XUE, SAVAGE, BRADSHAW, ZHANG, and IZATT

ether rings at suitable positions but also by the electronic properties of both the complexed cations and the merocyanine dipoles. In this section, various crowned spirobenzopyrans for use as alkali metal cation receptors have been presented. This new type of chromoionphore is conceptually different from the crown ether dyes because, in the case of the crown ether dyes, the absorption bands of the chromophones are shifted by complexation of cations. Crowned spirobenzopyrans might provide sophisticated ion-sensing systems such as ion sensors whose ion selectivity can be photochemically switched between cations and anions. 1~

CHEMOSENSORS FOR ALKALINE EARTH METAL CATIONS

0

3.1. Fluorescent Sensors for Alkaline Earth Metal Cations Of particular interest has been the measurement of the concentrations of alkaline earth metal cations in biological samples. Calcium and magnesium ions play important roles as intracellular messengers in the regulation of cell functions. 1~176 Multiple fluorescent sensors with structural elements of the EDTA type, 4'1~176 such as Ca 2+ - sensors 5558 and 56,11~ have been developed for use in measuring alkaline earth metal cations. However, there are fewer macrocycle-based fluorescent sensors for these cations than for the alkali metal cations. Compound 57,113 N,N'-bis(4-methylumbelliferone-8-methylene)diaza-18-crown-6, developed by Takagi and co-workers, was an early example whose fluorescence responded to alkaline earth metal cations, especially to Ca 2§ in a solvent extraction experiment. No crown ether based fluorescent PET sensors for alkaline earth cations a r e currently available, though 57 is an appropriate example of a fluorophore-spacerreceptor system. All examples of alkaline earth metal ion sensors are to be found among the fluoroionophores with integrated fluorophores and receptor moieties such as 738 and 8. 39'40'114 Compound 7 displayed enhanced fluorescence emission when forming a complex with Ca 2§ Interestingly, the sensitivity of the system is influenced by the counterion and water content of the matrix. The 7-Ca 2§ complex in acetonitrile is a useful probe for trace water determination in that solvent although

~o~ _- -o=c fc~ o=c\ ~ N

('~~NOH ,--O Me

X

55 X = H 56 X - R

O--,N~~HO

N

"O

~

57

O-

FunctionalizedMacrocyclic Ligands

113

the calibration curve is nonlinear due to multiple quenching mechanisms. For compound 8 in the presence of Ca(C104) 2, the fluorescence maximum shifts from 642 to 578 nm, and the fluorescence quantum yield increased from 0.33 to 0.64. Lithium and sodium perchlorates have only a slight effect upon the spectral parameters, whereas the potassium salt induces no detectable change. Delmond et al. 114 reported the synthesis of two nonpolar fluorescence cation probes 58 and 59 designed from stilbene and 1,4-diphenyl-l,3-butadiene, respectively, by substitution at the two ends with two electron-donor groups (dimethylamino and monoaza-15-crown-5), one of which is able to chelate a cation. The absorption and fluorescence spectra in several solvents of different polarities and the picosecond transient absorption spectra have given an estimate of the intramolecular charge transfer (ICT) strength in the excited state. When Ca 2§ was chelated by the macrocycles, the ICT process was increased and the fluorescence shifted to the red. These spectroscopic effects of cation chelation were enhanced in a third probe (60) derived from probe 58 by inserting an electron-acceptor group (CN) at the ortho position of the benzene ring attached to the macrocycle in order to increase the ICT process induced by the cation. More recently, an efficient fluorescent chemosensor 61 for Ba 2§ was developed by Prodi et al. 115'116The 1,8-dioxyxanthone-based crown ether 61 was not strongly fluorescent itself, but exhibited strong fluorescent enhancement and the appearance of a long-lived delayed fluorescence upon binding to alkaline earth metal cations (Ca 2§ Sr 2§ Ba2§ especially Ba 2§ while little or no change was observed on addition of Mg 2§ and the alkali metal ions. Furthermore, 61 showed good selectivity towards Ba 2§ (Kass = 8.9 x 104), so that this crown ether can act as a sensor for Ba 2§ in the presence of the alkali and alkaline earth cations Na § K § Cs § Mg 2§ Sr 2§ (10-fold excess) and Ca 2§ (100-fold excess). An additional relevant property for physiological applications is that complexation does not strongly depend on the pH of the solution.

"""

L o o/3 .e,,

(-o,__:o?3 59

M.,,

c,

60

\

/

114

XUE, SAVAGE,BRADSHAW, ZHANG, and IZATT Z--N

/--o o~

61

3.2. Cryptand and Bibrachial Lariat Type Crowned Spirobenzopyrans as Alkaline Earth Metal Cation Sensors In the previous examples (in Section 2.2) spirobenzopyrans 41-44 bearing a monoazacrown ether ring with a short linkage displayed distinctive color selectivities for Li § (n = 2) and Na § (n = 3). For large alkali cations, however, selective coloration could not be obtained even in the case of crowned spirobenzopyrans where n > 3. While the crowned spirobenzopyrans 48-51, in which the spirobenzopyran moiety was much further removed from the crown ether units by longer alkyl spacers than in 41-44, showed a small selective coloration for large alkali metal cations such as K § and Cs § the molar absorptivities in the presence of alkali metal iodides were considerably smaller than those of 41-44. The low coloring efficiency might result from unfavorable entropy effects: the probability of the existence of complexed cations in the neighborhood of the phenolate oxygen of the merocyanines is reduced, and the electrostatic interaction between the complexed large univalent cations and thep-nitrophenolate dipole of the merocyanines is weak. Taking into account the above points, Inouye and co-workers developed cryptand and bibrachial lariat type crowned spirobenzopyrans 62-64 and 65 and 66, respectively (Schemes 3 and 4). 117'118 Cryptand-containing spirobenzopyrans 62-64 (Scheme 3) exhibited no absorption bands above 400 nm in nonhydroxylic solvents indicating the closed spiropyran form. Indeed, this interpretation was verified by an NMR study. The absorption spectra were scarcely affected upon addition of any alkali metal iodides in CH3CN. In the 1H NMR spectra of 63 in CD3CN, however, downfield shifts (for aromatic and crown ring protons), splitting (crown ring and alkyl protons), and sharpening (aromatic protons) of the signals in the spiropyran form were observed after the addition of KI. This result clearly indicated that the alkali metal cations were bound to the macrocycle moiety of 63, and that the colorless form was a result of weak electrostatic interactions between the complexed univalent cations and the p-nitrophenolate dipole of the merocyanines. On the other hand, addition of alkaline earth metal iodides to these CH3CN solutions gave rise to changes in their spectra with 62 giving the most intense coloration for Ca 2§ and 63 for S1"2§ Titration

Functionalized Macrocyclic Ligands

115

Me_ Me NO2

0

NO=

Me Me,~ MIz

0

0

62' 63' 64'

62 n = l 63 n=2 64 n=3

Scheme 3.

experiments demonstrated that about 1 equiv of SrI 2 is enough to obtain the maximum coloration of 63. Unfortunately, 64 showed little change and poor selectivity in its absorption spectrum under the same conditions. This may be because the crown ether ring is too flexible to interact strongly with the cations. As expected, little change in the spectra of 41-44 and 48-51 occurred under the same conditions. Isomerization of cryptand-containing spirobenzopyran 63 (Scheme 3) to the open-colored merocyanine 63' in the presence of SrI 2 has been studied by 1H NMR spectroscopy at 500 MHz. In several deuterated solvents (e.g. CDC13, CD3CN, DMSO-d 6) the 1H NMR signals of 63 were considerably broadened at room temperature. High-temperature NMR (85 ~ in DMSO-d 6 succeeded in resolv-

Me

NO2 NO2

O

n

Me MI2

O

Me #

-=

O

62' 63' 64'

62 n = 1 63 n=2 64 n=3 Scheme 4.

116

XUE, SAVAGE, BRADSHAW, ZHANG, and IZATT

ing the identifiable species. (In contrast, the spectra of 62 and 64 at room temperature in CDCI 3 were well-resolved.) To a solution of 63 (20 lxmol) in CD3CN (0.7 mL) was added SrI 2 (40 lxmol), and the 1H NMR spectrum was measured as a function of time in the dark. After 30 rain, new resonances were detected which were assigned to merocyanine 63', and the remaining signals of spiropyran form 63 sharpened and shifted. Equilibrium was reached after 3 h (>80% conversion). These observations suggest fast and strong binding of Sr 2+ by 63, and a slow isomerization of the resulting complex (63-Sr 2§ to the merocyanine form 63'. The fast and strong binding of Sr 2§ by 63 was corroborated on the basis of FAB MS experiments. Before addition of SrI 2, ion peaks for (M + H) § and (M + Na) § were detected; after the addition, these signals decreased and finally disappeared, while peaks for (M - H + Sr § and (M + Sr + I) § appeared and increased. Coloration of bibrachial lariat type crowned spirobenzopyrans 65 and 66 (Scheme 4) was examined. The new crowned spirobenzopyrans were designed to recognize divalent cations in which the complexed cations would interact with the two phenolate oxygens of the open merocyocnines 65' and 66' and the crown ether tings (Scheme 4). Although the molar absorptivities of 65 and 66 in the presence of 1 equiv of alkaline earth metal iodides in CH3CN were small compared to those of 62 and 63, 65, and 66 revealed the highest coloration for Mg 2§ (~, max = 513 nm, e = 700) and Ca 2§ (~, max = 524 nm, e = 3500), respectively. These metal ions were smaller than those used for 62 and 63 which possess the same type of crown ethers. This observation was satisfactorily explained by CPK molecular models, which indicated that the radius of the cavity of (N,N'-diacetyl)diaza-crown ethers (in 65 and 66) was smaller than that of the corresponding (N-monoacetyl)diaza-crown ethers (in 62 and 63). More recently, a spirobenzopyran dimer bridged by a diaza-18-crown-6 moiety through the 8-position (67) was developed by Kimura and co-workers, llg-121 Crowned bis(spirobenzopyran) 67 shows a similar coloration selectivity to that of 63. Complexation of multivalent metals, especially Ca 2§ and La 3§ by 67 enhanced the isomerization of the spirobenzopyran moiety to the corresponding merocyanine form due to an effective intramolecular interaction between a crown-complexed cation and the two phenolate anions in the cation complexes of the merocyanine form. In this section, the sensing properties of the cryptand and bibrachial lariat-type crowned spirobenzopyrans were presented. In the spirobenzopyrans, coloration was efficiently induced in the presence of the alkaline earth metal cations. The cryptand and bibrachial lariat-type crowned spirobenzopyrans represent highly sensitive and selective chemosensors for alkaline earth metal cations.

FunctionalizedMacrocyclic Ligands

117

Me. Me

~

N

0 2 Me C["N

/---N O

O--~ O

67 3.3. AzacrownEthersFunctionalizedwith 8-Hydroxyquinolineand Its Derivativesas Metal-Ion Sensors Our interest in chemosensors has been focused on a series of azacrown ethers functionalized with 8-hydroxyquinoline and its derivatives. 122-125 Evaluation of stability constants for interactions of these ligands with metal ions indicates that the ligands form stable complexes with alkali, alkaline earth, and transition metal ions, which is a prerequisite for their use as chemosensors. 8-Hydroxyquinoline is an analytical reagent containing a phenol-like function wherein ligand fluorescence is moderated upon complexation with certain metal ions. 126 5-Chloro-8-hydroxyquinoline (CHQ)-substituted diaza- 18-crown-6 ligands prepared in our laboratory, 68 and 69, exhibit greatly improved ion-complexing ability and selectivity for certain metal ions compared to unsubstituted diaza-18-crown-6. ]22'123Ligands 68 and 69 have the CHQ sidearms attached through different CHQ positions to determine if positions of attachment would alter metal-ion complexation by 68 and 69. Indeed, the site of attachment has profound consequences for cation complexation properties of these ligands. Thermodynamic data in MeOH show that ligand 68 selectively binds Mg 2+ (log K = 6.82) over other alkali and alkaline earth cations (log K = 2.89-5.31).123 UV-vis

c, Co o_ 68

Cl

co

Cl

c~

X---ox_jo--/

OH

~

69

118

XUE, SAVAGE, BRADSHAW, ZHANG, and IZATT

spectra ofligand 68 and its complexes with Mg 2§ Ca 2§ and K § in MeOH are shown in Figures 3 and 4. The UV-vis spectral data for other metal ions have also been reported. 123 Among the alkali and alkaline earth metal ions, the UV-vis spectrum of the Mg 2+ complex is unique. A new peak at 265 nm is observed for the 68-Mg 2+ complex. Neither Mg 2+ nor the ligand itself has absorption in that region and the other alkali and alkaline earth metal cations do not cause such an absorption. ~23As observed in Figure 3, the presence of Mg 2+ causes both CHQ and ligand 69 to exhibit new peaks in the vicinity of 265 nm but the peak intensities are much weaker than that of the 68-Mg 2+ complex. Therefore, the unique UV peak at 265 nm for the 68-Mg 2+ system may provide a promising method for Mg 2§ analysis in samples where an excess amount of alkali and alkaline earth metal cations are present. Figures 5 and 6 show the UV-vis spectra of two analogues, 70 and 71124 of 68. In these cases, Mg 2+ also causes the new peaks in the vicinity of 265 nm (see curve c in both figures) but the peak intensities are lower than that of the 68-Mg 2+complex. The luminescence properties of ligand 68 and its complexes have been examined. 125 As shown in Figure 7, uncomplexed 68 exhibits a very weak luminescence band (@ < 5 x 10-5, t < 0.5 ns) centered at 540 nm in MeOH/H20 (1:1 v:v), which is consistent with the luminescence behavior of 8-hydroxyquinoline in protic solvents. Also, no appreciable luminescence intensity increase was observed from pH 2 to 13 with uncomplexed 68. However, addition of Mg 2+ to 68 (5 x 10-5 M) in a neutral (1:1 v:v) MeOH/H20 solution (pH 7.2) results in a strong enhancement of the luminescence band (~ = 0.042, t = 7.4 ns). Upon complexation with Mg 2+ and excitation at 393 nm, the fluorescence intensity of 68 is increased by a factor of 1000. The excitation spectrum of the complex strictly matches the absorption spectrum of 68-Mg 2+, suggesting that the observed fluorescence is due to neutral complex formation.

~

265

~1 200

397

3OO 400 Wavelength (X, rim)

Figure 3. UVovis spectra of Mg 2+ complexes with CHQ (dotted line), 68 (solid line) and 69 (point-dash line) in MeOH; [CHQ] = 3.1 x 10-5 M; [68] = 1.1 x 10-5 M; [69] = 2.0 x 10-s M; [Mg 2§ is 40 times the ligand concentration in each case.

119

F u n c t i o n a l i z e d /vlacrocyclic Ligands 1.25

1.00

0.75

0.50

0.25

200

250

300

350

400

450

Wavelength (;L,nrn) Figure 4. UV-vis spectra of free and complexed 68 in MeOH; [68] (solid line) = 1.0 x 10 -5 M; [K +] (point-dash l i n e ) = [Ca 2+] (dotted l i n e ) = 5.0 x 10-4 M.

Contrary to what was observed with Mg 2+, no luminescence increase was detected upon addition of K § Na +, Ca 2+, Sr 2+, or Ba 2+ to 68 (5 x 10 -5 M) at pH 7.2 in the 1:1 MeOH/H20 solvent. 125 The lack of luminescence increase can be attributed to the absence of formation of neutral complexes. Although Cu 2+ and Ni 2+ are able to form neutral complexes with 68, addition of Cu 2§ or Ni 2§ into a solution of 68 does not result in a luminescence increase. In these complexes, energy and electron transfer processes are accessible providing a fast deactivation route of the excited state to the ground state. The complexes of Cu 2§ and Ni 2+ with other 8hydroxyquinoline derivatives have also exhibited a lack of fluorescence. 127'128 Addition of Zn 2§ to a solution of 68 resulted in a luminescence complex. 125 However, the fluorescence quantum yield was 8 times lower than that of the

1.2

1

-e <

0.6

0.4 0.2

t

, .....

.~ :.%

""'"'"

29~

"~

,. i ~%..........". :I

Ii

Ii

#Z

s

-~

,." .d ',. ".l"" "'' ' ' /I// .,~" #/"" 1" ' :,

0.8 0 cn ,.D

#."

~

o

#

"

"..

'.. i

'..i

s

. . . .

,,

i

:,~

. . . .

1 . . . .

:~'rs

"..\

"',,

"'...

i

300

. . . .

i

325

. . . .

i

3so

Wavelength (rim) Figure 5. UV spectra of free and complexed 70 in MeOH; (a) 1.1 x 10 -5 M 70; (b) solution a with 1.0 x 10 -3 M Ba2+; (c) solution a with 1.0 x 10-3 M Mg2+; (d) solution a with 1.0 x 10 -3 M Zn2+; and (e) 3.4 x 10 -6 M 70 + 2.9 x 10 -4 M Cu z+.

120

XUE, SAVAGE, BRADSHAW, ZHANG, and IZATT

I

b

8 o.8-~

1 -J

...-' ""

....

./" .-.~. s

I:..,I,,." ....d ",. / / - ~ ., ",. ,. C

x

0.2

i.

_,. "..\

"~

,.,.

".

0

=s

2so

275

30o

Wavelength (nm)

a2s

3so

Figure 6. UV spectra of free and complexed 71 in MeOH; (a) 1.1 x 10 -s M 71; (b) solution a with 1.0 x 10 -3 M Ba2+; (c) solution a with 1.0 x 10 -3 M Mg2+; (d) solution a with 1.0 x 10 -3 M Zn2+; and (e) 3.4 x 10 -6 M 71 + 2.9 x 10-4 M Cu z+.

68-Mg 2+ complex. Thus, Zn 2+ did not interfere with the sensing of Mg 2+ via fluorescence unless the Zn 2+ concentration exceeded that of Mg 2+. The possibility of measurement of Mg 2+ concentration in a matrix complicated by alkali and alkaline earth metal ions has been explored by titration of 68 (5 x 10-5 M) in 1"1 MeOH/H20 solution (pH 7.2) containing Na + (5 x 10-3 M), K + (1 x 10-3 M), Ba 2+ (1 x 10-3 M), Sr2+ (1 x 10-3 M), and Ca 2+ (1 x 10-4 M). The titration was monitored via fluorescence. ]25 The fluorescence intensity reached a maximum at 1 equiv of Mg 2+, indicating that fluorescent intensity could be directly correlated to Mg 2+ concentration. The lack of interference from the other metal ions present can be a result of their lower binding constants with 68 and lower quantum yields of the charged complexes of 68 with Na +, K +, Ca 2+, Sr 2+, and Ba 2+ at this pH. Thus, ligand 68 possesses characteristics of an efficient fluorescent chemosensor for Mg 2+ and may find use in determining Mg 2+ in biological samples and, if immobilized on a solid support, may be incorporated into sensory devices for measurement of Mg 2+ concentrations in aqueous solutions.

CH3

~

x..._./

X...../

70

71

CH3

Functionalized Macrocyclic Ligands

121

250

~. 2OO f,f}

e

150

~

100-

~

50

~ 1 0

400

I

45O

I

5OO

550

I

600

65O

700

(mu)

Figure 7. Fluorescence spectra of 68 (I x Io-sM) in methanol-water (1"I vol:vol) with increasing amounts of Mg2+; ~ex = 393 nm. 4.

CHEMOSENSORS FOR T R A N S I T I O N METAL IONS 4.1. FluorescentSensorsfor Transition Metal Ions

The ability of macrocyclic ligands to bind heavy metal ions 129has been exploited by synthesizing fluorescent chemosensors for transition metal ions bearing polyazamacrocycles, 72-76. These materials were developed by Czarnik and co-workers. ]3~ The fluorescence of compounds 72-76 is pH-dependent showing low fluorescence under basic conditions. This behavior is consistent with the occurrence of a PET mechanism between the unprotonated amines and the chromophoric group. Fluorescence enhancement in water was detected upon complexation of 74 and 76 with Zn 2§ and Cd 2§ which was due to the prevention of the above-mentioned PET process. With 76, complexation with Hg 2§ resulted in a quenching of the residual fluorescence. Interestingly, complex 75-Cd 2§ uniquely displayed a different emission bandshape and maximum with respect to that of the free ligand. This feature can be used to discriminate between Zn 2§ and Cd 2§ This behavior is a result of the occurrence in this system of ground-state ligand-Cd (II) complexation. Fluorescent chemosensor 77, which signals the presence of only Hg 2§ and Cu 2§ in water, was reported by Czarnik and co-workers. 134This compound displayed chelation-enhanced quenching effects only with these two ions in water at pH 7. Titration of 77 with Hg 2§ (Kd _< 1 laM) and Cu 2+ (Kd = 56 ktM) gave overall emission changes of 18-fold (Ka > 106 M -1) and 4-fold (Ka = 1.8 x 104 M-l), respectively. As with the 75-Cd 2§ complex, ground-state Hg (II) rc-complexation has been proposed to explain the high association constant and the strong quenching observed for this complex.

122

XUE, SAVAGE, BRADSHAW, ZHANG, and IZATT

i

72-76

n

n = 1-5

The luminescence behavior of three macrocyclic ligands incorporating naphthylene fluorophores, 78-80, has been studied by Parker and co-workers. 135'136 Compound 78135 exhibits the typical fluorescence of naphthalene, although the intensity is partially quenched in CH3CN presumably via a PET mechanism involving the nitrogen atoms. In water, partial protonation of two of the four nitrogen atoms prevents PET from occurring, and more intense fluorescence can be observed. Addition of 5 equiv of zinc or cadmium triflates to a solution of ligand 78 in CH3CN (2 x 10-5 mol dm -3) led to a twofold increase in the intensity of the naphthalene fluorescence. A similar effect, but to a lesser extent (50% fluorescence enhancement), was observed when 50 equiv of CF3CO2H were added to 78 in the same solvent. This behavior is consistent with a chelation enhanced fluorescence effect. Different behavior was noted following addition of 5 equiv of Pb 2§ A 25% reduction in fluorescence intensity was observed accompanied by a 250% increase in absorbance at 270 nm. In this case, the chelation enhanced fluorescence effect is less significant than the quenching effect of the proximate "heavy atom." Cu 2§ is responsible for a 95% fluorescence decrease. A Cu 2§ to naphthalene PET was reported as the most likely mechanism for this quenching. Addition of nickel triflate in CH3CN had a similar effect. The same quenching with Cu 2§ was observed in water, while Pb 2§ in water caused a fluorescence decrease of about 60%. Less than 10% fluorescence enhancement was observed in the presence of Zn 2§ and Cd 2§ The fluorescence spectrum of 79 revealed not only the expected monomer emission at 337 nm, but also a broad band, sensitive to solvent polarity, at lower energy that is typical of naphthalene excimers due to intramolecular interactions. 135 While

r - - . ~H~ ~~

HN

NH

77

Functionalized Macrocyclic Ligands

O

123

H [~~

NMe

~

N"-/ ~""

N.Me

78

MeN I~I

/7"- ~ O r,,*N

o t,, ~

-N

N-" -,,i 0

_.J o

N~,,~

{

~Me , = ,

MeN

)

80

addition of Zn 2+ to 80 in CH3CN did not cause any effect, addition of Cd 2§ to this solution caused monomer emission to increase immediately, while the excimer emission decayed quickly. This effect could be due to conformational changes on the receptor moiety that involves the naphthalene groups which prevent excimer formation. Also Pb 2+ quenched the excimer fluorescence, in this case without a strong effect on the monomer emission. A heavy atom effect has been proposed as being responsible for the different behaviors caused by Cd 2§ and Pb 2§ Compound 80 is related to 79136 in that it contains four naphthalene groups, but with 80 they are arranged as two pairs bound to the trans-related nitrogen atoms of the macrocyclic ring. Again, 80 shows an excimer emission. Nonquenching ions Zn 2§ and Cd 2§ caused very weak effects both on the total intensity and on the ratio between monomer and excimer emission, while quenching ions Pb 2§ Ni 2§ and Cu 2§ caused some decrease (36, 37, and 86%, respectively) of the total emissions, with more effectiveness towards excimeric emissions. Chemosensor 81, possessing a dioxocyclam receptor, and its open-chain analogue, 82, were recently developed by Fabbrizzi et al. 137-139 For these chemosensors, the complexation mechanism involves deprotonation of the two amide groups. This very endoergonic process can take place only with metal ions (Cu 2§ and Ni 2§

124

XUE, SAVAGE, BRADSHAW, ZHANG, and IZATT o

CH2

CH2

81

82

NH2

which profit from a large ligand field stabilization. This makes ligands of this kind selective for these two cations. Their different ligand field stabilzations (higher for Cu 2§ with respect to Ni 2§ also allows one to distinguish between them. Complexation leads to an almost complete quenching of the luminescence of the appended chromophore, most probably via an energy transfer mechanism to a metal-centered state. Sensors 81 and 82 differ in the following aspects which depend on the cyclic or noncyclic structure of the receptor: (i) titration profiles for 81 for both Cu 2§ and Ni 2+ are shifted to higher pH values due to the thermodynamic macrocyclic effect; and (ii) the separation of Cu 2+ and Ni 2+ profiles is narrower for 81 because Ni 2+ profits from the macrocyclic effect to a larger extent than the larger Cu 2+. This effect causes a disadvantage in that for 81 the pH interval available for selective titration (in which the two ions can be discriminated) is reduced. In anthracenocryptand 23, exciplexes are formed between the anthracene and nitrogen ion pairs. 60'61'63'65'140In MeOH, the quantum yield dramatically decreases due to the formation, via exciplex intermediates, of nonfluorescent radical ions. Upon addition of an excess of K § Ag § or TI* to methanolic solutions of 23, 1:1 cryptate-type complexes are formed. 61'65 Complexation causes drastic changes in the spectroscopic properties. Cations such as Na +, for example, decrease the intensity of the exciplex emission and increase the intensity of the structured anthracene emission. Heavy-metal ions (Ag +, T1+) interact strongly with the central ring of anthracene as shown by an exciplex-type emission observed for the Ag + complex of 23. Complexation studies of sensors 83 and 84 with the alkali, alkaline earth, and transition metal ions Cu 2+, Ni 2+, Zn 2+, Cd 2§ and Pb 2+ were performed in Valeur's laboratory. 141'142 In these compounds, the chromophores are linked to the azacrowns in such a way that the cations interact directly with the carbonyl group that acts as an electron-withdrawing unit. Therefore, when an ion is coordinated,

CF3

~"~O'~

8 3 n - 1,2

CF3 ~'-

/--X o o

"-~

84 n.O,1

CF3

Functionalized Macrocyclic Ligands

125

the excited state is more stabilized than the ground state so that both absorption and luminescence spectra are red-shifted. In addition to the spectral shifts, large changes in fluorescence intensities have been observed upon ion binding. Complexation reduces the efficiency of PET from the nitrogen atom of the macrocycle that partially quenches the coumarin chromophore in 83 and 84. For 83, fluorescence enhancements were observed with Ni 2§ Zn 2§ Cd 2§ and Pb 2§ Two-component system 85,143'144in which an anthracene subunit has been linked through an ester bridge to a thiacyclam, was designed and investigated as a fluorescent chemosensor for Cu 2§ In ethanolic solution, 85 selectively incorporates Cu 2§ into its tetrathia crown component in the presence of other 3d metal ions and signals this recognition through quenching of the fluorescence of the anthracene fragment. Quenching of the photoexcited state takes place through a PET from the fluorophore to the metal ion. The d 1~ ion Ag § is shown to compete with Cu 2§ in ethanol, but not in CH3CN. However, complexation with Ag § can not be signaled since it does not cause sharp fluorescence changes due to a poor redox activity of nontransition ion Ag § Water-soluble cyclophane 86145 exhibited a well-defined fluorescence band at 290 nm with a 210 nm excitation. The emission intensity was markedly increased by complexation with Zn 2§ which forms a 2:1 (metal-ligand) complex. The fluorescence emission is pH-independent to pH 2. The fluorescence enhancement factor is 5.0 at pH 6 and 50 at pH 8.6 (due to the pH dependence of the free ligand). Ni 2§ and Cu 2§ ions quenched the ligand fluorescence via a PET mechanism. Furthermore, when cyclophane 86 was coordinated to Cu 2§ the molar absorptivity of the g - n * transition band observed at 260 nm was increased by a factor of about 10. Such a large spectral change was not observed for the Zn 2§ and Ni 2+complexes. In the Cu 2§ complex, the two phenyl rings of the cyclophane are expected to be

o

s/---ks

" iJ 86

126

XUE, SAVAGE, BRADSHAW, ZHANG, and IZATT

brought closer, as a result of the coordination of depronated amide nitrogens to the central metal ion. Macrotetracyclic cryptate 87 containing a Zn-porphyrin complex as a photoactive subunit and two lateral [18]-N20 4 macrocyclic receptors was developed by Lehn and co-workers. 146In a CHCI3/CH3OH (9" 1) solution, the two receptors were able to complex Ag § to form a polymetallic cryptate, 87-Ag § The Ag § ion in the polymetallic cryptate quenches the typical Zn- porphyrin fluorescence by a 10-fold factor via a PET process and by forming a long-lived charge-separated state. Three heteroditopic cryptands 88-90 with different cavity dimensions have been reported by Bharadwaj and co-workers. 147'148 Interestingly, the fluorophores in these systems do not show any fluorescence due to an efficient PET process from the nitrogen lone pairs. However, fluorescence can be recovered to different extents in the presence of different metal ions and with protons. Upon complexation by many transition metal ions (Mn 2§ Fe 3§ Co 2§ Ni 2§ Cu 2§ Zn 2§ and Pb 2§ or on protonation in THF, each ligand exhibits large fluorescence enhancements as the nitrogen lone pairs responsible for PET are engaged in bonding, and no exciplex emission could be observed. The more rigid receptor 90 shows lower enhancement, while the more flexible one (89) gives the highest enhancement. Among different metal ions, Zn 2§ induces the highest fluorescence enhancement. This is an expected result since complexation with this species makes the PET process thermodynamically forbidden without introducing other quenching mechanisms. Inner transition metal ions like Eu 3§ and Tb 3§ show remarkable discrimination and give high fluorescence enhancement only in the case of 88 where the cavity size is smaller than that of the other two receptors. Each system also exhibits a large fluorescence enhancement with Pb 2§ This study shows that transition metal ions and Pb 2§ which are well known for quenching, can indeed lead to fluorescence enhancement in cryptand-based systems. It was also reported for the first time that inner transition metal ions can also cause fluorescence. The noticeably high enhancement in each case was interpreted in terms of a communication gap between the metal ion and fluorophore that would avoid efficient energy or electron transfer processes. Silver

qo o]

[o RJ

87

FunctionalizedMacrocyclic Ligands

127

C

88

N

V N/

90

(I) and Hg 2§ ions were not able to increase the intensity of the fluorescence, probably because they do not enter into the cavity.

4.2. Macrocycles Functionalized with the Dansyl Chromophore as Zinc (11) Chemosensors The development of selective and sensitive chemosensors for quantitative analysis of Zn 2§ has become extremely important for environmental and biological applications. 149'15~While remarkable development has been made for other biologically important divalent metal ions such as Ca 2§ which has a few selective fluorophores (e.g. Fura-2, Quin-2), TM there are few Zn 2§ selective analytical reagents available. Recently, chemosensor 91 for Zn 2§ was developed by Kimura and co-workers 152'153 using the cyclen macrocyclic polyamine system. The success of this fluorophore is due to a high affinity of cyclen toward Zn § a strong affinity toward aromatic sulfonamides by the Zn 2§ complex, and a good luminescence property of the dansyl chromophone, all of which allow this system to sense Zn 2§ at low concentrations. Since metal-ion complexation of 91 requires the deprotonation of the sulfonamide group, the association constant is pH-dependent and it is very high at physiological levels (e.g. 7 x 109 M -1 at pH 7.0 and 2 x 1012 M -1 at pH 7.8). Complexation with Zn 2§ at physiological pH induces a blue shift of the

128

XUE, SAVAGE, BRADSHAW, ZHANG, and IZATT N(CH3)=

k_J

91 fluorescence peak from 582 to 540 nm, and increases the emission intensity by 5-fold at 540 nm and 10-fold at 490 nm. The Cd2§ complex similarly emits fluorescence. On the other hand, Cu 2§ linearly diminished the fluorescence emission until the fluorescence was completely quenched. Other quenching metal ions, including easily reducible Hg 2§ Pb 2§ or paramagnetic Co 2§ which tend to bind strongly with cyclen, also caused minor intramolecular quenching, although the effects were not so important as with Cu 2§ The Zn2§ fluorescence is unaffected by the presence of mM concentrations of the important biological metal cations Na § K § Ca 2§ and Mg 2§ Since Cd 2§ Hg 2§ and Pb 2§ are not present at significant concentrations and Cu 2§ can be removed by using bovine serum albumin (BSA), 91 may be used as a zinc fluorophore to quantify trace Zn 2§ concentrations. Nevertheless, the synthesis of dansylamidoethyl-pendent cyclen 91 seems to be impractical because it required four steps and the overall yield was low. 152 Very recently, we developed an efficient and convenient one-step methodology for the introduction of the dansyl chromophore onto macrocyclic polyamines. 154 Our synthetic strategy is based on ring opening of N-dansylaziridine caused by the secondary amine functions of a cyclic polyamine (Scheme 5). Thus, bis(dansylamidoethyl)-substituted cyclen, cyclam, diaza- 12-crown-4, diaza- 15-crown-5, diaza18-crown-6, and a series of diazadithia- and tetraza-15-(and-16)-crown-5 ligands 92-101 have been prepared in high yields (82-91%). Complexation and photophysical studies on these new fluorophores with transition metal ions are in progress.

4.3. Macrocycles Functionalized with 8-Hydroxyquinoline and Its Derivatives as Transition Metal Ion Sensors Studies of UV-vis spectra suggest that several bis-8-hydroxyquinoline-substituted aza-crown ethers are potential chemosensors for Cu 2§ As shown in Figure 8, addition of Cu 2§ to an aqueous acetate-buffered solution of a tetraaza-15-crown-5 ligand functionalized with two 8-hydroxyquinoline groups (102) leads to a n e w peak at 258 nm which develops gradually. 155The other transition and posttransition metal ions, Co 2§ Ni 2§ Zn 2§ Cd 2§ Hg 2§ Pb 2§ and Ag § (only Zn 2§ Cd 2§ and Pb 2§

FunctionalizedMacrocyclic Ligands

129

~ /N

CH3CN reflux

02

NMe2

92 n-m=l 93 n=l, 94

m=2

n=m=2

R-/-'XN-s~ H

Me"~Jn R 95

n=0

96 n = l

~NMe2

\ LO H 97 n=0

99 X=O

98 n = l

100 X=S

OH

101

Scheme5. are shown in Figure 8) do not produce such a peak. The visible spectra (to 800 nm) of ligand 102 in the presence of the above metal ions have been also examined. 156 Another unique absorption maximum of ligand 102 at 620 nm is caused by Cu 2+. In the presence of all of the other metal ions mentioned above, the solutions gave no absorption. In addition, an aqueoussolution of any one compound from 102 to 105 turns to blue as long as C u 2+ is added. The same solutions containing all other metal ions studied maintains a yellow color. Hence, compounds 102-105 could be used as chemosensors for Cu 2.. The UV-vis spectra of free and metal-ion complexed ligand 106 in MeOH are shown in Figure 9. TMK + and Ba 2+ ions cause no significant changes in the absorption bands. In the presence of Zn 2+ and Cu 2§ however, the ligand spectrum experienced a large change. As Zn 2+ is added to the ligand solution, the absorption maxima at 242 and 372 nm shift to 250 and 400 nm, respectively. The absorption peak of the Zn2+-106 complex at 400 nm may be used to monitor Zn 2+ concentrations in solution because the other metal ions studied do not have absorption at that wavelength.

o.

(,,

Ue"~"Me

o.

102 103 104 105

X=O,n=0 X==O, n - 1 X=S,n=0 X=S,n-1

130

XUE, SAVAGE, BRADSHAW, ZHANG, and IZATT

o..l <

m

~

C

0.6

0.4

"'7" 'i

<

02. "%..~""~' .. .~. ...2. . . . .

0 200

225

250

275

.. . . . . . . . 300

Way 9

,.

325

~)

Figure 8. UV-vis spectra of free and complexed 102 in an acetic acid buffer solution (pH = 4.7); (a and b) Cu 2+ (1 and 2 eguiv, respectively); (c) Zn 2+, (d) Pb 2+, and (e) Cd 2+ (2 equiv each); [102] = 0.977 x 10 -~ M and [buffer] = 5.0 x 10-2 M acetic acid.

In contrast to the behavior of 68, an efficient chemosensor for Mg2+, 125compound 107 has been shown to be a potential chemosensor for Hg2+.157 Ligand 107 forms stable 1:1 complexes with Hg 2+, CUE+, Cd2§ Zn2+, Ni2+, and Mg 2+ in methanolwater (1:1 vol:vol). Much lower association constants (log Ka < 2.5) are observed with the other alkaline earth ions, whereas no complexation is detected (log Ka < 1.5) for alkali metal ions. Complexes with Cu 2§ and Ni 2+ are not luminescent as ~,

-j"

b~:~ '..,.."

,~

o o.,-~ ...... ' I ; / I =u

...,

.f~

~''

i;

.,

'. . . . . . . . . . . . . . . . . . . . . . . ,,

id

"

'.

~ 0.4

~s

2so

275

~o

32s

~40

37s

4oo

,so

Wavelength (nm)

Figure 9. UV-vis spectra of free and complexed 106 in MeOH; (a) 1.3 x I 0 -s M 106, (b) solution a with 1.0 x10 -3 M K+, (c) solution a with 1.0 x 10 -3 M Ba 2+, (d) solution a with 1.0 x 10-3 M Zn2+; and (e) 5.0 x 10 -6 M 106 + 3.8 x 10 -4 M Cu 2+.

Functionalized Macrocyclic Ligands

131

Co o-

_

<,-o

X__J

106

expected, and small increases in the luminescence band can be detected upon addition of Cd 2+ (with an enhancement factor, EF - 3.0; ~'max- 466 nm), Zn 2+ (EF - 3.4; ~'max = 467 nm), or Mg 2+ (EF - 1.7; )~max -- 470 nm). A much higher luminescence increase (EF - 12; ~'max= 476 nm) is observed for Hg 2+ (Figure 10). Changes in the absorption and luminescence spectra concomitant with complexation allow determination of association constants of 107 in methanol-water (1"1 vol:vol) with Mg 2+ (1.2 x 105 M -1) and Ni 2+ (3.6 x 105 M-l), while with Zn 2+, Cu E+, Cd 2+, and Hg 2+ only a lower limit (1 x 108 M -l) can be measured. However, competition experiments with 107 in the presence of 1 equiv of Zn 2+ and Cd 2+ show that Hg 2+ ions can replace these ions with high efficiency, leading to the formation of the luminescent mercury complex, indicating that the association constant with the latter ion is likely at least 2 orders of magnitude higher than those with zinc and cadmium. In contrast, only a partial recovery of the luminescent band was obtained when 1 equiv of Hg 2+ is added to the 107-Cu 2+ complex, suggesting that 107 has

1.0 0.8

1.0 I -

II/"

~'x \ \

///,,t'-,,\\\

0.6

Iilll/:'..\\X\\

0.4

":'.

-t

t.""

o.oT,,,,,,

o.o

0.2

0.0

~176176176176176176

." o.,

9

....

, .... ,.o

q ofHg 2§

I

I

I

I

450

500

550

600

,.,

650

Figure 10. Fluorescence spectra (~.exc= 423 nm) of 107 (2.5 x 10 -5 M) in methanol-

water (1 "1 vol:vol)with increasing amounts of Hg2. ions. Inset: Fluorescence intensity values (~.exc = 423 nrn, )~em = 476 nm) vs. equivalents of added Hg 2+.

132

XUE, SAVAGE, BRADSHAW, ZHANG, and IZATT

NO2

co

/--k

OH

~--.O

O- /

N.02

OH

107 similar affinities for copper and mercury. Nevertheless, interference due to the presence of copper ions, which do not form a luminescent complex with 107, can be observed only when saturation of 107 occurs; otherwise, all of the mercury is complexed, giving its characteristic luminescence. Finally, no interference is observed while performing titrations with Hg 2§ in a complex matrix containing alkali and alkaline earth metal ions present in hundred-fold excess.

5. CONCLUSION Decades of research expanding the number of macrocyclic ligands that form complexes with metal cations has led to a wealth of information about selective macrocycle-ion interactions. This information has been applied in the development of multiple chromophore and/or fluorophore appended macrocycles that complex metal cations and respond to complexation via changes in absorbance or fluorescence properties. These appended macrocycles constitute metal ion chemosensors. Potential applications of these chemosensors range from measuring ion concentrations in biological samples to monitoring toxic metal-ion concentrations in industrial waste effluent streams to detecting metal ions in drinking water. Issues that impede the use of many of these chemosensors include lack of ion selectivity and matrix interference in the observation of absorbance and/or fluorescence changes. As selective chemosensors are prepared, and as the need for monitoring metal-ion concentrations is better recognized, it is likely that these derivatized macrocycles will find significant and widespread use.

REFERENCES 1. Czarnik, A.W. Chem. Biol. 1995, 2, 423. 2. Foulkes, E. Biological Effects of Heavy Metals; CRC Press: Boca Raton, FL, 1990, Vols. I and II. 3. Lester, J. Heavy Metals in Wastewater and Sludge Treatment Processes; CRC Press: Boca Raton, FL, 1987; Vol. I., pp. 105-124. 4. de Silva, A. P.; Gunaratne, H. Q. N.; Gunnlaugsson, T.; Huxley, A. J. M.; McCoy, C. P.; Rademacher, J. T.; Rice, T. E. Chem. Rev. 1997, 97, 1515. 5. Bakker, E.; B0hlmann, P.; Pretsch, E. Chem. Rev. 1997, 97, 3083. 6. B0hlmann, P.; Pretsch, E.; Bakker, E. Chem. Rev. 1998, 98, 1593. 7. Lehn, J.-M. Agnew. Chem., Int. Ed. Engl. 1990, 29, 1304.

Functionalized Macrocyclic Ligands 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. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49.

133

Balzani, V.; Scandola, E Supramolecular Photochemistry; Ellis Horwood: Chichester, 1991. Seitz, W. R. CRC Crit. Rev. AnaL Chem. 1988, 19, 135. Zhujun. Z.; Seitz, W. R. Anal. Chint Acta 1985, 171, 251. Chau, L. K.; Porter, M. D. Anal. Chent 1990, 62, 1964. Wang, K.; Seiler, K.; Rusterholz, P.; Simon, W. Analyst 1992, 117, 57. Oehme, I.; Prattes, S.; Wolfveis, O. S.; Mohr, G. J. Talanta 1995, 47, 595. lzatt, R. M.; Bradshaw, J. S.; Nielsen, S. A.; Lamb, J. D.; Christensen, J. J.; Sen, D. Chem. Rev. 1985, 85, 271. lzatt, R. M.; Pawlak, K.; Bradshaw, J. S.; Bruening, R. L. Chent Rev. 1991, 91, 1721. lzatt, R. M.; Pawlak, K.; Bradshaw, J. S.; Bruening, R. L. Chem. Rev. 1995, 95, 1231. Inoue, Y.; Gokel, G. W. Cation Binding by Macrocycles; Marcel Dekker: New York, 1990. Gokel, G. W. Chem Soc. Rev. 1992, 21, 39. Gokel, G. W.; Schall, O. E Comprehensive Supramolecular Chemistry; Gokel, G.W., Ed.; Pergamon Press: Oxford, 1996, Vol. 1, pp. 97-152. Czarnik, A. W.; Ed. Fluorescent Chemosensors for Ion and Molecule Recognition; ACS: Washington, DC, 1992. Bryan, A. J.; de Silva, A. P.; de Silva, S. A.; Rupasinghe, R. A. D. D.; Sandanayake, K. R. A. S. Biosensors 1989, 4, 169. Tan, W.; Shi, Z. Y.; Smith, S.; Bimbaum, D.; Kopelman, R. Science 1992, 258, 778. Xie, X. S. Acc. Chem. Res. 1996, 29, 598. Goodwin, P. M.; Ambrose, W. P.; Keller, R. A. Acc. Chem. Res. 1996, 29, 607. Moemer, W. E. Acc. Chem. Res. 1996, 29, 563. Moemer, W. E.; Basche, T. Agnew. Chem., Int. Ed. Engl. 1993, 32, 457. Helgeson, R. C.; Czech, B. P., Chapoteau, E.; Gebauer, C. R.; Kumar, A.; Cram, D. J. J. Am. Chem. Soc. 1989, 111, 6339. Reichardt, C.; Asharin-Fard, S. Angew. Chem., Int. Ed. Engl. 1991, 30, 558. l.~hr, H. G., V'ogtle, E Acc. Chem. Res. 1985, 18, 65. Hayashita, T.; Takaji, M.; In Comprehensive Supramolecular Chemistry; Gokel, G.W.; Ed.; Pergamon Elsevier Science: Oxford, Tarrytown, Tokyo, 1996; Vol. 1; pp. 635--669. Peterson, C. J. J. Am. Chem. Soc. 1967, 89, 7017. Sousa, L. R.; Larson, J. M. J Am. Chem. Soc. 1977, 99, 307. Shizuka, H.; Takada, K.; Morita, T. J. Phys. Chent 1980, 84, 994. Wolfbeis, O. S.; Offenbacher, H. Monatstt Chem. 1984, 115, 647. Shirai, M.; Tanaka, M. J. Chem. Soc., Chem. Commun. 1988, 381. Shinkai, S.; Ishikawa, Y.; Shinkai, H.; Tsuno. T.; Manabe, O. Tetrahedron Lett. 1983, 24, 1539. Shinkai, S.; lshikawa, Y.; Shinkai, H.; Tsuno, T.; Makishima, H.; Ueda, K.; Manabe, O. J. Am. Chem. Soc. 1984, 106, 1801. Street, K. W.; Krause, S. A. Anal. Len. 1986, 19, 735. Fery-Forgues, S.; Le Bris, M. T.; Guette, J. P.; Valeur, B. J. Phys. Chem. 1988, 92, 6223. Bourson, J.; Valeur, B. J. Phys. Chem. 1989, 93, 3871. Bissell, R. A.; de Silva, A. P.; Gunaratne, H. Q. N.; Lynch, P. L. M.; Maguire, G. E. M.; Sandanayake, K. R. A. S. Chent Soc. Rev. 1992, 21, 187. Gocmen, A.; Bulut, M.; Erk, C. Pure Appl. Chem. 1993, 65, 447. Blackburn, C.; Bai, M.; Le Compte, K. A.; Langmuir, M. E. Tetrahedron Lett. 1994, 35, 7915. Crossley, R.; Goolamali, Z.; Gosper. J. J.; Sammes, P. G.J. Chem. Soc., Perkin Trans. 2 1994, 513. Crossley, R.; Goolamali, Z.; Sammes, P. G. J. Chem. Soc., Perla'n Trans. 2 1994, 1615. Alonso, M. T.; Brunet, E.; Hemandez, C.; Rodriguez-Ubis, J. C. Tetrahedmn Lett. 1993, 34, 7465. Alfimov, M. V.; Gromov, S. P.; Lednev, I. K. Chem. Phys. Lett. 1991, 185, 455. He, G. X.; Wada, E; Kikukawa, K.; Shinkai, S.; Matsuda, T. J. Chem Soc., Chem. Commun. 1987, 1294. He, G. X.; Wada, E; Kikukawa, K.; Shinkai, S.; Matsuda, T. J. Org. Chem. 1990, 55, 548.

134

XUE, SAVAGE, BRADSHAW, ZHANG, and IZATT

50. Das, S.; Thomas, K. G." Thomas, K. J." Kamat, P. V.; George, M. V. J. Phys. Chem. 1994, 98, 9291. 51. Thomas, K. J.; Thomas, K. G.; Manojkumar, T. K.; Das, S.; George, M. V. Proc. Indian. Acad. Sci. (Chem. Sci.) 1994, 106, 1375. 52. Minta, A.; Tsien, R. Y. J. Biol. Chem. 1989, 264, 19449. 53. Smith, G. A.; Hesketh, T. R.; Metcalf, J. C. Biochem. J. 1988, 250, 227. 54. Golchini, K.; Mackovic-Basic, M.; Gharb, S. A.; Masilamani, D.; Lucas, M.; Furtz, I. Am. J. Physiol. 1990, 285, F438. 55. de Silva, A. P.; Gunaratne, H. Q. N.; Gunnlaugsson, T.; Nieuwenhuizen, M. J. Chem. Soc., Chem. Commun. 1996, 1967. 56. de Silva, A. P.; Rupasinghe, R. A. D. D. J. Chem. Soc., Chem. Commun. 1985, 1669. 57. de Silva, A. P.; Gunaratne, H. Q. N.; Sandanayake, K. R. A. S. Tetrahedron Lett. 1990, 31, 5193. 58. Tsien, R. Y. Biochemistry 1980, 19, 2396. 59. de Silva, A. P., Sandanayake, K. R. A. S. J. Chem. Soc., Chem. Commun. 1989, 1183. 60. Konopelski, J. P.; Kotzyba-Hibert, E; Lehn, J.-M.; Desvergne, J. P.; Fages, E; Castellan, A.; Bouas-Laurent, H. J. Chent Soc., Chem. Commun. 1985, 433. 61. Fages, E; Desvergne, J. P.; Bousa-Laurent, H.; Marsau, P.; Lehn, J.-M.; Kotzyka-Hibert, E; Albrecht-Gary, A. M.; AI-Joubbeh, M. J. Am. Chem. Soc. 1989, 11, 8672. 62. de Silva, A. P.; de Silva, S. A. J. Chem. Soc., Chem. Commun. 1986, 1709. 63. Fages, E; Desvergne, J. P.; Bouas-Laurent, H. J. Am. Chem. Soc. 1989, 111, 96. 64. Marsau, P.; Bouas-Laurent, H.; Desvergne, J. P.; Fages, E; Lamotte, M.; Hinschberger, J. Mol. Cryst. Lz'q. Cryst. Inc. Nonlin. Opt. 1988, 156, 383. 65. Fages, E; Desvergne, J. P.; Bouas-Laurent, H.; Hinschberger, J.; Marsau, P.; Petraud, M. New J. Chem. 1988, 12, 95. 66. Bissell, R. A.; de Silva, A. P.; Fernando, W. T. L. M.; Patuwathavithana, S. T.; Samarasinghe, T. K. S. D. Tetrahedron Lett. 1991, 32, 425. 67. Olsher, U.; lzatt, R. M.; Bradshaw, J. S.; Dalley, N. K. Chem. Rev. 1991, 91, 137. 68. Tosteson, D. Sci. Am. 1981, 244, 164. 69. Jefferson, J. W.; Greist, J. H.; Baudhuir, M. Lithium-Current Applications in Science, Medicine, and Technology; Bach, B. O., Ed.; Wiley: Interscience: New York, 1985. 70. Lazarus, J. H.; Collard, K. J. Endocrine and Metabolic Effects of Lz'thium, Plenum, New York, 1986. 71. Micheloni, M. Comments Inorg. Chem. 1988, 8, 79. 72. Bencini, A.; Bianchi. A.; Borselli, A.; Ciampolini, M.; Dapporto, P.; Garcia-Espafia, E.; Micheloni. M.; Paoli, P.; Ramirez, J. A., Valtancoli, B. Inorg. Chem. 1989, 28, 4279. 73. Bencini, A.; Bianchi, A.; Borselli, A.; Chimichi, S.; Ciampolini, M.; Dapporto, P.; Micheloni, M.; Nardi, N.; Paoli, P.; Valtancoli, B. Inorg. Chem. 1990, 29, 3282. 74. Bencini, A.; Bianchi, A.; Chimichi, S.; Ciampolini, M; Dapporto, P.; Garcia-Espafia, E.; Micheloni, M.; Nardi, N.; Paoli, P.; Valtancoli, B. lnorg. Chem. 1991, 30, 3687. 75. Ciampolini, M.; Micheloni, M.; Nardi, N.; Valtancoli, B. Coord. Chem. Rev. 1992, 120, 223. 76. Bazzicalupi, C.; Beucini, A.; Bianchi, A.; Ciampolini, M.; Fusi, V.; Micheloni, M.; Nardi, N.; Paoli, P.; Valtancoli, B. Supramol. Chem. 1994, 3, 279. 77. Bencini, A.; Fusi, V.; Giorgi, C.; Micheloni, M.; Nardi, N., Valtancoli, B. J. Chem. Soc., Perkin Trans. 2 1996, 2297. 78. Bardazzi, E.; Ciampolini, M.; Fusi, V.; Micheloni, M.; Nardi, N.; Pontellini, R.; Romani, P. J. Org. Chem. 1999, 64, 1335. 79. Formica, M.; Fusi, V.; Micheloni, M.; Pontellini, R.; Romani, P. Coord. Chem. Rev. 1999, 184, 347. 80. Ciampolini, M.; Formica, M.; Fusi, V.; Saint-Mauvicec, A.; Micheloni, M.; Nardi, N.; Pontellini, R.; Pina, E; Romani, P.; Sabatini, A. M.; Valtancoli, B. Eur. J. Inorg. Chem. 1999, 2261. 81. Kubo, K.; Ishige, R.; Yamamoto, E.; Sakurai, T. Heterocycles 1997, 45, 2365. 82. Kubo, K.; Ishige, R.; Sakurai, T. Heterocycles 1998, 48, 347.

Functionalized Ivlacrocyclic Ligands

135

83. Kubo, K.; Kato, N.; Sakurai, T. Bull, Chem. Soc. Jpn. 1997, 70, 3041. 84. De Wall, S. L.; Meadows, E. S.; Barbour, L. J.; Gokel, G. W. J. Chem. Soc., Chem. Commun. 1999, 16, 1553 85. Meadows, E. S.; De Wall, S. L.; Barbour, L. J.; Gokel, G. W. J. Chem. Soc., Chem. Commun. 1999, 16, 1555. 86. Meadows, E. S.; De Wall, S. L.; Barbour, L. J.; Fronczek, E R.; Kim, M.-S.; Gokel, G.W.J. Am. Chem. Soc. 2000, 122, 3325. 87. Bertelson, R. C. In Photochromism; Brown, G.H., Ed.; Wdey-lnterscience: New York, 1971. 88. Sunamoto, J.; lwamoto, K.; Mohri, Y.; Kominato, T. J. Am. Chem. Soc. 1982, 104, 5502. 89. Winkler, J. D.; Deshayes, K.; Shao, B. J. Am. Chem. Soc. 1989, 111,769. 90. Kato, S.; Aizawa, M.; Suzuki, S. J. Membr. Sci. 1976, 1,289. 91. Anzai, J.; Ueno, A.; Osa, T. J. Chem. Soc., Chem. Commun. 1984, 688. 92. lrie, M.; Menju, A.; Hayashi, K. Nippon Kagaku Kaislu" 1984, 227. 93. Ryba, O.; Petranek, J. Makromal. Chem. Rapid Commun. 1988, 9, 125. 94. lrie, M.; lwayanagi, T.; Taniguchi, Y. Macromolecules 1985, 18, 2418. 95. Ciardelli, E; Fabbri, D.; Pieroni, O.; Fissi, A. J. Am. Chem. Soc. 1989, 111, 3470. 96. lnouye, M.; Ueno, M., Kitao, T. J. Am. Chem. Soc. 1990, 112, 8977. 97. lnouye, M.; Ueno, M.; Kitao, T. J. Org. Chem. 1992, 57, 1639. 98. Inouye, M.; Ueno, M.; Tsuchiya, K.; Nakayama, N.; Konishi, T.; Kitao, T. J. Org. Chem. 1992, 57, 5377. 99. Kimura, K.; Yamashita, T.; Yokoyama, M. J. Chem. Soc., Chem. Commun. 1991, 147. 100. Kimura, K.; Yamashita, T.; Yokoyama, M. Chem. Lett. 1991, 965. 101. Kimura, K.; Yamashita, T.; Tokoyama, M. J. Chem. Soc., Perkin Trans. 2 1992, 613. 102. Kimura, K.; Yamashita, T.; Yokoyama, M. J. Phys. Chem. 1992, 96, 5614. 103. Kimura, K.; Kaneshige, M.; Yamashita, T.; Yokoyamo, M. J. Org. Chem. 1994, 59, 1251. 104. Dong, G.; Sakaki, T.; Kawahara, Y.; Shinkai, S. Supramol. Chem. 1993, 2, 71. 105. Tsien, R. Y. Annu. Rev. Biophys. Bioeng. 1983, 12, 91. 106. Grubbs, R. D.; Maguire, M. E. Magnesium 1987, 6, 113. 107. Henquin, C.; Tamagaw, T.; Nenquin, M.; Cogneau, M. Nature 1983, 301, 73. 108. Hoelzl Wallach, D. E; Steck, T. L. Anal, Chem. 1963, 35, 1035. 109. Rink, T. J. Pure Appl. Chem. 1983, 55, 1977. 110. Grynkiewicz, G.; Poenie, M.; Tsien, R. Y. J. Biol. Chem. 1985, 206, 3440. 111. Tsien, R. Y.; Minta, A. Eur. Patent 1989, 0 314480 A2. 112. Minta, A.; Kao, J. P. Y.; Tsien, R. Y. J. Biol. Chem. 1989, 264, 8171. 113. Nishida, H.; Katayama, Y.; Katsuki, H., Nakamura, H.; Takagi, M.; Ueno, K. Chem. Lett. 1982, 1853. 114. Delmond, S.; IMotard, J. E; Lapouyade, R.; Mathevet, R.; Jonusauskas, G.; Rulli6re, C. New J. Chem. 1996, 20, 861. 115. Prodi, L.; Bolletta, E; Zaccheroni, N.; Watt, C. I. E; Mooney, N. J. Chem. Eur. J. 1998, 4, 1090. 116. Prodi, L.; Bolletta, E; Montalti, M.; Pivari, S.; Zaccheroni, N. Sens. Microsyst., Proc. 3rd Ital. Conf., 1998 (Pub. 1999), p. 77. 117. Inouye, M.; Noguehi, Y.; Isagawa, K.Agnew. Chem. Int. Ed. Engl. 1994, 33, 1163. 118. Inouye, M. Coord. Chem. Rev. 1996, 148, 265. 119. Kimura, K.; Teranishi, T.; Tokoyama, M. Supramol. Chem. 1996, 7, 11. 120. Kimura, K.; Utsumi, T.; Teraniski, T.; Yokoyama, M.; Sakamoto, H.; Okamoto, M., Arakawa, R., Moriguchi, H.; Miyaji, Y. Agnew. Chem., Int. Ed. Engl. 1997, 36, 2452. 121. Kimura, K.; Teranishi, T.; Yokoyama, M.; Yajima, S.; Miyake, S.; Sakamoto, H.; Tanaka, M. J. Chem. Soc., Perka'n Trans. 2 1999, 199. 122. Zhang, X. X.; Bordunov, A. V.; Bradshaw, J. S.; Dalley, N. K.; Kou, X.; Izatt, R. M. J. Am. Chem. Soc. 1995, 117, 11507.

136

XUE, SAVAGE, BRADSHAW, ZHANG, and IZATT

123. Borduov, A. V.; Bradshaw J. S.; Zhang, X. X.; Dalley, N. K.; Kou, X.; lzatt, R. M. Inorg. Chem. 1996, 35, 7229. 124. Su, N.; Bradshaw, J. S.; Zhang, X. X.; Song, H.; Savage, P. B.; Xue, G. P.; Krakowiak, K. E.; Izatt, R. M. J. Org. Chem. 1999, 64, 8855. 125. Prodi, L.; Bolletta, E; Montalti, M.; Zaccheroni, N.; Savage, P. B.; Bradshaw, J. S.; Izatt, R. M. Tetrahedron Lett. 1998, 39, 5451. 126. Ueno, K.; Imamura, T.; Cheng, K. L. Handbook of Organic Analytical Reagents, 2nd ed.; CRC Press: Boca Raton, FL, 1992. 127. Sandell, E. B.; Onishi, H. In Photometric Determination of Trace Metals; Elving, P. J.; Winefordner, J. D., Eds.; John Wiley & Sons: New York, 1978, Vol. 3, pp. 415-447. 128. Soroka, K.; Vithanage, R. S.; Philips, D. A.; Walker, B.; Dasgupta, P. K. AnaL Chent 1997, 59, 629. 129. Hancock, R. D.; Martell, A. E. Chem. Rev. 1989, 89, 1875. 130. Akkaya, E. U.; Huston, M. E.; Czamik, A. W. J. Am. Chem. Soc. 1990, 112, 3590. 131. Huston, M. E.; Engleman, C.; Czamik, A. W. J. Am. Chem. Soc. 1990, 112, 7054. 132. Czamik, A. W. Trends Org. Chem. 1993, 4, 123. 133. Czamik, A. W. Acc. Chem. Res. 1994, 27, 302. 134. Yoon, J.; Ohler, N. E.; Vance, D. H.; Aumiller, W. D.; Czarnik, A. W. Tetrahedron Lett. 1997, 38, 3845. 135. Parker, D.; Williams, J. A. G. J. Chem. Soc., Perkin Trans. 2 1995, 13, 5. 136. Beeby, A.; Parker, D.; Williams, J. A. G. J. Chem. Soc., Perkin Trans. 2 1996, 1565. 137. Fabbrizzi, L.; Licchelli, M.; Pallavicini, P.; Perotti, A.; Sacchi, D. Agnew. Chem., Int. Ed. Engl. 1994, 33, 1975. 138. Fabbrizzi, L.; Licchelli, M.; Pallavicini, P.; Perotti, A.; Taglietti, A.; Sacchi, D. Chent Eur. J. 199@ 2,75. 139. Bolletta, E; Costa, I.; Fabbrizzi, L.; Licchelli, M.; Montalti, M.; Pallavicini, P.; Prodi, L.; Zaccheroni, N. J. Chem. Soc., Dalton Trans. 1999, 1381. 140. Desvergne, J.-P.; Fages, E; Bouas-Laurent, H.; Marsau, P. Pure Appl. Chem. 1992, 64, 1231. 141. Badaoui, E; Bourson, J.; Valeur, B. J. Fluorescence 1994, 4, 275. 142. Valeur, B.; Badaoui, E; Bardez, E.; Bourson, J.; Boutin, P.; Chatelain, A.; Devol, I.; Larrey, B.; Let6vre, J. P.; Soulet, A. In Fluorescent Chemosensors of Ion and Molecular Recognition, NATO-ASI Series, Desvergne, J.-P., Czamik, A. W., Eds.; Kluwer Academic Publishers: Dordrecht, 1996. 143. De Santis, G.; Fabbrizzi, L.; Licchelli, M.; Mangano, C.; Sacchi, D. Inorg. Chem. 1995, 34, 3581. 144. De Santis, G.; Fabbrizzi, L.; Licchelli, M.; Mangano, C.; Sacchi, D.; Sardone, N. Inorg. Chim. Acta 1997, 257, 69. 145. Inoue, M. B.; Merano, E; lnoue, M.; Raitsimring, A.; Femando, Q. Inorg. Chent 1997, 36, 2335. 146. Gubelmann, M.; Harriman, A.; Lehn, J.-M.; Sessler, J. U J. Chem. Soc., Chem. Commun. 1988, 77. 147. Ghosh, P.; Bharadwaj, P. K.; Mandal, S.; Ghosh, S. J. Ant Chem. Soc. 1996, 118, 1553. 148. Ghosh, P.; Bharadwaj, P. K.; Roy, J.; Ghosh, S. J. Am. Chem. Soc. 1997, 119, 11903. 149. Vallee, B. L.; Falchuk, K. H. Physiol. Rev. 1993, 73, 79. 150. Fratisto da Silva, J. J. R.; Williams, R. J. P. In The Biological Chemistry ofthe Elements; Clarendon Press: Oxford, 1994, pp. 299-318. 151. Haugland, R. P. Handbook of Fluorescent Probes and Research Chemicals, Ed. Spence, M.T.Z., Molecular Probes, Eugene, 6th ed., 1996, p. 503. 152. Koike, T., Watanabe, T.; Aoki, S.; Kimura, E.; Shiro, M.J. Am. Chem. Soc. 1996, 118, 12696. 153. Kimura, E.; Koike, T. Chem. Soc. Rev. 1998, 27, 179. 154. Xue, G. P.; Bradshaw, J. S.; Chiara, J. A.; Savage, P. B.; Krakowiak, K. E.; Izatt, R. M. Synlett. In press.

Functionalized Macrocyclic Ligands

137

155. Yang, Z.; Bradshaw, J. S., Zhang, X. X.; Savage, P. B., Krakowiak, K. E.; Dalley, N. K.; Su, N.; Bronson, R. T.; Izatt, R. M. J. Org. Chem. 1999, 64, 3162. 156. Zhang, X. X.; Bradshaw, J. S.; Lamb, J. D.; Savage, P. B.; Izatt, R. M. unpublished observations. 157. Prodi, L.; Bargossi, C.; Montalti, M.; Zaccheroni, N.; Su, N.; Bradshaw, J. S.; Izatt, R. M.; Savage, P. B. J. Am. Chem. Soc. In press.

This Page Intentionally Left Blank

CHIRALITY IN CALIXARENES AND CALIXARENE ASSEMBLIES

Myroslav Vysotsky, Christian Schmidt, and Volker BOhmer 1.

2.

3.

4.

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.1. Scope and Limits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.2. Properties and Modifications of Calixarenes . . . . . . . . . . . . . . . Chiral Derivatives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1. Calixarencs Substituted at the Wide Rim . . . . . . . . . . . . . . . . . 2.2. Calixarencs Substituted at the Narrow Rim . . . . . . . . . . . . . . . 2.3. Chiral Derivatives of Resorcarcnes . . . . . . . . . . . . . . . . . . . . 2.4. Biological Relevant Derivatives . . . . . . . . . . . . . . . . . . . . . Stereogenic Centers within the Macrocycle . . . . . . . . . . . . . . . . . . . 3.1. Stcreogcnic Bridges . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2. Spirodienones and Similar Derivatives . . . . . . . . . . . . . . . . . . 3.3. Sphcrand Type Calixarcncs . . . . . . . . . . . . . . . . . . . . . . . . 3.4. Further Macrocyclcs Similar to Calixarencs . . . . . . . . . . . . . . . Inherently Chiral Calixarencs (C1) . . . . . . . . . . . . . . . . . . . . . . . 4.1. Calix[4]arencs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.2. Calix[5]arcnes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Advances in Supramolecular Chemistry Volume 7, pages 139-233. Copyright O 2000 by JAI Press Inc. All rights of reproduction in any form reserved. ISBN: 0..7623-0678-5

139

140 140 141 143 143 145 149 152 158 158 160 162 165 166 166 176

140

5.

6.

7.

8.

MYROSLAV VYSOTSKY,CHRISTIAN SCHMIDT, and VOLKER BOHMER 4.3. Calix[6]arenes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.4. Calix[8]arenes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.5. Calixarene Analogs Incorporating Different Bridges . . . . . . . . . . . 4.6. Resorcarenes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Calixarenes with Higher Symmetry (Cn or Dn) . . . . . . . . . . . . . . . . . 5.1. Calixarenes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.2. Energy Barriers for Partially O-Methylated Calix[4]arenes . . . . . . . 5.3. Resorcarenes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Chiral Conformations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.1. The Parent Calixarenes and Resorcarenes . . . . . . . . . . . . . . . . . 6.2. Directionality in Hydrogen Bonded Systems . . . . . . . . . . . . . . . 6.3. Other Chiral Conformations . . . . . . . . . . . . . . . . . . . . . . . . Supramolecular Chirality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.1. Host-Guest Induced Chirality . . . . . . . . . . . . . . . . . . . . . . . 7.2. Tetraurea Dimers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.3. Melamine-Barbiturate Systems . . . . . . . . . . . . . . . . . . . . . . 7.4. Chiral Assemblies in the Crystalline State . . . . . . . . . . . . . . . . Conclusions and Outlook . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References and Notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

177 179 181 183 184 184 188 190 195 195 196 199 203 203 206 213 217 221 222

1. INTRODUCTION 1.1. Scope and Limits C. D. Gutsche coined the name calixarenes for macrocyclic molecules of type I, which are usually prepared by condensation of p-substituted phenols (mainly p-tert-butylphenol) with formaldehyde under alkaline conditions. 1 Procedures have been elaborated that make the cyclic tetra-, hexa-, and octamer-selectively available from t-butylphenol in large quantities and yields >50%. 2 The parent compounds are readily modified by chemical reactions and a variety of calixarenes are available by directed syntheses. 3 These facts have led to the description of calixarenes as "the third generation" of macrocyclic molecules, 4 which follow the crown ethers and cyclodextrins. Arguably, calixarenes are still gaining interest as starting materials or basic scaffolds for the preparation of more sophisticated host molecules and as building blocks for the construction of larger molecular architectures, that are either covalently linked or self-assembled. Resorcinol (and some 2-substituted derivatives such as 2-methyl resorcinol and phloroglucinol) react with aldehydes different from formaldehyde to form cyclic tetramers II, having the same [ 14]metacyclophane skeleton. 5 Among four possible stereoisomers, the rccc, rctt, and less frequently the rcct isomers are usually available. Various names have been used for these tetramers, which will be called resorcarenes 6 in this review. Chirality is a property that is not restricted to the molecular level. An object is chiral or dissymmetric if it is not superimposable upon its mirror image. This does

Chirality in Calb~arenes and Calixarene Assemblies

141

R

R'

R

R

R' H

~

H

R'

HO- "y" "OH

-

R~

I

II

not mean, however, that it must be asymmetric, i.e. without any symmetry element. Symmetry axes (C, axes) may be present (point groups C, and D,) but a chiral object can neither have a symmetry plane, nor an inversion center nor an alternating axis. Nature makes ample use of chirality at both the molecular and supramolecular levels. Examples are known in which two enantiomers have drastically different effects. 7 The search for enantioselective artificial host molecules is therefore not only inspired by natural receptors, it is also dictated by the need to distinguish enantiomers in various fields of chemical applications. In this chapter we present an overview of chiral structures realized on the basis of calixarenes and resorcarenes. 8 Clearly this cannot be an exhaustive survey, due to the enormous development calixarene chemistry has experienced during the last decade (mainly the last 5 years). Rather, it is a survey of the potential possibilities that are offered by this class of macrocyclic molecules--illustrated by selected examples of what has been done and indicating what might be done in the future.

1.2. Properties and Modifications of Calixarenes It seems appropriate at the outset to review some basic features of calixarenes and resorcarenes. Four basic conformations have been discussed for calix[4]arenes

R

O

_

R ~

~ RR

cone

R

R

R

R

partial cone

R

R

. .

R 1,2.alternate

1,3.alternate

Figure I. The four basic conformations of a calix[4]arene.

142

MYROSLAV VYSOTSKY,CHRISTIAN SCHMIDT, and VOLKER BOHMER

from the very beginning (Figure 1). However, the parent compounds with four hydroxy groups are found exclusively in the cone conformation, stabilized by a cyclic array of intramolecular O-H-..O-H.-. hydrogen bonds. Since OH groups can pass through the annulus, a more or less rapid interconversion between two identical cone conformations occurs with an energy barrier AG $ = 15-16 kcal/mol (depending on the solvent and the para-subsfituent). 9 T h e barrier is lower for calix[5]arenes and the conformational properties of the higher members (n = 6) are more complicated. 1~ Two obvious places for chemical modification exist in a calixarene molecule: 9 The hydroxy groups at the "narrow rim" which may be alkylated or acylated. 9 The para-positions at the "wide rim" where the t-butyl groups can be removed

and replaced by all kinds of substituents. While methoxy groups, like hydroxy groups, can pass the annulus, O-alkyl/acyl residues such as propyl (or larger residues) cannot. Even a single residue of this size prevents the complete ring inversion, while four residues fix the calix[4]arene molecule in one of the four basic conformations. Among the four possible stereoisomers of resorcarenes II, the rccc isomer, which assumes a cone (or crown) conformation with an axial position of the residues R, was mainly used as a building block for larger structures. Again, chemical modification is possible by O-alkylation or O-acylation and by (mild) electrophilic substitution in the 2-position of the resorcinol units. Intramolecular bridging of oxygen functions of adjacent units leads to molecules III with an enforced cavity; this explains the name cavitands chosen by D. J. Cram for these derivatives. II If bridges between the 2-positions of the resorcinol units connect two of these rigid, bowl-shaped molecules, container molecules IV with a closed cavity are formed. Depending on the size of the guest and the number and length of the bridges, this inclusion can be permanent or the escape/exchange is still possible under drastic conditions. These properties explain the names carcerand (carceplex for the complex) and hemicarcerand coined by Cram. 12

XIO'~lsO~x

/

O

N%N \

O

R

X~o~o~X III

RR

R

RR

R

R

RR

R

R

IV

Chirality in Calixarenes and Calb~areneAssemblies

143

2. CHIRAL DERIVATIVES Like every (macrocyclic) molecule, a calixarene can be converted into a chiral compound by the attachment of various chiral residues. 13 Although this approach does not make (in terms of chirality) any use of the nonplanar, spherical structure of the calixarene skeleton, it is often the most simple one, and can lead to very efficient enantioselective host molecules. Chiral groups may be attached to both the narrow and the wide rims of the calixarene skeleton I (or even to the bridges 14) and this may be done by chemical modification of the calixarene or by the use of appropriate chiral phenols for the calixarene synthesis. With resorcarenes H chiral groups may be introduced at the hydroxy groups or at the 2-position between the OH-functions. Their attachment at the bridges (CHR*) was also realized.

2.1. Calixarenes Substituted at the Wide Rim Recent examples of calixarenes having chiral substituents attached to the wide rim are calix[n]arenes l a and l b which were obtained by one-pot synthesis from the respective p-substituted phenol in moderate 15 to low yield. 16 Starting with the corresponding di- or tetraacid derivatives, calix[4]arenes 2 and 3a were synthesized 17 bearing two or four L-alanine or L-alanyl-L-alanine residues at the wide rim. In contrast to the flexible 3a the analogues 3b rigidified

1

8

5,6,8

la

lb

5

A'l

NHNI-I2 2

NH 0

2

Y

OM.

0

Y

3a

3b

Me O HN--I--~

144 Y Y

MYROSLAV VYSOTSKY, CHRISTIAN SCHMIDT, and VOLKER BOHMER Y Y

Y Y

Y Y

Y Y

C ~ C

t-Su ~ CH= \CH., / COOH O=C NH

CH= t-Bu t-Bu CI.I z

Y Y

0 O0 o

/ O=Ct NH

H=

/ O=C ~iH

cooc., cooc., cooc..

5a

5b Y Y b '

HOHO OHOH

Y Y ' '

CH2

/ NH eO

HN O ~.

CI-I=-NH-CO-CH=

%0

Y: C6H1=

H

6a

O OO

HN

H

6b

O

O OO O

H NH Me'~I=O

Me

O=~ H NH

HN

Me

IVle-'~ NH H HN" ~ H 0 ~.,,,N~.Ij 0

7a

oJ

O H NH

HN

IVleo~ NH

H

O Me

HN'~o H

~,.,,j N,,,~

7b

by the biscrown structure at the narrow rim showed some binding of ammonium ions and amino acid esters, but little chiral discrimination. 18 Ring-opening of cyclic anhydride 4 by aminolysis and subsequent formation of the second amide function led to the diamides, 5b. Either the meso-isomer or the C2-symmetrical chiral derivative is thus available in a controlled way.19 A calix[4]arene bearing one chiral boron Lewis acid at the wide rim was used with modest success as a catalyst in a Diels-Alder reaction. 2~ The cyclic amides 6a contain a chiral binaphthyl bridge 21 connecting two opposite p-positions. However, this fact has not been used in host-guest chemistry

Chirality in Calixarenes and Calbcarene Assemblies

145

(log K = 3.51 was reported for the complexation of Ag+). The wide rim crown-6 derivative 6b binds amino acid esters and shows a small enantioselectivity (KD/K L = 1.39 for Phe-OMe), 22 which is less pronounced than for the similar "narrow rim" crown-6 derivative (see below). More remarkable are the cyclic peptidocalixarenes 7a,b that have been prepared as vancomycin mimics. 23 In particular, compound 7b (which has the smaller ring) shows a minimum inhibitory concentration against some Grampositive bacteria, only twice as high as the concentration reported for vancomycin. For derivatives with chiral urea residues see Section 7.2.

2.2. CalixarenesSubstitutedat the Narrow Rim Probably the first examples of chiral calixarene derivatives were esters with camphorsulfonic acid 24 derived from t-butyl-calix[8]arene. ~ Later ethers obtained with (S)-2-methylbutyl bromide were synthesized and studied with respect to their CD spectra and their chiral recognition ability. 26 The calix[6]arene-derived cholesteryl hexaester $ was prepared only to inhibit the rotation of the oxygen functions through the annulus. 27 Thus, the coalescence observed by NMR for the Ar-CH2-Ar protons at higher temperature (a broad singlet is observed at 130 ~ while a pair of doublets and a singlet at 30 ~ are in agreement with a 1,2,3-alternate conformation), indicates the possibility that the p-substituents rotate through the annulus. Calixarene esters are easily available by alkylation with ethyl bromoacetate and are often used as starting materials for the introduction of chiral groups at the narrow rim. Their aminolysis by chiral amines led to chiral calixarene derivatives in high yields. Water soluble calix[4]arene amino acid derivatives 9a,b obtained in this way, were successfully used as a pseudostationary phase

0

0

~t.Bu "6

146

MYROSLAV VYSOTSKY,CHRISTIAN SCHMIDT, and VOLKER BOHMER

in capillary electrophoresis (CE) under neutral and basic conditions. 28'29 Separation of some racemic binaphthol derivatives (10) was achieved. A good enantio-separation of phenylglycinol was also reported for compound 9e, derived from (S)-di-2-naphthylprolinol in open tubular CE experiments (ge coated as stationary phase to the inner surface of the capillary). 3~ The determination of the enantiomeric compositions of some chiral amines was based on the quenching of the fluorescence of 9e and similar calixarenes derived from (R)-(-)-l-(9-anthryl)2,2,2_trifluoroethanol.3 z Y:.

4

"'"~"J"Cs fi 9b

9C

lOa

10c

lOb

An interesting application is the hydroformylation of styrene using calix[4]arene 11 as a homogeneous catalyst. 32 A highly regioselective formation of 2-phenylpropanal over 3-phenylpropanal (95:5) was observed. However, the possible enantioselectivity of the reaction due to the chiral amido groups was not even discussed.

H3 '1"1""~1,H

/\

t-Bu

H,I,I,N CH3

t-Bu

11

MoO- Z

O

k t-Bu

12a

Me<:)-

.:yo 2

~4

12b

Chiral calix[4]arene podands were made using N-benzyl histidine methyl ester. 33 These histidyl calixarenes 12a,b were studied in complexation experiments with Co(N), but no use was made of their chirality. The same is true for a chiral calix[4]arene capped tetraphenyl porphyrin which is Ca-symmetrical due to the four L-alanine derived linkers. 34 Etherification with (S)-(+)-glycidyl toluenesulfonate was also suggested to introduce chiral substituents. 35 The syn-1,3-diether and tetraethers in the partial cone, 1,2-alternate and 1,3-alternate were obtained from calix[4]arenes while calix[6]arenes gave in good yield the 1,2,4,5-tetraglycidylated product. Ring-opening reactions with NH 3 and dimethylamine leading to 13-aminoalcohols were described as examples for further derivatization.

Chirality in Calixarenes and Calixarene Assemblies

147

The reaction of 13, in which the (3-pyridyl)carbonyloxy groups can be either in syn or in anti position, with spiro[cyclopropene-3,9'-fluorene] creates two new stereogenic centers. 36 Thus two diastereomers are possible for each: the syn- and the anti-isomer which form a pair of C2-symmetrical enantiomers (R,R/S,S) and a C s- or Ci-symmetrical meso form (R,S). 37 T h e resulting calixarenes 14, bearing dihydroindolizine units, were studied as chromogenic compounds ("calixochromes") in quenching experiments not related to their chirality.

~

I,,

~1

t-Bu

13

t-Bu

") 2

14

Bridging similar to that devised for 6a and 6b was also achieved at the narrow rim (15a) 21 and the crown-6 derivative 15b with a dinaphthol segment 38 was already mentioned, showing KD/KL - 3.22 for the complexation of Phe-OMe.

O:C I HN X

t-Bu

/

O O ~

C=O I NH

~

X

I o HO OH O

t-Bu

t-Bu

t-Bu

t-Bu

C~176 0 HO OH 0

t-Bu

t-Bu

t-Bu

X = (CH2)~, (CH2)2-NH-CO-CH 2

15a

15b

More interesting are the chromogenic crown ethers 16 (Figure 2) derived from (S)-dinaphthol. For n - 0, strong spectral changes occur between 500 and 700 nm if the (R)-isomer of the guests 17 is added to an ethanolic solution. In a similar situation, the (S)-isomer exhibits no discernible change in the absorption spectrum. 39 Thus, a visual distinction of the two enantiomers is possible based only on changes in the color. Structural elements of the chiral pool have been used to synthesize the 1,3bridged calix[4]arenes 18a and 18b. In the former case the respective calix[4]arene

148

MYROSLAVVYSOTSKY,CHRISTIANSCHMIDT,and VOLKERBOHMER

N=~o Me

O~N X=

o,-,o,--,o I ~ I ~ ~

NH2

OH ~~ ] ~O H 17a

16

~ O

NH:,

OH

17b

n=0,1

Figure 2. Chromogenic chiral crown ethers 16 and guests 17 for which the enantiomers can be distinguished by the color.

was alkylated by the bis(chloroacetate) of various dialkyl tartrates, 4~while L-cystine dimethyl ester 41 or several bis(cysteins) 42 were reacted with the diacyl chloride of the calixarene derived 1,3-diacid in the latter case. Nothing is reported about the chiral properties of these compounds.

R'OOC COOR' ~CH"dH

O ~,0

0~0

/ R%

O~NH

HN .0

OHO OHOy

t-Bu t-Bu t-Bu t-Bu R = (CHz)2, (CH2)4

R = H, C(CH~3,CH2CH=CH2

CHz-~CH2 18b

18a

y. u N~ "'N

19a ,'~3

t

Bu

19b

Chirality in Calixarenes and Calixarene Assemblies

149

There was reported an interesting example in which copper(l) complexes of the bis(bipyridyl) ether 19a are chiral (C2-symmetry) and crystallize in form of their racemate. 43 The additional attachment of two chiral (S)-2-methylbutyl ether groups leads to the chiral ligand 19b which consequently can form two diastereomeric copper(I) complexes, from which the complex with the left-handed prohelical [CuI(bpy)2] substructure is found in 30% excess. 44 2.3.

Chiral Derivatives of Resorcarenes

The free 2-position of the four resorcinol subunits can be substituted by Mannich reaction under mild conditions with formaldehyde and appropriate amines. The use of chiral secondary amines, such as proline, led to the first known examples of resorcarenes bearing four chiral substituents. 45-47 The use of the water soluble derivative 20 as a non-lanthanide, chiral NMR-shift reagent for a chiral aromatic guest was proposed. 46 Under similar reaction conditions, the 1,3-oxazolidine derivatives 21 were formed exclusively with chiral o~-substituted aminoethanols. 48 Obviously the secondary amine, formed in the aminomethylation step, reacts with a further molecule of formaldehyde (present in excess) by ring closure between the amino function and the aliphatic hydroxy group. The introduction of only two chiral groups has been reported by starting with resorcarene tetrasulfonates. 49 Here the aminomethylation takes place only in the 2-position of the unsubstituted resorcinol rings yielding distally disubstituted resorcarenes 22. The resorcarene 23a, on the other hand, was obtained by O-alkylation of the corresponding octol. Its incorporation in a dimethylpolysiloxane backbone led to a stationary phase by which proteinogenic amino acids could be separated into their enantiomers by GC of their N(O,S)-trifluoroacetylmethylesters with separation factors CtL,D = 1.025-1.102. 50 The question remains in this case (and in similar cases), whether the chiral amide functions have to be attached to the resorcarene skeleton, or if a direct attachment to the polymer backbone via suitable spacers would lead to similar results. The chiral resorcarene octaamides 23b prepared by

R

20

C

R

21

R':

a Me d iBu b Et 9 Ph C iPr

22

Y: SO2

150

MYROSLAV VYSOTSKY,CHRISTIAN SCHMIDT, and VOLKER BC)HMER ........ -Si ........

O

HNR~z 1 H Ritz

1

R 1 = CH 3 . C-~Hs R z = Cell s , C6Hll, CH2OH

23b

23a

aminolysis of resorcarene octaethyl ester with chiral amines and aminoalcohols were recently described. 5] The reaction of resorcarene with (-)-bis(N,N-diethylamido)menthylphosphite (BAMP) led to various chiral organophosphorus derivatives. 52 By acid-catalyzed condensation of resorcinol with (4-formylphenyl)boronic acid, resorcarenes in the rccc- and rctt-configuration bearing four phenylboronic acids at the bridging carbon atoms are available. These can be easily converted in high yields into chiral boronate esters 24a using (-)-pinanediol. 53 Solid-liquid extraction experiments show D- over L-fucose selectivity for both isomers, while (surprisingly) the rctt-isomer generally exhibited a stronger affinity for the carbohydrates studied. If para-, meta- or ortho-tetraacetyl-glucoxybenzaldehyde was used for the condensation with resorcinol in the presence of aluminium chloride, the formation of rctt (major) and rccc (minor) isomers was observed. 54 These compounds, which have carbohydrates attached to the resorcarene skeleton, were characterized as their O-acetyl derivatives 24b. (For further calix sugars see Section 2.4.)

"•%.0 24a

OAc

24b

Chirality in Calixarenes and Calixarene Assemblies

151

E,ooc.cL f

NH

25

A direct synthesis of chiral resorcarene octamethyl ethers 25 starting with chiral monomers has been also reported. 55 The condensation of (E)-2,4-dimethoxycinnamic acid amides of D-, and L-valine catalyzed by BF3.Et20 leads to a mixture of the rccc-, rcct- and rctt-isomeres, which can be separated by chromatography. Cavitands III having chiral bridges, X, were studied in Langmuir monolayers and some pH-dependent enantiomeric recognition of amino acids in the subphase was reported. 56 Enantiomerically pure hemicarcerands in which the two cavitands are connected by chiral bridges derived from 1,1'-binaphthalene 26a 57 or from tartaric acid 26b,c 58 should be mentioned here. They show chiral recognition in complexation as well as a chiral differentiation for guest release (Figure 3). When the hemicarcerand/CHC13 complexes (Ra)-26a and ($4)-26a were heated in (S)-28a the diastereomeric hemicarceplexes (Ra)-26a/(S)-28a and (S4)-26a/(S)28a were obtained. Guest release was monitored by 1H NMR spectroscopy in CDC13 at 23 ~ and the ratio of the first-order rate constants for the hemicarceplexes 26a.28a was kR:ks = 7. On the other hand, when ($4)-26a was heated to 100 ~ in racemic alkyl bromides 28b or 28c, mixtures of the diastereomeric hemicarceplexes (S4)-26a/(S)-28b and (S4)-26a/(R)-28b (analogous with 28c) were obtained. The

Z: ~

O'~' ~:'
26a Y =z

26b Y=Z

27a

27b

Y = O-CHz-O

HO..T/-.O/", HOH'~O',/ 26c

Y: Z

Y = O-CHz-O

Br Br 28a

28b

28c

Figure 3. Chiral hemicarcerands 26, 27 and their guests 28.

152

MYROSLAV VYSOTSKY,CHRISTIAN SCHMIDT, and VOLKER BOHMER

ratio of the two diastereomers, determined by 1H NMR spectroscopy was 1.5:1 to 2:1 corresponding to AAG~ of--300 cal mo1-1 (at 100 ~ The guest release in CDCI 3 led to kfast:kslow= 5 for the diastereomeric mixtures (S4)-26a.28b (the less abundant diastereomer gave the faster rate) and to kf~t:kslow= 9 in the case of ($4)26a.28c (where the most abundant diastereomer gave the faster rate). Thus, the AAGr values (at 23 ~ in CDCI3) for the guest release from the diastereomeric hemicarceplexes are as follows: 26a.28a, 1.1; 26a.28b, 1.0 and 26a.28c, 1.3 kcal mo1-1. It is worth noting, that diastereomeric hemicarceplexes synthesized from the (S,S)-26b and racemic 2-butanol (in ratio 4:1) show quite different Rf values on TLC [0.8 for the major product and 0.5 for the minor one, eluant CHCI3-EtOAc 20:1 (v/v)]. Thus, even incarcerated chiral molecules provide different sorption properties that are transferred through the thick shelf of hemicarceplexes to the outside. Similar studies were carried out with the hemicarcerands 27a,b 59 containing only one chiral bridge in the molecule. Diastereomeric complexes were obtained in a similar way by heating the hemicarcerand/CHCl 3 complex with an excess of various racemic guests (in substance or in diphenyloxide, which is large enough to be excluded). The following ratios for guest inclusion were found for (S)-27a: >20(R): 1(S) for methyl p-tolyl sulfoxide, 2.5(S): I(R) for PhCH(OH)CH 3, 2.3 for 1,2-dibromopropane, 1.6(R): 1(S) for phenyl methyl sulfoxide, and 1:1 for 1,2-dichloropropane, 2-methyl-l-butanol, 1,2-dichloropropane, and 5-methylhexanol. This corresponds to AAGr > 2.4 kcal mo1-1 (at 398 K) for the most dramatic first example. The chiral recognition properties of the hemicarcerand (S,S)-27b were less pronounced: 1.4(R): 1(S) for 2-butanol, 1:1 for methyl phenyl sulfoxide. When the "sealing" method was used, which means that the final, chiral bridge is introduced in HMPTA in the presence of an excess of the guest, the selectivity slightly decreases. The difference in the rate constants for guest release from (S)-27a leads to AAG~ = 2.3 and 1.9 kcal mo1-1 for 1,2-dibromopropane and 1,2-dichloropropane, respectively. Again, the diastereomeric complexes of (S)-27a and (S,S)-27b with phenyl methyl sulfoxide gave quite different R F values on TLC (eluent: CH2C12, 2% EtOAc).

2.4. Biological Relevant Derivatives The vancomycin mimics, 7, discussed above are an important example of biologically active calix[4]arene derivatives. In this case the whole macrobicyclic structure including the calixarene skeleton with its cavity seems to be important for the pharmacological effect. In the following sections we discuss derivatives in which the calixarene serves more or less as an easily available scaffold for the assembly of biologically relevant fragments such as peptides or carbohydrates.

Chirality in Calixarenes and Calixarene Assemblies

153

Calixpeptides Conformationally rigid cavitands III have been used for the synthesis of four-helix bundle caviteins 29, 6~ which are de n o v o proteins composed of cavitands and proteins (Figure 4). The proximity of four ct-helical peptides significantly stabilizes their native-like structure, which was proved by denaturation experiments with guanidine hydrochloride. 62 Peptides having 0-3 Gly units in their spacer possess the native-like conformation while peptides having two or four methylene groups there appear more similar to molten globules. The experimentally estimated molecular weights determined by sedimentation techniques, predict a monomer-dimer equilibrium for the caviteins, which does not influence the stability of the bundle, while the cavitein with negatively charged phosphate groups at the methine bridge is monomeric. A similar stabilization of the native-like conformation of the helical strands was observed in caviteins derived from cyclotriveratrylene possessing only three helical peptides of the same structure. 63 Proteins 29 might be used as a drug delivery system due to the cavity of the resorcarene. Hamilton and co-workers used calix[4]arene as a platform for the synthesis of the proteins 30 having a large surface (approximately 450 ,~2). The capability of 30a to inhibit chymotrypsin 64'65 is probably due to this large peptide surface containing negatively charged carboxylate residues. This was supported by a lower activity of protein 30b with only one negative charge in each protein loop and the absence of activity in the case of 30c with positively charged ammonium groups. Thus, the inhibition by the synthetic compound 30a occurs in the manner found for natural proteinase inhibitors.

Calixsugars The biochemical importance of carbohydrates in the functions of living organisms, is well known and can be hardly overestimated. These molecules play a key

) |

;p Sp

P-S

I~.~Sp--P

sp P

29

P

0 (CH2)t4-CO--PI'~POEt 0 CH~-CO-GlyGly-OEt S CHz-CO-Gly~-peptide S (CHz)2.4-.CO-peptide CHzS CHz.-CO-peptide

peplide = NH-[GiuGluLeuLeuLysLysLeuGluGluLeuLeuLysLysGly]-CONHz

Figure 4. Schematic representation of the "cavitein" structure 29.

154

MYROSLAVVYSOTSKY, CHRISTIAN SCHMIDT, and VOLKER B(3HMER C4Ho

o

IHN~O 1 4 .a R1= R2= (S)-CH2CO2H O R ~ = (S)-CHzCO=H R2 = (S)-CH2-C6H4-OH

i...-H

o N. ~

#N.to

C R' = R2 = (S)-(CH2)4NH,*

RI~N~N'~R 2 O H 30

role in cell-recognition processes. 66 Thus, the attachment of such groups on the calixarene scaffold is undertaken in order to construct in this way smart artificial recognition systems of possible biochemical and medical interest. First successful synthetic approaches in this direction were made by the groups of Dondoni andUngaro, 67'68 who synthesized a variety of narrow- and wide-rim glycosylated calix[4]arenes. Mitsunobu condensations of calix[4]arene with OC-Dmannofuranose diacetonide or tetraacetyl-~,13-D-glucopyranose led to the formation of the corresponding mono 13-31a and 13,13-bisfuranose 31b as well as to the cz,cz- and oc,13-bispyranose derivatives. Deacetylation of the last two products gave deprotected calixsugars 31c,d with high yields, while all attempts to deprotect the hydroxy groups of the furanose derivatives 31a,b failed. For the glycosylation of the wide rim of the calix[4]arene core, the c o n e tetrapropyl ethers with two distal or four hydroxymethyl groups were used as

31a OH

HO ~

Hg,~ _

31b

OH

~

31c

31d

Chirality in Calixarenes and Calixarene Assemblies a b

R 1 = R2 = R3 = HO~._ O1"~H R 2 H, R 1 = R 3 - / ~ O , , , ~ ~ H

C R2=H,

32

155

HO.-,._ 0~ H

RI=R3= HO2C.~.O'~.~O ,'~

H

HO.~._ OH HO__ O~ H d R2=H,RI=R3= IvU.,Z..O~7~D -

HO..,_ OH OH e, f R:': H, R' =R3= X- - - ~ ~ O,,.f_.O~o R1 = R2 H, R3 ~ = = X~ O ,

g

H HO1u.>,,. n~HO../Z~ . O H H OH zr],,..,'-',, ~ v ~ . . ~ . - , H

starting materials. 67'68'69 Tetrakis-O and bis-O-galactosyl 32a,b, bis-heptulosonic acid and bis-O-lactosyl calixarenes 32c,d were obtained in 25-65% yield by reaction with the 1-thioethyl (32a, 32b, 32d) or 1-(2-thiazolyl) (32c) derivatives of the corresponding protected sugars with subsequent deprotection. A Suzuki coupling between the calixarene 1,3-diboronic acid and the 4-bromophenylether of tetraacetyl derivatives of corresponding glucosides, followed by deacetylation led to the derivatives 32e7~ and 32f. 71 The mono maltose derivative 32g was prepared in a similar way. So-called "C-linked calixsugars" were prepared by Wittig reaction of the calix[4]arene tetraaldehyde 33 with galactose-6- or ribose-5-phosphorus ylids (obtained in situ from the known phosphonium iodides, Figure 5) 70 followed by reduction of the double bonds and deprotection of the hydroxyl groups. Thus, the tetra- and trigalactosides 34e,d (all as 13-anomers) and tetra- and tri-Llyxosides 34 (as a mixture of ct- and 13-anomers) were obtained.

~O

73H7~ ' ~ "~00-~0~0 PPh3"I" ~O

4

33

73H7~ ' ~

MeO~-~ 0 0 PPh3*I" X

4 34a

o

R:

/ C3H7 -

3

H~.~.H H~'%O -OH 34b TM

~HOHr Figure 5. Survey on C-linked calixsugars and their synthesis.

0

/ C3H7

156

MYROSLAV VYSOTSKY,CHRISTIAN SCHMIDT, and VOLKER BOHMER

The t-butylcalix[4]arene with two 2-aminoethyl groups at the distal positions of the narrow rim was easily reacted with sugar derived thioisocyanates to give the corresponding thioureas containing two D-glucose 35a, D-mannose 35b, D galactose 35c, or D-lactose 35d moieties. 72 Of all the compounds described above only the tetrakis-O-galactosyl derivative 32a was used in complexation studies 68 with D-glucosamine hydrochloride and tetrabutylammonium dihydrogen phosphate. It forms 1"1 and 1"2 (host-guest) complexes with D-glucosamine and 1"1 complexes with the dihydrogen phosphate anion (Ka = 31 + 4 M -1, DMSO-d6). Recently Roy et al. 73described the dendritic calix[4]arene derivatives 36a-d with 4, 8, and 16 carbohydrate moieties, using as model carbohydrate the TN antigen (GalNAcctl ~ O-Ser/Thr) corresponding to one of the immunodominant epitopes found in human adenocarcinomas mucins. The calixarenes 36a-c interact strongly with lectin of Vicia villosa (VVA = Vicia villosa agglutinin) leading to the rapid formation of insoluble precipitates. Horseradish peroxidase labeled VVA (VVAHRP) was used for the quantitative determination by optical density measurements of the ability to inhibit the binding of asialoglycophorin, a natural glycoprotein, to VVA. The strongest inhibition was observed for the hexadecavalent conjugate 36r (IC50 = 13.4 ktM) a 12-fold increase in comparison to allyl Ct-D-GalNAc as monomeric model (IC50 = 158.3 ~M). Tetra and hexadecavalent glycocalixarenes 36a and 36c were successfully used as coating antigens on hydrophobic polystyrene surfaces. The sialylated derivative 36d TMshowed a strong cross-linking ability with tetrameric wheat germ agglutinin (WGA) with complementary recognition sites for sialosides. The resorcarene skeleton, with eight OH-functions pointing in one direction in the rccc-isomer was used by Aoyama's group as a molecular scaffold for the attachment of carbohydrates. "Sugar-clusters" bearing eight galactose 37a, 75 glucose 37b, 76 maltopentaose 37c, 77 or sialyl groups 37d 78 were obtained by reaction of the octaaminoethyl derivative with the corresponding sugar lactonolactone. Similarly, derivatives bearing four glucose 38a, maltose 38b and maltotriose 38c

R NH

s=~

NH

k, t-Bu

t-Bu "~2 35

R OH

a

OH

b

#"J~OH

c

H~O

H~OH~ d #

H

H

HO~,.~ O1-~) O

Chirality in Calixarenes and Calixarene Assemblies

Hoo.

HO.~

AcNHO-.L.NH

.o~~,

~N.

HO~Ad,~IH ~)._l...NH

.o OH

HO,~

HO OH

5

o~--

- L.]/21"1%O

HN~N""

O., NH

O.~.,.NH

HO, oOH

.AcNH~

AeNHO 7

NH

O ?NH

O~")

t-l ,.OHN,~o HN'-~......,,,J

.~

HN_

HO HOf~Ol"l

X,~O

AcNIt~)#

~

o oOH~ ~a-.N

'~,NI..IAc

HO'V'~OH

o-'

HN~O HN...~

36b

Figure 6.

~'

L,~_~o

O- NH

oy

4 36a

HO

O~,/

~~HH O

'~i~'~

157

/..o

,,NH

O.~.NH

o-'

4 36c

4 36d

Calixarene derived sugar dendrimers.

residues were prepared 79 from the more rigid cavitand skeleton functionalized in the 2-position of the resorcinol units by SH-groups. The binding properties of the sugar clusters were assessed first with 8-anilinonaphtalene-l-sulphonate (ANS), where binding constants of K = 2.2 x 105 M -1 and 2600 M -I were found for 37a and 38c, respectively, and with eosin Y (K = 7.5 x 105 M -1 for 37b). The octa sugars form complexes also with monoand dihydrogen phosphates, ribose-5-phosphate, and guanine monophosphate. Dynamic light scattering (DLS) and transmission electron microscopy (TEM) showed the formation of spheres/vesicles with diameters between 100 and 400 nm in the latter cases induced by multivalent anion binding. Lectins are capable of strong binding of both, the tetra- and the octa-derivatives, which might be exploited in the design of drug delivery systems. The amphiphilic character of the octa-substituted derivatives 37a-e was used to assemble these molecules on the hydrophilic surface of quartz. The driving force of this self-assembly process is obviously the hydrogen bonding of the sugar groups to the quartz surface. In this case the cavities still are able to bind ANS or vice versa complexes of 37d with ANS can be self-assembled on quartz. Alternately, the hydrophobic chains of these sugar clusters offer the possibility to immobilize them on hydrophobic sensor chips 8~ that could be used to monitor sugar binding lectins.

158

MYROSLAVVYSOTSKY,CHRISTIAN SCHMIDT, and VOLKERBOHMER X

X=

X

X

xo

~

ox

XO" ~

OH I~1 H~

"~'~dr~OH

I:I HO

H - - - ~~'~OH OH

b

"OX C

R = CloH21

~

H

H

37

OH HO,, OH

H 0

;R'

o

~

R = CH 3

COzH

~

H

~HO-~._ 9HE.. -

SR'

SR'

"3

RI

: I

R'S

OH

C

'

H ~ v " '

OH , , j OH

~.._.OH

~

H(X~ OH _OH ~ . ~ ) , ~ ~ O H ~P'"OH

38

Figure 7. Resorcarene derived sugar clusters. Finally we should mention various host molecules which were obtained by covalent connection of calixarenes with cyclodextrins, sl

3.

STEREOGENIC CENTERS W I T H I N

THE M A C R O C Y C L E

Stereogenic centers may obviously be attached to the calixarene skeleton but they may also be incorporated into the macrocycle itself.

3.1. Stereogenic Bridges Resorcarenes are usually prepared from aldehydes different from formaldehyde 82 and thus they are connected by CHR bridges as potential stereogenic centers. Calix[4]arenes with one or two opposite CHR bridges 14a and calix[5]arenes with a single CHR bridge 14b have been synthesized only recently. All four possible diastereomers of a resorcarene (rccc, rcct, rctt, rtct) are meso-forms having at least one symmetry plane. The situation is similar if two different residues R 1, R2 are attached to opposite bridges. 83 Here the rctt-isomer has no symmetry plane, but

Chirality in Calixarenes and Calixarene Assemblies

159

retains an inversion center. All these molecules become chiral, however, if one of the resorcinol units is modified. This has been demonstrated for one example, the monobenzyl ether 39, in its interaction with (R)-N,N,N-trimethyl-l-phenylethylammonium iodide. 84 A single regioisomeric syn-1,3-diether 40a or 1,3-crownether 40b can be formed from a calix[4]arene with one or two opposite alkylidene bridges. Two diastereomers with axial or equatorial position of R are possible for monoalkylidene or cisbisalkylidene calix[4]arenes. They have C 1 or C 2 symmetry. If the residues at the opposite bridges are in trans position only one Cl-symmetrical derivative results. For first examples these considerations have been confirmed, and the NMR spectra are in agreement with these symmetry considerations. 85 Chiral calix[4]arene amides 41 with CHR bridges were obtained recently also by homologous anionic ortho-Fries rearrangement from the corresponding diacylated dipropyl ethers of calix[4]arene in 11-67% yield. 86 The recently discovered access to thiacalix[4]arenes 42 in a one-pot procedure, 87 similar to the synthesis of t-butylcalix[4]arene, was the starting point for a rapid development exploring the chemistry of this new sprout of the calixarene family. In the context of this article, the oxidation of the A r - S - A r bridges to A r - S O - A r 88

HO

~~~/HO" ~

"OH OH

39

R'

R'= C(CH~)3

40 a Y = CHiCOOEt

160

MYROSLAV VYSOTSKY,CHRISTIAN SCHMIDT, and VOLKER BOHMER

R = C(O)NMe=, C(O)NEt=

41 t-Bu

t-Bu

42

(and subsequently to Ar-SO2-Ar ) is of interest, since these are potential stereogenic centers, analogous to the Ar-CHR-Ar bridges. As soon as both sides of a -SO-bridge are structurally different, the molecule must be chiral.

3.2. Spirodienones and Similar Derivatives t-Butylcalix[n]arenes can be converted into monospirodienones 43 by mild oxidation (tetrabutylammonium tribromide) under alkaline conditions. 89The structure of two examples 43a,b has been established by single-crystal X-ray analysis. 9~ Their chirality is mainly due to the asymmetrically substituted spiro carbon atom, however, compound 43a has also some similarity with calix[4]arenes having three different phenolic units in the order AABC (see below). Under neutral conditions the oxidation of t-butylcalix[4]arene leads to the formation of 44 which has been also confirmed by single-crystal X-ray analysis. 91 Further oxidation of 43 leads to bis- or tris-spirodienone derivatives of calixarenes. They possess with the directionality of each spirodienone moiety an additional, independent stereogenic element. Thus, for a calix[4]arene derived bis(spirodienone) 45 six isomeric forms are possible, two m e s o forms and two pairs of C2-symmetric enantiomers, see Figure 8. The number of possible isomers increases with increasing of the number of tings in the macrocycle. 92 Spirodienones are attractive intermediates for the preparation of calixarenes in which the intraanular hydroxy groups are partly replaced by hydrogen, 93 amine, 93 methyl, 94 or halogen. 95 The reaction with benzyne affords Diels-Alder adducts in high diastereofacial selectivity. 92 No special use has been made of their chirality, however.

161

Chirality in Calixarenes and Calixarene Assemblies

0 0

a n-1

b

(:~1

n=2

n=5

43

44

C2-symmetrical bis(6-nitrocyclohexa-2,4-dienone) derivatives 46 were found as side products of the selective ipso-nitration of 1,3-diether derivatives of t-butylcalix[4]arene. 96 Their structure was proved for one example by single-crystal X-ray analysis. Compounds of type 46 are formed by ipso-attack of the nitrating agent at the bridges and subsequent stabilization by deprotonation of the hydroxy groups. There was no indication that other stereo- or regioisomers were formed. Obviously, the structure of the macrocycle, which is conformationally restricted by the synposition of the two alkoxy groups, favors the regio- and stereochernical outcome

ii ! i i ! i i i !

i i =0

Ci i !

\

\

,,, \

\

"

~,,, ~0

! ! I aO

C=

Cs

Figure 8. Possible bis(spirodienones) 45 of a calix[4]arene and their symmetry

properties.

162

MYROSLAVVYSOTSKY, CHRISTIAN SCHMIDT, and VOLKER BOHMER

+

of the reaction. All attempts to further modify these compounds (e.g. by reduction of the nitrogroups) led again to the starting 1,3-diether.

3.3. SpherandType Calixarenes Condensation of 2,2"-dihydroxy-5,5'-di-t-butyl-biphenyl with formaldehyde under alkaline conditions leads to the macrocyclic trimer 47 or tetramer 48 depending on the hydroxide (NaOH or CsOH) used as catalyst. 97 According to their NMR spectra these compounds are flexible, showing singlets for t-Bu, CH 2, OH and a pair of m-coupled doublets for the aromatic protons at room temperature. At lower temperature (150 K in CDFCI2) an asymmetric (C1) conformation is frozen for the trimer 98 while the tetramer assumes a conformation that may be rationalized as C2v symmetric (with RSRS configuration of the diphenol units), with the symmetry planes intersecting opposite methylene bridges. The coalescence temperature of 4g suggests an energy barrier (AG *) of 13.5 kcal mo1-1 (15 ~ 97 O-Alkylation leads to derivatives in which rotation around Ar-Ar bonds is impossible, since the O-alkyl groups cannot pass each other. Hence, the single diphenol units become chiral and the macrocycle may be considered as resulting from R or S configured atropisomeric subunits. These compounds may also be regarded as inherently chiral since in a linear analogue the biphenyl unit can racemize not only via a cisoid transition state, where the O - Y groups have to pass each other (the only possibility in the macrocycle) but also via a transoid transition state. Based on residues Y that can pass the annulus 99 but not each other, two chiral diastereomers are possible for 47: namely a D3-symmetrical derivative (RRRISSS) and a C2-symmetrical derivative (RRS/SSR). While the D3-symmetrical hexamethylether was isolated (in low yield and impure form) by Yamato et al.,97 all our HO OH ~

B u

(CH20)'-[OH-]

tBu

1Bu

n=3 47 n=4 48

163

Chirality in Calixarenes and Calixarene Assemblies . ~.

e#

._

D3

C2 o

:'.'0. ~

"'T

C2

Y

C~

O "0

-o~-..b~ "2.

0,"

.YO

Y

o

D~

D4

c~

R,R,R,SI S,S,S,R

C2

R,5,R,S = S,R,S,R

R,R,S,S" S,S,R,R

Figure 9. Stereoisomers of O-alkylation products of macrocycles 47 (restricted rotation around Ar-Ar bonds above, and additionally around Ar-CH2 bonds below) and 48 (restricted rotation around At-At bonds). Twofold axes (---) and mirror planes (--) are indicated.

164

MYROSLAV VYSOTSKY,CHRISTIAN SCHMIDT, and VOLKER BOHMER

attempts to obtain a D3-symmetrical derivative have thus far failed. ]~176 Alkylation with ethyl bromoacetate, for instance, gives the C2-symmetrical hexaether in yields up to 92%; the structure was confirmed by X-ray analysis. ]~ When t-butyl bromoacetate was used as the alkylating agent, a Cl-symmetrical hexaether was isolated in addition to the C2-symmetrical product. Both compounds could be converted by transesterification into the same C2-symmetrical methyl ester derivative. The latter is available by direct alkylation with methyl bromoacetate or by methanolysis of the ethyl ester derivative. These results can be understood if the t-butyl ester residues are large enough to hinder rotation around the Ar-CH 2 bonds. ]~ Under the same assumption as above, the tetramer 48 may be fixed in four diastereomeric derivatives, two of which are chiral (D4: R R R R / S S S S and 6'2: RRRS/SSSR) while the other two isomers are meso-forms (D2d: R S R S = SRSR and C2h: R R S S = SSRR). According to Yamato et al. a single diastereomer is formed from 48, the ]H NMR spectrum of which (one singlet for t-Bu, one pair of doublets for Ar-CH2-Ar and two signals (m-coupled doublets?) for Ar-H) is in agreement only with the D2d-symmetrical isomer, although the authors consider also the Da-symmetrical derivative. 1~ However, this would show only a singlet for ArCH2-Ar, since the methylene bridges lie on twofold axes. It should be noted that the intriguing stereochemical properties of these spherand calixarenes are not yet entirely understood and hence, their potential as chiral host molecules is not yet fully developed. Compounds 47 and 48 may be considered as calix[6]- and calix[8]arenes in which three or four methylene bridges are replaced by a direct connection of the phenolic units (or as spherands enlarged by three or four methylene bridges). In the line of this analogy, the macrocycles 49 and 50 ]03 can also be regarded as 13-naphthol-derived calix[6/8]arene analogues with alternating C4- and C0-bridges. In this case the binaphthyldiol subunit is known as a structural element having "stable" chiral.. ,b

~

,,, o o ~ 8,1

9n

=, o o ~ ~

~.~

an = CH=~.~

~

.b

o I; o (-)o-i;-o o 6 ---

oO - ~ ~ an

80

O-b

5t

Chirality in Calixarenes and Calixarene Assemblies

165

ity, and the macrocycles have been prepared as all (R)-(-) enantiomers. The cyclic trimer with free OH groups shows binding constants between 90 and 370 M -1 for pyranosides in dry CDC13 while its enantioselectivity is modest. Complexation of disaccharides in protic solvent mixtures was also reported for an extended tetrameric tetraphosphate 51.1~

3.4. Further Macrocycles Similar to Calixarenes Calixarene analogues having L-cysteine 52,105 (R,R)-cystine 53154,1~ and other amino acid residues 551~ incorporated into one or two bridges of the macrocyclic skeleton have been synthesized by condensation of a bis(chloromethylated) dimer, trimer, or tetramer with the corresponding amino acid ester under basic conditions (Figure 10). The conformational mobility for compounds 52 was found to be higher than that for calix[4]arenes. A barrier of AGr - 11.7 kcal mo1-1 determined for R H is equal to the value found for trihydroxy-p-t-butyl-calix[4]arene (11.6 kcal mol-1). 1~ The pairs of broad doublets observed for the Ar-CHE-Ar protons (AS - 0.74 ppm) at - 6 0 ~ agree with a cone-like conformation of the macrocycle where the nitrogen is involved in the system of intramolecular hydrogen bonds. Due to the directionality of the bridge, this cone conformation is chiral. Diastereomeric cone conformations should exist therefore for the cysteine bridged t-Bu

t-

t-Bu t-

t-Bu

?O=CH3

Bu

/-Bu 53 R = H, CO)CH=

t-~ (

-Bu

R ~ = C H = bl~u 55

R = CH=, FBu

R', R* = H, CH(CH=)=

Figure l t). Calixarene analogues with chiral bridges.

54

I-Bu FI -- GO=OH=

166

MYROSLAVVYSOTSKY,CHRISTIANSCHMIDT,and VOLKERBOHMER

compounds, which might cause an additional splitting of signals in their 1H NMR spectra. A cone-shaped conformation was also proposed for the calixarenes 55,1~ owing to three pairs of doublets (AS = 0.67-0.76 ppm) 1~ for the methylene protons in their lH NMR spectra. In comparison to 52, this macrocycle is conformationally more stable. Several doublets for methylene protons with A8 = 0.73-0.89 ppm confirm the all-syn-like orientation of adjacent aromatic groups in compounds 53 and 54. On the basis of Cotton effects that are clearly observed in the CD spectra of 53 and 55 (and with much lower intensity in the case of 54), a chiral arrangement of the aromatic groups is proposed, which is induced by the presence of the chiral groups in these molecules. 1~ The stabilization of these chiral conformations by hydrogen bonds was proved by a much lower intensity of the Cotton effects in hydrogen-bond breaking solvents such as methanol. The binding of N,N,N-trimethyl-l-phenylethylammonium iodide by 55 was studied in chloroform, where 1"1 complexes having association constants in the range (K~-) 1 0 - 15 M -1 were found. The signals of the racemic salt were split into two patterns in the presence of the chiral calixarene analogues.

4.

INHERENTLY CHIRAL CALIXARENES (C1)

The chirality of the compounds discussed so far, especially of those in Section 2 is caused entirely by the chirality of their substituents. Due to their nonplanar shape, calixarenes offer various additional possibilities to produce chiral derivatives, the chirality of which is not based on a chiral group or subunit but on the absence of a symmetry plane, an inversion center, or an alternating axis in the molecule as a whole. This means, that opening of the macrocyclic structure would lead to an achiral linear molecule. Such molecules may therefore be called inherently chiral which should not be confused with the term intrinsically chiral. 11~A graph whose chirality is independent of its embedding in the three-dimensional space is intrinsically chiral, while the inherent chirality defined above is due to the three-dimensional structure. In this section we will discuss not only asymmetric calixarenes (Cl-symmetry), but also those C2-symmetrical derivatives, which are prepared in a similar way.

4.1. Calix[4]arenes It is immediately evident, that calix[4]arene molecules consisting of four or three (order AABC) different phenolic units ~ll have no symmetry element, as long as the molecules cannot be planar. As a result of the rapid ring inversion, the parent calix[4]arenes with free OH groups must be considered as "time-averaged" planar molecules. Consequently, to be permanently chiral, a complete ring inversion must be impossible, ll2 This can be achieved in principle by O-alkyl or O-acyl groups, which cannot pass through the annulus, or by bridging at the wide rim.

Chirality in Calixarenes and Calixarene Assemblies

167

Such molecules consisting of phenols with different substituents in p-position are available by fragment condensation of bisbromomethylated phenols or dimers with phenolic trimers or dimers ([3+1], 113 [2+2] 114) in yields up to 25-35%. However, a complete and conformationally clean derivatization of such an asymmetric calix[4]arene leading exclusively to the cone conformation 115 may be difficult due to this asymmetry (Figure 11).116 The chirality of 56 for instance, caused by the p-substituents Me versus n-Oct on the wide rim, is expressed also on the narrow rim by four signals for the four -O-CH2-COOEt groups. The annelated calixarene 57 may be regarded as an example for an AABC-type calix[4]arene, where the ring inversion is impossible due to the connection of the two A units by a bridge in p-position by which a second calix[4]arene system is created. 117 The phenolic units of a calixarene can also become structurally different by the attachment of different alkyl or acyl residues to the oxygens, which may simultaneously lead to the necessary conformational restriction. A recent paper describes a rational way to synthesize tetraethers with four different O-alkyl groups in the cone conformation, 118 which is outlined in Figure 12. The last four compounds in this sequence are inherently chiral. 119 A chiral triether (syn-syn) with two different ether residues (y1 = CH2CO2Et; y2 = CH2CO2CH2-pyrene, type AABH) has been obtained (55%) by direct alkylation of the corresponding monoether (K2CO3/THF), 12~and resolved by enantioselective HPLC (Chiracel OD). The excimer fluorescence of 58 increased up to twofold by addition of e.g. L-alanine methyl ester or L-phenylglycinol but no chiral discrimination was observed. Inherently chiral derivatives can be also obtained from calix[4]arenes if three different units are incorporated in the order ABAC or if only two different phenolic units are present, provided these derivatives are fixed in conformations having no symmetry plane and center. Figure 13 gives a survey of the possibilities. For such a classification, one should keep in mind that hydroxy groups (or methoxy groups but ethoxy groups are on the borderline) can pass the annulus. Their orientation may be necessary in a description of the actual conformation of such compounds. It must not be indicated, however, if different stable stereoisomers are to be OH

R ~

OH

OH

OH

+ HOR " ~ ~ 4

RI=R2 or ..... OH

Rs

R4

R2

+

R

,3

R4

3

Figure 11. Fragment condensation for the synthesis of calix[4]arenes.

168

MYROSLAV VYSOTSKY, CHRISTIAN SCHMIDT, and VOLKER BOHMER t-Bu

tBg am .CH2 CHz Y" -CH=COOEt n-Oct

Me

~00Et

56

OHOH

57

-

t.Bu t ~

~

t.Bu

58

distinguished. In such cases it is also not appropriate to describe these stereoisomers using the terms cone, partial cone, 1,2- or 1,3-alternate. These should be used only if all four phenolic units are conformationally blocked, since more than one conformation is still possible in the other cases. Examples for most of the possibilities summarized in Figure 13 have been realized (see Figure 12) some of which will be discussed subsequently.

OH

OH

^..

RI

R2

R3

propyl allyl butyl butyl propyl allyl allyl propyl butyl

Figure 12. Rational synthesis of cone calix[4]arenes with four different O-alkyl residues.

Chirality in Calixarenes and Calh~areneAssemblies Tetmether

A B C D, A A B C

169

alwayschiml

ABAC

A B

A

B

A

AABB B

AAAB A A BA B

Tdether

A

B

akvaysachiral

A

A

AAAH A

B

A

ABAH

~

AABH,

ABCH

A alwayschiral

A

Diether

AAHH A BHH AHAH,

A alwayschiral AHBH

alwaysachiml

Figure 13. Survey on inherently chiral calix[4]arenes. Capital letters (A-D) characterize phenolic units which are conformationally not mobile (e.g. by larger residues attached to the oxygen). Mobile units (hydroxy or methoxy) are not indicated.

Chiral di- and tripropyl ethers, 59b and 60b for example, have been synthesized as shown in Figure 14. TM Complete O-alkylation of the monobenzyl ether was possible with n-propyl iodide in THF/DMF with Nail as base. The cone and the partial-cone conformers were formed in a 1"1 ratio and could be separated chromatographically. Alkylation with n-propyl bromide in the presence of CszCO 3 in acetone gave the di-O-propylated compound 59a, in which both propoxy groups in anti orientation, in good yield. The cleavage of the benzyl ether with trimethylsilyl bromide led to the final products 59b and 60b. Etherfication in DMF using Ba(OH)2 as base usually gave syn-syn-triethers. 122 The racemic 61a thus obtained was converted in a pair of diastereomers by introduction of the chiral (-)-menthoxyacetyl group. These diastereomers 61b could be separated by HPLC using an achiral column (Zorbax ODS). Finally, the

170

MYROSLAV VYSOTSKY, CHRISTIAN SCHMIDT, and VOLKER BOHMER

0

Pr"O OH^ J Pr.^ OH O-pr

59a

O-pr

59b

OH OH=~ Pr-oPr o'Pr_~ O-pr

60a

Pr,,O Pr-O ~ 61a

O-pr

60b

Pr"oOZ~o'P~ 61b

Figure 14. Synthesisof some inherently chiral calix[4]arene ethers and separation of their enantiomers. menthoxyacetyl group was cleaved to give the single enantiomers 61a. 121 The picolyl analogue of 61a on the other hand was directly resolved by enantioselective HPLC (Sumipax OA 2000). Various inherently chiral tetraethers (partial cone AABB and AAAB) and triethers (syn-syn and syn-anti AABH) with a-picolyl and/or quinolylmethyl residues in combination with -CH2C(O)R or other alkyl ether groups have been derived from t-butyl calix[4]arene by Pappalardo et al. using strategies similar to Figure 14.123 Many of them were chromatographically resolved 123b and one example was confirmed by an X-ray structure. 123c Starting from calix[4]arene monoethers the reaction with ditosylates of oligoethylene glycols leads to the formation of inherently chiral 1,2-calix[4]crown ethers 62a in a more or less rational way. 124 Deprotonation of the distal hydroxy group gives the most stable anion and the first O-alkylation by the ditosylate will take place in this position. The resulting monotosylated intermediate then has only one possibility for an intramolecular cyclization. The yield for this final step is strongly dependent on the size of the first ether group, open chain structures and double calixarenes being isolated as side products.

Chirality in Calixarenes and Calixarene Assemblies

171

Compound 62a could be resolved by enantioselective HPLC (Chiralpak AD, ot - 1.20-3.57). 125 The remaining hydroxy group in 62a can be O-alkylated too. syn-Propoxy derivatives 62e show a smaller chiral discrimination by Chiralpak AD than the starting compound. The introduction of a second picolyl group in 62b led in good yield to the inherently chiral calix[4]crown ether 63, where the molecule is fixed in the partial cone conformation. 126 The interaction with (R)- and (S)-Iphenylethylammonium picrates was studied for the racemic 62c,d, showing a splitting of 1H NMR signals but no discrimination between the two enantiomers. 125 The complexation properties of 62 towards alkali, alkaline earth and heavy metal cations, by liquid/liquid extraction of the picrates are not as good as those of 1,3-calix[4]crown ethers. 125'126 Interesting chiral phosphates with a cyclic and an open phosphate group were obtained from calix[4]arenes either by direct phosphorylation with diethylchlorophosphate or by reaction with POCI 3 followed by hydrolysis and subsequent alkylation with triethyl orthoformate. The compounds 64 contain two stereogenic elements, the bridging phosphorus and the order of the residues attached to the phenolic oxygens. The bridging P = O group may be directed to the periphery (exo) or to the center (endo) of the macrocycle. Both diastereomers exist as pairs of enantiomers and all four stereoisomers could be separated by HPLC (the exo-isomers are stereoselectively formed by the direct phosphorylation) and confirmed by X-ray analysis of the racemates and of a single endo-isom e r . 127,128

A surprising O,O-phosphorotropic rearrangement was observed during the Oacylation or O-alkylation of the easily available 1,3-bis(diethoxyphosphoryl) t-BUrN

9 .,

•

a 2-Py H 4 b 2-Py H 5 t-

-

-

- Bu

d

Ph H 6 0 2-Py Pr 4

N"~

62

t-Bu t-Bu

63

O,. p E t E

R

R

(~OEt t

,

R

64

"

R

-Bu

172

MYROSLAV VYSOTSKY,CHRISTIAN SCHMIDT, and VOLKER BOHMER P

I

66a

Y

l

P

P

\

l

65

P

P

\

P=P(O)(OEt)z

I

P

I

Y

66b

calix[4]arenes 65. Instead of the expected Cs-symmetrical product 66a, asymmetric product 66b was mainly formed. Obviously in a first step (under the alkaline conditions) the deprotonated 1,3-phosphate 65 rapidly rearranges to the 1,2-phosphate (it can be isolated by acidification) which reacts in the second step with the electrophilic reagents. 129 The analogous rearrangement is also found for the 1,3monoethylmonophosphate 67.13~ The reason these reactions occur can be found in the relative stabilities of the anions involved. This approach could be of general use for the synthesis of inherently chiral calix[4]arenes. The stabilization of the monoanion of a monoO-alkyl/-acyl calix[4]arene by two intramolecular hydrogen bonds explains the usually easy access to 1,3-derivatives. However, upon further deprotonation the monoanion of a 1,2-O-alkyl (1,2-O-acyl) derivative is stabilized by an intramolecular hydrogen bond (unlike the analogous 1,3-derivative) and rearrangement occurs if there is a reaction pathway available. For the phosphorotropic rearrangement the authors assume a cyclic intermediate with five-coordinated phosphorus, which is not unreasonable although an intermolecular mechanism is not strictly ruled out. A similar rearrangement was observed for the 3,5-dinitrobenzoyl residue TM during O-alkylation reactions of 68 (refluxing THF, K2CO3) 132 and this was used similarly for the synthesis of inherently chiral O-alkyl/acyl derivatives 69. During the palladium-catalyzed rearrangement of 1,3-syn-bistriflates or bismethansulfonates to mono- and triesters (clearly an intermolecular process) a chiral trismethanesulfonate in a syn-anti-arrangement was isolated as side product. 133 Expectedly it shows splitting of NMR signals in the presence of Pirkle's reagent. If the phosphorotropic rearrangement is combined with selective electrophilic substitution of free phenolic units (e.g. bromination) and hydrolysis of ester or phosphate groups a variety of inherently chiral calixarenes becomes available, 13~ as demonstrated by the examples in Figure 15. Recently a novel strategy for the direct synthesis of a nonracemic inherently chiral calix[4]arene was reported by McKervey et al. 134 The trialcohol monophenols 70a,b, (70a being easily available by reduction of the corresponding tetraethylester derivative with LiAIH4) have three -CH2CH2OH functions in syn arrangement, two of which are enantiotopic. Their acetylation by acetyl chloride in THF/pyridine gives the racemic monoacetates 71a,b in 51 and 62% yield. Enzymatic acylation with vinyl acetate in toluene with three different lipases led to lower

173

Chirality in Calixarenes and Calixarene Assemblies P

P

P

P

Br

Et

/

p

Et P

P

Br

Br

Et P

Br

Et

Br

67

P

Br

Br

Br.

Br

Br

P ,, P(OXOEt)2

Figure 15. Pathwaysto inherently chiral calix[4]arenes. yields of 71a,b along with some (12-33%) nonspecified diacetate (one of the two possible diacetates would be chiral too) but with an enantioselectivity of up to 100% (Table 1). Remarkably, a given enzyme may have quite different effects on the two substrates (e.g. 82:18 vs. 50:50), while the enantioselectivity for a given substrate may be inverse for different enzymes (7:93 vs. 100:0). Calix[4]arene tetraethers can be converted into mono Cr(CO) 3 complexes 72, which makes one of the phenolic units different from the others (AAAB). 135Thus, a single chiral compound is obtained from the 1,2-alternate conformer, while for the partial cone conformer only one of the three possible products is asymmetric (compare Figure 13). Mono derivatives fixed in the cone and 1,3-alternate conformation contain a symmetry plane. An elegant synthesis of inherently chiral calix[4]arenes 75 in the 1,3-alternate conformation (type AABB) was recently described by Gutsche et al. 136 It consists of the ring opening aminolysis of calix anhydride 74, which is available from the corresponding 1,3-alternate tetraacid 73 by reaction with oxalyl chloride in yields up to 60%. This ring opening represents a desymmetrization from S4 to C 2 symmetry. A variety of inherently chiral derivatives, interesting as building blocks for the construction of larger molecules (e.g. a "molecular stick") via covalent linking 137 and especially via self-assembly 13s should be available in this way. Since the ring opening proceeds under mild conditions, there should also be a strong potential for an enantioselective catalysis including an enzymatic reaction.

-,9

"

"<"'~~~,.,.

o - o..-4,,, / X ' - " ' - <

"

"o-o.-(, //-o,.< Ii .:= ~,,;( \

o

x~,~,~

o--o..-...(, K _ . < = ~ o " "

:=

o"

~

z

,_- ~ : . , - ~

z

.<

/~o..<

.<~~-

o-o-oM

C}~C}

o

I

I ,~

o

0 --

,o

(~ o/ "0

,.F

0

M

~

//o

O"

',,4

m

-,4

o

o 0

o

"I0

"I0

o

o

"o

II

II

::0 ~

0

o/~/~

):=:0 0

~~176

m

-'-: "10

~,.~ ~176

0 ~ / 0

"T

m

O) ,-

0 .N)

Z

0

x

~

Z

9

O0

rn ;)0

o .-4

N "1"

> z

N n-

~

.-<

o

p-

0

"4

175

Chirality in Calixarenes and Calbcarene Assemblies Table 1. Enantiomer Ratio and Yield (in brackets) of Lipase-Catalyzed Esterification of 70a, b Lipase

70a

Candida cylindracea Mucor miehei Asperillus niger

70b

50:50 (18%) a 82:18 (13%)

7:93 (14%) b 50:50 (25%) c

93:7 (8%)

100:0 (19%)

Diacetate: =18%;b12%; c33%

Asymmetric calix[4]arenes result also by the incorporation of a single m-substituted phenolic unit. Early examples comprise compounds with a m-methyl phenol unit, 139 or a single resorcinol unit incorporated via the 2,6-positions, 14~and were obtained b3r 3+ 1 fragment condensation (Figure 11). One example was resolved by enantioselective HPLC, TM to our knowledge the first resolution of an inherently chiral calixarene. Later calixarenes 76 with various m-substituents were prepared by electrophilic substitution (bromination, nitration) of the acetamino unit, 142or by addition of various reagents to a calix[4]arene containing a singlep-quinone unit. 143 Remarkably this substitution in m-position (ortho to the N H - C O - C H 3) groups is strongly preferred even if free p-positions are available in the other propoxyphenyl units. Several examples of this type have been confirmed also by an X-ray structure. 144'142An apparently elegant formation of a single m-substituted phenolic unit is realized in 77 but unfortunately the yield of the final pyrolysis step is only moderate (19%). 145 y

R1

Rz

Pr H H H

NO2 NHCOCH3 OAc OAc O-CO-S SH OH

R2

76

H

HaC-'t-'OCH 3 OCH=

COCH3

77

176

MYROSLAV VYSOTSKY, CHRISTIAN SCHMIDT, and VOLKER BOHMER

A different principle was employed to obtain inherently chiral calix[4]arenes. This approach involves desymmetrizing C2v or Cs symmetrical calix[4]arenes or calix[4]arene analogues (see Section 4.5) by etherification. 1,3-Diethers 78a of AABB-type calix[4]arenes are (as usually) easily obtained in a stereochemically fixed syn-arrangement. 146 Since the symmetry plane of the calix[4]arene skeleton does not coincide with the two symmetry planes of the diether-pattern, the whole molecule is C 1 symmetrical, as revealed by the NMR spectrum. The chirality has been additionally demonstrated by further splitting of the signals in the presence of Pirkle's reagent. 147 For the same reasons 1,2-diether derivatives 78b of ABABtype calix[4]arenes are chiral, although less easily accessible. 148 4.2.

Calix[5]arenes

The heat-induced [3+2] fragment condensation 149 recently described for the synthesis of calix[5]arenes consisting of different phenolic units would, in principle, permit the synthesis of calix[5]arenes which have an asymmetric pattern due to different substituents in p-position (e.g. AAABC, ABABC etc.). This has not yet been done to our knowledge, but a calix[5]arene consisting of two t-butyl and three p-phenyl phenol units (ABAAB) has been converted into a chiral monoester and monoether.150 Surprisingly high yields (43-46%) are reported, considering the fact that three structurally different phenolic units are present in the starting calix[5]arene. Most of the inherently chiral calix[5]arenes described up to now, owe their chirality, however, to the asymmetric substitution pattern at the narrow rim, and due to the lack of other general, selective derivatization reactions are derived from 1,2or 1,3-crown ethers. Both compounds possess a symmetry plane and can be desymmetrized by a single O-alkyl or O-acyl residue in position 3 (=5) or 4 (=5). 151'152'153In practice, 1,2-crown ethers 79 were prepared from mono-O-alkyl derivatives by reaction with the appropriate ditosylates, while 80 was obtained from the 1,3-crown-ether by subsequent O-alkylation or O-acylation using a weak base to benefit from the fact, that the first deprotonation leads to a hydrogen-bonded (4/5)monoanion. The picolyl derivatives 79 (n = 2) were resolved by HPLC R~

R~

R1

R2

R2

|

78a

78b

Chirality in Calixarenes and Calixarene Assemblies t-Bu

t-Bu

t-Bu

t-Bu

t-Bu

t-Bu

79

177 t-Bu

t-Bu

--

X: (CH2OCH2)"

t-Bu

t-Bu

80

(Chiralpak AD or Chiralcel OD) with separation factors up to 19152 but further enantioselective interactions were not described.

4.3. Calix[6]arenes To hinder the ring inversion of calixarenes by sufficiently large groups introduced by O-alkylation or O-acylation becomes more and more difficult, the larger the macrocyclic ring is, and calix[6]arenes are on the borderline. It is not necessary for the intraannular oxygen functions at the narrow rim to pass the annulus. Some results indicate, that the wide rim may pass the macrocycle, a process that clearly depends on the size of the substituent in p-position. 27 The 1,3,5-trimethyl ether is obviously conformationally fixed on the NMR-time scale, 154due to a combination of steric fitting and intramolecular hydrogen bonding. However, since tetramethoxy-calix[4]arenes are conformationally mobile, methoxy groups clearly can pass the larger annulus of the calix[6]arene skeleton. On the other hand, even a p-xylylene bridge between opposite oxygen functions can pass the annulus, as evidenced by tetramethylether derivatives 81 obtained in a 1,2,3-up-4,5,6-down conformation ("self-anchored rotaxanes"). Only more bulky bridges derived from anthracene have to stay at one side of the macroring. 155 1,4-Di-p-methlybenzyl ethers have been converted to macrobicyclic 2,5-diesters 82 that are chiral (C 2 symmetry) in the all-syn conformer, while a 1,2,3-up-4,5,6-down conformer would be achiral (C i symmetry). A m-xylylene bridge connecting the oxygens in 1,3-position completely hinders the ring inversion. This was shown by Shinkai et al. for compound 83, which is chiral due to the methoxy group at the bridge. 156 The racemate could be resolved by enantioselective HPLC (Chiralpak AD). No racemization was observed when a single enantiomer was heated to 100 ~ (12 h, different solvents). Bridging with tri-functional reagents in 1,3,5-position at the wide, or at the narrow rim led to a more or less rigid skeleton. In such a series of 1,3,5-bridged calix[6]arenes 84 chirality is caused by the CTV moiety used as bridging element. 157'158 The calix[6]cryptands 85a and 85b on the other hand are inherently chiral due to the 1,2,4-tri-substitution pattern on the narrow rim. 159

178

MYROSLAV VYSOTSKY,CHRISTIAN SCHMIDT, and VOLKER BOHMER

X: CH2"~"=-CHs Me/`."

!kj"

"Me y= CO(CHz)z.(.5.eCO

81

82

0

,, -~EtO

Y:

0

/

(CH2)" ~O

84

However, non-bridged calix[6]arene derivatives may also be chiral. It was possible to prepare two diastereomeric 1,2-dibenzyl ethers from t-butylcalix[6]arenes with a syn and anti orientation of the ether residues. The free energy barrier for their mutual interconversion was found to be about 27 kcal mol-]. ]6~ This should be high enough to separate the enantiomers of the anti-isomer, which has effective C2 symmetry. The same should be true for another example of this type, a 1,2-p-nitrobenzoate with an anti-configuration ]6] and for 1,2,4-tri-O-alkyl or acyl derivatives. ]62

179

Chirality in Calixarenes and Calixarene Assemblies

Bu t-~'HOHOOH HOOHO

oP

Et ,

O,,~ f

r~ c'c~ ooo~

C~ 03 05

r--c>~ r " o o o- ~

co 03 05

85a

85b

Each of these examples is more or less singular. Most were not studied in the context of chirality. In general, the potential of inherently chiral calix[6]arenes has yet to be explored.

4.4. Calix[8]arenes Conformational fixing of calix[8]arenes is more difficult than in the case of the smaller calixarenes. Even bis-crown ether derivatives such as 1,3-5,7-calix[8]bis-

1,3-2,4 C=

1,4-2,5 C=

1,5-2,6 D=

1,3-2,5 Cl

1,4-2,6 Cl

1,5-3,7 Daa

1,3-2,6 C,

1,4-3,6 C=

Figure 16. Survey of O-O-bridged calix[8]arenes (schematic representation) with two identical intercrossing bridges and their symmetry classes.

180

MYROSLAV VYSOTSKY,CHRISTIAN SCHMIDT, and VOLKER BOHMER

86

crowns are conformationally mobile. 163 A complete conformational inversion becomes (topologically) impossible, however, for bis-crown ethers with intercrossing crown bridges. The possible structures are collected in Figure 16. With two exceptions, these molecules are chiral and have C 1, C 2, or D 2 symmetry. (Partial functionalization ofD2d-symmetrical 1,5-3,7-calix[8]bis-crown ethers leads also to chiral structures, see below). 164 For the 1,4-2,5-calix[8]bis-crown-4 86 (available directly from t-butylcalix[8]arene in 13-18% yield, or from the corresponding mono-crown derivative) this could be shown by splitting of NMR signals in the presence of Pirkle's reagent 165 and by resolution using enantioselective HPLC (Chiralpak AD, Chiralcel OD). 166 Separation factors up to 1.58 were obtained for 86, while the separation of its methyl ethers (two mono- and the tetramethyl ether were studied) is usually less pronounced. Obviously the OH groups are necessary for an effective interaction with the stationary phase. Alkylation of t-butylcalix[8]arene with 1,2-bis-(bromomethyl)benzene led to the doubly bridged derivative 87 in 19% yield. 167 In contrast to 86 it is achiral (D2d

87

Chirality in Calixarenes and Calixarene Assemblies

181

symmetry), but its mono- and trimethyl ether (available in about 50% yield) are asymmetric. 168 The two possible dimethyl ethers, having C2 and Cs symmetry, could be also obtained. (As in other c a s e s 169 splitting of the NMR signals for the enantiotopic methylene groups within a molecule of the Cs-symmetrical mesoform was observed in the presence of Pirkle's reagent.) Various inherently chiral complexes with C2 symmetry are also known from calix[8]arenes in solutions as well as in the crystalline state, but cannot be exhaustively reviewed here. 17~

4.5. Calixarene Analogs Incorporating Different Bridges The possible number of inherently chiral structures and conformers further increases if the calixarene contains both different phenolic units and different bridges in the macrocyclic skeleton. For example, two chiral monoethers 88a,b are available from dihomooxacalix[4]arenes (one-CH2-O-CH 2- bridge instead of -CH2-). 17188b is the preferred product of the mono-O-alkylation, since the negative charge of the respective phenoxide anion is better stabilized by intramolecular hydrogen bonds due to the smaller distance between the phenoxide anion and the hydroxy groups. Tetraketone derivatives (Y = CH2-C(O)-R) in the two possible partial cone conformations, have been prepared in moderate yields. 172 Calix[4]arenes containing, in alternating order, two different types of bridges are inherently chiral as their mono-, 1,3-di- and triethers. Some examples are described for calixarenes containing methylene and cyclobutane bridges 89a-r 173 They can be synthesized by direct etherification with alkyl halides or by stepwise regioselective cleavage of the tetramethyl ether derivative. The peralkylation of 89b leads to chiral tetraethers, in which two different ether groups are arranged in an alternate order (C2 symmetry). Owing to the additional connection of adjacent phenolic units by pentano bridges, enantiomerization by ring inversion is impossible here. 174 The etherification of calixarene analogues containing methylene and 1,3-propylene bridges leads in good yield and regioselectively to 1,3-di-O-substituted derivatives 90.175 To prevent enantiomerization the ether residues must be at least propyl

t-Bu

t-Bu

t-Bu--~\/)"0 u,,~/-~t-But - B u ~ t-Bu

t-Bu

88a

88b

t-Bu

182

MYROSLAV VYSOTSKY,CHRISTIAN SCHMIDT, and VOLKER BOHMER y1...

y2..

~)

y1 y2 y3 y4

NIeH H H MeH MeH c M e Me Me H

89

or larger, as with calix[4]arenes or with hexahomotrioxacalix[3]arenes 91. Their di-O-alkylated derivatives 92 may exist as a syn- or an anti-isomer with effective C s or C2 symmetry which are not easily distinguished by NMR since the protons of the - C H 2 - O - C H 2- bridge between the O-alkylated units are diastereotopic in both cases. 176 Introduction of a different O-alkyl group of sufficient size leads to a conformationally fixed Cl-symmetrical triether 93 (the Cs-symmetrical diether would give one of two possible Cs-symmetrical triethers), which allows one to determine the orientation of the ether residues by NMR. 177The anti-dibutyl methyl ether 93 (structure confirmed by X-ray analysis) was resolved by enantioselective HPLC (Chiralpak OP). 178 Association constants for the binding with picrates of chiral amines were determined spectroscopically for both enantiomers in CHC13/THF. The highest ratio of the association constants (6.7) as a measure for the chiral discrimination was obtained for the ethyl ester of L-phenylalanine.

t-Bu

t-Bu t-Bu

t-Bu

90

v

91

v,

v

92

7

93

Chirality in Calixarenes and Cal&arene Assemblies 4.6.

183

Resorcarenes

As pointed out before, chiral derivatives obtained from II by incomplete O-alkylation or O-acylation are not "inherently chiral" owing to the fact that the A r - C H R Ar bridges become stereogenic centers (asymmetric carbons) if the aryl residues attached to these bridges are different. In cavitands III, however, in which the nonplanar structure of the [ 14]metacyclophane skeleton is fixed by the introduction of covalent linkages - X - between OH groups of adjacent resorcinol units chirality would emerge also for Ar-CH2-Ar bridges and therefore some examples for chiral cavitands are discussed here. 179 In principle, two possibilities exist to create inherently chiral cavitands, namely the use of different linkers - X - or linkers having no symmetry plane. The first possibility was realized with cavitand 94.18~ It possesses two adjacent methylene bridges and one bridging quinoxaline unit while two hydroxyl groups remain unreacted. This combination results in an asymmetric structure comparable to the AABC calix[4]arenes. The second example (95a), synthesized by Dalcanale and co-worker 181 is asymmetric because of one substituted quinoxaline residue (similar to calix[4]arenes containing one m-substituted phenolic unit). Since the menthyl residue attached to the quinoxaline via an ester function is chiral itself a mixture of the two diastereomers was obtained which could be separated by chromatography on silica gel. The subsequent reduction of the ester function in 95a led to enantio-pure (+) and (-) alcohols 95b. The compound 95a was used in host-guest complexation studies 182 with benzene, fluorobenzene, 2-fluorotoluene, and isobutane in the gas phase by means of Desorption Chemical Ionization Mass Spectrometry; however, studies of chiral recognition have not yet been reported. An example similar to 95a was recently described. 183 Cavitand 96 contains three symmetrically substituted catechol-derived walls and one nonsymmetrical quinoxaline derived wall; it is asymmetric by structure. The seam of intramolecular

o/I ~

N,~O-X~OH ~....N

R = CsHll

94

H

NLoSoN 95a R - C6H13, R'= CO2(-)Mentyl 9 5 b (,), (-) R = C6H~3, R'= CH2OH

184

MYROSLAVVYSOTSKY,CHRISTIANSCHMIDT,and VOLKERBOHMER

HN-J ~ N.-~O

R' O"~NH O

O

O

O

HN-,~, R'

ON~:O R' 96

hydrogen bonds (see Section 6.2) should lead to two diastereomeric conformations, a picture that is complicated by the dimerization of these self-complementary molecules via mutual inclusion of the adamantyl residues in the cavities of the "self-folded" cavitands.

5. CALIXARENES WITH HIGHER SYMMETRY (C. or D.) Here we summarize calixarene derivatives having an n-fold molecular axis perpendicular to the molecular main plane, which is usually defined by the carbon atoms bridging the aromatic units. This includes C2-symmetrical molecules if this twofold axis is identical to the molecular axis. Some examples in which there is a twofold axis perpendicular to the molecular axis have been described above.

5.1. Calixarenes Calix[4]arenes 97 with inherent Ca symmetry could be obtained by condensation of o-hydroxymethylated 3,4-disubstituted phenols as illustrated in Figure 17.184 Whether the 2- or the 6-hydroxymethyl phenol is used as the starting compound depends mainly on its accessibility. 185 Sometimes a direct hydroxymethylation in the 6-position is possible in a selective way. In other cases, the 6-position is protected by bromination, to direct the hydroxymethylation to the 2-position. Reduction of the appropriate salicylic acid derivative is another possibility. The yield in the cyclization (TIC14 in dioxane) is usually in the range of 15-25%, depending on the substituents in 4-position. Surprisingly the reaction failed with 2-hydroxymethyl-4-t-butyl-5-methylphenol. 186 Dehalogenation of 97c (which should be possible also with 97d) enables the subsequent introduction of further substituents via amino- or chloromethylation or coupling with diazonium salts. 186

Chirality in Calixarenes and Calixarene Assemblies

185

a 1

OH

R2

2 --

OH H O ' ~ .

a

d e

f

z

R2

CH~

CH3

b CH(CH3)2 CH3

R~ C

R1

R1

Br CI

CH3 CH3 -(CH2)3-

-(CH3)4-

g -CH=CH-CH=CH

R'

R'

97 Figure 17. Synthesis of inherently chiral C4-symmetrical calix[4]arenes 97.

C2-symmetrical calix[4]arenes 98 having only two (opposite) m-substituted phenolic units were also available from a dimeric precursor, synthesized in a stepwise manner.

185b

The regioselective incorporation of the phenolic units (apparent from the simple NMR spectrum) was confirmed by X-ray analysis for several derivatives and for the parent 97a itself. This shows the molecules in a strongly pinched cone conformation in which two opposite phenolic units are nearly parallel (effective C 2 symmetry), a situation that would not be favorable for guest inclusion. This pinched cone conformation might be the energy minimum in solution but it could not be detected by NMR spectroscopy down to the lowest available temperature (150 K in CDFCI2). 187 However, all tetra-O-alkyl derivatives of 98 like the tetraesters show an effective C 2 symmetry in solution. An energy barrier of AG* = 13.3-13.4 kcal mo1-1 was determined by dynamic NMR for the C2-to-C 2 interconversion with the C4 cone as transition state. Due to the equivalence of all four phenolic units various derivatives (from mono to tetra) are available by O-alkylation. In all these derivatives with residues larger than ethyl the enantiomerization by ring inversion is blocked, and

a !

R2

OH

OH R3

R3

98

186

MYROSLAV VYSOTSKY, CHRISTIAN SCHMIDT, and VOLKER BOHMER

some of these derivatives 188 have been resolved by chromatography on chiral stationary phases 189 (see also Section 5.2 below). An elegant separation has been recently described 19~(Figure 18). The dichlorotungsten derivative 99a has been converted by reaction with a chiral diol into the diastereomeric cyclic alkoxides 99b. They could be separated chromatographically and subsequently decomposed to the pure enantiomers 99c,d in which an "open" cavity exists, resulting from capping by tungsten. Thus, these compounds should have some potential for liquid-crystal materials, as demonstrated for similar Cavsymmetrical tungsten derivatives. 191 2,4-Dihydroxy-benzophenone has been converted via the 3-hydroxymethyl compound into the resorcinol derived calix[5]arene 100 in which the resorcinol units have been incorporated in 2,6-positions. 192 Not only is this a very rare example of a molecule with inherent C 5 symmetry, it shows also a surprisingly high energy barrier (AG* = 17.3 kcal mo1-1 at 355 K) for the cone-to-cone ring inversion, the origin of which is still unknown. Recent results show that similar calix[5]arenes with C 5 symmetry are available analogously. 193 A permanent conformational fixation of such molecules (e.g. by alkylation or acylation of the endo-hydroxy groups) has not yet been achieved. Ph

)--(

CI

'~

w

A~

O0

R

/

0

CI

x

AN

O0

R

O0

R R

Q

R

R

R

/

O

CI

w O0

O0

R

R

Q

Q

w

A\\ O0

w

R

R

o0

R R

'R | i

i

R

Ph

O

0

O0

R

R R

R

R

R

R R

w

p

O O0

R

R

R

O

II /I~,,

O0

~

O0

II ,,'7 ~,,

w

O0

Ph

)---(

O

Q w

AN O0

s

Ph

w

OO

OO

R R

| |

R

R

OO

R R

' 99c

Figure 18. Preparation of enantiomerically pure, tungsten capped calix[4]arenes from Ca-symmetrical calix[4]arenes.

Chirality in Calixarenes and Calixarene Assemblies

187

HO OH

OH

HO

100

101

1-Naphthol may form cyclic tetrameric condensation products with formaldehyde among which 101 has a Ca axis. 194 However, since the OH groups are in exo positions in this case (in contrast to the calix[4]arene 97g derived from 2-naphthol) it seems difficult to fix a nonplanar conformation in these cases. 195 A possibility would be the connection of two opposite oxygens by a (crown) ether bridge which would lead to a C2-symmetrical derivative. Although calix[n]arenes with Cn symmetry are most "appealing" we will summarize finally some examples with inherent C 2 symmetry one of which 98 was already mentioned. 185bOther C2-symmetrical examples were prepared by m-substitution 142 or by addition reactions to calixbisquinones 143 with opposite quinone units along with the isomeric Cs-symmetrical compounds. The connection of adjacent phenolic units by directional bridges may be mentioned as an additional possibility. The C2-symmetrical lactone 1024 is obtained in 36% yield from the tetrachloromethylated calix[4]arene tetrapropyl ether by reaction with salicylic acid. The formation of the Cs-symmetrical isomer is not observed in this case, probably since the most nucleophilic phenolate attacks at opposite rings, due to a pinched cone conformation of the tetrapropyl ether. 196 In contrast, the similar cyclization with 3-hydroxymethyl-2-naphthol gave mainly the C s isomer (Cs/C 2 = 95/5). However, the C2 isomer 102b could be isolated and has been resolved by enantioselective HPLC (Chiralpak AD, ot = 3.17). 145 A potentially chiral metal analogue of a calix[4]arene 1034§ built by uracil units connected via coordination of the nitrogens to Pt(en) (en = 1,2-diaminoethane) should be mentioned here. It was formed along with other oligomers during spontaneous condensation of the monomer [(en)Pt(UH-N1)(H20)]NO3.H20 (UH = uracil monoanion) and could be isolated in 27% yield. 197'198The X-ray analysis revealed a 1,3-alternate conformation (S 4 symmetry) stabilized by four hydrogen bonds between the carbonyl oxygen of one uracil moiety and the hydroxyl group of the neighboring one. In solution not only the 1,3-alternate but also the cone conformation (C 4 symmetry) seems to be present (together

188

MYROSLAV VYSOTSKY,CHRISTIAN SCHMIDT, and VOLKER B(3HMER

el

102a

el

o~

102b

M = Pt j= (HzNCHzCHzNH =)

103

with further minor species). 199 A c o n e conformation was also suggested for complexes with cations like Ag(I), Zn(II), Be(II), and La(III). 2~176 The complexes with Zn(II) and Be(II) are capable to include organic anions like aryl- or alkyl-sulfonates into their cavities. No complexation with any chiral species was studied, however.

5.2. Energy Barriers for Partially O-Methylated Calix[4]arenes As pointed out, the cone-to-cone ring inversion of an inherently chiral calixarene converts one enantiomer into the opposite enantiomer. This was used recently to determine the energy barriers for the ring inversion of partially O-methylated calix[4]arenes. 2~ A surprising result obtained at the beginning of calixarene chemistry was that calix[4]arenes, as well as their tetramethyl ethers, were flexible on the NMR timescale at temperatures slightly above room temperature, while mono-, di- and trimethyl ethers were fixed in the c o n e conformation at least up to 120 ~ The sterically more demanding methoxy group, like the hydroxy group, can pass the annulus, but the combination of increased steric demands and the H-bond stabilized c o n e conformation obviously leads to higher energy barriers 2~ for the partially O-methylated derivatives. The mono-, 1,3-di-, and the trimethoxy derivatives of 97a assume the c o n e conformation according to their NMR spectra (see Figure 19) which show signals in the expected range and reflect the C 1 and C 2 symmetry. Their enantiomers could be separated by chromatography on Chiralpak AD and Chiralcel OD as stationary phases and their racemization rate could be determined by monitoring the decrease of their CD spectra. Kinetic runs at different temperatures led to the activation

189

Chirality in Calixarenes and Calixarene Assemblies

J

,

7.2

6.6

.._0__K_ I

4.2

3.6

2.0

2.4

Figure 19. Section of the I H NMR spectra of mono (above) and 1,3-dimethoxy (below) derivatives of 97a.

parameters collected in Table 2, which are the first experimental values obtained for the energy barrier of the ring inversion of partially O-methylated calix[4]arenes. As Table 2 shows, the values are distinctly lower than those calculated for the corresponding methyl ethers from t-butyl calix[4]arene, 2~ although the order was correctly predicted by the calculations. One might argue that the difference is due to the m-methyl groups, whose greater steric demand could raise the energy level of the c o n e conformation. However, this would be true also for the tetra-hydroxy compound 97a which shows a similar barrier (AG ~ = 13.4 kcal mo1-1 at 8 ~ to that for p-methylcalix[4]arene (AG ~ = 14.6 kcal mol -l at 50 ~ Partially O-methylated derivatives, which are chiral may be prepared from calix[4]arenes of the AABB-type where no m-methyl groups are present. However,

Table 2. Energy Barriers (kcal/moi) for the Ring Inversion of Partially O-Methylated Calix[4]arenes

monomethylether 1,2-dimethylether 1,3-dimethylether trimethylether

Experimental (97a)

Ca/c.203(MM3)

A/-/4

AG$ (303 K)

97a I (R = t-Bu)

20.7

24.3

15.6 15.7

23.3 22.7

28.8 25.3 27.3 23.3

24.4 23.1 20.3 20.4

Calc.2~ I (R = t-Bu)

35.1 32.2 30.3 27.0

190

MYROSLAV VYSOTSKY,CHRISTIAN SCHMIDT, and VOLKER BOHMER

t-Bu t-Bu

104a R-CH~ 104b R=C~

first attempts to determine energy barriers for the racemization of 104 failed. While the 1,3-dimethylether 104a could not be separated chromatographically, the CD band of 104b (base line separated on Chiralcel OD) was too weak to allow reliable kinetic measurements. 2~ 5.3.

Resorcarenes

The regioselective acylation (or alkylation) of one OH group per resorcinol unit would reduce the Cav symmetry of rccc-isomers of resorcarenes to C4.2~ Various interesting chiral host molecules should be available in this way. However, tetraphosphates obtained by selective acylation with CI-P(O)(OR)22~ were later shown to have the C2v-symmetrical structure 105a 2~ which was found (confirmed by various X-ray structures) also for tetrasulfonates 105b 49'2~ or tetrabenzoates 105c 2~ obtained selectively by partial acylation. Reaction of the octakis-trimethylsilyl ether with PF2C1 resulted in the tetrakis-difluorophosphite 106 for which the Ca-symmetrical structure is likely, based on its NMR data (i.e. six signals for aromatic carbons); however, this compound is of limited use due to its high reactivity. 2~~ The first convincing Ca-symmetrical derivative of a resorcarene, tetra-lactone 108, was obtained by acid-catalyzed rearrangement of the cavitand-derived tetra-

gF2

y.

a - -~(OR)2

Y-O ~ O-Y HO"~ -OH-105

~iMe~

6

Me3S

Fz

F=P-O ~ C -~--R O

O-SiMe3

~'iMe~ PF2 106

Chirality in Calixarenes and Calixarene Assemblies

191

OTOH

H

[H'] HO 0

0 HQ

D--~O

0

107

108

carboxylic acid 107. The structure of 108 was confirmed by single crystal X-ray analysis, but no further properties or applications were reported. 211 Condensation of resorcarenes with primary amines and formaldehyde leads to tetrabenzoxazines 109, usually in excellent yield. First examples were described by Matsushita and Matsui a5 without any discussion of the stereochemical aspects. In principle, four regioisomers can be formed in this reaction (see Figure 20) from which the C 1- and the Cs-symmetrical isomers can be easily excluded on the basis of the 1H NMR spectra. The distinction between the C2v- and the Ca-symmetrical derivatives relies only upon the difference of the CHR 1 groups in the C2v isomer which could be isochronous by chance. Several X-ray structures have proved that under various conditions (e.g. condensation under alkaline or slightly acidic conditions) exclusively the Ca-symmetrical isomer is formed, 212'213 in agreement with semiempirical calculations. Most probably this is due to its stabilization via four intramolecular O-H-..O hydrogen bonds (thermodynamic aspect), while a rapid isomerization (kinetic aspect, see Figure 21) remedies wrong structures formed during the synthesis. If this isomerization comprises all four resorcinol units, a Ca-symmetrical molecule is converted in its enantiomer. This racemization may be the reason for the difficulty encountered with the chromatographic separation of the enantiomers. 214 In fact, on-column racemization could be demonstrated for slightly elevated temperatures. In principle, this racemization should be prevented by acylation or alkylation (see below) of the remaining hydroxy groups, which however, proved to be rather difficult. 215 Surprisingly, a regioselective formation of two benzoxazine rings is possible for the tetratosylates 105b, leading to molecules 110 with an overall propeller-like distortion, as shown by X-ray analysis. 216 The regioselectivity with which the benzoxazine structures are formed is kept for the reaction with diamines, which leads (under high dilution conditions) selectively to the C2-symmetrical cleft-like molecules 111 (confirmed by X-ray analysis in one case), although seven different regioisomers are possible in this case. 217 For tetratosylates the C2-symmetrical, transcavity-bridged derivatives 112 are formed

192

MYROSLAV VYSOTSKY, CHRISTIAN SCHMIDT, and VOLKER B(gHMER R~

R~

f,N

I

I

, O 1 " ~ O.H

O"~O.H

..9 ""H. o~,,~O-''H

H.O,~O

,H

"N p

109

I

I'N P I

R~

R1

R~

R~

I

I

I..N.,J

I-.N,.J

Figure 20. Possible regioisomers for resorcarene-derived benzoxazines. Only the C4-symmetrical compound 109 is formed. unless the bridge is too short. With ethylendiamine (but unfortunately not with other 1,2-diamines) a one-step synthesis led to carcerand-like double resorcarenes 113 having eight benzoxazine rings, for which the NMR spectra do not allow a distinction between a Cab- or a Da-symmetrical arrangement. 218

+H

-H

g

O

-H

/,

,,.._

HO

R,.N~.cH2

,,....._

+H

/,

Figure 21. Acid-catalyzed isomerization of the 3,4-dihydro-2-H-1,3-benzoxazine ring. When all four rings undergo this reorientation it means enantiomerization for the tetrabenzoxazine, and epimerization if the amino residues R1 contain a chiral center.

Chirality in Calixarenes and Calixarene Assemblies

193

RI

,OH

111

110

112

0

0

f,.,N HO

N,~I 0

113

The Ca-symmetrical arrangement of the benzoxazine rings should lead to two diastereomers when a chiral primary amine is used. Surprisingly, only one of these possible diastereomers (epimers) 114 is obviously formed with 1phenylethylamine 219'22~ and with p-substituted phenylethylamines. 222 This may be concluded already from the 1H NMR spectra which show only one pair of doublets for the diastereotopic methylene protons of the O - C H 2 - N groups. 222 It was additionally shown for two examples by single crystal X-ray analysis, 219'22~which allowed also the determination of the absolute directionality of the oxazine rings, since the absolute configuration of the asymmetric carbon is known. These epimers, readily obtained under various conditions (alkaline as well as slightly acidic) are stable in solution only in the absence of acids. With traces of TFA the optical rotation changes (to reach an equilibrium value) and the 1H NMR spectrum shows the appearance of a second set of signals, probably due to an acid-catalyzed epimerization, as shown in Figure 21. This clearly limits the use of compounds such as 109 or 114 as chiral host molecules. Recently, conditions were found to methylate the four remaining OH groups in 114 (deprotonation by butyllithium in THF a t - 7 8 ~ followed by reaction with methyltriflate). 223 This may be considered as a breakthrough in this area,

@

<

~o

i~

e~

_.. 1"3

t~ .,<

I

P

0 0

_~. <

0

m

"

C/

"

%/

F-o

,L/

"

C/

,,,-= /

\

,L/

~ 2--~/

',-o / ' ~

~:-o 2-=/

"

,L/

~:-o 2--=/

"

,'~

F-~ 2--=/

, =

"" o ~ o " "

6~o.,.

.<

O

"T"

II

jil

l:

~z"

I

"

Z

"1 ~'

..<

0, "'r

o. r ,,,

zJ

I

z'-,O~ I~0 z"

oo ~

Z~

F

"<

I'M

O: "1"

!'1"1

o iI -

Q,. <

"r

Z

--I

c~ "r

..<

-4

o

< .-<

o

...<

~D

Chirality in Calixarenes and Calixarene Assemblies

195

since evidently an epimerization is impossible in these tetramethyl ethers. Thus, various stable enantiomers are available in this way; they remain chiral after the removal of the chiral auxiliary groups. The reaction sequence outlined in Figure 22 gives some examples for this enantioselective preparation of C4-symmetrical resorcarene derivatives that are directly obtained as pure enantiomers, the chirality of which is determined by the configuration of the asymmetric carbon of the initially used amine. It should be mentioned that conditions have been found recently to obtain C4-symmetrical tetraethers 115 directly from resorcinol monoethers by condensation with aldehydes. The regular incorporation in the macrocyclic skeleton was confirmed for one example by X-ray analysis. 224 However, in contrast to Heaney's enantioselective synthesis, the racemic mixture of the two enantiomers is formed in this case.

6. CHIRAI. CONFORMATIONS For conformational isomers (conformers), whether they can be isolated as separate species is mainly a question of the energy barrier. This means that only a small difference may exist between chiral and nonchiral compounds, which is best illustrated by atropisomers of biaryl compounds. We therefore discuss in the following sections some interesting examples for chiral conformations found with calixarenes. Intuitively, one might expect the order of a system to increase with decreasing temperature and that the minimum energy conformation of a molecule should be the one having the highest symmetry. In many cases, this is not the situation. 6.1. The Parent Calixarenes and Resorcarenes The low-temperature 1H NMR spectra for calix[4]- and calix[5]arenes reveal that all four or five phenolic units are identical. 225This is compatible with a cone conformation with Cav and Csv symmetry, respectively. 226'227For calix[8]arenes the same pattern is interpreted by a D4d-symmetrical "pleated loop" conformation, with a regular "up-anddown" of the methylene bridges. Calix[7]arenes, however, show seven singlets for the OH protons and seven (overlapping pairs of doublets) for the methylene protons. This total lack of any symmetry element (C1) makes this conformation chiral. 228 An interesting situation has been found for calix[6]arenes. The low temperature NMR spectrum of t-butylcalix[6]arene shows three singlets for OH and t-Bu and three pairs of doublets for the Ar-CHE-Ar groups. This means that a symmetry plane, a twofold axis, or an inversion center must be present. A calix[6]arene consisting of two different phenolic units in alternating sequence (116) has by constitution a threefold axis. Due to this "mismatch" of symmetry elements it must be chiral (C l) at low temperature, which has been found in fact for 116. 229 In addition, the authors conclude that the direction of the cyclic array of OH...OH hydrogen bonds (clockwise or

196

MYROSLAV VYSOTSKY, CHRISTIAN SCHMIDT, and VOLKER BOHMER COOR

J 3 116 COOR 117

counterclockwise) is frozen, and they derive an energy barrier of 10.6 kcal mo1-1 for the reversal of this direction. The C2-symmetrical winged cone conformation (suggested on the basis of NMR studies) in which all methylene bridges point away from the cavity, has been questioned, since two O-O-distances are too large for intramolecular hydrogen bonding. Instead apinched cone conformation, also C2 symmetrical but more similar to the X-ray structure(s), 23~has been suggested (on the basis of MM-calculations), in which two methylene bridges point towards the cavity. TM Similar observations of the type (ABBB) 2 have recently been described for calix[8]arenes 117. Four singlets for OH protons are found at-50 ~ (and three singlets for the t-Bu groups), two of which coalesce at a higher temperature. 232 This has been interpreted as a reversal of the direction of the hydrogen bonds although the energy bamer of 14.1 kcal mo1-1 seems rather high for such a process. (A similar value follows from the t-Bu groups which appear as two singlets, ratio 2:1, at normal temperature.) 233 On the other hand, no indication of a directionality of hydrogen bonds TM could be observed for the C4-symmetrical calix[4]arenes 97 with m-methyl groups (or their partial ether derivatives) where such a directionality should lead to two diastereomeric species. 235 The rccc-isomers of resorcarenes are usually stabilized in a more or less regular cone (crown) conformation by four intramolecular hydrogen bonds between the hydroxyl groups of adjacent resorcinol units. The remaining four hydroxyls are available for intermolecular hydrogen bonding. Two possible arrangements may be envisaged, both of which have been found for various assemblies in the crystalline state: a chiral C4-symmetrical form and an achiral C2v-symmetrical meso form (Figure 23).

6.2. Directionality in Hydrogen Bonded Systems There are other examples in which the directionality of hydrogen bonds follows more stringently from the NMR spectra. Aminomethylation of resorcarenes leads to 118, which was found in a C4-symmetrical conformation in the crystalline state.

197

Chirality in Calixarenes and Cal&arene Assemblies

.-o:'

,H'Ox~IO"H, R R_O~O --H

H ~o~o

~

H . . . 6 ~ ,o~,

,-o. v _ _ . , . y o ' . N ~T .H

~o~9.-" H

Figure 23. C4- and C2v-symmetricalhydrogenbonding in rccc-resorcarenes.

This conformation also exists in solution, as deduced from a single set of signals for the repeating unit, including a pair of doublets for the diastereotopic protons of the Ar-CH2-N groups and two singlets for the aromatic carbon atoms carrying the OH groups. 236 If chiral secondary amines are used, two diastereomeric Ca-symmetrical conformations should exist. This has been observed for tetraoxazolidines 21 at low temperature. 237 Strong intramolecular hydrogen bonds (O-H-.-O-H-.-O--C) were also found in tetraamides 119 in solution (NMR) as well as in the crystalline state. 215 Cavitands with quinoxaline bridges between the oxygen functions may exist in an open, C2v-symmetrical, kite conformation and in the closed Cav-symmetrical vase conformation. 238 This latter conformation, most interesting for the inclusion of guests, can be stabilized by a seam of intramolecular hydrogen bonds in cavitands 120, which have been named self-folded cavitands for this reason. 239 Due to the direction of these hydrogen bonds the closed (or "folded") vase form is no longer Cav, but C a symmetric, which can be unambiguously deduced from the NMR spectrum. In addition to the

R2 \

9

/

R~

HO

118

119

198

MYROSLAV VYSOTSKY,CHRISTIAN SCHMIDT, and VOLKER B(3HMER

RI

R~

O4~.NoH .......

., ".

o

O~,N.H.. O o

a ~

R ~

R'

N

O.... H.N .,'~'O .......

H..N/~"O

'k

N

! !

g v

.

b

o'-V" ,N-H .......

r"J--'O

,, ,,

N-H-"o -R'

or

120

two signals for the two different NH protons, it shows four singlets for the aromatic protons, while the Ca,, symmetry would require one NH signal and three aromatic signals in the ratio 2:1:1. From the coalescence temperature of the NH signals a barrier of AG * = 17.4 + 0.5 kcal mo1-1 for the reversal of these cooperative hydrogen bonds was deduced in toluene-d 8 or p-xylene-dl0 .24~ Such a high value is understandable here, since the change in the hydrogen bond direction requires a reorientation of the amide functions (rotation around the A r - N H bond similar to the tetraurea capsules described below in S e c t i o n 7.2), while the reversal of the O - H . . . O - H system in calixarenes (see above) could occur just by converting hydrogen bonds in covalent bonds and vice versa. A similar energy barrier was found for the guest exchange (AG ~ = 16.9 + 0.4 kcal tool -1) which occurs with the unusually (for an open cavity) slow rate of k = 2 + 1 s-1. An unfolding/folding mechanism is supposed for the guest exchange therefore. For a similar water-soluble cavitand, a folded vase conformation is induced also in water by an appropriate guest, which shows again slow exchange kinetics. TM However, the NMR spectrum is in agreement with C4v symmetry, which can be rationalized by a seam of hydrogen bonds stabilizing the vase conformation but rapidly changing its direction. The diastereoselective formation of a cyclic array of intramolecular hydrogen bonds between amide and carbonyl functions at the narrow rim was also observed for the tetraamides 121 and for similar 1,3-di- or monoamides. 242

Chirality in Calixarenes and Calixarene Assemblies

o

199

I

o ]

H

NH4..'Ph b-Ol~ 0 iBu

',,h.,~=o,O...H

NH~HoM e

I

o

121

6.3.

Other Chiral Conformations

Chirality has been found in various bridged calixarenes, which by their constitution should have C s or even C2v symmetry. An example is the 1,3-phthaloyl bridged calix[5]arene 122. In contrast to the more flexible 1,3-crown ethers its NMR spectrum shows five singlets for the t-butyl groups with equal intensity (1.38, 1.30, 1.24, 1.14, and 1.11 ppm), which proves the absence of any symmetry element. 243 Other spectral features, e.g. five pairs of doublets for the methylene bridges, are in agreement with this. The expected three signals (1.26, 1.21, and 1.10 ppm, ratio 2:1:2) of a time averaged Cs-symmetrical conformation are found only at higher temperatures (120 ~ A similar observation was made for the 1,2-phthaloyl t-butylcalix[6]arene, for which the asymmetric conformation was demonstrated also by an X-ray structure. Intramolecular hydrogen bonding of the remaining OH groups may contribute to the stabilization of the chiral conformers in these cases. A C2-symmetrical conformation was deduced for the bisamides 123a and 123b in solution on the basis of their NMR spectra 244 (e.g. two doublets for OCH2CO, four doublets for ArCH2Ar and four doublets for ArH protons), and further established by X-ray analysis for 123a. C 2 symmetry was also found at low temperature in solution for the calix[4]arenederived pyrophosphate 124 and the calix[6]arene-based syn-bisphosphite 125a 245 and even at room temperature for the analogous bis-phosphate 125b. 9~ In these

Y:

122

200

MYROSLAV VYSOTSKY, CHRISTIAN SCHMIDT, and VOLKER BOHMER O

o

II

Y:

),

f,..,,~C-Y-C

OHOOHd

a

H3c

~N

b--N t-Bu

t-Bu t-Bu t-Bu

O..p,,O~p.O

/"-'N oCH3

N~

L... 0.~/0J

d "o d "o

N-t-Bu

t-Bu

123

t-Bu

t-Bu

124 0

I1\

I/"

P

P

I-Bu t-Bu I-Bu t-Bu t-Bu t-Bu

125a

0

II

II

~1'\

I/"

P

P

t-Bu I-Bu t-Bu t-Bu t-Bu t-Bu

125b

cases, intramolecular hydrogen bonding cannot contribute to the stabilization of these chiral conformers, but dipole/dipole interactions might. A time averaged C2v-symmetrical conformation results in all cases at higher temperatures and the energy barriers for the C 2 --->C2 transition were determined from variable temperature NMR to be AG $ = 10.1-10.2 kcal mol -l for 124 andAG $ = 12.7-12.9 and 14.4 kcal mo1-1 for 1254 and 125b, respectively. A C2-symmetrical conformation stabilized by intramolecular S--O...H-O hydrogen bonds was also found in C2v-symmetrical tetratosylates 105b as the energy minimum (AG $ = 11.6 kcal mo1-1 at 247 K in CDCI 3 for the interconversion C2 --~ C2). 49

Quite another example for a chiral conformation concerns the carcerands IV described first by Cram et al. 12 They usually consist of two C4,,-symmetrical cavitands linked together via four (identical) bridges. Thus, under permanent inclusion of an appropriate guest 247 a cage molecule with Dah symmetry 248 is formed that is again time-averaged. All X-ray structures of such carceplexes show that the two cavitands are twisted towards each other by angles of 13 ~ o (rotation around the fourfold molecular axis). 249 This chiral environment ( 0 4s y m m e t r y ) makes enantiotopic groups of an included guest diastereotopic (e.g. the methyl groups in DMSO). If the included guest is nonsymmetric with respect to the C2 axes, the symmetry of the whole assembly is further reduced to Ca. Then protons of two structurally identical cavitands, which are enantiotopic with a symmetrical guest (such as pyrazine) become heterotopic with a nonsymmetrical guest (such as DMSO). The existence for such "twistomers" was first described by Chapman and Sherman for a self-assembled hydrogen-bonded capsule formed by the hydroxy

201

Chirality in Calixarenes and Calixarene Assemblies

4

126

127

cavitand 126 in the presence of a base (usually DBU = 1,8-diazabicyclo[5.4.0]undec-7ene). 25~ An energy barrier of AG ~ - 13.4 +_ 0.2 kcal tool -1 was determined from the coalescence of the two singlets of the methyl groups in DMSO and attributed to the interconversion of the two "twistomers." A lower energy barrier of AGr - 12.6 _+0.2 kcal mo1-1 derived from the coalescence of the inward-pointing O-CH2-O protons of the cavitand was attributed to the rotation of DMSO around the C2 axes. This is schematically shown in Figure 24. Although the first example of Chapman and Sherman involved a self-assembled capsule bound via O-H...O C-) hydrogen bonds, as such an excellent example for "supramolecular chirality" (see Figure 38), similar energy barriers were found in a corresponding carcerand 127 with O-CH2-O bridges between the two cavitands, 249a and this type of isomer seems to be quite general in capsules of this type.E51,252 The calix[6]arene 128 with a syn-arrangement of the ct-picolyl groups is C3v symmetric by its constitution. It forms a copper(I) complex [Cu128X] where the tetrahedral coordination sphere of the copper ion is built up by the three picolyl

/~G,

r

*

Z~ c-.-~

Figure 24. Schematic representation of conformational interconversions in carcerands and structurally similar self-assembled capsules. Left:. Interconversion of "twistomers" with a still freely rotation guest. Right:. Rotation of the guest (symbolized by an arrow)in a frozen twistomer.

202

MYROSLAV VYSOTSKY, CHRISTIAN SCHMIDT, and VOLKER BOHMER

CN

CH3

128

(models for metalloenzymes), shown by a singlet for the O-CH2-Pic protons and two doublets for the Ar-CH2-Ar protons changes to C 3 at low temperature (two doublets for O-CH2-Pic and four doublets for Ar-CH2-Ar). 254An energy barrier of AG* = 12.8 + 0.2 kcal mo1-1 was derived from variable temperature NMR for Cu § 128 Cl- in CDC13. Palladium and platinum complexes of syn-diphosphite 125a (125a PdC12, 125a Pd(CH3)CI, 125a Pt(CH3)CI and 125a2Pd(0)), which are interesting for CO2/CH2CH 2 copolymerization, show a conformational behavior similar to that of the free ligand 245 (see above). Conformational freezing may lead to additional stereoisomers in calixarenes beating chiral substituents, as demonstrated for one example. A 1,3-syn-diether with two enantiomeric residues R/S is a mesoform (C s symmetry) regardless of its conformation 36 (cone, partial cone, and 1,3-alternate being possible). The analogous 1,3-anti-diether appears as a single mesoform only on the time average. A single partial cone conformation forms a pair of enantiomers, while two diastereomeric 1,2-alternate conformations with Ci symmetry are possible (see Figure 25).

Cs

~

Cs

)

,

(~

CI

:

(R)

OH

Ci

!

(

OH

Cs OH

CI (S)

(~

OH (3")

(R)

~

Ci

OH C~

Ci

Figure 25. Calix[4]arene 1,3-diethers with chiral ether residues; possible conformations and their symmetry.

Chirality in Calixarenes and Calixarene Assemblies

203

The same considerations are valid when methoxy groups replace the hydroxy groups. 36 The dibutylmethyl ether 93a has C 1 symmetry (at low temperature), but since the methyl group can easily pass the annulus two identical C l conformers are rapidly interconverting at room temperature and a time averaged C 2 symmetrical structure results, for which the authors created the expression "pseudo-C2-symmetrical." 178

7. SUPRAMOLECULAR CHIRALITY The examples discussed so far have been concerned with single molecules: assemblies of atoms held together by covalent bonds. They may be chiral due to their constitution or they may assume a chiral conformation at their energy minimum. Chirality may also be found in assemblies of nonchiral molecules held together by noncovalent forces such as hydrogen bonds or metal coordination. Such bonds are reversible; errors in connecting single parts of such an assembly may be corrected, which helps in the construction of larger structures by "self-assembly".

7.1. Host-Guest Induced Chirality Chiral molecules show different absorption coefficients for circularly polarized light of different handedness. This circular dichroism (CD) is one of the main optical properties by which they are distinguished from nonchiral compounds. CD spectroscopy measures this difference in the UV-vis region as a function of the wavelength. Usually the only molecules that show CD spectra are those that contain suitable light-absorbing group(s) [chromophore(s)]. 255 If a chiral molecule possesses two or more chromophores, its CD spectrum may show several peaks of different sign (split Cotton effect), due to the interaction of the dipole moments of chromophores (exciton coupling). Their signs are dependent on the orientation of the chromophores in space and thus on the absolute configuration of the molecule. This exciton chirality method was used to determine the absolute configuration of various organic molecules using appropriate derivatives (Figure 26). 256 A host-guest complex may be considered as one of the simplest supramolecular structures. If a nonchiral host interacts with a chiral guest, this interaction may be reflected by some properties of the host; for instance a CD band may be observed in a spectral region, where only the host absorbs. Such phenomena have been called "induced circular dichroism" or "induced optical activity". 257'258It is more or less a semantic problem, if this is due to a geometrical distortion of the host into a chiral conformation, or if the host becomes effectively chiral due to contacts with the guest. 259 Calixarenes and resorcarenes consist of aromatic tings that absorb light in the UV region. Thus, host-guest interactions with chiral guests lacking chromophores in their structures lead to induced CD spectra, which has been shown for many examples.

204

MYROSLAVVYSOTSKY, CHRISTIAN SCHMIDT, and VOLKER B(gHMER

.oC

.o

,

00.

o.

o

o

x

x

Figure 26. Exciton chirality method for the determination of the absolute configuration of steroidal diols. The clockwise (+) and counterclockwise (-) directionality of two dipoles is determined by the configuration of the diol.

Aoyama and co-worker 26~ studied the interaction of resorcarene 129 with a variety of chiral glycols, steroidal polyols, and sugars in chloroform and carbon tetrachloride. They showed that binding in these cases occurred via hydrogen bonding between the alcohol and the resorcinol hydroxyl groups leading to 1:1 complexes 261 for the examples shown in Figure 27. It can be conveniently monitored by induced CD spectroscopy where split Cotton effects are observed. The association constants for the binding of various alcohols increased with increasing number of hydroxyl groups in the guest molecule from 8 M -1 for 130e 262 over 28 to 67 M -l for 130a-d, and 130 M -l for 130f to 690 M -1 for 130g (in CDCI3).

OH

OH

(R,R)-a

(S,S)-b

OH OH P h " ~ OH P h ' ~ O H

HO

(R)..d

129

OH

C H 3 0 " ~ " " ~ i OCH, CH3OzC'~i cOzCH, OH OH OH

(S)-d

Ph O~

(R,R)-c

OH Ph'l~

HO

H f

OCH3 HO*" v

130

v

g

"OH

Figure 27. Survey on guests complexed by resorcarene 129.

OzCH3

205

Chirality in Calixarenes and Calixarene Assemblies

The signs of the first and second Cotton effects in complexes with diols are opposite to those determined by the classical dibenzoate chirality rule (compare Figure 26). 263 This means that the chiral guest induces an opposite chirality in the resorcarene host 129. Thus, 129 can be used as a supramolecular CD probe for the assignment of the absolute configurations of glycols and alcohols. The complexation properties of resorcarene 129 were also studied in chiral hydrocarbons such as limonene. 264'265 This gave an opportunity to follow the complexation of 129 with a variety of nonchiral guests by the decreasing intensity of Cotton effects. One-to-one complexes having association constants in the range of K a = 2-270 M -1 were observed for methyl esters of monoacids and of dicarboxylic acids such as malonic or succinic acid. Strongly hydrogen bonded 2:1 complexes (host to guest) were found for the methyl esters of longer dicarboxylic acids with Ka = 8.5 x 104-1.8 • 106 M -2, while dimethyl ethers of tri- and tetraethyleneglycol and 18-crown-6 form 1"3 (host:guest) complexes with 129 (Ka = 1.5 x 101~M-3). Shinkai et al. reported induced CD spectra for the interaction of the achiral dyes 132a,b with the chiral calix[6]arene hexasulfonate 1314 and for the recognition of the chiral (R)-biphenyl phosphoric acid 1324:264"266by the nonchiral nitrocalix[6]and -[8]arenes 13lb. The 1"1 stoichiometry of the complexes formed, was deduced from Job plots, while 1"2 (host-guest) stoichiometry was found for the CD active complex of 1314 with 1324. 26a The same group studied the complexation of the chiral ammonium salts 267 132d,e and of methyl esters of amino acids 268 by the nonchiral sulfonates of calix[4]-, -[6]-, and -[8]arenes 131c. In these cases the ICD spectra were dependent not only on the chirality of the guest molecule but also on the size of the macrocycle and the binding motif of the guest. This clearly limits (or at least complicates) the use of these systems as supramolecular probes for the determination of the absolute configuration of chiral guests.

E NO N=NO

131a R = SO3Na, Y = I 131b

~ n

a

R = CN

b

R=NO 2

R i , ~ ~

o" "OH

C

=6

R = NO2, Y = H, n = 6, 8

1 3 1 C R= SO3Na, Y = H, n =4, 6, 8

(R)-d

(S)-d

(R)-e 132

(S)-e

206

MYROSLAV VYSOTSKY, CHRISTIAN SCHMIDT, and VOLKER B(gHMER yI.y4

R:

4

O-C NH I R 133

7.2.

d

C.H3 C.H3 C-CH 2-C-CH 3 ell3 ell3

Y: CH 3, (CH~.-CH3, Bz, CH2COOEt

Tetraurea D i m e r s

Calix[4]arenes 133, substituted at their wide rim by four urea-functions, are self-complementary. In aprotic, apolar solvents such as benzene or CDCI 3, they exist as dimers held together by hydrogen bonds between the urea functions of opposite molecules as indicated in Figure 28. A suitably sized guest (often the solvent) is included in the cavity thus created. 269

Stereochemistry of Dimers This dimerization reduces the symmetry of the single calix[4]arene from a time-averaged Car symmetry to Ca symmetry stable on the NMR timescale. This can be deduced for instance from a pair of m-coupled doublets for the aromatic protons of the calixarene skeleton, while these protons appear as a singlet in DMSO, where the tetraurea exists in the form of single molecules (see Figure 29).

Figure 28. Schematic representation of the belt of hydrogen-bonded urea functions and shape of a dimer (R = Y = Me) with included benzene as modeled by MM3.

Chirality in Calixarenes and CalL~areneAssemblies

207 ArH

CsHI I

NH HN

No

HN

NH Nil

.

i-

i

9i-

v.

i-

ArH

ArH

.

.

i.

.

.

v-

,. . . . . . .

i - l -

9.5

v . i -

.. . . . .

i - | - l - l -

9.0

i

8.5

.

-

l

..

-

l

-

-

l

8.0

-

i

-

1 - 1 -

|-

i-

i-

i - 1 - | - i -

7.5

_~

~.

i - i ' - i -

7.0

.

l-

i"

I ' ' l V l "

6.5

I ' 1 "

6.0

I ' -

ppm

Figure 29. Section of the NMR spectra of a tetraurea derivative 133c in DMSO-d6 (above) and CDCI3 (below). The loss of the symmetry planes intersecting the single phenolic units can be seen also by the signals of enantiotopic protons or groups in the residues Y or R, which become diastereotopic in the dimer (Figure 30). The dimerization was established also by the observation of upfield shifted signals (AS =-2.5-3.0 ppm) for the included guest, 27~by ESI-MS studies on dimers 0~0~

benzene%, methanol-d, (5%)

sill1

HN

,,

3=0

HN

.;

"%o

"2,.o"',.;'" !

6

.

.

.

.

v

.

.

.

.

I

4

.

.

.

.

w

9

ppm

9

9

~i

i i i I iIi

3.0

2.5

i i i i i ill

2.0

i ii

1.5

i I i ii'i

1.0

i | i'

ppm

Figure 30. Splitting of signals for enantiotopic protons/groups in 133a,d upon dimeri-

zation.

208

MYROSLAV VYSOTSKY, CHRISTIAN SCHMIDT, and VOLKER BOHMER

including tetraalkylammonium ions 271 and the shape was additionally confirmed by a single-crystal X-ray analysis 272 shown in Figure 3 1, which confirms entirely the calculated shape. The stereochemical properties of tetraurea dimers are intriguing and challenging as well. As already mentioned for the simplest case, the directionality of the C--O groups makes each single calix[4]arene in the dimer chiral (C4-symmetry), while the dimer as a whole is an achiral m e s o form (Ss-symmetry) composed of a pair of enantiomers. Usually in a mixture of two tetraureas A and B, differing for instance by R or Y, the three dimers AA, AB, and BB are formed in an approximately statistical ratio of 1:2:1, regardless of whether R is aliphatic or aromatic. 273 However, if tosyl substituted tetraureas (R = SO2-C6Ha-CH3) are combined with aryl substituted ones (e.g. R = C6H4-CH3) only heterodimers are formed, 274 while the combination of tosyl with aliphatic residues leads to less than 10% of the heterodimer. The exact reason for this serendipitous observation is not really known, but a favorable combination of NH acidity and O--C basicity may be assumed. Hetero dimerization may be caused also by steric interactions. Tetraureas having bulky groups such as R = C ( C 6 H 5 ) 3 do not homodimerize, but form (nearly exclusively) heterodimers

Figure 31. Single-crystal X-ray structure of 133a (Y = CH2COOEt).

Chirality in Calixarenes and Calixarene Assemblies

209

with a tetraurea having R = C 6 H 4 - C ( C 6 H 5 ) 3. This can be understood as a compromise between a minimum of steric strain and a maximum of hydrogen bonded urea functions. 275 For a heterodimer formed by two different C4v-symmetrical tetraureas A and B, the directionality of the carbonyl groups leads to an overall C4-symmetrical dimer I (Figure 32), existing as a pair of enantiomers (AP.BM and AM.BP). Let us now consider tetraurea derivatives possessing two different phenolic units A and B. This difference may be due to different residues R or Y or both. Tetraurea derivatives of the type ABAB are C2v-symmetrical. Their dimerization leads to chiral capsules II with an overall C 2 symmetry. The situation is even more complicated for a Cs-symmetrical derivative AAAB. Two regioisomeric dimers are possible here, with a proximal (IIIa) or distal (IIIb) arrangement of the phenolic unit B. Both are asymmetric and consequently chiral. (The same is true for a derivative AABB, not shown in Figure 32.) It is important to note, that in the heterodimer I the chirality is due to the directionality of the carbonyl groups. The change of the C - ' O direction creates the opposite enantiomer for each tetraurea part (AAAA and BBBB) and consequently also for the whole dimer I. In the homo-dimers II and III the chirality stems from the mutual arrangement of the both calixarenes. Although both calix[4]arene parts form a pair of enantiomers (due to the directionality of the C - - O groups) if they are considered alone, they are diastereomers due to their combination within the capsule, giving two sets of signals in the NMR spectrum. The change of the C = O direction converts each single calixarene into its enantiomer and thus leads to the same enantiomer for the homo-dimers II and III as a whole. I

II B

Ilia A

B

A

/

Y B

-%

A

C4v

B

B

B

D2

A

C4

A /

A

A

lUb B

B

A Y A

~

A

C2

B

C2

A

f

A Y'8 A C2

B

C1

A

C1

Figure 32. Schematic representation of the stereochemical properties of some heteroand homodimers. The symmetry classes are indicated for the whole dimer without and with directionality of the C--O groups (indicated by arrows).

MYROSLAV VYSOTSKY, CHRISTIAN SCHMIDT, and VOLKER BOHMER

210

Figure 33 shows the aromatic part of the 500 MHz IH NMR spectrum of such a dimer of type II (R = C6Ha-CH 3, y1 = CH3, y2 = CsHI1)" As pointed out a C 2 axis is the only symmetry element of the dimer the two calixarene parts of which are nonequivalent. Thus, for the aromatic protons of each of the two different phenolic units four different positions exist, while only two positions for A r - H exist in the usual Ss-symmetrical dimer. This allows the study not only one but three independent exchange reactions (I-III) by NMR (a fourth one leading to the identical situation cannot be detected). Two of them (II, III) are possible only by dissociation/recombination while the third one, the change of the direction of the C--O functions could occur also within the dimer. An identical rate for all three processes (0.055-0.069 s -1) suggests the same mechanism, and leads to an overall first order rate constant for the dissociation/recombination process of kd/r = 0.26 + 0.06 s -1 (in benzene at room temperature). This is in reasonable agreement with the rate constant obtained for the exchange between free and encapsulated benzene ke = 0.47 + 0.1 s -1 (determined by NOESY in a mixture of 77% C6H6 and 23% C6D6).

"a" \

,,,II

,,,II

I ~=~'~

\

I, o

I, o

I/=9~,~'~,~

.

0

0 "CH

zoo'

9.8

A'

9

fa,

"C~HI

'

e~

~a' 3a ~I~ 31Y

13' 2Gt

9 ~ , !

3

"~|

8.4

o

. . . .

B'

!

. . . .

|

8.0

. . . .

!

. . . .

'a' 'a

!

. . . .

7.6

"a" "13 "a '13'

't3" '13

,

. . . .

!

B o'A

. . . .

7.2

!

. . . .

!

6.8

. . . .

!

. . . .

|

6.4

. . . .

i

ppm

Figure 33. Section of the 1H NMR spectrum (500 MHz) of 133a (Y = CH3/CsH11) in benzene-d6 (bottom) and schematic representation of the possible homo-oligomerization pathways (top). The schematic formula explains the peak assignment (based on NOEs/ROEs) and shows the exchange cross-peaks in NOESY experiments.

211

Chirality in Calixarenes and Calixarene Assemblies

If additionally chiral centers (e.g. asymmetric C-atoms, characterized here by D and L) are incorporated into the ether or urea residues two diastereomeric forms of this single tetraurea LA (LAP, LAM) are possible in a dimer due to the orientation of the carbonyl groups (indicated by P and M). Both forms must be present in a homodimer of such a tetraurea and for a heterodimer formed with a "normal" tetraurea B two diastereomeric forms (LAP.BM, LAM.BP) should exist. Eventually, this additional chiral center could even induce a certain direction of the C = O groups (e.g. LAP being more stable than LAM, and consequently DAM more stable than DAP). This has been studied recently in some detail by Rebek et al. who prepared a series of tetraureas 134 derived from various L-amino acids. 276 Among these only the isoleucine (134c) and the valine (134d) derivative show fairly sharp NMR spectra in C6D 6 which reveal the symmetry expected for homodimers. The latter, for example, shows four singlets for NH protons and four m-coupled doublets for the calixarene Ar-H protons (two from each half) as well as two singlets for the ester methyl group. (One might expect that the formation of heterodimers LAP.DAM or LAM.DAP should be preferred in a (racemic) mixture of tetraureas LA and its enantiomer DA over the homodimers LAP.LAM and D A M . D A P . However, at best a statistical heterodimerization has been observed for 134c or 134d.) On the other hand, if the L-isomers 134 are mixed with the tolylurea 133a, 100% heterodimerization is observed for the 13-branched amino acid derivatives 134c,d (derived from isoleucine and valine) in benzene-d 6. This preference for the heterodimerization is less pronounced for the other amino acid derivatives and also in the more polar CDC13 as a solvent (which is also known to be a less favorable guest). Moreover, only one of the two diastereomeric heterodimers LAP.BM or LAM.BP is observed in the case of 134c and 134d, and again this diastereoselectivity is less pronounced for the other examples. .01oH21 0

IN~H1 O=C\/

4

NH

O~C/C'HI~,

I ,,O H3C"

R

134

a -CH2.CH

b

C d

-C4H9 ~CH3 "CH3 ,.CH3 -CH "C2Hs -cHCH3 "CH3

e

-CHz'O'C(CH3)3

f

-CH2 O

h

-CH3

212

MYROSLAV VYSOTSKY,CHRISTIAN SCHMIDT, and VOLKER B(3HMER

CD-spectroscopy has been used to assign the absolute stereochemistry of the heterodimer of 134r with the tetratolylurea 133a. While 134c alone shows only a weak CD maximum (Ae < 10 at 260 nm in CHC13) addition of 133 leads to a large bisignate CD response (up to Ae - 100), which disappears on addition of methanol. From this exciton-coupled CD spectrum (in combination with molecular modeling) it has been concluded that the C--O groups point in a clockwise direction if the heterodimer is regarded from the tolylurea side.

Chiral Guests Chiral discrimination of guests by inclusion in hydrogen-bonded capsules of tetraureas is a highly interesting and challenging problem. The analysis of such systems by NMR spectroscopy is complicated by two facts: 1. The formation of these capsules is reversible, and thus, chiral discrimination (a given chiral host distinguishes between two enantiomeric guests) is superimposed by chiral induction (a certain chiral guest induces the formation of one of two possible enantiomeric host capsules). 2. More than a single set of signals may be observed for a nonchiral guest (e.g. tricyclene) in a "racemic" capsule of S8 symmetry; this can be understood by assuming that there are different orientations of the guest that exchange slowly on the NMR timescale. Inclusion of a chiral guest (like (1R)-(+)-nopinone) in a Ss-symmetrical meso-dimer leads to a doubling of the NMR signals for the calixarene since two diastereomeric host-guest pairs exist now within a dimer.277 They may be characterized by (R)-P and (R)-M, if the chirality of the calix[4]arene part is indicated by P/M and (R)/(S) is used for the included guest. This holds as long as the direction of the C-'-O groups is fixed on the NMR time scale, even if the guest is rapidly tumbling inside the cavity. It should even be true if the guest is offered as a racemic mixture, since only one guest molecule is included. Then (S)-P and (S)-M pairs would be present in addition, the mirror images of the above mentioned species. (The whole capsules may be symbolized by P-(R)-M and P-(S)-M.) Inclusion of a chiral guest (R) in a chiral heterodimer of type I should lead to two diastereomeric host-guest complexes AP-(R)-BM and AM-(R)-BP indicated by the same doubling of the NMR signals (for both tetraurea parts A and B) as before. However, the diastereomeric host-guest pairs (R)-AP/(R)-AM and (R)-BPI(R)-BM exist now in different (enantiomeric) capsules and must not appear in the ratio 1:1. (In an extreme case, one of the two diastereomeric complexes AP-(R)-BM and AM-(R)-BP could be (strongly) favored, or in other words, the guest could induce the chirality of the reversibly formed capsule.) This guest-induced chirality is discussed more easily with a dimer of type II (or III), that is chiral also when the direction of the carbonyl groups changes rapidly.

Chirality in Calixarenes and Calixarene Assemblies

213

This chirality may be indicated by [R]/[S] for the whole capsule (M[R]P/M[S]P if additionally the direction of the C--O groups should be differentiated for the two calixarenes). With a chiral guest (R), the two diastereomeric capsules [R]-(R) and [S]-(R) may be formed in a ratio (x) different from 1. If a racemic guest mixture (R)/(S) is offered, the capsules [S]-(S) and [R]-(S) (as mirror images) are formed in the same ratio x, and signals for the host and for the guest should be split in this ratio in both cases. If the direction of C--O is "fixed" on the NMR timescale, an additional splitting occurs for the host signals of each diastereomeric complex, and the situation becomes even more complicated, if the guest is not freely tumbling within the cavity. Experimentally a splitting of the downfield NH-resonances in the ratio 1.3:1 could be observed for the inclusion of racemic norcamphor in the chiral capsules formed by 134c and tolylurea 133a and for the racemic heterocapsule (type I) formed by tolyl- and tosyl-ureas 133a and 133b. This must be interpreted as chiral discrimination in the first case (since obviously the directionality of the C = O groups is not changed by the guest) and chiral induction in the second case. A 1.3:1 mixture of two diastereomers is also observed for (R)-(+)-3-methylcyclopentanone with the tolyl/tosyl heterodimer while only one assembly is formed with 134c/133a as deduced from a single doublet for the methyl group at -2.08 ppm (in p-xylenedl0). This is split in the ratio 1:1 if racemic 3-methylcyclopentanone is offered as guest, indicating the absence of any detectable enantioselectivity in this case.

7.3. Melamine-Barbiturate Systems Melamines and barbiturates or cyanurates are complementary with respect to their ability to form hydrogen bonds. Three intermolecular hydrogen bonds between adjacent molecules may lead to hydrogen-bonded polymeric or oligomeric structures in the crystalline state. 278 It was demonstrated that hydrogen bonded hexamers ("rosettes") can be formed not only in the crystalline state but also in apolar solvents and that various well-defined self-assembled species can be obtained from building blocks in which melamine and barbiturate/cyanurate fragments are covalently attached in an appropriate way to a molecular skeleton. 279 Reinhoudt et al. have chosen tetraether derivatives of calix[4]arenes fixed in the cone-conformation as such a molecular skeleton. 28~B is-melamine derivatives 135 are easily prepared in a huge diversity (different residues Y, R l, R 2) from 1,3diamino calix[4]arenes by reaction with cyanurchloride, followed by stepwise substitution of the remaining chlorine atoms by ammonia and an aliphatic amine. In fact, as expected, the interaction of these calixarene derivatives with barbiturates e.g. 136 (or cyanurates 138) in the ratio 1:2 results in the formation of "box-like" assemblies consisting of nine particles (3 x 135 and 6 x 136) held together by 36 hydrogen bonds. These aggregates are stable on the NMR timescale at ambient temperature in apolar solvents such as CDCI 3 or toluene-d 6. However,

214

MYROSLAV VYSOTSKY, CHRISTIAN SCHMIDT, and VOLKER BOHMER

R R ..H'N~N~N'H.. O" .N N. "O X~N H'" ~N . "'H.N~X 9 .H" H. I

O~N~O -" i

9

i

I

.H'N'.I~N,~t~'H.. .H't~"1~N,~tq'H.. R

"-O~N~ o i

i

-NyNyNH.. .H'N~N~N"R N~N, "O'" ,N~N R.N.H "'HN~N'H'" H.N.R "'-o~x~o,'"

I

R

R

R

"'O" .N.N. "O'" .N...N.. "-H.a.J,l,,a.H-" ' ~ "-H.N.JJ,.N.H-'H.NT ..O~,X,,~O...n H...O~,.X.,~.O.. "H,. X = CR z,

NR

R = H, aryt,alkyl

Figure 34. Hydrogen-bonded cyclic hexamers I ("rosettes") or infinite tapes II by self-assembly of melamines and barbitures/cyanurates. they are completely destroyed by the addition of small amounts of polar solvents such as DMSO or MeOH, which can act as hydrogen bond donor or acceptor. In order to form such an aggregate, the bis-melamine calix[4]arene 135 may assume either a C2-symmetrical (staggered) conformation, in which the residues R 1 are in an identical environment, or a Cs-symmetrical (eclipsed) conformation with two different environments for the residues R 1. As illustrated in Figure 35, three different diastereomeric boxes may be formed by combination of three molecules of 135 with six molecules 136. TM Usually the 1H NMR spectra of the aggregates (135)3.(136) 6 are simple, exhibiting for instance only two sharp singlets for the imide protons of the barbiturate (at 13-14 ppm in CDC13). This strong downfield shift is characteristic for the melamine-barbiturate hexamer, while the simple pattern in general is only in agreement with the formation of the chiral assembly with an overall D 3 symmetry. However, the simultaneous formation of all three supramolecular diastereomers was detected by 1H NMR spectroscopy in the case of sterically hindered cyanurates (e.g. R = CH2-CH2-C(CH3) 3, C6Ha-C(CH3)3, CH2-CH2-C(C6Hs)3). TM

Chirality in Calixarenes and Calixarene Assemblies

215

Figure 35. Three possible diastereomeric boxes and their symmetry. The combination of C2- and G-symmetrical bis-melamines is not possible in one aggregate.

An additional structural proof for the chiral, D3-symmetrical arrangement was found for one example by single crystal X-ray analysis (Figure 36). The calix[4]arene molecules assume a strongly pinched C2-symmetrical cone conformation, in which the two opposite aryl rings bearing the melamine fragments are nearly parallel. This conformation, dictated by the planar, disklike shape of the rosette precludes the inclusion of guests in the boxes. The D3-symmetrical boxes exist in solution as a pair of enantiomers and (rapid) enantiomerization is likely due to the reversibility of the hydrogen bonds. If an

Figure 36. Single crystal X-ray structure of 135 (Y =

C3H7, R1 = N O 2 , R 2 = C4H9). Butyl and propoxy groups of 135 as well as ethyl groups of 136 are omitted for clarity.

216

MYROSLAVVYSOTSKY, CHRISTIAN SCHMIDT, and VOLKER BOHMER

o

N,.~I-N O~-N~'O Io

R

N='~

/~'N

NH2 NH2

.~ CH3

R"

\CH3 oII

138a

II C R" = -Cl~l c'OCH3

138

b"~ oII

,/,c'OCH3 R"= - c~'C"oCH3 R" = --CM

CH3

137

"

/C.ocH3 R'= -CH

R'= -C R'.N_x/NH HN .'~)r-N

O

138c

CH(CH3)2

138d

CH2CH(CH3)~

Figure 37. Chiral bis-melamine derivatives and chiral cyanurates used in self-assem-

bly studies.

additional stereogenic center is introduced in R 2 or R (Figure 37) two diastereomers with D 3 symmetry could be formed. However, the 1:2 mixture of (R,R)-I37a or (S,S)-137a with 136 in CD2C12 leads to a single assembly only, as judged on the basis of a single set of signals in the IH NMR spectrum, from which a diastereomeric excess >98% is estimated. 282ROESY experiments allow correlation of the absolute configuration of the substituents R* with the helicity of the assembly: an (R)-substituent induces M-helicity and an (S)-substituent P-helicity. While the individual compounds 137a show no significant CD activity (AEma x < 8 cm 2 mmol -]) a strong CD band with opposite sign (AEma x X 100 cm 2 mmo1-1) is found for the boxes 13731366 . Complete induction ofchirality was also observed for (R,R)-137b and (R,R)-137c in combination with 136 and for the combination of the achiral bis-melamine 138b with the chiral cyanurates (R)- or (S)-138a, (S)-138b, (S)-138c and (S)-138d which all form well defined D3-symmetrical assemblies. The meso compound (R,S)-137a on the other hand forms only non-defined oligomeric aggregates with 136. Mixtures of the enantiomeric assemblies (M)-(R,R)-I37a31366 (=I) and (P)-(S,S)-137a31366 (=II) show no additional NMR signals which proves that no heteromeric assemblies involving both enantiomers (R,R)-137a and (S,S)-137a are formed. This is further Y Y

y

Y

R

i

-,N.X.N.H

oA ', o H

139

Chimlity in Calixarenes and Calixarene Assemblies Rz H "N )='-N

Pr ~

O~.~' Pr ~ - ~

H

N .)--N

0"~'~ k Pr-o.~. ~"N

pr-O--<2

217

~-N H

H

,);'-N H N ')'-- N )='N N

/H

X

140 H

N ~_N

H

Pr~o.~~"

N~..N)" NH

Pr.o.~~'~

NH

o~.~

"~N

H

)-:-N

H

Pr

~

N

Rz-N H

')'-"N

X = (CH2)6 , CH2

CHz

confirmed by the linear decrease of the CD intensity to zero in going from 100% I (or II) to the racemic mixture. Heteromeric D3-symmetrical assemblies are formed, however, for mixtures of 135b and (R,R)-137a with 136. The non-linear dependence of the CD on the composition of such mixtures shows that the achiral units 135b reinforce the chirality induced by the chiral units (R,R)-137a. Stereogenic centers in the bis-melamine may also control the formation of the different diastereomeric combinations shown in Figure 35. While 135b and 138e form simultaneously the three assemblies with D3-, C3h-, and Cs-symmetry, only the C3h- and the Cs-symmetrical assemblies are formed by (R,S)-137a and 138e and exclusively the D 3 isomer by (R,R)-137a and 138e. It should be mentioned, that larger assemblies consisting of nine calixarenes are available with the calixarene-monocyanurate 139, 2~ while the double-calixarene 140 forms a double-box with 136 consisting of 15 molecules (3 x 140 and 9 x 136). 283 Polymeric aggregates are formed between bis-melamine and bis-cyanurate calix[4]arenes. 284 In all these cases, chiral induction still remains to be explored.

7.4. Chiral Assemblies in the Crystalline State To discuss self-assembly in the crystalline state is somehow "problematic" since all crystals are self-assembled. (For example in sodium chloride, six chloride anions

218

MYROSLAV VYSOTSKY, CHRISTIAN SCHMIDT, and VOLKER B(3HMER

Figure 38. Single crystal X-ray structure of 126 with included pyrazine (DBU cations are not shown). are "self-assembled" in an octahedral fashion around a sodium cation and vice versa.) Therefore the expression should be used only if a certain, special (structural) property is connected with or found in these self-assembled structures. Chirality in an assembly of nonchiral subunits is clearly such a property, and we will discuss in the following sections some examples in the field of calixarenes. As already mentioned, cavitand 126 forms a hydrogen-bonded capsule in the presence of a base. A single crystal X-ray structure could be obtained for the inclusion complex with pyrazine. 285 Figure 38 shows its overall D 4 symmetry (not regarding the guest) due to the twisting of the two cavitands. As mentioned before, this chiral arrangement is also found at low temperature in solution. A chiral arrangement of six molecules of 141a and eight water molecules held together by 60 hydrogen bonds in an octahedral-cubic assembly was found by McGillivray and Atwood (Figure 39). 286 Although the hydrogen atoms of the hydroxyl groups could not be determined experimentally, it was postulated that in order to reach the optimal saturation of hydrogen-bond donors and acceptors four molecules should have a Ca-symmetrical arrangement of the hydrogen bonds (chiral sites) while the two remaining ones should be C2v-symmetrical (achiral sites; compare Figure 23). Under this assumption two types of contacts between three molecules of 141a are possible in which the water molecule forms two hydrogen bonds as a donor and one as an acceptor and vice versa. If such an assembly of six molecules 141a would exist also in solution, as the authors suggest, this chirality would not persist on a longer time scale. A similar octahedral arrangement of six molecules was observed for the pyrogallol-derived resorcarene 14lb. 287 The molecules are held together by 72 hydrogen bonds between the 72 phenolic hydroxyl groups and again a Ca-symmetrical arrangement of the hydrogen bonds must be assumed to reach the mutual saturation of all hydrogen bond donors and acceptors (Figure 39). Calix[4]arene 142 in the 1,3-alternate conformation has two pairs of pyridine residues (hydrogen-bond acceptor) pointing in opposite directions. Together with

Chirality in Calixarenes and Calixarene Assemblies

219

Figure 39. Self-assembled octahedral "containers" formed by six molecules 141a (left) or 141b (right). 4,4'-biphenol (as hydrogen-bond donor) it forms a one-dimensional network via N...HO bonding. 288 Only two pyridine residues of adjacent phenolic units are involved in these hydrogen bonds (the other two are in close proximity to nitromethane each), a situation similar to the chiral calixarenes of the AABB type in the 1,3-alternate conformation. This chiral situation of the single molecule is translated along the hydrogen-bonded chain of molecules into a single stranded helix. Five of this equally handed helical strands are arranged together to form a fight handed, quintuple helical strand. The pitch of the helix is composed of four molecules 142, four molecules of 4,4'-biphenol and eight nitromethane molecules (Figure 40). Very appealing supramolecular structures have been obtained 289 with the threecomponent system consisting of the p-sulfonatocalix[4]arene 143290 (used as its pentasodium salt), pyridine N-oxide and La(NO3)3.6H20. If these components are mixed in a 2:2:1 molar ratio 12 calixarene pentaanions, shaped like a truncated pyramid, form spherical supramolecules with the phenolic OH functions inside and the sulfonato groups outside. Coordination of the sulfonato groups and of pyridine N-oxide included in the calixarene cavity to the La 3§ stabilizes this assembly. Each sphere is in contact with six other spheres via C-shaped dimers, as shown in Figure 41. If the components are mixed in a 2:8:1 molar ratio, a molecule of pyridine N-oxide intercalates between the pentaanions, as shown in Figure 42, and these

220

MYROSLAVVYSOTSKY,CHRISTIAN SCHMIDT, and VOLKER BC)HMER

~ 3H7 O

4

142

Figure 40. Helical arrangement of hydrogen bonded chains of 142 and 4,4'-biphenol. Five of these strands form a super-helix.

143

o3s~ ~ " o - -

-

'

~i"

~r--_(

o o~O

6

o.~.o o.~.o

9

so3

H~C~//OHz HzO~L~"O~ H20~L~'OH2 O~S

o;~ ~

'~

"

III r~, "~'o,

H ~ SO3 HOHoO./q

Figure 41. Section of the spherical arrangement of 143 (right side; six of twelve molecules are shown) and contact between spheres (left side).-289

Chirality in Calixarenes and Calixarene Assemblies

H,O 9H, o 0 0 Hz \ I / "S O--Na-. 0 \

221

S.Oa SO, SO, ~ I "1 "

HzOTNa--O, oo

o.o o.o.

Figure 42. Helical arrangement of 143 (right side; twelve molecules of the helix are shown) and contact between molecules in the helix. 289

structural elements form helical tubular assemblies with 4,5 pairs of 143 and Py-NO in alternating order per turn. The helical tubes are organized in a hexagonal array connected by C-shaped dimers similar to Figure 41. Analogous supramolecular structures have been obtained with a variety of other lanthanide cations proving some generality of this "noncovalent synthetic approach". The future will show if, by appropriate modifications, useful functions will emerge.

8. CONCLUSIONS AND OUTLOOK We have tried to give an overview, general, and comprehensive, but not necessarily complete and exhaustive, of chiral structures associated with macrocyclic compounds known in the broadest sense as "calixarenes." Necessarily, the picture arising in this way is complex and heterogeneous, perhaps even confusing. We hope, however, that we have at least left the reader with a feeling for the fascination that may be connected with such chiral structures and for the various possibilities that exist for calixarenes. The area covered is very widespread and the role of the calixarene molecules reach from a simple platform or skeleton on which to assemble chiral centers to an inherent part of the chiral structure. Biologically active molecules or derivatives are involved as well as artificial ligands and their metal complexes. Chiral calixarenes have been used as stationary phases in analytical separations or as host molecules in sensors. Basic properties of calixarenes, such as their conformational stabilities, have been studied with chiral derivatives as well as more

222

MYROSLAV VYSOTSKY,CHRISTIAN SCHMIDT, and VOLKER BOHMER

general questions such as induced chirality or chiral induction in supramolecular assemblies. Many of the molecules described have been clearly prepared without any immediate purpose, but this is often the case, if a new class of compounds is to be explored. During the first stages it is just important to establish what is possible. However, after about 20 years of modern calixarene chemistry and after the accumulation of a huge body of knowledge concerning their properties and reactivities, the time seems ripe to use the potential of their chirality in a more active sense. What could be further developments? Enantioselective catalysts may be named as one important field. The construction of calixarene-based, eventually biomimetic, catalysts is still in its infancy. 291 The simultaneous attachment of catalytically active groups in combination with chiral functions or with the inherent chirality of the calixarene skeleton itself opens numerous possibilities that are far from being exhausted. Even the combination of achiral calixarenes with chiral catalysts may lead to useful results. 292 Water-soluble calixarene derivatives have increasingly been synthesized with the aim to use them for biological and medical purposes. Here "chirality" is clearly one of the key factors. Strategies of combinatorial chemistry, often used in this area, may increase the number of available structures. 293 Supramolecular structures obtained via "self-assembly processes" may be named as another future topic, which just starts an increasingly rapid development. 294 By some examples it was already demonstrated how molecular chirality may by translated into supramolecular chirality in such self-assembled structures and this may lead not only to new, improved host-guest systems but also eventually to "new materials" in a more general sense. Liquid-crystalline systems or materials for nonlinear optics may be mentioned just as two exampies. In addition to (potential) applications, calixarenes will serve in the future as models or model compounds in fundamental or basic research. In that sense their potential in general as well as their potential to create chiral structures remains inexhaustible.

REFERENCES A N D NOTES 1. For reviews on calixarenes see: (a) Gutsche, C. D. Calixarenes Revisited; The Royal Society of Chemistry, 1996. (b) B~hmer, V. Angew. Chem., Int. Ed. Engl. 1995, 34, 713-745. 2. For the synthesis of t-butyl-calix[4], [6], and [8]arenes see: (a) Gutsche, C. D.; Iqbal, M. Org. Synth. 1990, 68, 234-237. (b) Gutsche, C. D.; Dhawan, B.; Leonis, M.; Stewart, D. Org. Synth. 1990, 68, 238-242. (c) Munch, J. H.; Gutsche, C. D. Org. Synth. 1990, 68, 243-246. Even the pentamer (t-butyl-calix[5]arene) can be prepared in about 20% yield, astonishing still for a reaction in which 10 molecules react forming 20 new covalent bonds, see: Stewart, D. R.; Gutsche, C. D. Org. Chem. Proc. Int. 1993, 25, 137-139. 3. Brhmer, V. Liebigs Ann./Receui11997, 2019-2030.

Chirality in Calixarenes and Calixarene Assemblies

223

4. Shinkai, S. Tetrahedron 1993, 49, 8933-8968; Shinkai, S. In Adv. Supramol. Chem.; Gokel, G. W., Ed.; 1993, Vol. 3, pp. 97-130. 5. For a review see: Timmerman, P.; Verboom, W.; Reinhoudt, D. N. Tetrahedron 1996, 52, 2663-2704. 6. This name, used throughout this article, is justified mainly by its similarity to "calixarenes" which indicates the cup-like shape (calix) and the aromatic units (arenes) of the macrocycle. In the line with newly created names such as calixpyrroles, calixfuranes, calixindoles, etc. it would be reasonable to distinguish between calixphenols (I) and calixresorcinols (II) or calixpyrogallols etc. and to use calixarenes as the general name for the whole class of [ ln]-metacyclophanes (or -cyclophanes). 7. See for example: Chirality and Biological Activity; Holmstedt, B., Frank, H., Tesla, B., Eds.; Liss, New York, 1990. 8. For earlier reviews on inherently chiral calixarenes and related compounds see: (a) B6hmer, V.; Kraft, D.; Tabatabai, M. J. Incl. Phenom. Mol. Recogn. 1994, 19, 17-39. (b) Otsuka, H.; Shinkai, S. Supramolecular Science 1996, 3, 189-205. 9. The exact value depends on the solvent and slightly also on the p-substituent. For first experimental values see: (a) Happel, G.; Mathiasch, B.; K~immerer, H. Makromol. Chem. 1981,176, 3317-3334. (b) Gutsche, C. D.; Bauer, L. J. Am. Chem. Soc. 1985, 107, 6052-6059. (c) Araki, K.; Shinkai, S.; Matsuda, T. Chem. Lett. 1989, 581-584. 10. See ref. 1a, pp. 63-66. 11. Moran, J. R.; Karbach, S.; Cram, D. J. J. Am. Chem. Soc. 1982, 104, 5826-5828. 12. (a) Cram, D. J.; Cram, J. M. Container Molecules and Their Guests; The Royal Society of Chemistry, 1994. (b) Quan, M. L. C.; Cram, D. J. J. Am. Chem` Soc. 1991, 113, 2754-2755. (c) Quan, M. L. C.; Knobler, C. B.; Cram, D. J. J. Chem. Soc., Chem. Commun. 1991, 660-662. (d) Jasat, A.; Sherman, J. C. Chem. Rev. 1999, 99, 931-967. 13. In the sense of the present discussion it is not relevent if the chiral centers are introduced into the phenolic units after the calixarene has been formed or before. 14. There is no example known yet from the calix[n]arene series, where compounds with -CHRbridges are scarce anyway: (a) Biali, S. E.; B6hmer, V.; Cohen, S.; Ferguson, G.; Grtittner, C.; Grynszpan, E; Paulus, E. E; Thondorf, I.; Vogt, W. J. Am` Chem` Soc. 1996, 118, 12938-12949. (b) Biali, S. E.; B6hmer, V.; Columbus, I.; Ferguson, G.; Griitmer, C.; Grynszpan, E; Paulus, E. E; Thondorf, I. J. Chem. Soc., Perkin Trans. 2 1998, 2261-2269. 15. Jauch, J.; Schurig, V. Tetrahedron Asym. 1997, 8, 169-172. 16. Soi, A.; Pfeiffer, J.; Jauch, J.; Schurig, V. Tetrahedron Asym. 1999, 10, 177-182. 17. Sansone, E; Barboso, S.; Casnati, A.; Fabbi, M.; Pochini, A.; Ugozzoli, E; Ungaro, R. Eur. J. Org. Chem` 1998, 5, 897-905. 18. Sansone, E; Barboso, S.; Casnati, A.; Sciotto, D.; Ungaro, R. Tetrahedron Lett. 1999, 40, 4741-4744. 19. Xie, D.; Gutsche, C. D. J. Org. Chem. 1997, 62, 2280-2284. 20. Hunter, R.; Koch, K. R.; Verblun, B. S. Afr. J. Chem. 1998, 51,200-206. 21. Pinkhassik, E.; Stibor, I.; Casnati, A.; Ungaro, R. J. Org. Chem. 1997, 62, 8654-8659. See also: Pinkhassik, E.; Stibor, I.; Havlicek, V. Coll. Czech. Chem. Commun. 1996, 61, 1182-1190. 22. lhm, H.; Paek, K. Bull. Korean Chem. Soc. 1998, 19, 492-495. 23. Casnati, A.; Fabbi, M.; Pelizzi, N.; Pochini, A.; Sansone, E; Ungaro, R. Bioorg. Med. Chem` Lett. 1996, 6, 2699-2704. 24. Muthukrishnan, R.; Gutsche, C. D. J. Org. Chem. 1979, 44, 3962-3964. 25. For the X-ray structure of a diester derived from t-butylcalix[4]arene see: Bavoux, C.; Perrin, M.; Chino, E Supramol. Chem. 1994, 4, 63-68. 26. (a) Arimura, T.; Kawabata, H.; Matsuda, T.; Muramatsu, T.; Satoh, H.; Fujio, K.; Manabe, O.; Shinkai, S. J. Org. Chem. 1991, 56, 301-306. (b) Ikeda, A." Nagasaki, T." Shinaki, S. J. Phys. Org. Chem. 1992, 5, 699-710.

224

MYROSLAV VYSOTSKY,CHRISTIAN SCHMIDT, and VOLKER BOHMER

27. Otsuka, H.; Araki, K.; Shinkai, S. Chem. Express 1993, 8, 479-483. 28. Pefia, M. S.; Zhang, Y.; Thibodeaux, S.; McLaughlin, M. L.; Mufioz de la Pefia, A.; Warner, I. M. Tetrahedron Lett. 1996, 37, 5841-5844. 29. Pefia, M. S.; Zhang, Y.; Warner, I. M. Anal. Chem. 1997, 69, 3239-3242. 30. Grady, T.; Joyce, T.; Smyth, M. R.; Harris, S. J.; Diamond, D.Anal. Commun. 1998, 35, 123-125. 31. Grady, T.; Harris, S. J.; Smyth, M. R.; Diamond, D. Anal. Chem. 1996, 68, 3775-3782. 32. Loeber, C.; Wieser, C.; Matt, D.; De Cian, A.; Fischer, J.; Toupet, L. Bull. Soc. Chim. Fr. 1995, 132, 166-177. 33. Molard, Y.; Bureau, C.; Parrot-Lopez, H.; Lamartine, R.; Regnouf-de-Vains, J.-B. Tetrahedron Lett. 1999, 40, 6383-6387. 34. Nagasaki, T.; Fujishima, H.; Shinkai, S. Chemistry Lett. 1994, 989-992. 35. Neri, P.; Bottino, A.; Geraci, C.; Piattelli, M. Tetrahedron." Asymmetry 1996, 7, 17-20. 36. Weber, C.; Diirr, H.; Bonneau, R. Synthesis 1999, 6, 1005-1012. 37. The situation would be more complicated, if the methoxy groups would be replaced by substituents, which cannot pass the annulus. 38. Kim, T.; Ihm, H.; Paek, K. Bull. Korean Chem. Soc. 1997, 19, 681-684. 39. (a) Kubo, Y.; Maeda, S.; Tokita, S.; Kubo, M. Nature 1996, 382, 522-524. (b) Kubo, Y.; Maeda, S.; Tokita, S.; Kubo, M. Enantiomer 1997, 2, 287-292. (c) Kubo, Y.; Hirota, N.; Maeda, S.; Tokita, S. Anal. Sci. 1998, 14, 183-189. 40. Yuan, H.-S.; Huang, Z.-T. Tetrahedron Asymm. 1999, 10, 429-437. 41. Hu, X.; Chan, A. S. C.; Han, X.; He, J.; Cheng, J.-P. Tetrahedron Lett. 1999, 40, 7115-7118. 42. Hu, X.; He, J.; Chan, A. S. C.; Han, X.; Cheng, J.-P. Tetrahedron Asymm. 1999, 10, 2685-2689. 43. Regnouf-de-Vains, J.-B.; Lamartine, R.; Fenet, B.; Bavoux, C.; Thozet, A.; Perrin, M. Helv. Chin~ Acta 1995, 78, 1607-1619. 44. Regnouf-de-Vains, J.-B.; Lamartine, R.; Fenet, B. Helv. Chim. Acta 1998, 81, 661-669. 45. Matsushita, Y.; Matsui, T. Tetrahedron Lett. 1993, 34, 7433-7436. 46. Yanagihara, R.; Tominaga, A. M.; Aoyama, Y. J. Org. Chem. 1994, 59, 6865-6967. 47. Schneider, U.; Schneider, H.-J. Chem. Ber. 1994, 127, 2455-2469. 48. Iwanek, W.; Wolff, C.; Mattay, J. Tetrahedron Lett. 1995, 36, 8969-8972. 49. Lukin, O.; Shivanyuk, A.; Pirozhenko, V. V.; Tsymbal, I. E; Kalchenko, V. I. J. Org. Chem. 1998, 63, 9510-9516. 50. Pfeiffer, J.; Schurig, V. J. Chromat. A, 1999, 840, 145-150. 51. Iwanek, W. TetrahedronAsym. 1998, 9, 3171-3174. 52. Kazakova, E. K.; Makarova, N. A.; Zotkina, V. V.; Burilov, A. R.; Pudovik, M. A.; Konovalov, A. I. Mendeleev Commun. 1996, 157-159. 53. Lewis, P. T.; Davis, C. J.; Saraiva, M. C.; Treleaven, W. D.; McCarley, T. D.; Strongin, R. M. J. Org. Chem. 1997, 62, 6110-6111. 54. Curtis, A. D. M. Tetrahedron Lett. 1997, 38, 4295-4296. 55. Botta, B.; Delle Monache, G.; Salvatore, P.; Gasparrini, E; Villani, C.; Botta, M.; Corelli, F.; Taft, A.; Gacs-Baitz, E.; Santini, A.; Carvalho, C. F.; Misiti, D. J. Org. Chem. 1997, 62, 932-938. 56. Pietraszkiewicz, M.; Prus, P.; Bilewicz, R. Polish J. Chem. 1999, 73, 2035-2042. 57. Judice, J. K.; Cram D. J. J. Am. Chem. Soc. 1991, 113, 2790-2791. 58. Park, B. S.; Knobler, C. B.; Eid, C. N.; Warmuth, R.; Cram, D. J. Che~ Commun. 1998, 55-56. 59. Yoon, J.; Cram, D. J. J. Am. Chem. Soc. 1997, 119, 11796-11806. 60. Gibb, B. C.; Mezo, A. R.; Sherman, J. C. Tetrahedron Lett. 1995, 36, 7587-7590. 61. Gibb, B. C.; Mezo, A. R.; Causton, A .S.; Fraser, J. R.; Tsai, E C. S.; Sherman, J. C. Tetrahedron 1995, 51, 8719-8732. 62. Mezo, A. R.; Sherman, J. C. J. Am. Chem. Soc. 1999, 121, 8983-8994. 63. Causton, A. S.; Sherman, J. C. Bioorg. Med. Chem. 1999, 7, 23-27. 64. Hamuro, Y.; Calama, M. C.; Park, H. S.; Hamilton, A. D. Angew. Chem., Int. Ed. Engl. 1997, 36, 2680-2683.

Chiralityin Calixarenes and Calixarene Assemblies

225

65. Park, H. S.; Lin, Q.; Hamilton, A. D. J. Am. Chem. Soc. 1999, 121, 8-13. 66. Carbohydrate-Protein Interactions; Clarke, A. E., Wilson, I. A., Eds.; Springer-Verlag: Heideberg, Germany, 1988. 67. Marra, A.; Scherrmann, M.-C.; Dondoni, A.; Casnati, A.; Minari, P.; Ungaro, IL Angew. Chem., Int. Ed. Engl. 1994, 33, 2479-2481. 68. Dondoni, A.; Marra, A.; Scherrmann, M.-C.; Casnati, A.; Sansone, E; Ungaro, R. Chem. Eur. J. 1997, 3, 1774-1782. 69. Marra, A.; Dondoni, A.; Sansone, E J. Org. Chem. 1996, 61, 5155-5158. 70. Dondoni, A.; Kleban, M.; Marra, A. Tetrahedron Lett. 1997, 38, 7801-7804. 71. Felix, C.; Parrot-Lopez, H.; Kalchenko, V.; Coleman, A. W. Tetrahedron Lett. 1998, 39, 81718174. 72. Saitz-Barria, C.; Torres-Pinedo, A.; Santoyo-Gonzales, F. Synlett 1999, 1891-1894. On the basis of NOE experiments the authors assigned a 1,2-alternate conformation (anti-orientation of the ether groups) to the final products. This is surprising, since in the starting amine the ether groups were in syn-position (cone-conformation) and the 2-aminoethyl groups should be too large to pass through the annulus during the reaction. 73. Roy, R.; Kim, J. M. Angew. Chem., Int. Ed. Engl. 1999, 38, 369-372. 74. Meunier, S. J.; Roy R. Tetrahedron Lett. 1999, 37, 5469-5472. 75. Fujimoto, T.; Shimizu, C.; Hayashida, O.; Aoyama, Y. J. Am. Chem. Soc. 1997, 119, 6676-6677. 76. Fujimoto, T.; Shimizu, C.; Hayashida, O.; Aoyama, Y. J. Am. Chem. Soc. 1998, 120, 601-602. 77. Hayashida, O.; Kato, M.; Akagi, K.; Aoyama, Y. J. Am. Chem. Soc. 1999, 121, 11597-11598. 78. Fujimoto, K.; Hayashida, O.; Aoyama, Y.; Guo, C.-T.; Hidari, K. I.-P. J.; Suzuki, Y. Chemistry Lett. 1999, 1259-1260. 79. Hayashida, O.; Nishiyama, K.; Matsuda, Y.; Aoyama, Y. Tetrahedron Lett. 1999, 40, 3407-3410. 80. Hayashida, O.; Shimizu, C.; Fujimoto, T.; Aoyama, Y. Chemistry Lett. 1998, 13-14. 81. (a) D'Allessandro, E; Gulini, E G.; Impellizzeri, G.; Pappalardo, S.; Rizzarelli, E.; Sciotto, D.; Vecchio, G. Tetrahedron Lett. 1994, 35, 629-632. (b) van Dienst, E.; Snellink, H. B. M.; van Piekartz, I.; Engbersen, J. E J.; Reinhoudt, D. N. J. Chem. Soc., Chem. Commun. 1995, 1151-1152. (c) BUgler, J.; Engbersen, J. E J.; Reinhoudt, D. N. J. Org. Chem. 1998, 63, 5339-5344. (d) Biigler, J.; Sommerdijk, N. A. J. M.; Visser, A. J. W. G.; van Hoek, A.; Nolte, R. J. M.; Engbersen, J. F. J.; Reinhoudt, D. N. J. Am. Chem. Soc. 1999, 121, 28-33. 82. For methylene-bridged resorcarenes see: Falana, O. M.; AI-Farhan, E.; Keehn, P. M.; Stevenson, R. Tetrahedron Lett. 1994, 35, 65-68. 83. Rumboldt, G.; B6hmer, V.; Botta, B.; Paulus, E. E J. Org. Chem. 1998, 63, 9618-9619. 84. Konishi, H.; Tamura, T.; Ohkubo, H.; Kobayashi, K.; Morikawa, O. Chem. Len. 1996, 685-686. 85. Griittner, C.; B6hmer, V., unpublished results. 86. Middel, O.; Greff, Z.; Taylor, N. J.; Verboom, W.; Reinhoudt, D. N.; Snieckus, V. J. Org. Chem. 2000, 65, 667-675. 87. Kumagai, H.; Hasegawa, M.; Miyanari, S.; Sugawa, Y.; Sata, Y.; Hori, T.; Ueda, S.; Kamiyama, H.; Miyano, S. Tetrahedron Lett. 1997, 38, 3971-3972. 88. Mislin, G.; Graf, E.; Hosseini, M. W.; De Cian, A.; Fischer, J. Tetrahedron Lett. 1999, 40, 1129-1132. 89. Litwak, A. M.; Biali, S. E. J. Org. Chem. 1992, 57, 1943-1945. 90. Aleksiuk, O.; Grynszpan, E; Biali, S. E. J. Chem. Soc., Chem. Commun. 1993, 11-16. 91. Grynszpan, E; Aleksiuk, O.; Biali, S. E. J. Org. Chem. 1994, 59, 2070-2074. 92. (a) Grynszpan, E; Aleksiuk, O.; Biali, S. E. Pure & Appl. Chem. 1996, 68, 1249-1254. (b) Litwak, A. M.; Grynszpan, E; Aleksiuk, O.; Cohen, S.; Biali, S. E. J. Org. Chem. 1993, 58, 383-402; similar bis(spirodienone) derivatives with C 2 and C i symmetry as well as a mono(spirodienone) calix[4]arenes with naphthol moieties were described: Georghiou, P. E.; Ashram, M.; Clase, H. J.; Bridson, J. N. J. Org. Chem. 1998, 63, 1819-1826. 93. Aleksiuk, O.; Cohen, S.; Biali, S. E. J. Am. Chem. Soc. 1995, 117, 9645-9652.

226 94. 95. 96. 97. 98. 99.

MYROSLAV VYSOTSKY,CHRISTIAN SCHMIDT, and VOLKER BOHMER

Van Gelder, J. M.; Brenner, J.; Thondorf, I.; Biali, S. E. J. Org. Chem. 1997, 62, 3511-3519. Van Gelder, J. M.; Aleksiuk, O.; Biali, S. E. J. Org. Chem. 1996, 61, 8419-8424. Mogck, O.; Btihmer, V.; Ferguson, G.; Vogt, W. J. Chem. Soc., Perkin Trans. 1 1996, 1711-1715. Yamato, T.; Hasegawa, K.; Saruwatari, Y.; Doamekpor, L. K. Chem. Ber. 1993, 126, 1435-1439. Biali, S. et al., unpublished results. The sterochemical situation would be even more complicated, if O-Y cannot pass the annulus, since then both O-Y groups of a given biphenyl unit may be at one or the other side (with respect to the other units) or on both sides of the macrocycle. 100. The single crystal X-ray structure of a tetradecylether revealed, that already at this stage of the alkylation process a configuration is fixed, which cannot lead to the D3-symmetrical hexaether; Paulus, E. E; Jakobi, R. A.; Bthmer, V., unpublished results. 101. O'Sullivan, P.; Bthmer, V.; Vogt, W.; Paulus, E. E; Jakobi, R. A. Chem. Ber. 1994, 127, 427--432. 102. Yamato, T.; Yasumatsu, M.; Saruwatari, Y.; Doamekpor, L. K.J. Incl. Phenom. Mol. Recogn. 1994, 19,315-331. 103. Anderson, S.; Neidlein, U.; Gramlich, V.; Diederich, E Angew. Chem., Int. Ed. Engl. 1995, 34, 1596-1600. 104. Neidlein, U.; Diederich, E Chem. Commun. 1996, 1493-1494. 105. Ito, K.; Ohba, Y.; Sone, T. Chem. Lett. 1996, 183-184. 106. lto, K.; Ohba, Y.; Sone, T. Chem. Lett. 1996, 783-784. 107. Ito, K.; Kida, A.; Ohba, Y.; Sone, T. Chem. Lett. 1998, 1221-1222. 108. Grynszpan, E; Goren, Z.; Biali, S. E. J. Org. Chem. 1991, 56, 532-536. 109. Strictly speaking, the chiral carbon is not incorporated but attached to the macrocycle in this case. 110. Flapan, E. J. Mol. Struct. (Theochem) 1995, 336, 157-164. 111. Only the single phenolic unit is considered in this discussion. Taking into account the neighborhood, the two units A, adjacent to B or C are different of course. 112. Not necessarily they must be entirely blocked in one conformation. 113. Bthmer, V.; Marschollek, F.; Zetta, L. J. Org. Chem. 1987, 52, 3200-3205. 114. (a) Bthmer, V.; Merkel, L.; Kunz, U. J. Chem. Soc., Chem. Commun. 1987, 896-897. (b) No, K.; Kim, J. E.; Kim, S. J. Bull. Korean Chem. Soc. 1996, 17, 869-872. 115. This is necessary, not to end up with an untractable mixture of isomeric or similar compounds, since there are four monoethers or four tetraethers in the partial cone conformation possible, to mention just two examples. 116. (a) Zetta, L.; Wolff, A.; Vogt, W.; Platt, K.-L.; Bthmer, V. Tetrahedron 1991, 47, 1911-1924; for a similar example see: (b) No, K.; Kim, J. E.; Kim, S. J. Bull, Korean Chem. Soc. 1996, 17, 869-872. 117. Bthmer, V.; Dtrrenb~icher, R.; Frings, M.; Heydenreich, M.; de Paoli, D.; Vogt, W.; Ferguson, G.; Thondorf, I. J. Org. Chem. 1996, 61,549-559. 118. Shu, C.-M.; Chung, W.-S.; Wu, S.-H.; Ho, Z.-C.; Lin, L.-G. J. Org. Chem. 1999, 64, 2673-2679. 119. For similar examples see also: (a) Kim, J. M.; Chun, J. C.; Nam, K. C. Bull, Korean Chem. Soc. 1997, 18, 409-415. (b) Nam, K. C.; Kim, J. M.; Kim, Y. J. Bull, Korean Chem. Soc. 1998, 19, 132-134. 120. Jin, T.; Monde, K. Chem. Commun. 1998, 1357-1357. 121. lwamoto, K.; Shimizu, H.; Araki, K.; Shinkai, S. J. Am. Chem. Soc. 1993, 115, 3997--4006. 122. (a) Gutsche, C. D.; Dhawan, B.; Levine, J. A.; No, K. N.; Bauer, L. J. Tetrahedron 1983, 39, 409-426. (b) lwamoto, K.; Fujimoto, K.; Matsuda, T.; Shinkai, S. Tetrahedron Lett. 1990, 31, 7169-7172. (c) lwamoto, K.; Yanagi, A.; Arimura, T.; Matsuda, T.; Shinkai, S. Chem. Lett. 1990, 1901-1904. (d) Iwamoto, K.; Araki, K.; Shinkai, S. J. Org. Chem. 1991, 56, 4955-4962. (e) Barrett, A.; McKervey, M. A.; Malone, J. E; Walker, A.; Arnaud-Neu, E; Guerra, L.; SchwingWeill, M.-J.; Gutsche, C. D.; Stewart, D. R. J. Chem. Soc., Perkin Trans. 2 1993, 1475-1479.

Chirality in Calixarenes and Calixarene Assemblies

227

123. (a) Pappalardo, S.; Caccamese, S.; Giunta, L. Tetrahedron Lett. 1991, 32, 7747-7750. (b) Pappalardo, S.; Caccamese, S. Chirality 1993, 5, 159-163. (c) Ferguson, G.; Gallagher, J. E; Giunta, L.; Neff, P.; Pappalardo, S.; Caccamese, S.; Parisi, M. J. Org. Chem. 1994, 59, 42-53. 124. Pappalardo, S.; Parisi, M. E Tetrahedron Lett. 1996, 37, 1493-1496. 125. Amaud-Neu, E; Caccamese, S.; Fuangswasdi, S.; Pappalardo, S.; Parisi, M. E; Petringa, A.; Principato, G. J. Org. Chem. 1997, 62, 8041-8048. 126. Amaud-Neu, E; Ferguson, G.; Fuangswasdi, S.; Notti, A.; Pappalardo, S.; Parisi, M. E; Petringa, A. J. Org. Chem. 1998, 63, 7770-7779. 127. Gloede, J.; Keitel, I.; Costisella, B.; Kunath, A.; Schneider, M. Phosphorus, Sulfur and Silicon 1996, 117, 67-88; see also: Byme, L. T.; Harrowfield, J. M.; Hockless, D. C. R.; Peachey, B. J.; Skelton, B. W.; White, A. H. Aust. J. Chem. 1993, 46, 1673-1683. 128. For a short review see: Gloede, J. Phosphorus, Sulfur and Silicon 1997, 127, 97-111. 129. Markovsky, L. N.; Visotsky, M. A.; Pirozhenko, V. V.; Kalchenko, V. I.; Lipkowski, J.; Simonov, Y. A. Chem. Commun. 1996, 69-71. 130. Vysotsky, M. O.; Tairov, M. O.; Pirozhenko, V. V.; Kalchenko, V. I. Tetrahedron Lett. 1998, 39, 6057-6060. 131. Kim, J. M.; Nam, K. C. Bull. Korean Chem. Soc. 1997, 18, 1327-1330. 132. See also: Park, Y. J.; Shin, J. M.; Nam, K. C.; Kim, J. M.; Kook, S.-K. Bull. Korean Cheat Soc. 1996, 17, 643-647. 133. Gonz~ilez, J. J.; Nieto, E M.; Prados, E; Echavarren, A. M.; de Mendoza, J. J. Org. Chem. 1995, 60, 7419-7423. 134. Browne, J. K.; McKervey, M. A.; Pitarch, M.; Russell, J. A.; Millership, J. S. Tetrahedron Lett. 1998, 39, 1787-1790. 135. Iki, H.; Kikuchi, T.; Shinkai, S. J. Che~ Soc., Perkin Trans. 1 1993, 205-210. 136. Sharma, S. K.; Gutsche, C. D. J. Org. Chem. 1999, 64, 3507-3512. 137. For a first non-chiral example see: Ikeda, A.; Shinkai, S. J. Chem. Soc., Chem. Commun. 1994, 2375-2378. 138. ll~tsabianca, H.; Royer, J.; Satrallah, A.; Taty-C, A.; Vicens, J. Tetrahedron Lett. 1987, 6595-6596. 140. Tabatabai, M.; Vogt, W.; BiShmer, V. Tetrahedron Lett. 1990, 31, 3295-3298. 141. (a) Shinkai, S.; Arimura, T.; Kawabata, H.; Murakami, H.; Araki, K.; lwamoto, K.; Matsuda, T. J. Chem. Soc., Chem. Commun. 1990, 1734-1736. (b) Shinkai, S.; Arimura, T.; Kawabata, H.; Murakami, H.; Iwamoto, K. J. Chem. Soc., Perkin Trans. 1 1991, 2429-2434. 142. Verboom, W.; Bodewes, E J.; van Essen, G.; Timmerman, E; van Hummel, G. J.; Harkema, S.; Reinhoudt, D. N. Tetrahedron 1995, 51,499-512. 143. Reddy, E A.; Gutsche, C. D. J. Org. Chem. 1993, 58, 3245-3251. 144. Ueda, Y.; Fujiwara, T.; Tomita, K.-I.; Asfari, Z.; Vicens, J. J. Incl. Phenom. Mol. Recogn. 1993, 15, 341-349. 145. Ikeda, A.; Yoshimura, M.; Lhotak, E; Shinkai, S. J. Chem. Soc., Perkin Trans. 1 1996, 1945-1950. 146. Btihmer, V.; Wolff, A.; Vogt, W. J. Chem. Soc., Chem. Commun. 1990, 968-970. 147. A 1,3-dimethylether of this type has been resolved by enantioselective HPLC: Schmidt, C.; Okamoto, J.; Btihmer, V., unpublished results; see below. 148. For one example see: Marschollek, E; B6hmer, V.; Vogt, W., unpublished results; see also ref. 8a. 149. No, K. H.; Kwon, K. M. Synthesis 1996, 1293-1295. 150. No, K.; Kwon, K. M.; Kim, B. H. BulL Korean Chem. Soc. 1998, 19, 1395-1398. 151. Amecke, R.; B6hmer, V.; Ferguson, G.; Pappalardo, S. Tetrahedron Lett. 1996, 37, 1497-1500. 152. Caccamese, S.; Notti, A.; Pappalardo, S.; Parisi, M. E; Principato, G. Tetrahedron 1999, 55, 5505-5514. 153. Any directionality in the crownether bridge would also make a calix[5]-crown asymmetric in contrast to 1,3-crowns of calix[4]arenes. 154. Casnati, A.; Minari, P.; Pochini, A.; Ungaro, R. J. Chem. Soc., Chem. Commun. 1991, 1413-1414. 155. Kanamathareddy, S.; Gutsche, C. D. J. Am Chem. Soc. 1993, 115, 6572-6579.

228

MYROSLAV VYSOTSKY,CHRISTIAN SCHMIDT, and VOLKER Bt~HMER

156. Otsuka, H.; Shinkai, S. J. Am. Chem. Soc. 1996, 118, 4271-4275. 157. Janssen, R. G.; Verboom, W.; van Duynhoven, J. P. M.; van Velzen, E. J. J.; Reinhoudt, D. N. Tetrahedron Lett. 1994, 35, 6555-6558. 158. For reviews on cyclotffveratrilenes and cryptophanes see: (a) Collet, A. Tetrahedron 1987, 43, 5725-5759. (b) Collet, A.; Dutasta, J.-P.; Lozach, B.; Canceill, J. Top. Curr. Chent 1993, 165, 103-129. 159. Li, J.-S.; Chen, Y.-Y.; Lu, X.-R. Eur. J. Org. Chem. 2000, 485-490. 160. Neff, P.; Rocco, C.; Consoli, G. M. L.; Piattelli, M. J. Org. Cheat 1993, 58, 6535-6537. 161. Rogers, J. S.; Gutsche, C. D. J. Org. Chem. 1992, 57, 3152-3159. 162. For examples see for instance: (a) Janssen, R. G.; Verboom, W.; Harkema, S.; van Hummel, G. J.; Reinhoudt, D. N.; Pochini, A.; Ungaro, R.; Prados, P.; de Mendoza, J. J. Chem. Soc., Chem. Commun. 1993, 506-508. (b) Neff, P.; Pappalardo, S. J. Org. Cheat 1993, 58, 1048-1053. 163. Geraci, C.; Piattelli, M.; Neff, E Tetrahedron Lett. 1995, 36, 5429-5432. 164. Geraci, C.; Bottino, A.; Piattelli, M.; Gavuzzo, E.; Neff, E J. Cheat Soc., Perka'n Trans. 2 2000, 185-187. 165. Geraci, C.; Piattelli, M.; Neff, E Tetrahedron Lett. 1996, 37, 7627-7630. 166. Caccamese, S.; Principato, G.; Geraci, C.; Neff, E Tetrahedron Asymmetry 1997, 8, 1169-1173. 167. Ikeda, A.; Akao, K.; Harada, T.; Shinkai, S. Tetrahedron Lett. 1996, 37, 1621-1624. 168. Ikeda, A.; Suzuki, Y.; Shinkai, S. Tetrahedron Asymm. 1998, 9, 97-105. 169. It should be kept in mind, however, that this splitting indicates enantiotopic groups, which not necessarily must be present in different molecules (enantiomers). They can be found also within a molecule having (at least) a symmetry plane as shown for the enantiotopic methylene bridges of 1,3-diethers of calix[4]arenes: See, K. A.; Fronczek, E R.; Watsot, W. H.; Gutsche, C. D. J. Org. Chem. 1991, 56, 7256-7268; Gutsche, C. D.; See, K. A. J. Org. Chem. 1992, 57, 4527-4539. 170. E.g. (a) Hofmeister, E.; Alvarado, E.; Leafy, J. A.; Yoon, D. I.; Pedersen, S. E J. Am. Cheat Soc. 1990,112, 8843-8851. (b) Furphy, B. M.; Harrowfield, J. M.; Kepert, D. L.; Skelton, B. W.; White, A. H.; Wilner, E R. Inorg. Chem. 1987, 26, 4231-4236. (c) Harrowfield, J. M.; Ogden, M. I.; White, A. H. Aust. J. Cheat 1991, 44, 1237-1247. (d) Harrowfield, J. M.; Ogden, M. I.; White, A. H. Aust. J. Chem. 1991, 44, 1249-1262. (e) Btinzli, J.-C. G.; Froidevaux, E; Harrowfield, J. M. Inorg. Chem. 1993, 32, 3306-3311. 171. Marcos, E M.; Ascenso, J. R.; Lamartine, R.; Pereira, J. L. C. J. Org. Cheat 1998, 63, 69-74. 172. Ftlix, S.; Ascenso, J. R.; Lamartine, R.; Pereira, J. L. C. Tetrahedron 1999, 55, 8539-8546. 173. Okada, Y.; Mizutani, M.; Ishii, E; Nishimura, J. Tetrahedron Lett. 1997, 38, 9013-9016. 174. For derivatives of 89 with chiral ether residues, derived from L-amino acids see: Okada, Y.; Kasai, Y.; Nishimura, J. Tetrahedron Lett. 1995, 36, 555-558. 175. Yamato, T.; Saruwatari, Y.; Yasumatsu, M. J. Chem. Soc., Perkin Trans. 1 1997, 1731-1737. 176. Araki, K.; Inada, K.; Otsuka, H.; Shinkai, S. Tetrahedron 1993, 49, 9465-9478. 177. Matsumoto, H.; Nishio, S.; Takeshita, M.; Shinkai, S. Tetrahedron 1995, 51, 4647-4654. 178. Araki, K.; Inada, K.; Shinkai, S. Angew. Chem., Int. Ed. Engl. 1996, 35, 72-74. 179. This shows, that the definition of inherent chirality may be somehow "artificial" or "arbitrary". 180. Cram, D. J.; Tunstad, L. M.; Knobler, C. B. J. Org. Chem. 1992, 57, 528-535. 181. Soncini, P.; Bonsignore, S.; Dalcanale, E. J. Org. Chem. 1992, 57, 4608-4612. 182. Vincenti, M.; Dalcanale, E.; Soncini, P.; Guglielmetti, G. J. Ant Cheat Soc. 1990, 112, 445-447. 183. Renslo, A. R.; Rudkevich, D. M.; Rebek, Jr., J. J. Am. Cheat Soc. 1999, 121, 7459-7460; for a correction of the original structural assignment see: Renslo, A. R.; Tucci, E C.; Rudkevich, D. M.; Rebek, Jr., J. J. Am. Cheat Soc. 2000, 122, 4573-4582. 184. For an early example of the synthesis of a structurally similar compound in a partial cone conformation (C l symmetry) see: Wu, T.-T.; Speas, J. R. J. Org. Cheat 1987, 52, 2330-2332.

Chirality in Calixarenes and Calixarene Assemblies

229

185. (a) Wolff, A.; Btihmer, V.; Vogt, W.; Ugozzoli, E; Andreetti, G. D. J. Org. Chent 1990, 55, 5665-5667. (b) Andreetti, G. D.; Btihmer, V.; Jordan, J. G.; Tabatabai, M.; Ugozzoli, E; Vogt, W.; Wolff, A. J. Org. Chem. 1993, 58, 4023-4032. (c) Mizyed, S.; Georghiou, E E.; Ashram, M. J. Chem. Soc., Perkin Trans. 2 2000, 277-280. 186. Fu, D.-K.; Xu, B.; Swager, T. M. J. Org. Chem. 1996, 61,802-804. 187. Biali, S.; B~hmer, V., unpublished results. 188. It should be mentioned that a tetraether in the 1,3-alternate conformation (S4) or 1,2-alternate conformation (Ci) or an anti-1,3-diether (Ci) are not chiral. 189. Pickard, S.; Pirkle, W. H.; Tabatabai, M.; Vogt, W.; B6hmer, V. Chirality 1993, 5, 310-314. 190. Xu, B.; Carroll, E J.; Swager, T. M. Angew. Chem., Int. Ed. Engl. 1996, 35, 2094-2097. 191. (a) Xu, B.; Swager, T. M.J. Ant Chem. Soc. 1993, 115, 1159-1160. (b) Swager, T. M.; Xu, B. J. Incl. Phenom. Mol. Recogn. 1994,19, 389-398. (c) Harvey, E D.; Gagnon, J.; Provencher, R.; Xu, B.; Swager, T. M. Can. J. Chem. 1996, 74, 2279-2288. 192. Tabatabai, M.; Vogt, W.; Btihmer, V.; Ferguson, G.; Paulus, E. E Supramol. Cheat 1994, 4, 147-152. 193. Schmidt, C.; B6hmer, V., unpublished results. 194. Georghiou, E E.; Li, Z. J. Incl. Phen. Mol. Recogn. 1994, 19, 55-66. 195. This applies also to calixindoles, see: (a) Black, D. S.; Craig, D. C.; Kumar, N. Aust. J. Chem. 1996, 49, 311-318. (b) Black, D. S. C.; Craig, D. C.; Kumar, N.; Rezaie, R. Tetrahedron 1999, 55, 4803-4814 and references cited therein. 196. Ikeda, A.; Shinkai, S. J. Chent Soc., Perkin Trans. 1 1993, 2671-2673. 197. Rauter, H.; Hillgeris, E. C.; Lippert, B. J. Chem. Soc., Chem. Commun. 1992, 1385-1386. 198. Rauter, H.; Hillgeris, E. C.; Erxleben, A.; Lippert, B../. Am. Chent Soc. 1994, 116, 616-624. 199. For covalently linked heterocalix[4]arenes with two uracil units incorporated via methylen bridges see: Kumar, S.; Hundal, G.; Paul, D.; Hundal, M. S.; Singh, H. J. Org. Chent 1999, 64, 7717-7726. So far all examples described have Cs-symmetry, but chiral examples (C l, C2) with one or two uracil units should be also possible. 200. Navarro, J. A. R.; Janik, M. B. L.; Freisinger, E.; Lippert, B. lnorg. Chem. 1999, 38, 426-432. 201. Kusano, T.; Tabatabai, M.; Okamoto, Y.; B6hmer, V. J. Am. Chem. Soc. 1999, 121, 3789-3790. 202. van Hoom, W. E; Morshuis, M. G. H.; van Veggel, E C. M.; Reinhoudt, D. N. J. Phys. Chent A 1998, 102, 1130-1138. 203. Thondorf, I. et al., unpublished. 204. Schmidt, C.; Okamoto, Y.; B/Shmer, V., unpublished results. 205. Such a C4-symmetrical derivative is not "inherently chiral" in the strict sense of the definition given above, since after opening the macrocycle the -CHR-groups remain asymmetric. 206. Markovsky, L. N.; Kalchenko, V. I.; Rudkevich, D. M.; Shivanyuk, A. N. Mendeleev Commun. 1992, 106-108. 207. Kalchenko, V. I.; Rudkevich, D. M.; Shivanyuk, A. N.; Pirozhenko, V. V.; Tsymbal, I. E; Markovsky, L. N. Zhurn. Obshch. Khim. 1994, 64, 731-742. 208. Lukin, O. V.; Pirozhenko, V. V.; Shivanyuk, A. N. Tetrahedron Lett. 1995, 42, 7725-7728. 209. Shivanyuk, A.; Paulus, E. E; B6hmer, V.; Vogt, W. J. Org. Chem. 1998, 63, 6448-6449. 210. Vollbrecht, A.; Neda, I.; Schmutzler, R. Phosphorus, Sulfur and Silicon 1995, 107, 173-179. 211. Choi, H.-J.; Btthring, D.; Quan, M. L. C.; Knobler, C. B.; Cram, D. J. J. Chent Soc., Cheat Commun. 1992, 1733-1735. 212. Amecke, R.; B6hmer, V.; Paulus, E. E; Vogt, W. J. Am. Chem. Soc. 1995, 117, 3286-3287. 213. Airola, K.; BiShmer, V.; Paulus, E. E; Rissanen, K.; Schmidt, C.; Thondorf, I.; Vogt, W. Tetrahedron 1997, 53, 10709-10724. 214. B6hmer, V.; Caccamese, S.; Principato, G.; Schmidt, C. Tetrahedron Lett. 1999, 40, 5927-5930. 215. Acylation by acetanhydride, for instance led under cleavage of the benzoxazine structures to a tetra-acetamide: Schmidt, C.; Paulus, E. E; B6hmer, V.; Vogt, W. New J. Chem. 2000,24, 123-126.

230

MYROSLAV VYSOTSKY,CHRISTIAN SCHMIDT, and VOLKER BOHMER

216. Shivanyuk, A.' Schmidt, C." Bthmer, V.; Paulus, E. E" Lukin, O." Vogt, W. J. Am. Cheat Soc. 1998, 120, 4319-4326. 217. Schmidt, C.; Airola, K.; Bthmer, V.; Vogt, W.; Rissanen, K. Tetrahedron 1997, 53, 17691-17698. 218. Schmidt, C.; Thondorf, I." Kolehmainen, E.; Bthmer, V.; Vogt, W.; Rissanen, K. Tetrahedron Lett. 1998, 39, 8833-8836. 219. Iwanek, W.; Mattay, J. Liebigs Ann. 1995, 1463-1466. 220. El Gihani, M. T.; Heaney, H.; Slawin, A. M. Z. Tetrahedron Lett. 1995, 36, 4905-4908. 221. Arnecke, R.; Bthmer, V." Friebe, S.; Gebauer, S.; Krauss, G. J.; Thondorf, I.; Vogt, W. Tetrahedron Lett. 1995, 36, 6221-6224. 222. Schmidt, C.; Vogt, W.; Bthmer, V., unpublished results. Careful inspection of the NMR spectra of tetrabenzoxazine derivatives obtained with 1-cyclohexylethylamine, 1-(1-naphthyl)ethylamin and 1-indylamine show on the other hand a double set of signals compatible with two diastereomers. This is in contrast to results reported by Arnecke, R. Ph.D. thesis, Mainz 1996, published in ref. 221. 223. Bulman Page, E C." Haeney, H.; Sampler, E. E J. Ant Chem. Soc. 1999, 121, 6751-6752. 224. McIldowie, M. J." Mocerino, M." Skelton, B. W.; White, A. H. Cheat Commun. 2000, submitted. 225. It shows a singlet for OH, ArH, t-Bu protons and a pair of doublets with geminal coupling for Ar-CH2-Ar protons. 226. MM-calculations lead to an energy minimum with CEv-symmetry ("pinched cone" conformation) for calix[4]arenes: (a) Harada, T.; Ohseto, E; Shinkai, S. Tetrahedron 1994, 50, 13377-13394. (b) Harada, T.; Shinkai, S. J. Chem. Soc., Perkin Trans. 2 1995, 2231-2242. 227. The global minimum calculated for calix[5]arenes on the other hand has Csv-symmetry: Thondorf, I." Brenn, J. J. Cheat Soc., Perkin Trans. 2 1997, 2293-2299. 228. The same is true for the spherand-type calixarene 47, for which six singlets for OH and t-Bu are found at -150 ~ Biali, S." Bthmer, V.; Thondorf, I., unpublished results. 229. Molins, M. A.; Nieto, P. M.; Sanchez, C.; Prados, P." de Mendoza, J." Pons, M. J. Org. Cheat 1992, 57, 6924-6931. 230. Halit, M." Oehler, D.; Perrin, M." Thozet, A." Perrin, R.; Vicens, J.; Bourakhoudar, M. J. Incl. Phenoat 1988, 6, 613-623. 231. van Hoorn, W. P." van Veggel, F. C. J. M.; Reinhoudt, D. N. J. Org. Chem. 1996, 61, 7180-7184. 232. Tsue, H.; Ohmori, M.; Hirao, K. J. Org. Chem. 1998, 63, 4866-4867. 233. A minimum energy conformation with C2-symmetry (or Ci) might be considered as an alternative explanation in this case too. 234. For a solid state NMR study at low temperature suggesting reversal of the OH-directionality by tunneling see: Brougham, D. F.; Caciuffo, R.; Horsewill, A. J. Nature 1999, 397, 241-243. 235. Pons, M.; Bthmer, V., unpublished results. 236. Leigh, D. A." Linnane, P.; Pritchard, R. G.; Jackson, G. J. Chem. Soc., Cheat Commun. 1994, 389-390. 237. Schmidt, C." Straub, T." Falhbu, D.; Paulus, E. F.; Kolehmainen, E." Bthmer, V." Rissanen, K.' Vogt, W. Eur. J. Org. Cheat, submitted. 238. Moran, J. R.; Ericson, J. U" Dalcanale, E.' Bryant, J. A.; Knobler, C. B." Cram, D. J. J. Am. Cheat Soc. 1991, 113, 5707-5714. 239. (a) Rudkevich, D. M.; Hilmerrson, G.; Rebek, Jr., J. J. Am. Cheat Soc. 1997, 119, 9911-9912. (b) Rudkevich, D. M.; Hilmerrson, G.; Rebek, Jr., J. J. Am. Cheat Soc. 1998, 120, 12216-12225. 240. Recently, Rebek et al. have used the expression "cycloenantiomers" in this connection (see also the hydrogen bonding in tetraurea dimers). Originally this name was coined for compounds in which stereogenic centers (e.g. asymmetric carbons) are attached to or incorporated in a macrocycle with an additional directionality in its structure: Prelog, V.; Gerlach, H. Helv. Chim. Acta 1964, 47, 2288. Due to this directionality the sequence of R and S configurated centers may be the reason for additional isomers (enantiomers as well as diastereomers). The expression was later slightly extended to conformational cycloenantiomers: Singh, M. D."

Chirality in Calixarenes and Calixarene Assemblies

241. 242. 243. 244. 245.

246. 247. 248. 249. 250. 251. 252. 253. 254. 255.

256. 257. 258.

259.

260. 261.

262. 263. 264. 265.

231

Siegel, J.; Biali, S. E.; Mislow, K.J. Am. Chem. Soc. 1987,109, 3397. In the present case, however, the directionality of the hydrogen bonds is the only stereogenic element and consequently all inherently chiral calixarenes would be cycloenantiomers. The expression "cycloenantiomers" was also used for instance for rotaxanes composed of wheel and an axle, which both are achiral by themselves but have a structural directionality: Yamamoto, C.; Okamoto, Y.; Schmidt, T.; J~iger, R.; Vogtle, V. J. Am. Chem. Soc. 1997, 119, 10547-10548. Haino, T.; Rudkevich, D. M.; Rebel Jr., J. J. Am. Chent Soc. 1999, 121, 11253-11254. Frkanec, L.; Vi~njevac, A." Koji6-Prodi6, B.; 7.ini6, M. Chem. Eur. J. 2000, 6, 442-453. Kraft, D.; B6hmer, V.; Vogt, W.; Ferguson, G.; Gallagher, J. E J. Chent Soc., Perkin Trans. I 1994, 1221-1230. B6hmer, V.; Ferguson, G.; Gallagher, J. E; Lough, A. J.; McKervey, M. A.; Madigan, E.; Moran, M. B.; Philips, J.; Williams, G. J. Chem. Soc., Perkin Trans. 1 1993, 1521-1227. (a) Parlevliet, E J.; Olivier, A.; de Lange, W. G. J.; Kamer, E C. J.; Kooijman, H.; Spek, A. L.; van Le~uwen, E W. N. M. Chem. Commun. 1996, 583-584. (b) Parlevliet, E J.; Zuideveld, M. A.; Kiener, C.; Kooijman, H.; Spek, A. L.; Kamer, E C. J.; van Leeuwen, E W. N. M. OrganometaUics 1999, 18, 3394-3405. Aleksiuk, O.; Grynszpan, E; Biali, S. E. J. Incl. Phenom. 1994, 19, 237-256. The expression "guest" deprives not a certain irony for a permanently included subject; "carcer" being the latin word for "prison." If two different cavitands are combined C4v-symmetry results. (a) Sherman, J. C.; Knobler, C. B.; Cram, D. J.J. Am. Chem. Soc. 1991,113, 2194-2204. (b) Cram, D. J.; Tanner, M. E.; Knobler, C. B. J. Am. Chem. Soc. 1991, 113, 7717-7727. Chapman, R. G.; Sherman, J. C. J. Am. Chem. Soc. 1999, 121, 1962-1963. For a recent example see: Paek, K.; Ihm, H.; Yun, S.; Lee, H. C. Tetrahedron Lett. 1999, 40, 8905-8909. See also: Sherman, J. C.; Chapman, R. G. J. Org. Chem. 2000, 65, 513-516. Blanchard, S.; L. Clainche, L.; Rager, M.-N.; Chansou, B.; Tuchagues, J.-E; Duprat, A. E; Le Mest, Y.; Reinaud, O. Angew. Chem. Int. Ed. EngL 1998, 37, 2732-2735. Blanchard, S.; Rager, M. N.; Duprat, A. E; Reinaud, O. New. J. Chem. 1998, 22, 1143-1146. However, a Cotton effect at 200 nm was recently observed for [4]triangulanes, molecules consisting of four cyclopropane tings only: de Meijere, A.; Khlebnikov, A. E; Kostikov, R. R., Kozhushkov, S. I.; Schreiner, E R.; Wittkopp, A.; Yufit, D. S. Angew. Chem., Int. Ed. Engl. 1999, 38, 3474-3477. Harada, N.; Nakanishi, K. Circular Dichroic Spectroscopy--Exciton Coupling in Organic Stereochemistry; University Sciences Books: Mill Valley, CA, 1983. Buckingham, A. D.; Stiles, P. J. Acc. Chem. Res. 1974, 7, 258-271. For ICD phenomena in cyclodextrin complexes, see: (a) Zhdanov, Yu. A.; Alekseev, Yu. E.; Kompantseva, E. V.; Vergeichik, E. N. Russ. Chent Rev. 1992, 61,563-575. (b) Connors, K. A. Chent Rev. 1997, 97, 1325-1357. In some cases, conformational chirality of the guest included into cyclodextrin cavity can be clearly deduced from the ICD spectra: Lighter, D. A.; Gawronski, J. K.; Gawronski, K. J. Ant Chem. Soc. 1985, 107, 2456-2461. Kikuchi, Y.; Kobayashi, K.; Aoyama, Y. J. Am. Che~ Soc. 1992, 114, 1351-1358. However, in the case of 1-alkylglucopyranoses 1:2 (G:H) and 4:1 complexes with 2 were characterized: Kikuchi, Y.; Tanaka, Y.; Sutarto, S.; Kabayashi, K.; Toi, H.; Aoyama, Y. J. Am. Chem. Soc. 1992, 114, 10302-10306. Kobayashi, K.; Asakawa, Y.; Kikuchi, Y.; Toi, H.; Aoyama, Y. J. Am. Chem. Soc. 1993, 115, 2648-2654. Harada, N.; Chen, S.-M. L.; Nakanishi, K. J. Am. Chem. Soc. 1975, 97, 5345-5352. Kikuchi, Y.; Aoyama, Y. Bull. Chem. Soc. Jpn. 1996, 69, 217-220. Kikuchi, Y.; Aoyama, Y. Supramol. Chem. 1996, 7, 147-152.

232

MYROSLAV VYSOTSKY,CHRISTIAN SCHMIDT, and VOLKER BOHMER

266. Arimura, T.; Edamitsu, S.; Shinkai, S.; Manabe, O.; Muramatsu, T.; Tashiro, M. Chem. Lett. 1987, 2269-2272. 267. Morozumi, T.; Shinkai, S. J. Chem. Soc., Chem. Commun. 1994, 1219-1220. 268. Morozumi, T.; Shinkai, S. Chem. Lett. 1994, 1515-1518. 269. Shimizu, K. D.; Rebel Jr., J. Proc. Natl. Acad. Sci. USA 1995, 92, 12403-12407. 270. Diffusion coefficients derived from pulsed gradient spin echo NMR-studies show lower values for the encapsulated guest identical to those of the host, see: Frish, L.; Matthews, S. E.; B6hmer, V.; Cohen, Y. J. Chem. Soc., Perkin Trans. 2 1999, 669-671. 271. Schalley, C. A.; Castellano, R. K.; Brody, M. S.; Rudkevich, D. M.; Siuzdak, G.; Rebek, Jr., J. J. Am. Chem. Soc. 1999, 121, 4568-4579. 272. Mogck, O.; Paulus, E. E; B6hmer, V.; Thondorf, I.; Vogt, W. Chem. Commun. 1996, 2533-2534. 273. Mogck, O.; B6hmer, V.; Vogt, W. Tetrahedron 1996, 52, 8489-8496. 274. Castellano, R. K.; Rebel Jr., J. J. Am. Chem. Soc. 1998, 120, 3657-3663. 275. Vysotsky, M. O.; Thondorf, I.; B6hmer, V. Angew. Chem., Int. Ed. Engl. 2000, 39, 1264-1267. 276. Castellano, R. K.; Nuckolls, C.; Rebek, Jr., J. J. Am. Chem. Soc. 1999, 121, 11156-11163. 277. Castellano, R. K.; Kim, B. H.; Rebel Jr., J. J. Am. Chem. Soc. 1997, 119, 12671-12672. 278. (a) Zerkowski, J. A.; Mathias, J. E; Whitesides, G. M.J. Am. Chem. Soc. 1994, 116, 4305-4315. (b) Zerkowski, J. A.; Whitesides, G. M. J. Am. Chem. Soc. 1994, 116, 4298-4304. 279. Whitesides, G. M.; Simanek, E. E.; Mathias, J. E; Seto, C. T.; Chin, D. N.; Mammen, M.; Gordon, D. M. Acc. Chem. Res. 1995, 28, 37-44. 280. (a) Vreekamp, R. H.; van Duynhoven, J. E M.; Hubert, M.; Verboom, W.; Reinhoudt, D. N.Angew. Chem., Int. Ed. Engl. 1996, 35, 1215-1218. (b) Timmerman, E; Vreekamp, R.; Hulst, R.; Verboom, W.; Reinhoudt, D. N.; Rissanen, K.; Udachin, K. A.; Ripmeester, J. Chem. Eur J. 1997, 3, 1823-1832. 281. Prins, L. J.; Timmerman, E; Reinhoudt, D. N. Pure & Appl. Chem. 1998, 70, 1459-1468. 282. Prins, L. J.; Huskens, J.; de Jong, E; Timmerman, E; Reinhoudt, D. N. Nature 1999, 398, 498-502. 283. Jolliffe, K. A.; Timmerman, P.; Reinhoudt, D. N. Angew. Chem., Int. Ed. Engl. 1999, 38, 933-937. 284. Klok, H.-A.; Jolliffe, K. A.; Schauer, C. L.; Prins, L. J.; Spatz, J. E; M611er, M.; Timmerman, E; Reinhoudt, D. N. J. Am. Chem. Soc. 1999, 121, 7154-7155. 285. Chapman, R. G.; Olovsson, G.; Trotter, L; Sherman, J. C.J. Am. Chem. Soc. 1998,120, 6252-6260. 286. McGillivray, L. R.; Atwood, J. L. Nature 1997, 389, 469-472. 287. Gerkensmeier, T.; lwanek, W.; Agena, C.; Fr6hlich, R.; Kotila, S.; N~ither, C.; Mattay, J. Eur. J. Org. Chem. 1999, 2257-2262. 288. Jaunky, W.; Hosseini, M. W.; Planeix, J. M.; De Cian, A.; Kyritsakas, N.; Fischer, J. Chem. Commun. 1999, 2313-2314. 289. Orr, G. W.; Barbour, L. J.; Atwood, J. L. Science 1999, 285, 1049-1052. 290. Usually the p-sulfonatocalix[4]arene forms bilayered structures. For a recent example with L-lysine spanning the bilayer see: Selkti, M.; Coleman, A. W.; Nicolis, I.; Douteau-Gu6vel, N.; Villain, E; Tomas, A.; de Rango, C. Chem. Commun. 2000, 161-162. 291. For recent examples see: (a) Ozerov, O. V.; Lapido, E T.; Patrick, B. O. J. Am. Chem. Soc. 1999, 121, 7941-7942. (b) Molenveld, E; Engbersen, J. E J.; Reinhoudt, D. N. Angew. Chem., Int. Ed. Engl. 1999, 38, 3189-3192. (c) Zheng, Y. S.; Shen, Z. Q. Eur. Polym. J. 1999, 35, 1037-1042. (d) Paciello, R.; Siggel, L.; R6per, M. Angew. Chem., Int. Ed. Engl. 1999, 38, 1920-1923. 292. A spectacular activation of the chiral zirconium-BINOL Lewis acid complex was achieved by the addition of the (achiral !) t-butyl-calix[4]arene. Less than 2% of the catalyst were sufficient in the enantioselective allylation of various aldehydes by allyltributyltin to reach enantiomeric excesses of more than 90%, see: Casolari, S.; Cozzi, P. G.; Orioli, P.; Tagliavini, E.; Umani-Ronchi, A. Chem. Commun. 1997, 2123-2124. 293. For a peptide library based on a calix[4]arene see: Hioki, H.; Yamada, T.; Fujioka, C.; Kodoma, M. Tetrahedron Lett. 1999, 40, 6821-6825.

Chirality in Calixarenes and Calixarene Assemblies

233

294. Combinatorial approaches are reported also here: Calama, M. C.; Timmerman, P.; Reinhoudt, D. N. Angew. Chem., Int. Ed. Engl. 2000, 39, 755-758. Cardullo, E; Calama, M. C.; Snellink-Ru'ee, B. H. M.; Weidmann, J.-L.; Bielejewska, A.; Fokkens, R.; Nibbering, N. M. M.; Timmerman, P.; Reinhoudt, D. N. Che~ Commun. 2000, 367-368.

This Page Intentionally Left Blank

FROM MOLECULES TO CRYSTALS SUPRAMOLECULAR SYNTHESIS OF SOLIDS

Brian Moulton and Michael J. Zaworotko

1.

2.

3.

4.

5.

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 236 1.1. The Birth of Crystal Engineering . . . . . . . . . . . . . . . . . . . . . 236 1.2. Crystal Engineering Today . . . . . . . . . . . . . . . . . . . . . . . . 237 1.3. The Conception of Crystal Engineering . . . . . . . . . . . . . . . . . 238 1.4. Coordination Polymers . . . . . . . . . . . . . . . . . . . . . . . . . . 240 The Modular Approach to Building Crystals: From Molecules to N e t w o r k s . . 242 2.1. Self-Assembly . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 242 2.2. Modular Self-Assembly . . . . . . . . . . . . . . . . . . . . . . . . . 242 Supramolecular Isomerism . . . . . . . . . . . . . . . . . . . . . . . . . . . 245 3.1. Structural Supramolecular Isomerism . . . . . . . . . . . . . . . . . . 246 3.2. Conformational Supramolecular Isomerism . . . . . . . . . . . . . . . 248 3.3. Catenane Supramolecular Isomerism . . . . . . . . . . . . . . . . . . . 250 3.4. Optical Supramolecular Isomerism . . . . . . . . . . . . . . . . . . . . 251 Supramolecular Synthesis of 2D Structures . . . . . . . . . . . . . . . . . . . 253 4.1. Hydrogen Bonded 2D Networks . . . . . . . . . . . . . . . . . . . . . 254 4.2. Networks Sustained by Organic Ions . . . . . . . . . . . . . . . . . . . 258 4.3. Metal-Organic Networks . . . . . . . . . . . . . . . . . . . . . . . . . 263 Supramolecular Synthesis of 3D Structures . . . . . . . . . . . . . . . . . . . 268

Advances in Supramolecular Chemistry Volume 7, pages 235-283. Copyright O 2000 by JAI Press Inc. All rights of reproduction in any form reserved. ISBN: 0-7623-0678.5 235

236

BRIAN MOULTON and MICHAEL I. ZAWOROTKO

5.1. Hydrogen-Bonded Networks . . . . . . . . . . . . . . . . . . . . . . . 5.2. Metal-Organic Networks . . . . . . . . . . . . . . . . . . . . . . . . . 5.3. Hybrid Structures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6. Potential Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7. Conclusion and Future Directions . . . . . . . . . . . . . . . . . . . . . . . . References and Notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

268 270 275 278 279 280

1. INTRODUCTION Organic chemistry nowadays almost drives me mad. To me it appears like a primeval, tropical forest full of the most remarkablethings, a dreadful endlessjungle into which one does not dare enter for there seems to be no way out. Friedrich W6hler (1800-1882) If one were to change just one word, organic to supramolecular, this commentary by W6hler on the state of organic chemistry at the end of the 19th century could have been used to describe the state of supramolecular chemistry as recently as the 1970s. W6hler was probably correct in believing that there was "no way out" of the jungle; after all, organic chemistry remains a thriving discipline. In recent decades chemists have forged the new discipline of supramolecular chemistry and today, at the beginning of the 21 st century, it seems entirely appropriate to see it as being a "tropical forest full of the most remarkable things" and a motivation for chemists and molecular scientists to bungle in the jungle. This chapter concerns an aspect of supramolecular chemistry that offers particular promise and opportunity: the development of strategies for the design of functional solids, crystal engineering.

1.1. The Birth of Crystal Engineering One might assert that W6hler's seminal paper on the preparation of urea from ammonium cyanate I marked an evolution in scientific thought from the existing theory of vitalism (that the products of living organisms originated from a special "vital force" which made them distinct from inorganic substances) to what we now know so well as synthetic chemistry. Similarly, Schmidt's seminal contribution on the subject of organic solid-state photochemistry, which defined and coined the term crystal engineering, 2 had much broader implications and marked a thought evolution in at least two important ways: 9 As implicit by use of the term crystal engineering, it became clear that, in appropriate circumstances, crystals could be thought of as the sum of a series of molecular recognition events--self-assemblymrather than the result of the need to "avoid a vacuum". However, it should be noted that there remained severe restrictions concerning the range of molecular components that could be used in this context. Indeed, Maddox's provocative comments almost 20 years later 3 still ring true, even today, for many simple organic compounds: "One of the continuing scandals in the physical sciences is that it remains in

Supramolecular Synthesis of Solids

237

general impossible to predict the structure of even the simplest crystalline solids from a knowledge of their chemical composition." 9 Schmidt's work emphasized the point that the physical and chemical properties of crystalline solids are as critically dependent upon the distribution of molecular components within the crystal lattice as the properties of its individual molecular components. In other words, crystal engineering is not only a scientific challenge, there are significant technological implications, in particular in the context of the development of solvent free chemistry ("green chemistry") and materials science. It is surely no coincidence that almost concurrent with the birth of crystal engineering, Lehn defined the new field of supramolecular chemistry, chemistry beyond the molecule. Lehn defined a "supramolecular assembly" as a chemical species of greater complexity than molecules themselves that is held together by intermolecular (i.e. noncovalent) interactions, 4 and "supramolecular chemistry" as the study of the chemical, physical, and biological properties of supramolecular assemblies. In this context, crystals might be regarded as being single chemical entities, and as such are perhaps the ultimate examples of supramolecules. Indeed, Dunitz has referred to organic crystals as "supermolecule(s) par excellence. ''5 As revealed in the next section, although the meaning is consistent with the approach to crystal engineering delineated herein, the terminology is perhaps inappropriate.

1.2. Crystal Engineering Today From the above, it should be clear that crystal engineering has evolved into much more than crystal structure analysis and prediction. Indeed, as we assert in this contribution, crystal engineering can be regarded as a subdivision of supramolecular chemistry that has developed a paradigm for design and synthesis of all solid-phase architectures, many of which are simple yet have no precedent in existing natural or synthetic compounds. In a sense, crystal engineering has become synonymous with the concept of supramolecular synthesis in the solid state. It should be noted that the adjective supramolecular has to be used judiciously as it imparts special relevance and meaning. Supramolecular synthons and supramolecular isomers represent two terms that are relevant to crystal engineering and they will be addressed herein. These terms are based upon analogies with their molecular level analogues, but their context and meanings are very different. The concept of supermolecular chemistry, chemistry of very large molecules, is also of great current interest and is relevant to both crystal engineering and supramolecular chemistry. When one considers the development of "supermolecules", the 1990s will probably be regarded as the key time period in which the basic concepts were delineated and experimental results were forthcoming. Indeed, we have witnessed synthesis and characterization of the largest hydrocar-

238

BRIAN MOULTON and MICHAEL I. ZAWOROTKO

bon 6 and coordination compounds 7 known to mankind. Interestingly, the structures of these compounds are at least an order of magnitude larger than their precursors, yet they are often simple to prepare because they are made by strategies that rely on self-assembly or, in the case ofdendritic architectures, divergent s or convergent 9 iterative synthetic methods. It seems apparent that we are witnessing a change in the conceptual approach that chemists have towards synthesis and it is not based upon new techniques, rather new goals and dreams. Scheme 1 illustrates how supramolecular chemistry has diversified from its beginnings in host-guest chemistry (i.e. biomimetic chemistry) and now encompasses crystal engineering and, most recently, suprasupermolecular chemistry. 1~

1.3. The Conception of Crystal Engineering If one were to view Schmidt's introduction of the term crystal engineering as its "birth", then perhaps its "conception" might have occurred via the delineation of noncovalent bonding by Linus Pauling in 1939.11 In his treatise on the nature of chemical bonding, Pauling focused upon bonding between molecules at least as much as he did on bonding within molecules. For example, he defined values for van der Waals radii that are still used today as well as the hydrogen bond (Chapter IX) and the metallic bond (Chapter XI). Such interactions are, of course, at the very heart of supramolecular chemistry and crystal engineering. Although Werner had earlier defined and developed the field of coordination chemistry, 12 Pauling's comprehensive evaluation of noncovalent bonding was developed in the context of the structure and properties of crystals and surely laid the foundation for supramolecular chemistry and crystal engineering as we know them today. It should be unsurprising that alternate approaches to understanding crystal structures have evolved. Kitaigorodskii asserted that the primary factors influencing

Bio Host/Guest

t

Chemistry 4mm~ SupramolecularmqJpSaprasupermolecular Chemistry Chemistry Crystal Engineering

Materials Science Scheme 1.

Supramolecular Synthesis of Solids

239

crystal structure are "packing forces," or van der Waal's forces. 13 He suggested that the crystal structures of organic substances represent manifestations of molecules adopting the closest possible packing of particles, regardless of the intermolecular forces. He applied the same principles to the analysis of all metallic, inorganic, and polymeric structures. This approach and crystal engineering are not necessarily mutually exclusive, especially for species that generate discreet or one-dimensional supramolecular aggregates. It is also understandable since for many years X-ray crystallography was applied primarily to the determination of the structure of molecules rather than crystal packing patterns. Most authors tended to focus on analysis of intramolecular distances and data on intermolecular parameters within crystals and cell-packing diagrams all but disappeared from crystallographic papers in the 1980s. Fortunately, the information and tools needed to more completely understand and analyze crystal structures were forthcoming in a timely manner. The importance of the role of that the Cambridge Crystallographic Data Centre, which was established in 1965, has played in the development of crystal engineering cannot be overstated. Its objective is to compile crystallographic data for all organic compounds studied by X-ray and neutron diffraction, provided by authors of the scientific literature, by means of a computerized database: the Cambridge Structural Database (CSD). 14 Initially the database contained 1500 structures and several hundred were added per year. Currently there are more than 200,000 structures with almost 20,000 additions per year. However, it was only in the past decade that sufficient entries and appropriate search engine tools became available to draw meaningful conclusions from analyses of the CSD. Even today, the CSD is by no means a representative sample of supramolecular forces in crystals because chemical crystallography was so focused upon molecular structure determination. Fortunately, as crystal engineering itself generates large numbers of new structures, each subsequent release of the CSD becomes increasingly more an appropriate tool to comprehensively define the forces behind crystal packing. The field of crystal engineering lay somewhat dormant in the 1970s but sprung to life in the 1980s thanks to a series of papers by Desiraju 15and Etter 16that concentrated upon using the CSD for analysis and interpretation of noncovalent bonding patterns in organic solids. Their seminal work in solid-state organic chemistry helped to characterize the hydrogen bond as being a prototypical supramolecular synthon and led to hydrogen bonds being perhaps the most widely exploited of the noncovalent interactions in the context of crystal engineering. Their research programs addressed the use of hydrogen bonding as a design element in crystal design, and delineated the nature (strength and directionality) of the interaction. Although Professor Desiraju continues his valuable contributions to the discipline, Professor Etter died prematurely in 1992. However, crystal engineering did not become synonymous with supramolecular synthesis until the 1990s. As noted by J. S. Maddox in 19883 and more recently by Gavezzotti 17 and Ball, is crystal structure prediction remains in its infancy. However, prediction is fundamentally very different from engineering and design. Predicting a crystal structure requires an analysis of the recognition features present in the molecular component in such

240

BRIAN MOULTON and MICHAEL J. ZAWOROTKO

a way that the manner in which the components will assemble can be converted to crystallographic symmetry operations, i.e. it requites space group determination. The existence of polymorphs, solvates, clathrates, and cocrystals, and our limited understanding of crystal nucleation and growth suggest that crystal structure prediction, even today, is an ambitious goal. Engineering and design are far less restrictive from a conceptual perspective since they can focus broadly upon network architectures. Design of network architectures is in many ways a less challenging proposition in crystal structure prediction and has the added advantage of affording new types of structures that had not hitherto been encountered. 19 The principles of design require a blueprint and allow the designer to select components in a judicious manner, thereby promoting the desired network structure, i.e. one can limit research to chemical moieties that are predisposed to a successful outcome. One might use the following analogy: it would be imtx~ssible for an architect to look at a brick and predict what structure would spontaneously result from putting many bricks together. However, an architect is able to draw a blueprint of a structure that can be built from those very same bricks.

1.4. Coordination Polymers Coordination polymers exemplify how crystal engineering has become a paradigm for the design of new supramolecular structures. In this context, the work of Wells is seminal. Wells was primarily concerned with the overall structure of solids, particularly inorganic compounds. 2~He described structures in terms of their topology by reducing crystal structures to a series of points (nodes) of a certain geometry (tetrahedral, trigonal planar, etc.) that are connected to a fixed number of other points. The resulting structures, which can also be calculated mathematically, are either discreet (zerodimensional) polyhedra, or infinite (one-, two-, and three-dimensional) periodic nets. In Structural Inorganic Chemistry, 2~ Wells effectively described one of the primary goals of modern crystal engineering well before experiment confirmed his assertions: Ideally we should be able to set out the possible types of structure for a molecule or crystal composed of certain numbers of atoms with known bonding requirements so that the observed structures can be compared with all (geometrically and topologically) possible structures, up to some arbitrary limit of complexity. It is important to know why some apparently reasonable structures are never adopted by known molecules or crystals. Systematic studies of possible structure types have not been numerous...

Given this intuition, it is perhaps surprising that it took until the 1990s and the work of Robson 21 for the field of coordination polymers to develop rapidly alongside crystal engineering of organic solids. Robson extrapolated Wells work on inorganic network structures into the realm of metal-organic compounds and coordination polymers. The resulting "node and spacer" approach has been remarkably successful at producing predictable network architectures in this context. Scheme 2 illustrates some of the simple architectures that can be generated using commonly available metal moieties. Whereas diamondoid networks 22 represent a

5upramolecular Synthesis of Solids

241

class of structure that could be described as mineralomimetic because there are many naturally occurring analogues, that is not the case for any of the other architectures illustrated in Scheme 2. The nature of these novel structures, their organic analogues, and their supramolecular isomers 23 represents the primary focus for the remainder of this chapter. In particular, their relevance in terms of inclusion chemistry will be featured because a recurring artifact that is observed in these network structures is the presence of voids or cavities that are inherently present because of the architecture itself. Furthermore, the size, shape, and molecular recognition features of these cavities are essentially predetermined by the selection of node and spacer. Simply put, we are dealing with a new generation of functional porous solids that complement zeolitic and clay-like inorganic compounds. This aspect of crystal engineering has been one of the main driving forces for its continued growth. In this context, it should be noted that we have only very recently seen the first examples of purely organic 24 and metal-organic 25 networks that are completely stable to loss of guest and exhibit porosity (as shown by reversible gas adsorption isotherms). Indeed, in an even more recent report, a metal-organic compound that dramatically supercedes any existing compound in terms of relative porosity was presented. 26

Scheme 2.

242

BRIAN MOULTON and MICHAEL J. ZAWOROTKO

2. THE MODULAR APPROACH TO BUILDING CRYSTALS: FROM MOLECULES TO NETWORKS There are two basic and complementary approaches to building crystalline networks that are constituted from organic or metal-organic components" Self-assembly of a single molecular component or "molecular tectonics. ''27 In such a situation, all the molecular recognition features that lead to supramolecular synthons 15d are present in a single molecule. 9 Self-assembly of two or more complementary components, a node and spacer, or "modular self-assembly". The molecular recognition features are present in two or more complementary molecules (e.g. cocrystals) or ions (e.g. salts or metal-organic polymers).

9

These approaches and their differences are well illustrated if one considers the subject of diamondoid networks.

2.1. Self-Assembly 1,3,5,7-Adamantanetetraacetic acid 28 and methanetetraacetic acid 29 can be regarded as being prototypal for self-assembled diamondoid architectures. Both structures are sustained by one of the most well-recognized supramolecular synthons--the carboxylic acid dimer. 3~ Pyridone dimers have been used in a similar fashion to build diamondoid networks, in this case from tetrahedral tetrakispyridones. 27a It should be noted that a number of well-known inorganic structures also represent examples of self-assembly (e.g. ice, potassium dihydrogenphosphate) and one might even consider covalent bonds as conceptually related: diamond, Si, Ge, ZnS, BE GaAs, ZnSe, CdS, CulnSe 2, CuFeS 2 (chalcopyrite). 2.2. Modular Self-Assembly Coordination polymers and hydrogen-bonded structures with multiple complementary components can be regarded as being the consequence of modular selfassembly. 22 Whereas self-assembled architectures occur in minerals, there is to our knowledge only one example of a naturally occurring modular structure: cristobalite, which forms an interpenetrated diamondoid network sustained by linear S i - O - S i linkages. However, the number of modular systems that have been synthesized and characterized now exceeds those formed by self-assembly. This is perhaps unsurprising when one considers that a single tetrahedral node can easily generate dozens of diamondoid compounds and that tetrahedral metal moieties such as Ag I and Cu I can serve as the node. Scheme 3 illustrates the two types of structure and should help to rationalize the significant differences in the types of tetrahedral moiety that can sustain the two types of network. The most fundamental difference between the two types of

Supramolecular Synthesis of Solids

243

Scheme 3. Diamondoid networks.

structure is that the tetrahedral component that sustains self-assembled architectures would not ordinarily be able to sustain modular architectures, and vice versa. In the case of the former, the tetrahedral moieties must be self-complementary and there is only one component necessary for self-assembly to occur. This means, for

Figure 1. The tetrahedral structure of [Mn(CO)g(OH)]4.

244

BRIAN MOULTON and MICHAEL J. ZAWOROTKO

example, that both hydrogen bond donors and hydrogen bond acceptors must be present in the same molecule. In the case of the latter, the tetrahedral node must be either an acceptor or a donor of hydrogen bonds and the linker or spacer must be complementary. Both components are necessary in a 1:2 ratio in order for the diamondoid architecture to self-assemble. A good example of such a node would be a cubane cluster such as [Mn(~-OH)(CO)3]4, 31 which possesses perfect Td symmetry (Figure 1). The tetrahedral hydrogen bond donor forms diamondoid cocrystals with a wide range of obvious and, in some cases, not so obvious spacer molecules. A "not so obvious" structure is that formed when [Mn(I.t3-OH)(CO)3] 4 is cocrystallized with benzene. A two-fold diamondoid structure is sustained by OH...~ hydrogen bonds and the tetrahedral symmetry of the node is observed in the crystallographic sense since [Mn(~-OH)(CO)3]4.2 benzene crystallizes in the cubic space group Pn-3m with z = 2. 32 As revealed by Figure 2, which illustrates an adamantoid portion of the structure, a large cavity is generated and this facilitates the interpenetration of a second diamondoid network. The use of transition metals or transition metal clusters to act as nodes for the modular self-assembly of diamondoid networks that are sustained by coordinate covalent bonds is also well established. Such architectures are of more than aesthetic appeal. Indeed, such structures have resulted in a class of compound with very interesting bulk and functional properties. Metal-organic diamondoid structures in which the spacer moiety has no center of inversion are predisposed to generate polar networks since there would not be any inherent center of inversion. Pyridine-4-carboxylic acid is such a ligand and bis(isonicotinato)zinc exists as a three-fold diamondoid structure that is thermally stable and inherently polar. 33

Figure 2.

The structure of [Mn(CO)3(OH)]4.2 benzene.

Supramolecular Synthesis of Solids

245

Indeed, it exhibits SHG (second harmonic generation) activity that is three times higher than that of commercially relevant NLO (nonlinear optical active) materials such as potassium dihydrogenphosphate. Diamondoid networks represent just one example of the types of architecture that can be generated in a rational manner using the modular self-assembly approach. The remainder of this chapter will focus upon the diversity of simple architectures that are accessible using crystal engineering strategies, many of which have no precedent in minerals yet are afforded via one-pot reactions and commercially available components.

3. SUPRAMOLECULAR ISOMERISM An aspect of crystal engineering that is becoming increasingly relevant and is closely related to polymorphism in crystalline solids is the existence of supramolecular 23 or architectural 34 isomerism in polymeric network structures. Simply put, supramolecular isomerism is the existence of more than one network structure for the same building blocks. In some cases this results directly in polymorphism, but where guest or solvent molecules are involved, and especially in open framework structures, polymorphism is not an appropriate term. Indeed, in a sense polymorphism can be regarded as being a type of supramolecular isomerism. The subject of supramolecular isomerism is important for a number of reasons: 9 The relationship between supramolecular isomerism and polymorphism must be further investigated when one considers that bulk properties of solids are critically dependent upon architecture and that crystal structure confmns composition of matter from a legal perspective. Polymorphism in molecular crystals is a commonly encountered phenomenon that is particularly relevant in the context of pharmaceuticals and is receiving increasing attention from a scientific perspective. 35 It should also be noted that McCrone was prompted to suggest that the "number of forms known for a given compound is proportional to the time and money spent in research on that compound. ''36 However, the generality of McCrone's statement remains ambiguous. 37 For example, Desiraju 15b'd asserted that the frequency of occurrence of polymorphic modifications is not uniform in all categories of substance. He suggested that the phenomenon is probably more common with molecules that have conformational flexibility and/or multiple groups capable of hydrogen bonding or coordination. Coincidentally and importantly, this is inherently the situation for many pharmaceuticals. Conformational polymorphism is a subject in its own right. 38 Desiraju also suggested that polymorphism can be strongly solvent-dependent. 9 Similar analogies can be drawn to other forms of isomerism that occur at the molecular level. This matter is addressed later in this section.

246

BRIAN MOULTON and MICHAEL J. ZAWOROTKO

9 Control over supramolecular isomers and polymorphs lies at the very heart of the concept of crystal engineering (i.e. design of solids). However, there is presently very little understanding concerning even the existence of supramolecular isomers, never mind how to control them. 9 Supramolecular isomerism also lies at the heart of gaining a better understanding of supramolecular synthons and, by inference, how they develop and occur in other solid phases and even solution. The Cambridge Structural Database remains a very powerful tool in this context but it must be remembered that even such a large database will not necessarily be reflective of the full range of compounds that will be isolated and characterized in future years. The remainder of this section summarizes the current state-of-the-art with particular reference to the conceptual link between polymorphism and supramolecular isomerism in organic and metal-organic networks. Specifically, since polymorphs can be rationalized on the basis of supramolecular interactions, we suggest that polymorphism can be regarded as a type of supramolecular isomerism. Implicitly, all sets of polymorphs can therefore be regarded as being supramolecular isomers of one another but the reverse is not necessarily the case. Supramolecular isomerism as seen in metal-organic and organic networks may be classified based upon analogies drawn with isomerism at the molecular level. Thus far it is appropriate to categorize the following classes of supramolecular isomerism: The node (metal moiety or exofunctional organic molecule) remains the same but there is a different network architecture. 23 9 Conformational: Different conformation in flexible ligands such as bis(4pyridyl)ethane generates a different network architecture. 23 Conformational polymorphism is a closely related subject. 38 9 Catenane: The different ways and levels in which networks interpenetrate or interweave can afford significant variations depending upon the basic architecture involved. 39 9 Optical: Networks crystallize in chiral (enantiomorphic) space groups and can therefore be related to homochiral compounds. 9 Structural:

3.1. Structural Supramolecular Isomerism Structural supramolecular isomerism is exemplified by the range of structures that has thus far been observed in coordination polymers for some of the simplest building blocks and stoichiometries. We shall illustrate this by looking at two of these situations.

1:1 $toichiometry, cis.Metal and Linear Spacer Ligand Scheme 4 illustrates the possible structures that can result from self-assembly of a cis-octahedral or cis-square-planar metal and a linear "spacer" ligand. There are

Supramolecular Synthesisof Solids

247 9J

iS[

.('o/~"

square box

zig-zag chain Ligand

helix

9 metal

Scheme4. Possiblesupramolecular isomers for compounds formed from a cis-octahedral or square-planar metal and a "spacer" ligand.

three obvious architectures that might result and they are quite dramatically different from one another. The "square box" architecture represents a discreet species that has been developed extensively in recent years by the groups of Fujita, 4~ Stang, 41 and Hupp. 42 The other architectures are 1D polymers but are quite different from one another. The zig-zag polymer has been widely encountered 43 but the helix remains quite rare. 44 The helix architecture is also of interest because it is inherently chiral regardless of what its components might be. The inherent chirality of this architecture comes from spatial disposition rather than the presence of chiral atoms. This particular issue is addressed later in this section because of its relevance to optical isomerism.

1"1.5 Stoichiometry, mer.Metal and Linear Spacer Ligand These building blocks can be regarded as being based upon self-assembly of T-shape nodes. There already exists a surprisingly diverse range of structures that have been observed in this context. Scheme 5 below illustrates the supramolecular isomers that have thus far been observed: ladder (A); 45 brick wall (B); 46 3D frame or "Lincoln Logs" (C); 47 bilayer (D); 48 herring bone (E); 49 and another version of a 3D frame (F). 5~Interestingly, three of the isomers A, 45a, D, 48a and F 5~ have been observed for the same asymmetric unit for metal = Co(NO3) 2 and ligand = bipy and the other three have been seen in similar compounds which use bipy or extended analogues as "spacer ligands." The following points should be noted about such structures: 9 These compounds are not true polymorphs since guest or solvent molecules are present in the lattice. However, neither are they solvates. They are probably

248

BRIAN MOULTON and MICHAEL J. ZAWOROTKO

Scheme 5. Structural supramolecular isomers for self-assembly of T-shaped nodes.

best regarded as host-guest compounds for which the host framework exhibits profoundly different structural supramolecular isomers. 9 The diversity of network structures and hence bulk properties is remarkable. 9 None of these architectures occurs naturally in minerals. 9 It seems likely that the use of appropriate templates might facilitate recipes for all supramolecular isomers that are possible for a given node (cf. McCrone's assertion). A similar approach based upon networks can be used to analyze true polymorphs. A recent paper highlighted this situation in the context of 2-amino-5-nitropyrimidine, a compound that exhibits three readily available polymorphs, all of which have distinct hydrogen-bonded networks. 51

3.2. Conformational Supramolecular Isomerism Flexibility in ligands can lead to subtle or dramatic changes in architecture. For example, 1,2-bis(pyridyl)ethane, dipy-Et, can readily adapt gauche or anti conformations. In the case of [Co(dipy-Et)l.5(NO3)2] n, which contains a T-shape node, infinite molecular ladders which contain six molecules of chloroform per cavity exist as the most commonly encountered architecture (Figure 3A). 45b In such a situation all "spacer ligands" are necessarily anti. However, under certain crystal-

Supramolecular Synthesis of Solids

249

Figure 3. Two supramolecular isomers of [Co(dipy-Et)1.s(NO3)2]n. lization conditions (e.g. solvent MeCN or dioxane), a bilayer architecture is obtained with two anti- and one gauche-spacer ligand per metal atom (Figure 3B). Interestingly, the bilayer architecture can contain solvent molecules such as MeCN or can collapse on itself in the absence of solvent. 23 This more subtle form of supramolecular isomerism occurs if [Co(dipy-Et)l.s(NO3)2] n is crystallized in the absence of a suitable guest or solvent. 52 Figure 4 reveals how [Co(dipy-

Figure 4. The "open framework" and collapsed forms of [Co(dipy-Et)l.s(NO3)2]n.

250

BRIAN MOULTON and MICHAEL J. ZAWOROTKO

Et)l.5(NO3)2] n collapses to close the cavity that exists when crystallized from MeCN. 23 Note the difference in torsion angles between the two compounds.

3.3. Catenane Supramolecular Isomerism The existence of independent interpenetrating networks is surprisingly common if relatively large cavities are generated within a network. A thorough review of the subject in the context of coordination polymers was recently published by Batten and Robson. 39 Interestingly, the existence of interpenetration was until recently regarded as a factor that mitigates strongly against the generation of stable openframework structures. However, it is now becoming clear that appropriate use of templates can afford either open framework or interpenetrated structures for the same network. Furthermore, some interpenetrated structures can also be regarded as open framework since they might still contain channels large enough to hold, for example, aromatic guests. Such is the case for square-grid networks based upon ligands such as dipy-Et. Figure 5 reveals how either open-framework, square-grid or interpenetrated square-grid structures can be readily generated for the same network. 53 Both compounds contain square grids of formula [Ni(dipy-Et)2(NO3)] n. As would be expected, the compound illustrated in Figure 4A, [Ni(dipyEt)2(NO3)] n, exhibits clay-like properties and can desorb and adsorb guests without irreversible collapse of the square-grid architecture. The compound illustrated in Figure 4B is effectively a 3D architecture built by interpenetration of square grids. This compound has a more rigid structure and behaves more like a zeolite than a clay. Interestingly, both compounds are stable to loss of guest but the former loses crystallinity upon loss of guest.

Figure 5. The open-framework non-interpenetrated (A) and interpenetrated (B) architectures exhibited by [Ni(dipy-Et)2(NO3)]n.

Supramolecular Synthesis of Solids

251

Figure 6. The triply interpenetrated structure formed bythe 1"1.5 cocrystal oftrimesic acid and 4,4'-bipyridine. Organic networks are also capable of exhibiting interpenetration. Figure 6 illustrates the threefold interpenetration that occurs in the cocrystal formed between trimesic acid (benzene-1,3,5-tricarboxylic acid) and bipy. This structure exhibits interpenetration in 2D to form a novel carpet-like architecture. 54The non-interpenetrated form of this compound has not yet been isolated.

3.4. Optical Supramolecular Isomerism Crystals that are built from homochiral components are not necessarily preordained to crystallize in chiral space groups, although polarity (i.e. absence of a center of inversion) is guaranteed in such situations. Fortunately, there exist at least four other strategies for the design of polar crystals that are independent of the need for homochiral molecular components: 1. Use of achiral building blocks that crystallize in a chiral space group. 2. Use of achiral molecular building blocks to build a chiral framework. 3. Use of achiral host framework built from achiral molecular components with chiral guest(s). 4. Use of achiral host framework built from achiral molecular components with achiral guest(s). Whereas the direct exploitation of homochiral tectons would seem to be the most obvious approach because the absence of a center of inversion is guaranteed, it in no way implies or affords any control of molecular orientation and therefore optimization of bulk polarity. Furthermore, reliance upon the use of pure enantiomers raises the substantial problem of needing to control the stereochemistry at the molecular level without yet solving the problem of controlling the supramolecular

252

BRIAN MOULTON and MICHAEL J. ZAWOROTKO

stereochemistry. Indeed, strategy 1, which basically relies upon serendipity, offers just as much chance of optimal control of crystal packing as the use of homochiral tectons. However, a review of recent literature reveals three types of polar architecture that do not need to be sustained by homochiral molecular components: helical networks; 55 1D acentric networks sustained by head-to-tail stacking of complementary molecules; 56 and host-guest networks which are polar because of the presence of acentric guest molecules or guest aggregates. 57 Although the crystallization process for strategies 1-4 inherently affords homochiral single crystals, only the use of homochiral tectons guarantees that all crystals will be of the same enantiomorph. Indeed, batches of crystals will often be heterochiral as both enantiomers tend to be formed equally during crystallization. Fortunately, it has been demonstrated 58 that formation of homochiral bulk materials can be afforded by seeding with the desired enantiomer. To illustrate the potential for generation of chiral architectures from simple achiral building blocks, let us consider how one might build a helix from simple components. Earlier in this section we illustrated how linking of octahedral or square-planar metal moieties might reasonably afford a helical architecture in a spontaneous fashion. Indeed, for [Ni(bipy)(benzoate)2(MeOH) 2] such an architecture persists in the presence of several guests, even if 2-hydroxybenzoate ligands (i.e. ligands that are capable of forming strong hydrogen bonds) are employed. The crystal structure of the nitrobenzene clathrate is presented in Figure 7 and reveals how large chiral cavities that induce the guest molecules to form chiral dimers are created. The guest molecules are trapped in a closed environment since helices from adjacent planes close off the 500/~3 cavities.

Figure 7. Crystal structure of [Ni(4,4'- bipyridine)(benzoate)2(MeOH)2] nitrobenzene. 9

Supramolecular Synthesis of Solids

253

As suggested above, the attractiveness of using crystal engineering for generation of polar architectures is several-fold: 9 There is no prerequisite for homochiral molecular components. Host-guest compounds have the potential to be modular and fine-tunable. 9 In principle, all existing achiral moieties can be incorporated into polar structures.

9

We shall now focus upon 2D and 3D structures that will be organized by dimensionality of architecture and chemical components and draw mineralomimetic analogies. In terms of chemical composition, we shall focus upon purely organic networks, which are typically sustained by hydrogen bonds and stacking interactions, and metal-organic structures that are based upon coordinate covalent bonds.

4. SUPRAMOLECULAR SYNTHESIS OF 2D STRUCTURES In order to simplify the complex and as yet unrealized goal of developing reliable and broadly applicable methods of crystal structure prediction, it is only natural to analyze existing crystal structures by breaking them down, at least as far as reasonably possible, into discreet 1D, 2D, or 3D networks. At the very least, rationalization of crystal structures would then become greatly simplified since even the presence of a reliable 1D network for a given set of molecular recognition features or supramolecular synthons significantly restricts the number of possible packing modes. An exciting by-product of this thought process is that design of new crystal structures can also be achieved by thinking in terms of network design. In this context, supramolecular synthesis of new generations of 2D and 3D structures offers enticing targets for both scientific and technological reasons: 9 From a scientific perspective, the greatest degree of predictability will occur if 2D and 3D architectures can be generated from first principles. Indeed, in the case of the latter, the only degree of unpredictability would appear to relate to whether subtle conformational effects occur or, in the case of open framework structures, as to whether or not interpenetration occurs. 9 A primary goal of crystal engineering since its inception has been to generate linkages between synthetic chemistry and materials science. This link is already becoming apparent in the context of 2D and 3D networks since, from a technological perspective, interesting things are expected in terms of bulk properties and functionality. In particular, at the very least one might expect to be able to draw analogies with naturally occurring minerals such as clays and zeolites, both of which enjoy widespread application because of their ability to intercalate and adsorb guests, respectively. Crystal engineering offers the intriguing concept of inherent control over the dimensions and

254

BRIAN MOULTON and MICHAEL J. ZAWOROTKO molecular recognition features that are present in laminated and porous structures. This is not really the case in naturally occurring structures.

The remainder of this chapter will focus upon an overview of developments in the context of the crystal engineering of 2D and 3D structures.

4.1. Hydrogen Bonded 2D Networks Hydrogen-bonded 2D networks are best exemplified by organic molecular networks which, as the name implies, are constituted by organic moieties with multiple complementary terminal functional groups that assemble into 2D arrays. Scheme 6 illustrates some of the more common supramolecular synthons that have been used by crystal engineers to design and build organic molecular networks. The carboxylic acid moiety represents perhaps the most widely observed and exploited of these supramolecular synthons. This should be unsurprising since the carboxylic acid moiety is ubiquitous in organic chemistry and the carboxylic acid dimer represents one of the most widely recognized examples of hydrogen bonding. The carboxylic acid moiety also represents an ideal illustration of the problems associated with crystal structure prediction and design if one concentrates upon a molecular unit with a limited number of dimensions and molecular recognition sites. Monocarboxylic acids have been widely studied crystallographically and a study by Frankenbach and Etter3~ revealed that there are two common motifs observed in the solid state for carboxylic acids. These supramolecular isomers, which profoundly influence the crystal packing and bulk properties, are illustrated in Scheme 7. As might be expected, the carboxylic dimer, A, is the most commonly observed supramolecular isomer and, being inherently centrosymmetric, it tends to afford nonpolar crystals. However, since it is effectively a discreet 0D aggregate there is little information that might be used in order to facilitate prediction of the

--N

O--"-H

/

\

O----H

\ N-

/

--N

/O~ ~0 t "

|1

~u_k,--il-|

----N----Br--

H

--

_

H"O

--No__. /N--

~N--H-----F---

- H H ~, , H

/~N~, H_~ H --N N-~N/-"H-~H H

O--"-H--O

CO--H----O

~"~.N-H-" --O..,.,,~

_H..O

Scheme 6. Examples of supramolecular synthons.

Supramolecular Synthesis of Solids A. dimer

255

/~Oe ~ e H - O \ R'C \OH e eO~c'R

fC,o

B. polymer

R

R

Scheme 7. Carboxylic acid motifs/supramolecular isomers.

overall crystal structure. The other motif, B, affords a 1D chain that is polar because it is the result of "head-to-tail" self-assembly. This simplifies to some extent the problem of crystal structure prediction but it still does not address how these 1D chains pack. Indeed, there is almost random choice of parallel or antiparallel packing of these chains in observed crystal structures. 3~ This is a subtle but important distinction since, in the case of parallel packing of chains, there is no crystallographic center of inversion within the 2D layer and crystallization in polar space groups occurs in approximately 50% of structures that have thus far been characterized, i.e. centers of inversion between layers appear to be randomly generated. We can summarize the situation in the context of monocarboxylic acids as follows: 9 A lot of information can be generated by analyzing the crystal packing in monocarboxylic acids and we confirm that the presence of centrosymmetric supramolecular synthon (i.e. the carboxylic acid dimer) affords centrosymmetric crystals. 9 Unfortunately, we still have only really learned about what not to do (i.e. one should avoid a centrosymmetric supramolecular isomer if one wants to design polar crystals) rather than what to do to ensure optimized bulk polarity. 9 Multiple molecular recognition sites will be necessary in order to gain a greater degree of predictability over organic crystal structures. Based on the above, the well-known structure of trimesic acid (1,3,5-benzene tricarboxylic acid, H3TMA ), a polyfunctional carboxylic acid that is inexpensive and chemically robust, has long intrigued crystal engineers and represents a more ideal prototype for crystal structure prediction and design. Its trigonal exodentate functionality imparts self-assembly into two dimensions. Figure 8 illustrates how the hydrogen-bonding pattern in the 2D networks formed by H3TMA generates cavities of predictable size (approximately 14/~ diameter). In pure H3TMA the grid

256

BRIAN MOULTON and MICHAEL J. ZAWOROTKO

Figure 8. 2D H3TMA Network. (a) Illustration of hydrogen bonded network, hydrogens omitted for clarity; (b) space filling model. is puckered and the cavities are filled by self-inclusion, or interpenetration, of other networks. However, subsequent reports revealed that there are methods for preparing the non-interpenetrated or open framework form of H3TMA. If crystallized in the presence of alkanes, H3TMA forms open framework honeycomb layers that align in such a manner that adjacent sheets are almost eclipsed with respect to each other (Figure 9). The resulting architecture observed in the crystal structure is essentially identical to that depicted in Figure 8b. H3TMA represents an example of a self-assembled motif, meaning that there are limits in terms of supramolecular synthesis when compared to modular systems.

Figure 9. Trimesic acid isooctane, 9 an organic open framework structure.

Supramolecular Synthesis of Solids

257

The H3TMA network can be extended to generate larger holes, and therefore larger channels and cavities, by employing the modular self-assembly approach. The structure of [HaTMA][bipy]l.5 would be expected to exist as an expanded form of H3TMA since the pyridine-carboxylic acid supramolecular synthon appears to be more stable than the carboxylic acid dimer itself, s4 As shown in Figure 6a, the anticipated structure indeed occurs and the cavities are large (ca. 26 x 35 /~). However, these cavities are filled by the interpenetration of three independent networks (Figure 6b), thereby affording a close-packed structure with no cavities. This type of interpenetration, which resembles weaving, is facilitated by puckering of the pseudohexagons that form the network. Rao et al. recently reported a related structure that is based upon modular self-assembly: 59 an organic network formed by trithiocyanuric acid (TCA) and bipy. Adjacent layers are aligned parallel to each other and there is no interpenetration. The resulting open framework structure exhibits channels with an effective diameter of 10/~. An interesting feature of this compound is that the cavities in the layers, and therefore the channels, can vary in size depending on the solvent of crystallization that is used to template the modular self-assembly process. (Scheme 8). It should be noted that the two architectures are not simply distorted or stretched variants of one another, they have distinct hydrogen-bonding patterns. Another salient feature is that there are sulfur atoms accessible in the cavities, which could

a)

9

b)

._.

_jr--\

Scheme & Cavities in TCA-bipy complex resulting from (a) xylene template; (b) all other templates.

258

BRIAN MOULTON and MICHAEL J. ZAWOROTKO

promote selective sorption, or could facilitate desired chemical reactivity in the context of green chemistry. 6~

4.2. Networks Sustained by Organic Ions Salts that are based upon organic ions with complementary hydrogen-bonding sites represent an alternate approach to modular self-assembly of molecular components. H3TMA also represents an appropriate precursor in this context, via formation of ammonium salts of deprotonated forms of H3TMA. 56$ Although it might not at first be obvious how the ammonium moiety could extend anionic forms of H3TMA into honeycomb networks, Scheme 9 illustrates two such motifs, the supramolecular isomers A and B, both of which facilitate linear propagation of carboxylate anions. An important feature of architectures that are sustained by A and B is that some of their components and features can be fine-tuned without destroying the basic architecture. For example, the ammonium cation substituents can be changed without influencing the basic molecular recognition properties in the context of motifs A and B. For secondary amines, organic substituents would extend above and below the network, and in appropriate circumstances would preclude interpenetration. Depending upon the nature of the substituents, adjacent layers might interdigitate and/or adopt clay-like intercalates in the presence of appropriate guest molecules (Scheme 10). Both HTMA 2- and TMA 3- can sustain laminar structures that result from crystallization of H3TMA in the presence of primary and secondary amines (RNH 2 and R2NH). Related laminar architectures have also been synthesized using other polyfunctional carboxylic acids such as trimellitic and pyromellitic acids. 61Interestingly, stoiochiometry has little influence over whether or not laminar structures are obtained but it has a profound influence over the local hydrogen-bonding patterns and the molecular recognition features of the "organic clays" that are formed.

........H

H.......

0""" I ~

/ H ........O

ck

/" R,,C--Q\

,.,

H

,,R

H//

.o-.c.

~0

N Motif A

Motif B

Scheme 9. Likely motifs or supramolecular isomers for primary or secondary ammonium carboxylates.

Supramolecular Synthesis of Solids

259

Scheme 10. Interdigitation and Guest Inclusion in "Organic Clays".

$toichiometry 1:2 (1"t3TMA: Amine) Our observations indicate that the hydrogen bond network within each sheet is hardly predictable but that it can be reproducible over a wide range of ammonium cations. This basic network is illustrated in Figure 10 and it is composed entirely of ionic hydrogen bonds. If alkyl substituents are present on the ammonium cation, then the typical result is a laminated material with poor ability to adsorb molecules because of interdigitation of the alkyl substituents 62 (Figure 11). However, use of dibenzylamine (DBA), [(PhCh2)2NH], mitigates against interdigitation and promotes reversible incorporation of aromatic guest molecules. The resulting compounds are structurally related to clays, but they are inherently hydrophobic and have affinity for aromatic guests over alcohols or water. In this series of compounds, there is some variation in the geometry of the hydrogen bond layer and in the manner in which guest molecules are incorporated. In general, the benzyl groups form a plethora of aromatic C-H...n interactions to the surrounding guest molecules. The unit cell lengths are typically multiples of ca. 12 x 17 x 21 ,/k (stacking axis, short axis and long axis, respectively). The length of the stacking axis represents the interlayer separation and a doubling of the length of the stacking axis occurs when adjacent layers are not related by translation. Multiples of short and long axes also occur because of differences in the arrangement of guest molecules between benzy! groups. In effect, guest molecules and/or benzyl groups do not necessarily repeat with the asymmetric unit of the hydrogen-bonded layer. The crystal structures might be classified based upon the stacking axis as being of one of two types: (a) identical packing of adjacent layers (i.e. related by translation); and (b) adjacent layers which are different from each other. The hydrogen-bonded sheets can be either flat or corrugated. In effect, the host matrix is a flexible, genetic host material for aromatic molecules. A representative structure is illustrated in Figure 12 and, as should be

260

BRIAN MOULTON and MICHAEL J. ZAWOROTKO

Figure 10. Hydrogen bond network in [Htrimesate][dibenzylammonium]2. clear, there is no interdigitation of benzyl groups. Interestingly the guest molecules interact with walls of the channels only and the asymmetric unit is very unusual: 3:3:1 for host-guest-solvent. In the presence of primary ammonium cations similar structures are obtained but they are more appropriately termed bilayer architectures

Figure 11. Crystal packing in [HTMA][NR2H2]2, R = alkyl.

Supramolecular Synthesis of Solids

261

Figure 12. Intercalation of EtOH and veratrole in [HTMA][N(benzyl)2H2]2. since there are alternating hydrophobic and hydrophilic regions. A typical structure is illustrated in Figure 13. Similar structures are obtained for both alkyl and benzyl ammonium cations. It might be reasonable to describe such structures as being cytomimetic since there is a resemblance to the type of supramolecular synthons that exist in phospholipid membranes and in the solid phases of surfactants. The ancillary organic groups orient in the same direction and interdigitation with adjacent layers to generate hydrophobic regions. The hydrophilic faces of adjacent bilayers also face one another and tend to adsorb water molecules. The thickness of hydrophilic layers ranges from 3.2 to 3.4/~, while the thickness of interdigitated layers increases with the size of the organic group.

Stoichiometry 1:3 (1"13TMA:Amine) In principle, 1:3 stoichiometry offers the opportunity to generate honeycomb networks. As revealed by Figure 14a, motif A or B should be capable of propagating the trimesate anion into a honeycomb structure. Figure 14b reveals that the crystal

Figure 13. Crystal structure of 8[HTMA] 16[HPhenethylamine]2.10 9 H20.

262

BRIAN MOULTON and MICHAEL J. ZAWOROTKO

structure of [TMA][dicylohexylammonium]3 does indeed exist as the anticipated honeycomb array. The cyclohexyl moieties, which are omitted for the sake of clarity, effectively prevent interpenetration by closing the 1.3 nm cavities that are present within the honeycomb structure. The modular nature of this structure permits replacement of the cyclohexyl moieties by other moieties. In this context, alkyl groups that are less sterically demanding (e.g. n-alkyl) have also been incorporated into the motif in Figure 14(a). Interpenetration occurs in these structures.

Guanadinium Sulfonates A series of related structures that are based upon two-dimensional layers resulting from hydrogen bonding of the trigonal guanidinium cation, C(NH2) ~, and organic sulfonate ions RSO 3 has been extensively studied by the Ward group (Scheme 11). Interdigitation of the organic substituent of the sulfonate ions on adjacent layers and ionic hydrogen bonding predictably leads to a broad series of laminar architectures. It should be noted that there are several key differences between guanadinium sulfonates and alkylammonium trimesates: 9 There exists only one ancillary organic functional group per sulfonate ion compared to up to two ancillary functional groups per ammonium cation. 9 The anion is functionalized rather than the cation. 9 In one sense, the alkylsulfonates are more versatile since they can exhibit architectural (i.e. supramolecular) isomerism so as to generate either bilayer or clay-like architectures. In order to generate a clay-like architecture, organic

Figure 14. Modular honeycomb grids from trimesic acid.

Supramolecular Synthesis of Solids

263

Scheme 11. (a) Guanidinium sulfonate 2D layer; (b) supramolecular isomers of guanidinium sulfonates (dependent on orientation of R on the sulfonate group, i.e. up or down). groups must orient above and below each layer as illustrated by Scheme 12. The steric demands of the organic group appear to determine whether they orient in the same direction (i.e. a bilayer structure) or alternate above and below the layer (i.e. a clay-like structure).

4.3. Metal-Organic Networks A broadly studied alternative to 2D hydrogen-bonded organic networks is exemplified by a number of classes of metal-organic coordination polymers that also represent examples of modular self-assembly. The most commonly used design strategy exploits known coordination geometries of metals to propagate 2D sheets via coordination with linear bifunctional spacer ligands. The ratio between metal and ligand, and the nature of the coordination of terminal ligands (i.e. degree of chelation) are the primary factors that determine the topology of the network.

264

BRIAN MOULTON and MICHAEL J. ZAWOROTKO

Scheme 12. Illustration showing the effect of the orientation of the ancillary functional group on suprarnolecular isomerism.

Square Grids In this context, square grid networks exemplify a particularly simple and commonly reported example of a predictable 2D metal-organic network. Square grid coordination polymers based upon 1:2 metal-ligand complexes with linear bifunctional spacer ligands were first reported using cyano ligands and have recently been expanded in terms of chemical type and cavity size to include pyrazines 63 and bipy.21b,57,64 Square grid networks generated with bipy spacer ligands were first reported by Fujita et al. 46aFujita's structures are based on Cd(II) and numerous other examples have subsequently been reported based on a number of other transition metals, including Ni(II) and Co(H). Although these 2D coordination networks are essentially identical within the coordination grid (square dimensions ca. 11.5 x 11.5 ~), the crystal structures of compounds differ in the mode that networks stack with each other (interlayer separations range from 6-8/~). The compounds [M(bipy)E(NO3)E].guest have also been studied extensively by us and, interestingly, we have only observed three basic crystal structure types (Figure 15). Type A compounds crystallize in C2/c with similar cell parameters (monoclinic; a = 21.5, b = 11.5, c = 13; 13 = 102~ have 2:1 guest-host stoichiometry and interplanar separations of ca. 6 A. The crystal packing appears to be influenced by C-H...O hydrogen bond interactions between the bipy ligands of one square grid and the nitrate anions of adjacent square grids. The square grids do not align with a unit cell face and adjacent grids are slipped in one direction by ca. 20%, i.e. every sixth layer repeats. The crystal packing of Type B compounds is also controlled by weak interactions between adjacent layers. They generally crystallize in P211nwith 2.5 guest molecules per metal center and cell parameters are fairly consistent

5upramolecular Synthesis of Solids

265

Figure 15. Modes of stacking in [M(dipy)2(NO3)2]n.

(monoclinic; a = 16, b = 14.75, c = 16; 13= 100~ The interlayer separation is ca. 0.8 nm. Type C compounds have interlayer separations that are similar to those seen for Type B compounds. Four examples of type C compounds have 3:1 stoichiometry: three crystallize in C2/c (monoclinic; a = 16, b = 11.5, c = 23; 13= 100~ the other crystallizes in Cc and has similar cell parameters except there is a tripling of the a-axis and the cell volume. Another example of a type C grid 65 crystallizes in Pn (monoclinic; a = 11.4, b = 22.8, c = 15.9; 13= 93.3~ Although these cell parameters are inconsistent with the previous four structures, the packing of the grids is appropriate for a type C grids. The positioning of the grids facilitates inclusion of one guest molecule in the center of each grid. The other guest molecules lie between the grids and engage in stacking interactions between the bipy ligands and themselves. In all of these compounds the proportion of the crystal that is occupied by guest molecule is ca. 50% by volume. In such a situation it becomes reasonable to question whether interactions between the guest molecules determine the cavity shape and crystal packing of the square grid polymers rather than vice-versa. This issue is addressed later. As these square grid architectures are inherently modular, it should be possible to extend their dimensions by simply using longer spacer ligands. An example of such a structure, [Ni(1,2-bis(4-pyridyl)ethane )2(NO3)2]n-2 veratrole, is illustrated in Figure 16. This structure has grid dimensions ca. 20% larger than the smaller grids (square dimensions ca. 13.6 x 13.6 ~) and an effective cavity size of 11 x 11 /~, large enough to enclathrate more than one aromatic guest.

Other 2D Architectures Another metal geometry or node that is of particular interest because of its potential range of supramolecular isomers is the T-shaped geometry, i.e. a m e r substituted metal moiety with 1:1.5 metal : spacer ligand ratio. This node has thus far produced three distinct 2D supramolecular isomers: brick wall, 46 herringbone, 49

266

BRIAN MOULTON and MICHAEL J. ZAWOROTKO

Figure 16. [Ni(1,2-bis(4-pyridyl)ethane)2(NO3)2]n.2 veratrole. and bilayer. 48 It is interesting to note that, if one calculates the possible tiling patterns (i.e. all points lie in the same plane) that are possible for T-shaped nodes (Scheme 13), three of the five possibilities have already been realized: ladder, brick, and herringbone. The brick architecture is observed as the product of the reaction between heptacoordinate Cd(II) and 1,4-bis((4-pyridyl)methyl)-2,3,5,6tetrafluorophenylene. 46a The T-shape geometry is the result of two nitrate ligands chelating in a bidentate manner, thereby occupying four of the seven coordination sites. The structure is triply interpenetrated and, as such, does not have channels or cavities. Interestingly, in a similar system using the non-fluorinated pyridyl based ligand, a 1D ladder structure was observed. The brick architecture was also seen in [Ni(4,4'-azopyridine)l.5(NO3)2] n, which interpenetrates with two perpendicular [Ni(4,4'-azopyridine)2(NO3)2] n square-grid networks. 46b The herringbone or "parquet floor" architecture has recently been observed by several groups: 9 In these structures, the coordination sphere is similar to that of the brick architectm'es: heptacoordinate Cd(H), or Co(ID, with two terminal bidentate nitrate ligands and coordination to one end of three 4,4'-azopyridine bridging ligands; an isostructural example has also been reported with 1,2-bis(4-pyTidyl)ethyne as the bridging ligand. 4s The bilayer architecture has been observed in at least three compounds. 4s Interestingly, it has been observed as the product from the reaction of Co(NO3) 2 and bipy, which also generates ladder, square grid, and herringbone architectures. The bilayer form of [Co(bipy)l.5(NO3)2] is observed if crystallization occurs in the presence of CS 2 or H20. The bilayers pack by partial interdigitation, which allows

Supramolecular Synthesis of Solids

267

Scheme 13. Tiling patterns of T-shaped moieties.

1D channels to run through the structure. This structure is particularly relevant since it represents the first reported example of a compound that might be regarded as a metal-organic zeolite, i.e. the structure is porous and stable to loss of guest. 25 The bilayer architecture has also been reported for systems using 1,2-bis(4,-pyridyl)ethane. The number of supramolecular isomers already observed in the Co(NO3)2/bipy system indicates how important selection of template and crystallization conditions are. It is conceivable that by careful consideration of template size and stoichiometry, the weave and long-and-short brick motifs will also be realized in the near future. In terms of topology, it should be noted that brick and herringbone motifs are both examples of (6,3) nets and can therefore be regarded as being closely related to honeycomb (6,3) nets. 2~ Honeycomb networks are quite common in organic structures because of the availability of trigonal nodes (i.e. 1,3,5-trisubstituted benzenes such as trimesic acid and species such as the guanadinium cation) but they seldom occur in the context of metal-organic polymers because trigonal and trigonal b i p y r a m i d a l c o o r d i n a t i o n g e o m e t r i e s are quite rare. H o w e v e r , [Cu(pyrazine)l.5]BF466 is based upon trigonal Cu(I) and it should therefore be unsurprising that it crystallizes as a honeycomb (6,3) net. That there now exist a number of ligands with trigonal geometry means that it is likely that a wider range of honeycomb structures will be generated in the near future.

5. SUPRAMOLECULAR SYNTHESIS OF 3D STRUCTURES The design of 3D network architectures normally presents an added level of complexity in comparison with 2D architectures but it does in many ways represent

268

BRIAN MOULTON and MICHAEL J. ZAWOROTKO

the ultimate challenge to crystal engineers since it leads more directly to crystal structure control and prediction. For example, in most situations, a finite number of structural isomers can be calculated for the scenario where all nodes lie in the same plane. However, a larger number of possibilities might exist when that limitation is not imposed. It is therefore somewhat ironic that two of the simplest examples of predictable networks are exemplified by 3D networks generated via assembly of tetrahedral or octahedral nodes. Tetrahedral nodes are predisposed to generate diamond-like (diamondoid) architectures, whereas octahedral nodes are expected to afford octahedral networks. These architectures can be obtained for both organic (typically hydrogen bonded) and metal-organic (i.e. coordination polymer) systems. Interpenetration can occur in these compounds, thereby mitigating against enclathration and porosity. However, interpenetration can also be exploited as a potentially important design paradigm for rational transformation of some of the 2D networks described earlier into 3D frameworks. This principle is discussed with respect to interpenetration of identical networks (homocatenation), and interpenetration of different networks (heterocatenation). 5.1. Hydrogen-Bonded Networks

Diamondoid Networks A report 28 by Ermer on the structural characterization of adamantane-l,3,5,7tetracarboxylic acid and its implications represented a watershed for crystal engineering. Ermer's study was followed by a flurry of activity into design from first principles of both organic diamondoid networks and metal-organic diamondoid coordination polymers. The carboxylic acid groups of adamantane- 1,3,5,7-tetracarboxylic acid are tetrahedrally oriented. It is therefore not surprising that they self-assemble via the hydrogen-bonded carboxylic dimer supramolecular synthon to afford an infinite diamondoid network. Each network possesses cavities that could accommodate a large roughly spherical guest, or guest aggregate, of roughly 12/~ in diameter. However, five independent networks interpenetrate in such a way that the crystal structure is densely packed and, consequently, guest inclusion is precluded. As subsequent studies revealed, interpenetration is a widespread phenomenon in diamondoid networks and, indeed, occurs in many other organic and metal-organic structures that would otherwise have large cavities or channels. An interpenetrated diamondoid architecture is also exhibited by methanetetraacetic acid, for which the cavities generated are approximately 10/~ in diameter. 29 As would be expected, methanetetraacetic acid exhibits a lower degree of interpenetration: threefold. Interestingly, 2,6-dimethylidine-adamantane-l,3,5,7-tetracarboxylic acid also forms a hydrogen-bonded diamondoid structure but it exhibits a much lower degree of interpenetration than its unsubstituted precursor. The twofold "double diamondoid" architecture is not as densely packed, it can therefore act as a host and it enclathrates guest molecules. 28

SupramolecularSynthesisof Solids

269

Wuest has demonstrated that the pyridone moiety also generates a hydrogenbonded supramolecular synthon that is suitable for building extended arrays. 67 Remarkably, methanetetra(6-phenylethynyl-2-pyridone) exhibits a diamondoid network, sevenfold interpenetration and cavities large enough to enclathrate butyric or valeric acid. 27aWuest introduced the concept of"tectons" to describe molecules that inherently possess the molecular structure and intermolecular recognition features to predictably self-assemble into crystalline networks. He followed this study with several other examples of diamondoid networks sustained by the pyridone moiety.27c-d Modular self-assembly relies upon two molecular components that are not individually capable of self-assembly, 22 and was discussed in detail in Section 2. The modular self-assembly approach is well illustrated by the cubane-like cluster [Mn(CO)3(~t3-OH)]4, which is a tetrahedral hydrogen bond donor. [Mn(CO)3( ~ OH)] 4 (Figure 1) does not self-assemble if crystallized pure 31 but forms hydrogenbonded diamondoid structures if crystallized with 2 equiv of a linear bifunctional H-bond acceptor. For example, 1,2-diaminoethane yields diamondoid networks with threefold interpenetration. Numerous other diamondoid structures can be prepared with [Mn(CO)3(~-OH)] 4 or its Re analogue, some of which exhibit high levels of interpenetration and channels that contain solvent molecules. 31'32 The single component (tectonic) and modular approaches to building diamondoid networks are illustrated schematically in Scheme 3. Finally, it should be noted that there are supramolecular synthons that do not rely upon hydrogen bonds. In this context, N...Br interactions were exploited to propagate a diamondoid network in the cocrystal formed by carbon tetrabromide and hexamethylenetetraaamine. This structure also represents a different but equally effective form of the modular approach: two tetrahedral nodes with one possessing donor functionality and the other acceptor functionality only. The structure of the cocrystal formed by carbon tetrabromide and hexamethylenetetraaamine exhibits twofold interpenetration and does not enclathrate solvent or guest.

Other 3D Hydrogen-Bonded Networks Although there are many examples of organic crystals that can be defined as 3D networks, few of them are predictable or even rational in the same sense as diamondoid networks. Trimesic acid, H3TMA, is an interesting exception and was discussed earlier in the context of 2D structures. Indeed, H3TMA represents a prototypal molecule in the context of hydrogen bonding and generates extended structures when pure, partially deprotonated, in coordination polymers or in cocrystals. Anionic derivatives of H3TMA self-assemble into honeycomb grids via O H---O- hydrogen bonds. 62 However, pyromellitic acid, 1,2,4,5-benzenetetracarboxylic acid, H4PMA, has been less widely explored than H3TMA. It has been utilized as a ligand in coordination polymer networks 6s and very few organic structures containing H4PMA or its derivatives are known. We anticipated that

270

BRIAN MOULTON and MICHAEL J. ZAWOROTKO O"

0

O"

"Q

8 0

0

"Q

O"

.0

0D

""O"

o~H"" O"

2D o

D

A

o%o.%. "x

b

o

.

.

o.

o" ..HIO

ID

.

O'

0~,..

& 3D

"A

b

.

q

D,

,,

q

"x . . . .

A

2D isomer

D

A

3D isomer

Scheme 14.

doubly deprotonated H4PMA would self-assemble via dicarboxylate hydrogen bonds to form 0D (two intramolecular hydrogen bonds), 1D (one intramolecular and one intermolecular hydrogen bond), or 2D/3D (two intermolecular hydrogen bonds) networks (Scheme 14). Interestingly, H2PMA 2- anions exhibit all four of these supramolecular isomers depending upon the polymorph or the counterion. 52 The 3D structure occurs because H2PMA 2- moieties orient in such a manner that they form hydrogen bonds to the next layer. The network can be described as a framework built from square building blocks which alternate parallel and perpendicular with respect to one another. This network can therefore be regarded as an organic analogue of NbO. 2~

5.2. Metal-Organic Networks Diamondoid Networks As we have tried to stress throughout this chapter, the diversity of components that are available for crystal engineering and the means by which they assemble spans the full range of chemistry. Furthermore, metal-organic or coordination polymer networks are conceptually related to hydrogen-bonded systems since they can also be regarded as "donor-acceptor" networks that are the result of selfassembly of at least two self-complementary components. The breadth of chemical moieties that might be used for crystal engineering is particularly well illustrated by the case of diamondoid networks. Diamondoid architectures using a tetrahedral metal (Zn or Cd) as the node and cyanide ligands (CN-) as the spacer represent

Supramolecular Synthesis of Solids

271

prototypal examples of diamondoid coordination polymers. Zn(CN) 2 and Cd(CN) 2 form diamondoid networks with twofold interpenetration. 21a However, Cd(CN) 2 can also be obtained as a single network with CC14filling the cavity. 21r This result illustrates two principles that have broad implications: 9 Interpenetration can be avoided in the presence of an appropriate template or guest molecule! 9 Such compounds might be regarded as catenated and noncatenated supramolecular isomers of each other. A diamondoid architecture also results when Zn(CN) 2- is reacted with Cu(I). The resulting anionic network might be viewed as tetrahedral zinc nodes linked to tetrahedral copper nodes by cyanide spacers, however the nature of coordination at the copper and zinc ions remains unclear. Analysis of the structural data indicated that it is most appropriate to consider the coordination of the copper as 100% organometallic (Cu-C) and the coordination of the zinc 100% metal-organic (Zn-N). The ionic nature of this particular framework means that the presence of a counterion in the resulting cavities is required. Indeed, N(CH3) ~ fits comfortably inside the adamantoid cavity and precludes interpenetration. A report 69 on the crystal structure and p r o p e r t i e s of [Cu(2,5-dimethylpyrazine)2(PF6)] represented one of the first examples of a metal-organic diamondoid structure and the related compound [Cu(4,4'-bipy)2](PF6)] was reported shortly thereafter. 66 Both structures exemplify the modular assembly design strategy and contain anions in the cavities generated by the diamondoid structure. In the case of the latter, the intermetallic separations are 11.16 tt, and result in cavities that are sufficiently large to facilitate fourfold interpenetration as well as inclusion of the counterions. A diamondoid architecture propagated by silver(I) and bipy, {[Ag(4,4'-bipy)2](CF3SO3) }, was reported shortly thereafter 7~ and it also exhibits fourfold interpenetration with anions in cavities. The Ag-Ag separations are 11.6/~. The 4-cyanopyridine analogue was reported in the same article and exhibits metal-metal separations of 9.93 ]k. Despite the variations observed in the dimensions of these networks, both exhibit fourfold levels of interpenetration. Subsequent studies resulted in a plethora of diamondoid metal-organic structures and two-, three-, four-, five-, seven-, and ninefold levels of interpenetration. 71 It should be noted that although interpenetration reduces or eliminates porosity, there are at least two important properties that can be addressed with such structures. 9 They are predisposed to form acentric networks since there is no inherent center of inversion at a tetrahedral node. Indeed, an odd level of interpenetration and an unsymmetrical ligand will definitely generate a structure that exhibits polarity. 22 9 These structures could be useful for selective anion exchange.

272

BRIAN MOULTON and MICHAEL I. ZAWOROTKO

In the context of the former, a series of neutral diamondoid architectures have been prepared with bridging ligands of varying size. 21a'66,71a'72These compounds are of general formula ML 2 (M = T d metal; L = bridging anionic ligand) and it follows that a neutral network will be generated if a +2 metal is coordinated to two -1 anionic ligands. Zn(isonicotinate)2 and Cd(trans-4-pyridylacrylate) 2 exhibit three- and fivefold degrees of interpenetration, respectively, and possess interesting properties in the context of polarity. 33 In the former compound, the Zn-Zn distance is ca. 8.8/~. This is consistent with the previous structures that exhibit fourfold interpenetration. The Cd-Cd distance is ca. 11.5/~, similar to the intermetallic distances observed in the fourfold interpenetrated structures that also contain counterions.

Octahedral Networks Prototypal examples of octahedral networks are exemplified by iron cyano compounds. Such compounds are very well documented since they have been used for centuries as pigments. An early X-ray study 73 of Berlin green, [FelIIFelII(CN)6], Prussian blue, [KFelIFelII(CN)6], and Turnbull's blue, [K2FeIIFelI(CN)6], demonstrated that the iron cations act as the node in octahedral arrays in which the iron cations are linked by linear cyano ligands. These compounds form isostructural networks that vary only in the degree of potassium inclusion and the oxidation states of the iron atoms. Berlin green can therefore be regarded as being the prototypal example of an open framework octahedral network; however, the limited length of the cyano means that it has no relevance in the context of porosity. Synthetic metal-organic octahedral networks were first reported in 1995. [Ag(pyrazine)a](SbF6) 74 is sustained by octahedral Ag(I) cations and relatively short pyrazine ligands. The framework is necessarily cationic. A neutral framework is exemplified by [Zn(bipy)2(SiF6) ]. In this structure (Figure 17), the SiF 2- counterions cross-link the square grids that are formed by Zn and bipy to form a rigid octahedral polymer with very predictable shape and dimensions. The structure cannot interpenetrate because the walls of the channels are blocked by bipy ligands. The resulting channels have an effective cross section 8 x 8/~ and represent ca. 50% of the volume of the crystal. Solvent molecules are readily eliminated but the framework collapses irreversibly upon loss of solvent. Perhaps the most salient feature of this structure is that not only is the structure rational, it is entirely predictable. Indeed, [Zn(bipy)2(SiF6) ] crystallizes in space group P4/mmm with Z = 1. In other words, the point group at Zn, D4h, is propagated into space group symmetry. Furthermore, the cell parameters are determined by the intermetallic separations. Octahedral coordination polymers remain much less common that their diamondoid counterparts, but a recent report revealed a novel metal-organic coordination polymer, Zn40(BDC)3 (BDC = benzenedicarboxylate) that suggests an exciting future for such compounds. 26 ZnaO(BDC)3 is a relatively simple and inexpensive

Supramolecular Synthesis of Solids

273

Figure 17. [Zn(4,4'-dipy)2(SiF6)]rr--an octahedral coordination polymer. material to prepare and is remarkably stable after loss or exchange of guest, remaining crystalline at temperatures above 300 ~ The key feature that makes ZnaO(BDC)3 special is that it exhibits a relative degree of porosity that is hitherto unprecedented in crystalline solids. As revealed by Figure 18, the octahedral framework exhibits a large amount of surface area that remains accessible to guest molecules because it contains pores and cavities that are large enough to accommodate and release organic molecules such as chlorobenzene and dimethylformamide. Calculations and experimental data indicate that ca. 60% of the structure

Figure 18. A portion of the crystal structure of Zn40(BDC)3. The accessible surfaces are highlighted.

274

BRIAN MOULTON and MICHAEL J. ZAWOROTKO

is available and accessible. This compares to the typical value of ca. 30% seen in zeolites, l~

Other 3D Networks In addition to the obvious, i.e. diamondoid and octahedral networks, there are numerous examples of novel 3D networks that have been observed in recent years. Many can be described as supramolecular isomers of low-dimensional structures. Two such structures are supramolecular isomers formed by self-assembly of T-shaped nodes. As discussed in Section 2, such self-assembly can afford 3D architectures that have not been seen in naturally occurring compounds. Scheme 15 reveals one of these structures. [Co(bipy)l.5(NO3)2] n. 1.5 benzene generates this particular architecture. The cavities are revealed in Figure 19 and they are exceptionally large, having effective cross section 8 x 40 ]k. These large cavities are capable of sustaining both threefold interpenetration and inclusion of guest molecules in channels (Figure 20). Interestingly, although the networks are inherently centrosymmetric, the crystal is polar because the guest molecules align in such a manner that their supramolecular structure cannot contain a center of inversion. [Ag(bipy)(NO3)] n generates another type of supramolecular isomer for selfassembly of T-shaped components. It self-assembles into linear Ag-bipy chains that cross-link via Ag-Ag bonds. This particular 3D structure has been described as a "Lincoln Log" type structure and exhibits a threefold level of interpenetration that is open enough to facilitate ion exchange of the loosely bound nitrate anions. 47a

5.3. Hybrid Structures An alternate approach to building 3D structures that seems to offer considerable potential is to manipulate already existing 2D structures. There are two relatively simple strategies in this context.

Scheme 15. A supramolecular isomer for self-assembly of T-shape modules.

Supramolecular Synthesis of Solids

275

Figure 19. [Co(4,4-bipyridine)1.s(NO3)2]n.1.5 benzene.

Cross-Linking of 2D Structures Such a strategy is feasible if one selects an appropriate 2D structure that has functionality in the vertical direction. Such an approach has been widely used by clay chemists and hence the term "pillaring" might be applied to describe such a process. [Zn(bipy)2(SiF6) ] might be used as a prototype in the context of coordination polymers since it can be regarded as having been generated from square-grid coordination polymers that are cross-linked by la-SiF 6 anions. In the context of hydrogen-bonded structures, guanidinium sulfonates represent a class of compounds that have been cross-linked in a rational manner so as to generate infinite 3D structures. 75

Figure 20. Threefold interpenetartion in [Co(4,4-bipyridine)1.5(NOg)2]n"1.5 benzene.

276

BRIAN MOULTON and MICHAEL J. ZAWOROTKO

Interpenetration of Identical or Different 2D Networks As discussed earlier in this chapter, interpenetration is a widely encountered phenomenon that mitigates against the existence of frameworks with very large cavities. However, Scheme 16 reveals that there are situations in which interpenetration can occur, generate porosity and afford 3D structures. Square-grid polymers that are based upon long spacer ligands such as dipy-Et can interpenetrate is such a fashion. 53 However, an even more intriguing situation that could have important implications for design of new hybrid materials is exemplified by the crystal structure of the square grid coordination polymer {[Ni(bipy)2(NO3)2].2pyrene },. Careful examination of the crystal packing in this compound reveals that the pyrene molecules form an independent noncovalent network that is complementary from a topological perspective with the square grid. Indeed, the resulting crystal represents what is to our knowledge the first compound in which it has been revealed that two very different types of 2D net interpenetrate. The square-grid coordination networks (Figure 21a) possess inner cavities of ca. 8 x 8/~ and stack in such a manner that they lie parallel to one another with an interlayer separation of ca. 7.9 /~. The pyrene nets (Figure 21b) are sustained by edge-to-face interactions and contain cavities of ca. dimensions 6.5 x 3.5 A. The planes of the neighboring molecules intersect at an angle of ca. 60 ~ and there are no face-to-face stacking interactions between the molecules. The pyrene nets can be regarded as distorted (4,4) nets if the node is the point in space at which the vectors of the four pyrene planes intersect. An alternate interpretation is that nodes exist at the point of the edge-to-face interactions. The pyrene net could then be regarded as a distorted brick wall form of a (6,3) net. It is important to note that either a (4,4) or a (6,3) planar

Scheme 16. From 2D to 3D via catenane supramolecular isomerism.

Supramolecular Synthesis of Solids

277

Figure 21. Complementarycoordination polymer and pyrene nets. net is complementary from a topological sense with the (4,4) coordination polymer net and ensures that the coordination polymer nets must pack in a staggered manner. Given that cavity size within the pyrene nets is complementary with the width and height of a single aromatic ring, it should be unsurprising that the pyrene nets thread orthogonally with the bipy ligands of the coordination polymer via face-to-face and edge-to-face interactions and that the calculated volumes of the two nets are similar. Indeed, this is to be expected based upon the observation that bipy square grids are self-complementary as they can interpenetrate in a twofold fashion. 39 The interpretation of this crystal structure as interpenetrating covalent and noncovalent nets is potentially important in the context of understanding the structure and stoichiometry of compounds that are based upon interpenetrated covalent and noncovalent nets. The structure also illustrates how polarity in crystals can be generated from subtle packing of achiral components, since the pyrene molecules form chiral nets. It should be noted that this type of packing is not unique to {[Ni(bipy)2(NO3)2]2pyrene}n. Indeed, its naphthalene analogue, {[Ni(bipy)2(NO3)2].3naphthalene }n, can be interpreted as being the result of interpenetration of hexagonal and square nets 65 and a study of a series of more than twenty related compounds has revealed the presence of noncovalent nets in every one of these compounds. 76 It should also be noted that a study on the structure of the cocrystal formed between ferrocene and pyrene 77 reveals the presence of pyrene 2D nets that are almost identical to those observed in {[Ni(bipy)2(NO3)2]-2Pyrene }n. In other words, the noncovalent nets even exist in the absence of covalent nets.

278

BRIAN MOULTON and MICHAEL J. ZAWOROTKO

6. POTENTIAL APPLICATIONS A considerable amount of fundamental research into understanding the nature and predictability of supramolecular synthons remains to be completed. Indeed, an enhanced database concerning supramolecular synthons in the broad context would also assist our fundamental understanding of solution chemistry and biochemistry. Nevertheless, it is clear that there are a number of applications of crystal engineering that could be realized in the short term. Several of these are summarized below. 9 Supramolecular synthesis of new classes of cocrystal and modular solid offers the potential to increase the known range of crystalline materials by 2 or 3 orders of magnitude and to facilitate combinatorial approaches to materials science. For example, cocrystals could serve as supramolecular derivatives of drugs and functional materials (i.e. modification of bulk properties without changing the molecular structure of the active species) or they could serve as precursors to covalent products, including polymers. Such an approach has already been effective in formulation of polaroid film. 78 In this context, we have recently shown that it is possible to rationalize certain types of host-guest structures as being based upon topologically complementary networks. 19b 9 Solvent-free synthesis, green chemistry, offers many potential advantages, including cost and environmental benefits. CocrystaUization of substrates and subsequently conducting reactions in the solid state offers the opportunity of very careful control over regio- and stereochemistry. Indeed, it should be noted that the concept of crystal engineering was first introduced by Schmidt in the context of solid-state organic photochemistry. It is also possible that supramolecular arrays could act as precursors to new classes of 2D and 3D covalent polymers. 79 9 New classes of adsorbent, "organic and metal-organic clays and zeolites", represent an area in which considerable progress has already been made. Such compounds offer clear potential for the following: efficient, cost-effective alternatives to current methods of enantiomeric separations; new materials for separation of gases, liquids, and solutes; new industrial heterogeneous catalysts; new drug delivery matrices (e.g. matrix for oral delivery of otherwise unstable drugs); a new generation of chemical sensors; and new storage matrices for gases such as methane. 9 The rational design of polar materials for use in materials science also represents a concept in which there have already been promising developments. Unfortunately, in most organic crystals, antiparallel architectures predominate, thereby canceling dipoles of highly polarizable molecules and mitigating against bulk polarity. Fortunately, there already exists a range of modular, open-framework organic and metal-organic solids. Many of these compounds contain architectures (e.g. square grid, honeycomb, octahedral) that favor incorporation of polar strands into channels, thereby reducing the

5upramolecular Synthesis of Solids

9

279

driving force for antiparallel alignment. Results obtained by Hulliger et al.80 suggest that such compounds, in particular channel type inclusion compounds, hold considerable promise in the context of the design of solids that possess fine tunable bulk polarity. Metal-organic polymers offer considerable potential in the context of molecular magnetism, semiconductors, and conductors. 8~

7. CONCLUSION AND FUTURE DIRECTIONS The peasantwho wants to harvestin his lifetime cannotwait for the ab initio theoryof weather. H.G. von Schnering(1981). The fundamental precept of crystal engineering is that all information necessary for design of extended one-, two-, and three-dimensional structures is already present at the molecular level in existing chemical species. Recent advances in our understanding of supramolecular chemistry and supramolecular synthons have been aided by the advent of CCD diffractometers coupled with ever more powerful visualization and analysis tools. It should therefore be unsurprising that control over supramolecular architectures, also known as molecular tectonics, 27 has advanced rapidly. That these tools are now routinely available means that an even more concerted and systematic approach to gaining an understanding of the subtle factors that control architectures in the solid state is feasible. Such rational design of supramolecular structure relies upon invoking the concepts of self-assembly, in effect supramolecular synthesis, and exploits noncovalent forces as varied as the following: 1. Hydrogen bonding of diverse types: strong hydrogen bonding (e.g. O-H---O) and weak hydrogen bonding (e.g. C-H---O and even C-H---g) can be exploited. 2. Coordinate covalent bonds. 3. Electrostatic and charge transfer attractions. 4. Aromatic n-stacking interactions. These principles of crystal engineering and supramolecular synthesis have thus far been used to design, isolate, and characterize network structures from relatively small molecular components. In the context of coordination polymer networks, a recent review indicates how wide the range of chemical components and accessible network motifs has become. 48b However, the scale of these structures is such that cavities and channels are on the order of 1 nm and, to date, each cavity is identical. Careful selection of appropriate substrates or components and ever more control over crystal packing will offer the potential for rational design of an even more extensive array of modular (i.e. binary, ternary, or even higher order) structures than those that are currently available. In particular, judicious choice of supermolecules

280

BRIAN MOULTON and MICHAEL J. ZAWOROTKO

or biomolecules as templates and nodes should afford composite materials with nanoscale dimensions and cavities. Such composite materials would represent "uncharted territory" but they now appear to be at hand. In essence, suprasupermolecular synthesis in the solid state is likely to develop and, whereas prediction of crystal structures remains an elusive goal that will continue to be addressed, it does not preclude short-term applications of crystal engineering in a number of important areas. H.G. von Schnering's comments therefore seem particularly appropriate to summarize the current situation. 82

REFERENCES AND NOTES 1. 2. 3. 4. 5. 6. 7.

8.

9. 10. 11. 12. 13. 14. 15.

16. 17. 18. 19. 20.

W~ihler, E Annalen der Physik und Chemie 1828, 88, Leipzig. Schmidt, G. M. Pure Appl. Chem. 1971, 27, 647. Maddox, J. S. Nature 1988, 335, 201. (a) Lehn, J.-M. Pure Appl. Chem. 1978, 50, 871. (b) Lehn, J.-M. Supramolecular Chemistry; VCH: Weinheim, 1995. (a) Dunitz, J. D. Pure Appl. Chem. 1991, 63, 177. (b) Dunitz, J. D. Perspectives in Supramolecular Chemistry; Wiley: New York, 1996, Vol 2, Chap. 1. C1398H1278:Xu, Z. F.; Moore, J. S. Angew. Chem., Int. Ed. Engl. 1993, 32, 1354. (a) Cu146Seq3(PPh3)3o: Krautscheid, H.; Fenske, D.; Baum, G.; Semmelmann, M. Angew. Chem., Int. Ed. Engl. 1993, 32, 1303. (b) [Mo154(NO)14Oa2o(OH)28(H20)7o]25-: Muller, A.; Krickemeyer, E.; Meyer, J.; Bogge, H.; Peters, E; Plass, W.; Diemann, E.; Dillinger, S.; Nonnenbruch, E; Randerath, M.; Menke, C. Angew. Chem., Int. Ed. Engl. 1995, 34, 2122. (c) C2900H230oN60P1205600200FIgoPt60: Olenyuk, B.; Levin, M. D.; Whiteford, J. A.; Shield, J. E.; Stang, E J. J. Am. Chem. Soc. 1999, 121, 10434. (a) Tomalia, D. A.; Naylor, A. M.; Goddard, W. A. Angew. Chem., Int. Ed. Engl. 1990, 29, 138. (b) Tomalia, D. A.; Baker, H.; Dewald, J.; Hall, M.; Kallos, G.; Martin, S.; Roeck, J.; Ryder, J.; Smith, E Polym. J. 1985, 17, 117. (c) Newkome, G. R.; Guther, R.; Moorefield, C. N.; CarduUo, E; Echegoyen, L.; Perezcordero, E.; Luftmann, H. Angew. Chem., Int. Ed. Engl. 1995, 34, 2023. (d) Newkome, G. R.; Yao, Z. Q.; Baker, G. R.; Gupta, V. K. J. Org. Chem. 1985, 50, 2003. (a) Frechet, J. M. J. Science 1994, 263, 1710. (b) Hawker, C. J.; Frechet, J. M. J. J. Am. Chem. Soc. 1990, 112, 7638. Newkome, G. R. Pure Appl. Chem. 1998, 70, 2337. Pauling, L. The Nature of the Chemical Bond; Comell University Press: London, 1939. Wemer, A. Z Anorg. Chem. 1893, 3, 267. Kitaigorodsldi, A. I. Mixed Crystals; Springer-Verlag: New York, 1984. Allen, E H.; Davies, J. E.; Galloy, J. J.; Johnson, O.; Kennard, O.; Macrae, C. E; Watson, D. G. J. Chem. Inf. Comput. Sci. 1991, 31, 187. (a) Desiraju, G. R.; Gavezzotti, A. Chem. Commun., 1989, 10, 621. (b) Desiraju, G. R. Crystal Engineering: The Design of Organic Solids; Elsevier: Amsterdam, 1990. (c) Desiraju, G. R. Acc. Chem. Res. 1991, 24, 290. (d) Desiraju, G. R. Angew. Chem., Intl. Ed. Engl. 1995, 34, 2311. (a) Etter, M. C. Acc. Chem. Res. 1990, 23, 120. (b) Etter, M. C.; Macdonald, J. C.; Bemstein, J. Acta Crystallogr B 1990, 46, 256. (c) Etter, M. C. J. Phys. Chem. 1991, 95, 4601. Gavezzotti, A. Acc. Chem. Res. 1994, 27, 309. Ball, E Nature 1996, 381,648. (a) MacGillivray, L. R.; Atwood, J. L. Nature 1997, 389, 469. (b) Biradha, K.; Domasevitch, K. V.; Moulton, B.; Seward, C.; Zaworotko, M. J. Chem. Commun. 1999, 1327. (a) Wells, A. E Structural Inorganic Chemistry; Oxford University Press: Oxford, 1975. (b) Wells, A. E Three-Dimensional Nets and Polyhedra; Wiley: New York, 1977.

Supramolecular Synthesis of Solids

281

21. (a) Hoskins, B .E; Robson, R. J. Ant Chent Soc. 1990, 112, 1546. (b) Gable, R. W.; Hoskins, B. E; Robson, R. J. Ant Chent Soc. 1990, 112, 1677. (c) Abrahams, B. E; Hoskins, B. E; Liu, J. P.; Robson, R. J. Am. Chent Soc. 1991, 113, 3045. (d) Abrahams, B. E; Hoskins, B. E; Robson, R. Chent Commun. 1990, 60. 22. Zaworotko, M. J. Chent Soc. Rev. 1994, 23, 283. 23. Hennigar, T. L.; MacQuarrie, D. C.; Losier, P.; Rogers, R. D.; Zaworotko, M. J. Angew. Chem., Int. Ed. Engl. 1997, 36, 972. 24. Endo, K.; Sawaki, T.; Koyanagi, M.; Kobayashi, K.; Masuda, H.; Aoyama, Y. J. Ant Chem. Soc. 1995, 117, 8341. 25. Kondo, M.; Yo~2htomi, T.; Seki, K.; Matsuzaka, H.; Kitagawa, S. Angew. Chent, Int. Ed. Engl. 1997, 36, 1725. 26. Li, H.; Eddaoudi, M.; O'Keeffe, M.; Yaghi, O. M. Nature 1999, 402, 276. 27. (a) Su, D.; Wang, X.; Simard, M.; Wuest, J. D. Supramol. Chem. 1995, 6, 171. (b) Hajek, E; Graf, E.; Hosseini, M. W.; Delaigue, X.; De Cian, A.; Fischer, J. Tetrahedron Lett. 1996, 37, 1401. (c) Vaillancourt, L.; Simard, M.; Wuest, J. D. J. Org. Chem. 1998, 63, 9746. (d) Brunet, P.; Simard, M.; Wuest, J. D. J. Ant Chent Soc. 1997, 119, 2737. (e) Wang, X.; Simard, M.; Wuest, J. D. J. Am. Chent Soc. 1994, 116, 12119. 28. Ermer, O. J. Ant Chent Soc. 1988, 110, 3747. 29. Ermer, O.; Eling, A. Angew. Chem. Int. Ed. Engl. 1988, 27, 829. 30. Frankenbach, G. M.; Etter, M. C. Chent Matter. 1992, 4, 272. 31. Holman, K. T.; Zaworotko, M. J. J. Chent Cryst., 1995, 25, 93. 32. Copp, S. B.; Holman, K. T.; Sangster, J. O. S.; Subramanian, S.; Zaworotko, M. J. J. Chem. Soc., Dalton Trans. 1995, 2233. 33. Evans, O. R.; Xiong, R.-G.; Wang, Z.; Wong, G. K.; Lin, W. Angew. Chent, Int. Ed. Engl. 1999, 38, 536. 34. Evans, C. C.; Sukarto, L.; Ward, M. D. J. Ant Chent Soc. 1999, 121,320. 35. (a) Haleblian, J.; McCrone, W. J. Pharnt Sci. 1969, 58, 911. (b) Grunenberg, A. Regelweiterbildimgsseminar Pharmazeutische Analytik; Bonn, 1992. 36. McCrone, W. C. Physics and Chemistry of the Organic Solid State, Fox, D.; Labes, M. M.; Weissberger, A., Eds.; Wiley-Interscience: New York, 1965, p. 726. 37. (a) Dunitz, J. D.; Bernstein, J. Acc. Chent Res. 1995, 28, 193. (b) Gavezzotti, A.; Filippini, G. J. Ant Chent Soc. 1998, I 17, 12299. 38. Bemstein, J.; Sarma, J. A. R. E; Gavezzotti, A. Chent Phys. Lett. 1990, 174, 361. 39. Batten, S.; Robson, R. Angew. Chem., Int. Ed. Engl. 1998, 37, 1460. 40. Fujita, M.; Sasaki, O.; Mitsuhashi, T.; Fujita, T.; Yazaki, J.; Yamaguchi, K.; Ogura, K. Chem. Commun. 1996, 1535. 41. Manna, J.; Kuehl, C. J.; Whiteford, J. A.; Stang, E J.; Muddiman, D. C.; Hofstadler, S. A.; Smith, R. D. J. Ant Chent Soc. 1997, 119, 11611. 42. Slone, R. V.; Yoon, D. I.; Calhoun, R. M.; Hupp, J. T. J. Ant Chent Soc. 1995, 117, 11813. 43. (a) Colacio, E.; Dominguez-Vera, J. M.; Ghazi, M.; Kivekas, R.; Klinga, M.; Moreno, J. M. Eur. J. Inorg. Chent 1999, 441. (b) Stocker, E B.; Troester, M. A.; Britton, D. lnorg. Chent 1996, 35, 3145. 44. (a) Withersby, M. A.; Blake, A. J.; Champness, N. R.; Hubberstey, E; Li, W.-S.; Schroder, M. Angew. Chem. Int. Ed. Engl. 1997, 36, 2327. (b) Biradha, K.; Seward, C.; Zaworotko, M. J. Angew. Chent, Int. Ed. Engl. 1999, 38, 492. 45. (a) Fujita, M.; Kwon, Y. J.; Sasaki, O.; Yamaguchi, K.; Ogura, K. J. Ant Chent Soc. 1995, 117, 7287. (b) Losier, E; Zaworotko, M. J. Angew. Chent, Int. Ed. Engl. 1996, 35, 2779. (c) Blake, A. J.; Champness, N. R.; Khlobystov, A.; Lemenovkii, D. A.; Li, W.-S.; Schroder, M. Chent Commun. 1997, 2027. (d) Venkataraman, D.; Lee, S.; Moore, J. S.; Zhang, E; Hirsch, K. A.; Gardner, G. B.; Covey, A. C.; Prentice, C. L. Chent Mater. 1996, 8, 2030. (e) Yaghi, O. M.; Li, H.; Groy, T. L.

282

46.

47. 48.

49.

50. 51. 52. 53. 54. 55.

56.

57. 58. 59. 60.

BRIAN MOULTON and MICHAEL J. ZAWOROTKO Inorg. Chem. 1997, 36, 5090. (f) Carlucci, L.; Ciani, G.; Proserpio, D. M. Chem` Commun. 1999, 449. (a) Fujita, M.; Kwon, Y. J.; Washizu, S.; Ogura, K. J. Am. Chem. Soc. 1994,116, 1151. (b) Carlucci, L.; Ciani, G.; Prosperio, M. New J. Chem. 1998, 1319. (c) Choi, H. J.; Suh, M. P. J. Am. Chem. Soc. 1998, 120, 10622. (a) Robinson, E; Zaworotko, M. J. Chem. Commun. 2413, 1995. Co) Yaghi, O. M.; Li, H. J. Am. Chem. Soc. 1996, 118, 295. (a) Power, K. N.; Hennigar, T. L.; Zaworotko, M. J. New J. Chem. 1998, 177. (b) Kitagawa, S.; Kondo, M. Bull. Chem. Soc. Jpn. 1998, 71, 1739. (c) Rujiwatra, A.; Kepert, C. J.; Rosseinsky, M. J. Chem. Commun. 1999, 2307. (a) Wlthersby, M. A.; Blake, A. J.; Champness, N. R.; Cooke, P. A.; Hubberstey, P.; Schroder, M. New J. Chem. 1999, 23, 573. (b) Dong, Y.-B.; Layland, R. C.; Pschirer, N. G.; Smith, M. D.; Bunz, U. H. E; zur Loye, H.-C. Chem. Mater. 1999, 11, 1413. (c) Masse, R.; Nicoud, J. E; BagieuBeucher, M.; Bourgogne, C. Chem. Phys. 1999, 245, 365. Gudbjartson, H.; Biradha, K.; Poirier, K. M.; Zaworotko, M. J. J. Am. Chem. Soc. 1999,121, 2599. Aakeroy, C. B.; Nieuwenhuyzen, M.; Price, S. L. J. Am. Chem` Soc. 1998, 120, 8986. Atencio, R.; Biradha, K.; Hennigar, T. L.; Poirier, K. M.; Power, K. N.; Seward, C. M.; White, N. S.; Zaworotko, M. J. Cryst. Eng. 1998, 1,203-212. (a) Park, S. H.; Kim, K. M.; Lee, S.; Jung, O. S. Bull. Kor. Chem` Soc. 1998,19, 79. (b) Zaworotko, M. J., unpublished results. Sharma, C. V. K.; Zaworotko, M. J. Chem. Commun. 1996, 2655. (a) Bowyer, P. K.; Porter, K. A.; Rae, A. D.; Willis, A. C.; Wild, S. B. J. Chem. Soc., Chem. Commun. 1998, 1153. (b) Rowan, A. E.; Nolte, R. J. M. Angew. Chem., Int. Ed. Engl. 1998, 37, 63. (c) Zaworotko, M. J. Angew. Chem., Int. Ed. Engl. 1998, 37, 1211. (d) Masciocchi, N.; Ardizzoia, G. A.; LaMonica, G.; Maspero, A.; Sironi, A. Angew. Chem., Int. Ed. Engl. 1998, 37, 3366. (e) Wu, B., Zhang, W.-J.; Yu, S.-Y.; Wu, X.-T. J. Chem. Soc., Dalton Trans. 1997, 1795. (f) Piquet, C.; Bemardinelli, G.; Hopfgartner, G. Chem. Rev. 1997, 97, 2005. (g) Moore, J. S. Curt. Opin. Solid State Mater. Sci. 1996, 1,777. (h) Williams, A. E Chem. Eur. J. 1997, 3, 15. (i) Constable, E. C. Tetrahedron 1992, 48, 10013. (j) Biradha, K.; Seward, C.; Zaworotko, M. J. Angew. Chem., Int. Ed. Engl. 1999, 38, 492. (k) Soghomonian, V.; Chen, Q.; Haushalter, R. C.; Zubieta, J.; O'Connor, C. J. Science 1993, 259, 1596. (1) Gelling, O. J.; van Bolhuis, E; Feringa, B. L. Chem. Commun. 1991, 917. (m) Dai, Y.; Katz, T. J.; Nichols, D. A. Angew. Chem`, Int. Ed. Engl. 1996, 35, 2109. (n) Kaes, C.; Hosseini, M. W.; Rickard, C. E. E; Skelton, B. W.; White, A. H. Angew. Chem., Int. Ed. Engl. 1998, 37, 920. (o) Fleming, J. S.; Mann, K. L. V.; Couchman, S. M.; Jeffrey, J. C.; McLeverty, J. A.; Ward, M. D. J. Chem. Soc., Dalton Trans. 1998, 2047. (p) Endo, K. The 62nd Okazaki Conference: Structural Hierarchy in Molecular Science From Nano- and Mesostructures to Macrostructures; Okazaki, Japan, 1999 (Abstract L6). (a) Etter, M. C.; Huang, K.-S. Chem. Mater 1992, 4, 272. (b) Panunto, T. W.; UrbanczykLipkowska, Z.; Johnson, R.; Etter, M. C. J. Am. Chem. Soc. 1987,109, 7786. (c) Pecaut, J.; LeFur, Y.; Masse, R.Acta. Crystallogr. 1993, B49, 535. (d) Frankenbach, G. M.; Etter, M. C. Chem` Mater. 1992, 4, 272. (e) Leiserowitz, L. Acta. Crystallogr. 1976, B32, 775. (f) Gorbitz, C. M.; Etter, M. C. J. Chem. Soc., Perkin Trans. 1992, 2, 131. (g) Melendez, R.; Robinson, E; Zaworotko, M. J. Supramolecular Chem` 1996, 7, 275-293. (h) Pepinsky, R.; Vedam, K.; Hoshino, S.; Okaya, Y. Phys. Rev. 1958, 111, 1508. (i) Pepinsky, R.; Vedam, K. Phys. Rev. 1960,117, 1502. (j) Payan, E; Haser, R. Acta~ Crystallogr. 1976, B32, 1875. MacGillivray, L. R.; Groeneman, R. H.; Atwood, J. L. J. Am` Chem` Soc. 1998, 120, 2676. Ezuhara, T.; Endo, K.; Aoyama, Y. J. Am. Chem` Soc. 1999, 121, 3279-3283. Ranganathan, A.; Pedireddi, V. R.; Chatterjee, S.; Rao, C. N. R. J. Mater. Chem. 1999, 9, 2407. (a) Amato, I. Science 1993, 259, 1538. (b) Brennecke, J. F. Nature 1997, 389, 333. (c) Horton, B. Nature, 1999, 400, 797.

Supramolecular Synthesis of Solids

283

61. (a) Biradha, K.; Zaworotko, M. J. Cryst. Eng. 1998,1, 67. (b) Biradha, K.; Dennis, D.; MacKinnon, V. A.; Sharma, C. V. K.; Zaworotko, M. J. J. Am. Chem. Soc. 1998, 120, 11894. 62. Sharma, C. V. K.; Bauer, C. B.; Rogers, R. D.; Zaworotko, M. J. Chem. Commun. 1997, 1559. 63. (a) Groenman, R. H.; MacGillivray, L. R.; Atwood, J. L. Chent Commun. 1998, 2735. (b) Batten, S. R.; Hoskins, B. E; Robson, R. New J. Chent 1998, 22, 173. (c) Abrahams, B. F.; Hardie, M. J.; Hoskins, B. E; Robson, R.; Sutherland, E. E. Chem. Commun. 1994, 1049. 64. (a) Lu, J.; Paliwala, T.; Lim, S. C.; Yu, C.; Niu, T.; Jacobsen, A. J. Inorg. Chent 1997, 36, 923. (b) Subramanian, S.; Zaworotko, M. J. Angew. Chem., Int. Ed. Engl. 1995, 35, 2127. (c) Hagrman, D.; Hammond, R. P.; Haushalter, R.; Zubieta, J. Chem. Mater. 1998, 10, 2091. 65. Biradha, K.; Dennis, D.; MacKinnon, V. A.; Seward, C.; Zaworotko, M. J. Current Challenges on Large Supramolecular Assemblies; Tsoucaris, G., Ed; Kluwer: Dordrecht, 1999. 66. Macgillivray, L. R.; Subramanian, S.; Zaworotko, M. J. Chent Commun. 1994, 1325. 67. (a) Simard, M.; Su, D.; Wuest, J. D. J. Ant Chent Soc. 1991, 113, 4696. (b) Boucher, E.; Simard, M.; Wuest, J. D. J. Org. Chem. 1995, 60, 1408. 68. (a) Ward, D. L.; Luehrs, D. C. Acta Crystallogr. Sect. C. 1983, 39, 1370. (b) Robl, C.; Hentschel, S. Z Naturforsch. B. 1991, 46, 1188. (c) Jaber, E; Charbonnier, E; Faure, R. J. Chent Crystallogr. 1997, 27, 397. (d) Bodgie, S. G.; Rogers, R. D.; Blackstock, S. C. Chent Commun. 1997, 1669. 69. Otieno, T.; Rettig, S. J.; Thompson, R. C.; Trotter, J. Inorg. Chent 1993, 32, 1607. 70. Carlucci, L.; Ciani, G.; Proserpio, D. M.; Sironi, A. Chem. Commun. 1994, 2755. 71. (a) Hirsch, K. A.; Venkataraman, D.; Wilson, S. R.; Moore, J. S.; Lee, S. Chent Commun. 1995, 2199. (b) Lopez, S.; Kahraman, M.; Harmata, M.; Keller, S. W. Inorg. Chent 1997, 36, 6138. 72. (a) Yaghi, O. M.; Li, G. Angew. Chem., Int. Ed. Engl. 1995, 34, 207. (b) Sinzer, K.; Hunig, S.; Jopp, M.; Bauer, D.; Bietsch, W.; von Shultz, J. U.; Wolf, H. C.; Kremer, R.; Metenthin, T.; Bau, R.; Khan, S. I.; Lindbaum, A.; Lengauer, C. L.; Tillmanns, E. J. Ant Chent Soc. 1993, 115, 7696. 73. Heggin, J. E Nature 1936, 137, 577. 74. Carlucci, L.; Ciani, G.; Prosperpio, D. M.; Sironi, A. Angew. Chem., Int. Ed. Engl. 1995, 34, 1895. 75. Russell, V. A.; Evans, C. C.; Li, W. J.; Ward, M. D. Science 1997, 276, 575. 76. Zaworotko, M. J., unpublished results. 77. Atencio, R.; Domasevitch, K. V.; Zaworotko, M. J. Cryst. Eng., in press. 78. Cheng, C. M.; Egbe, M. I.; Grasshoff, J. M.; Guarrera, D. J.; Pai, R. P.; Warner, J. C.; Taylor, L. D. J. Polynt Sci. Pol. Chent 1995, 33, 2515. 79. Ranford, J. D.; Vittal, J. J.; Wu, D. Angew. Chent, Int. Ed. Engl. 1998, 37, 1114-1116. 80. (a) Hoss, R.; Konig, O.; Kramer-Hoss, V.; Berger, U.; Rogin, P.; Hulliger, J. Angew. Chent, Int. Ed. Engl. 1996, 35, 1664-1666. (b) Konig, O.; Burgi, H.-B.; Armbruster, T.; Hulliger, J.; Weber, T. J. Am. Chent Soc. 1997, 119, 10632-10640. (c) Hulliger, J.; Langley, P. J.; Konig, O.; Roth, S. W.; Quintel, A.; Rechsteiner, P. Pure Appl. Opt. 1998, 7, 221-227. 81. Sra, A. K.; Andruh, M.; Kahn, O.; Golhen, S.; Ouahab, L.; Yakhmi, J. V. Angew. Chem., Int. Ed. Engl. 1999, 38, 2606-2609. 82. Quast, H.; Ross, K. H.; Spiegel, E.; Peters, K.; Schnering, H. G. V. Angew. Chent, Int. Ed. Engl. 1977, 16, 177.

This Page Intentionally Left Blank

INDEX

partially O-methylated calix[4]arenes, energy barriers for, 188-190 resorcarenes, 190-195 inherently chiral (C1), 166-184 analogs incorporating different bridges, 180-183 calix[4]arenes, 166-176 calix[5]arenes, 176-177 calix[6]arenes, 177-179 calix[8]arenes, 179-180 and intrinsically chiral, difference between, 166 resorcarenes, 183-184 survey, 169 introduction, 140-142 carcerand, 142 hemicarcerand, 142 properties and modifications of, 141-142 rccc isomer, 142 resorcinol and resorcarenes, 140 scope and limits, 140-141 as "third generation" of macrocyclic molecules, 140 stereogenic centers within macrocycle, 159-166 bridges, 159-160 macrocycles, other, similar to calixarenes, 164-166 spherand type calixarenes, 161-1 64

"C-linked calixsugars," 155 (see also Calixarenes) Calixarenes and calixarene assemblies, chirality in, 139-233 chiral conformations, 195-203 directionality in hydrogen bonded systems, 197-199 other, 199-203 parent calixarenes and resorcarenes, 196-197 "twistomers," 201,202 chiral derivatives, 143-158 biological relevant, 153-158 "C-linked calixsugars," 155 calixarenes substituted at narrow rim, 145-148 calixarenes substituted at wide rim, 143-145 calixpeptides, 153-154 calixsugars, 154-158 caviteins, 153-154 of resorcarenes, 148-152 conclusions and outlook, 222-223 enantioselective catalysts in future, 222 self-assembly processes in future, 222-223 water-soluble calixarene derivatives, 222 with higher symmetry (Cn or Dn), 184-195 calixarenes, 184-188 285

286

spirodienones and similar derivatives, 160-161 supramolecular chirality, 203-222 CD spectroscopy, 203-204, 212 chiral assemblies in crystalline state, 217-222 Cotton effect, 204-205 host-guest induced, 203-205 "induced circular dichroism" 204 "induced optical activity," 204 melamine-barbiturate systems, 213-217 "rosettes" (hydrogen bonded hexamers), 213-214 self-assembly, 217-223 tetraurea dimers, 206-213 Calixpeptides, 153-154 Calixsugars, 154-158 Cambridge Crystallographic Data Centre, 239 Cambridge Structural Database (CSD), 239, 246 Carbohydrates,calixarenes and, 154-158 (see also Calixarenes) Carcerand, 142 Caviteins, 153-154 CD spectroscopy, 203-204, 212 Chromoionphore, 112 Complexed PHBs, 53 (see also Ion recognition) Condensed phosphates, 50-53 (see also Ion recognition) Coordination chemistry, 238 Cotton effect, 204-205 Cristobalite, 242 Crystal engineering, 235-283 (see also Solids) Desorption Chemical Ionization Mass Spectrometry, 183 Dynamic light scattering (DLS), 158 Freeze-fracture electron microscopy, 66, 67 Gating, 3--4, 5 Goldman-Hodgkin-Katz equation, 10

INDEX

Green chemistry, 237, 278 Hemicarcerand, 142 Inorganic polyphosphates, ion recognition and transport by poly-(R)-3hydroxybutanoates and, 49-98 (see also Ion recognition) Ion channels, natural and artificial, supramolecular assemblies in, 1-47 artificial (non-peptide) ion channel models, classification of, 15-44 "barrel-stave" model, 17 examples, early, 15-16 functional modes, two, 15 hydrogen-bonded chain (HBC) mechanism, 18 ion channel forming polymers, 41-42 macrocycle based mimics, 18-33 macrocyclic peptides, 33-36 patch-clamp technique 28-29, 31 pharmaceutical and drug-delivery systems, application to, 15 planar lipid bilayer method, 28-30, 34-36, 39 potassium channel selectivity, modeling, 42-44 proton channels, 16-18 redox-active centers, introducing, 25-28 squalamine, 40-41 synthetic channels via aggregate formation, 36-41 tris-macrocycles, 22-25 conclusion, 44 engineered natural ion channels, 13-14 alamethicin, 13-14 introduction, 2-3 hydrophobic barrier, 2 natural ion channel proteins, 3-8 differentiation between K§ and Na§ 6 functions of ion channels, 3 gating, 3-4, 5 gramicidin as model for ion channel conduction, 8, 13-14 KscA K+ channel, 6-7

Index

model, proposed general, 5 potassium selectivity, 4-8 selectivity filter, 3-4, 6-8 voltage-activated sodium channel (VASC), 4 non-peptide channel models, 15-44 (see also.., artificial) single channel measurements, 8-11 analysis of date, 9 Goldman-Hodgkin-Katz equation, 10 ion selectivity, determining, 10 patch-clamping, 10-11 planar lipid bilayer method, 9-10, 23-24 synthetic peptide ion channel models, 11-12 template-assembled synthetic proteins (TASP), 12 Ion recognition and transport by poly-(R)3-hydroxybutanoates and inorganic polyphosphates, 49-98 complexes of PHB and PolyP, 63-79 in bacterial membranes, 63-66 block of ion channels, 76 as DNA channels? 90-93 E. coli complexes as ion channels, 66-68, 70, 78 gating of ion channels, 76-79 lanthanum opn, 76 NPN, 63-66, 91 selectivity of PHB/polyP ion channels, 72-76 synthetic ion channels from OHBt9/23 and polyPs, 70-72 synthetic ion channels from PHB 128 and polyPs, 68-70 voltage dependence of ion channels, 79 evolutionary aspects and concluding remarks, 93-94 inorganic polyphosphates (PolyPs), 50-53 biological functions proposed for, 52 condensed phosphates, 50-53 flexible polyelectrolytes, 53 ion selection and transport, 52

287

occurrrence and synthesis, 50-52 orthophosphates, 52 polyphosphate kinase (PPK), 52 residues, chain of, 51 introduction, 50 PHBs, 50 polyPs, 50 poly-(R)-3-hydroxybutyrates (PHBs), 53-63 amphiphilic and solvent properties, 55-58 backbone structure of salt-solvating polymers, 57 channels in planar lipid bilayers, 58-63 "dissolving" salts, capacity for, 56, 57 occurrence and synthesis, 53-55 OHBs, 58-61 PHAs as homologues of, 54-55 cPHBs, 53-55 polyethylene oxide, 56-57 polymer electrolytes, 56, 57-58 residues, chain of, 54 protein, PHB, and polyP, supramolecular complexes of, 83-90 in CaATPase pump, 83-85 in potassium channel, 85-90 structure and mechanism of ion transport, 79-83 Reusch model, 82, 83, 92 Seebach model, 82 Metal ions, functionalized macrocyclic ligands as sensory molecules for, 99-137 chemosensors for alkali metal cations, 101-112 chromoionphore, 109-112 coordination-induced ligand rigidification, 102 fluorescent sensors, 101-108 Li§ and Li § ionophores, 106-107 photoinduced electron transfer (PET) system, 105 spin-orbit coupling effects, 102

288

spirobenzopyrans, crowned, 108-112 chemosensors for alkaline earth metal cations, 112-120 azacrown ethers functionalized with 8-hydroxyquinoline and derivatives, 117-121 cryptand and bibrachial lariat type crowned spirobenzopyrans, 114-117 fluorescent, 112-113 fluorophore-spacer-receptor system, 112 chemosensors for transition metal ions, 121-132 fluorescent, 121-127 macrocycles functionalized with dansyl chromophore as zinc (II) chemosensors, 127 macrocycles functionalized with 8-hydroxyquinoline and derivatives, 128-132 conclusion, 132 introduction, 100-101 chemosensor, schematic, 101 sensor efficiency, factors in, 100 OHBs, 58-61 Patch-clamping, 10-1 l, 28-29, 31 Pauling, Linus, 238 PHAs, 54-55 (see also Ion recognition) PHBs, 50, 53-63 (see also Ion recognition) complexes of, 63-79 cPHBs, 53-55 Photoinduced electron transfer (PET) system, 105, 121-122 Pirkle's reagent, 172,176, 179, 180 Planar lipid bilayer technique, 9-10, 23-24, 28-30, 34-36, 39 Poly-(R)-3-hydroxybutanoates, ion recognition and transport by, 49-98 (see also Ion recognition) Polyethylene oxide, 56-57 Polymer electrolytes, 56 (see also Ion recognition)

INDEX

PolyPs, 50-53 (see also Ion recognition) complexes of, 63-79 Polyphosphate kinase (PPK), 52 (see also Ion recognition) Resorcarenes, 140, 148-152 (see also Calixarenes) Selectivity filter, 3-4, 6-8 Solids, supramolecular synthesis of, 235-283 applications, potential, 278-279 conclusion and future directions, 279-280 introduction, 236-241 Cambridge Crystallographic Data Centre, 239 Cambridge Structural Database, 239, 246 coordination polymers, 240-241 crystal engineering, birth of, 236-237 crystal engineering, conception of, 238-240 crystal engineering today, 237-238 green chemistry,237 hydrogen bonding, 239 "node and spacer" approach, 240-241 supramolecular chemistry, definition, 237 supramolecular isomers, 237, 245-253 supramolecular synthons, 237,239 suprasupermolecular chemistry, 238 modular approach to building crystals, 242-245 cristobalite, 242 diamondoid networks, 242-245 modular self-assembly, 242-245 molecular tectonics, 242 self-assembly, 242 supramolecular isomerism, 237, 245-253 catenane, 246, 250-251 classes of, four, 246 conformational, 246, 248-250 definition, 245

Index

importance of, four reasons for, 245-246 optical, 246, 251-253 structural, 246-248 of 2D structures, 253-267 bilayer architecture, 266-267 brick architecture, 266, 267 herringbone architecture, 266, 267 hydrogen bonded 2D networks, 254-258 metal-organic networks, 263-267 modular self-assembly, 257 networks sustained by organic ions, 258-263 "organic clays" formation of, 258-259 potential of, 253-254 square grid networks, 264-265 T-shaped geometry, 265-266 trimesic acid, structure of, 255-256, 262, 269 of 3D structures, 267-277 cross-linking of 2D structures, 275 diamondoid networks, 268-269, 270-272 hybrid structures, 274-277 hydrogen-bonded networks, 268-270 interpenetration of identical/different networks, 276-277 iron cyano compounds, 272-274

289

"Lincoln log" type structure, 274 metal-organic networks, 270-274 "pillaring" 275 octahedral networks, 272-274 T-shaped nodes, self-assembly of, 274 tectons, 269 Spirodienones, 160-161 (see also Calixarenes) Squalamine, 40-41 Supramolecular chemistry, definition, 237 Supramolecular synthesis of solids, 235-283 (see also Solids) TASP, 12 Tectons, 269 Template-assembled synthetic proteins, (TASP), 12 Transmission electron microscopy (TEM), 158 Trimesic acid, structure of, 255-256, 262, 269 "Twistomers;' 201,202 VASC, 4 Vitalism, 236 Voltage-activated sodium channel (VASC), 4 W6hler, Friedrich, 236