Strategies and Tactics in Organic Synthesis, Volume 7 (Strategies and Tactics in Organic Synthesis)

Dedication This volume is dedicated to Professor Paul A. Wender on the occasion of his 60th birthday and in recognitio...

Author: M. Harmata

169 downloads 1532 Views 30MB Size Report

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

Report copyright / DMCA form

DOWNLOAD PDF

Dedication

This volume is dedicated to Professor Paul A. Wender on the occasion of his 60th birthday and in recognition of the beautiful chemistry he has created.

CONTRIBUTORS

JEFFREY AUBI~, Department of Medicinal Chemistry, University of Kansas, 1251 WescoeHall Drive, Lawrence, KS 66045- 7852 VERONIQUE BELLOSTA, Laboratoire de Chimie Organique, ESPCI, CNRS 10 rue Vauquelin, 75231 Paris Cedex 05, France KAYM. BRUMMOND, Department of Chemistry, University of Pittsburgh, Pittsburgh, PA 15260 CAMERON M. BURNETT, Department of Chemistry, Colorado State University, Fort Collins, CO 80523 JANINE COSSY, Laboratoire de Chimie Organique, ESPCI, CNRS 10 rue Vauquelin, 75231 Paris Cedex 05, France SIMON R. CRABTREE, Research School of Chemistry, Australian National University, Canberra, Australia 0200 XING DAI, Department of Chemistry, University at Buffalo, The State University of New York, Buffalo, NY 14260 HUW M. L. DAVIES, Department of Chemistry, University at Buffalo, The State University of New York, Buffalo, NY 14260 SCOTT E. DENMARK, Department of Chemistry, University of lUinois, at Urbana-Champaign, Urbana, IL 61801 KELLY A. FAIRWEATHER, Research School of Chemistry, Australian National University, Canberra, Australia 0200 KEVIN J. FRANKOWSKI, Department of Medicinal Chemistry, University of Kansas, 1251 WescoeHall Drive, Lawrence, KS 66045-7852 SHINJI FUJIMORI, Department of Chemistry, University of Illinois, at Urbana-Champaign, Urbana, IL 61801

XV

xvi

CONTRIBUTORS

BAUDOUIN GERARD, Department of Chemistry, Boston University and

Center for Chemical Methodology and Library Development, Boston, MA 02215 FRANK-GERRIT ~ E R , Institut fiir Organische Chemie, Fachbereich Chemie, Universitiit Duisburg-Essen, Essen, Germany MIREIA CAMPANA KUCHENBRANDT, Institut fiir Organische Chemie, Fachbereich Chemie, Universitiit Duisburg-Essen, Essen, Germany LEONARD R. MACGILL1VRAY, Department of Chemistry, University of Iowa, Iowa City, IA 52242 LEWIS N. MANDER, Research School of Chemistry, Australian National University, Canberra, Australia 0200 BRANKO MITASEV, Department of Chemistry, University of Pittsburgh, Pittsburgh, PA 15260 THOMAS PETTUS, Department of Chemistry and Biochemistry, University of California at Santa Barbara, Santa Barbara, CA 93106-9510 JOHN A. PORCO, Jr., Department of Chemistry, Boston University and Center for Chemical Methodology and Library Development, Boston, MA 02215 JON D. RAINIER, Department of Chemistry, University of Utah, Salt Lake City, UT 84112 CATHERINE TAILLIER, Laboratoire de Chimie Organique, ESPCI, CNRS 10 rue Vauquelin, 75231 Paris Cedex 05, France TODD WENDERSKI, Department of Chemistry and Biochemistry, University of California at Santa Barbara, Santa Barbara, CA 93106-9510 DAVID R. WILLIAMS, Department of Chemistry, Indiana University, Bloomington, IN 4 7405 ROBERT M. WILLIAMS, Department of Chemistry, Colorado State University, Fort Collins, CO 80523 AARON WROBLESKI, Department of Medicinal Chemistry, University of Kansas, 1251 WescoeHall Drive, Lawrence, KS 66045- 7852

Preface

I believe I first saw Paul Wender speak at the University of Illinois-Champaign/Urbana on December 8, 1981. Having an intense love affair with organic chemistry at the time, I recall being utterly thrilled by the chemistry I saw. Photochemistry to do a [5+2]-cycloaddition, taking simple starting materials to complex structures in one step!! It remains very powerful chemistry. Though not an easy thing to do, with the help of Scott Denmark and the NIH, I eventually became a postdoc in Paul's lab, and was, I believe, the first person he introduced to the neocarzinostatin chromophore. There was great science in those labs and many very talented people. It was fun. Paul was a very interesting mentor. It became clear very early that he wanted his co-workers to think and think deeply about their research. More important, he often seemed willing to "wait it out" as someone struggled with a problem he believed they could handle. I always felt as though he knew an answer, but expected us to know, learn or discover an answer too. This is a rather daring way to run a research group, but it produces real thinkers. Some things that really shocked me about Paul can be related in two stories. A few weeks into my postdoc I decided to head up to Muir Woods one weekend to see what I could see. While sitting and enjoying some ice cream, I looked up to see Paul standing right in front of me. Ouch! I should have been in lab, or so I thought. Paul happily introduced me to his companions and we went our separate ways. I expected some comments on Monday regarding my absence from lab. They never came.

xvii

xviii

PREFACE

Much later, I informed Paul that my girlfriend would be visiting me for a week and I would not be doing my 80 reactions per week as he had come to expect from me (editor's privilege !). Without hesitation, he suggested that I take the week off and do some touting of California. I accepted his offer. Paul has always impressed me as someone who has an incredibly deep interest in chemistry and science in general. So deep, in fact, that he is willing to spend a great deal of time with people talking about it. I am one of those people. His ability to see in ways that are often unique, from my point of view at least, have no doubt opened new roads, not only for me, but for others as well. Especially important from my perspective is my observation that at meetings, unless a very tight travel schedule calls him away, Paul will visit with people at posters, empty tables, you name it, and pass along insights and give encouragement, whether they are doing a complex natural product synthesis or relatively simple chemistry. He does not have to do that; he chooses to and the chemistry community at large benefits from it. I have heard physical chemists refer to synthetic organic chemists as reptiles who eat their own children. Paul is not that way. He takes great pleasure in seeing his "children" reach their highest potential and is willing to take steps to help them achieve that goal. Happy Birthday, Paul!

As the editor of this book, I have many people to thank. The authors did a spectacular job in presenting some very nice chemistry. One of the perks of being an editor is getting to enjoy all of that science. My wife, Judy Snyder, (the "girlfriend" above) helped with proofing and for that I am extremely grateful. Thanks to Joan Anuels and all at Elsevier for their encouragement. It is time to think about another volume in this series, but I also have papers and especially proposals to write to keep my own research alive. I hope to see another volume in 2009 or so. Until then, take care of your science and yourselves.

Michael Harmata

Foreword

This 7 th volume of Strategies and Tactics in Organic Synthesis presents an extraordinary range of superb chemistry in the 13 chapters that follow. It also makes clear the great human effort involved in the construction of complex targets, and teaches us how a combination of perseverance and imagination can conquer the problems that always arise as the price of creativity. For some 33 years, the combination of great art with great teaching has been the hallmark of Professor Paul Wender's chemistry. It is most fitting that this volume is dedicated to him. Gilbert Stork

Columbia University, NY May 2007

xix

STRATEGIES AND TACTICS IN ORGANIC SYNTHESIS, VOL. 7 9 2008 Elsevier Ltd. All rights reserved.

Chapter 1 TOTAL SYNTHESIS OF RK-397 Scott E. Denmark and Shinji Fujimori Department of Chemistry University of Illinois at Urbana-Champaign Urbana, IL 61801, USA

I. II. III.

Introduction and Background Retrosynthetic Analysis Synthesis of RK-397 A. Preparation of Fragment 41 B. Preparation of the Fragment 42 C. Aldol Addition of Trichlorosilyl Enolate Derived from 41 D. Synthesis of C(11)-C(33) Fragment E. Preparation of Polyene Fragment 40 IV. Completion of the Synthesis V. Summary Acknowledgements References and Footnotes

1 8 10 10 13 16 20 26 28 30 32 32

I. Introduction and Background

RK-397, a member of a large family of polyene macrolides that include amphotericins, nystatin, mycoticins, and roxaticins, was isolated and structurally characterized by Osada et al. in 1993 (Figure 1). ~ Most of these macrolides are produced by soil actinomycetes, mainly belonging to the genus Streptomyces, and are isolated by extraction. 2 The structural features include a large lactone ring (20-44-membered), three to eight conjugated double bonds and, in most cases, a sugar moiety. Because of the presence of the polyene, these compounds show strong UV absorption, and characteristic absorption maxima are observed depending on the length of polyene, allowing classification of over 200 macrolides into subcategories. This class of compounds is particularly known for their antifungal properties. Nystatin was the first polyene macrolide antibiotic to be isolated and continues to be used in antifungal therapy. 2 Amphotericin B has

SCOTT E. DENMARK AND SHINJI FUJIMORI

O~ ~

Me

HO'"L'~Meo

~

~

?H ?H

Me

~

~

O: OH _ OH NH2le

Me M2~~v-~~~v~.~,,,OH

?H ?H ? ' ~ '''CO2H Me) ....~ ~ r

-OH Amphotericin B

OH

OH

~"

_~ _

~H o ~

Me"' '-,,,~""-,~ " ~ __'-,,,~" , ~ V : OH OH OH OH OH Roxaticin

.,/k--..0.~--.-.-Me ~ O H " OH NH2

O

-

Me

~

~

HO"'L~']~Meo

~

~

~

~

?H ?H ?H

M2 ~ i ~,~,,,O

?H ?,,,~,,,CO2H

Me

-OH

OH

OH

R

Nystatin

M

E

O

W

,

,

Me...-J..... 0 M

e

,

M

~

~

~

~

,''OH i Med,~,,,OH

e ' " ~ OH OH OH OH OH OH Mycoticin A (R = H) g (R= Me) OH

L.,. ,,,OH " ' ~ OH OH OH OH OH OH RK-397

FIGURE 1. Representative polyene macrolides.

been used extensively in treatment of serious infections for the past 40 years and remains drug of choice for systemic fungal infections. 3 These antifungal agents have gained growing interest in the last decade since the occurrence of fungal infections has increased due to AIDS and immunosuppression associated with organ transplants. 4 However, prolonged parenteral administration of amphotericin B often leads to various adverse effects, especially in the kidney. 5 RK-397 possesses relatively weak antifungal activity. However, unlike amphotericin B and nystatin, it exhibits potent antibacterial activity along with some cytotoxicity against human leukemia cells. ~The mechanism of action for amphotericin B and nystatin has been studied extensively, and these antibiotics are known to associate with sterols and to form channels on cell membranes, inducing leakage of cellular contents. 6 However, due to the great difference in biological activity, RK-397 and several other polyene macrolides are believed to have different mechanisms of action. 7 The structural elucidation and configurational assignment of these macrolides initially relied on degradation studies combined with spectroscopic analysis. Amphotericin B was the first polyene macrolide to have

1

TOTALSYNTHESISOF RK-397

its three-dimensional structure established by X-ray crystallography. 8 More recently, the advances in NMR spectroscopy greatly improved its ability to analyze structures, and Rychnovsky's 9,1~ acetonide analysis is particularly useful in determining the configuration of polyols. These techniques allowed determination of molecular structure and absolute configuration of several other polyene macrolides including RK-397. Because of their potent biological activity and structural complexity, these natural products have attracted interest as targets for total synthesis by many synthetic chemists. ~~ So far, total syntheses of amphotericin B, 12 mycoticin A, 13 roxaticin, ~4 filipin III, ~5 roflamycoin, ~6 dermostatin ~7 and RK-397 ~a have been reported. The obvious challenges in the synthesis of polyene macrolides lie in the stereocontrolled assembly of the polyol chain and construction of the polyene backbone. Various approaches are available for construction of the 1,3-polyol: iterative Sharpless asymmetric epoxidation/hydride reduction, 12two-directional chain synthesis, 13aalkylation of cyanohydrin acetonides 1~ and aldol/anti (or syn) reduction sequence (Figure 2). 14d Nicolaou: SAE/reduction (+)-DET Ti(OiPr)4 O R ~ O H t-BuOOH R ~ O H CH2Cl2 1

1. Swern 2. Ph3PCHCO2Et

2

R i~/..O ""~ " ' ~ ~COOEt

R~

OH

1. DibaI-H 2. PivCl 3.TBDPSCI 4. DibaI-H 1. Red-AI 2. TBSCI

(+)-DET Ti(Oi-Pr)4 t-BuOOH

R'.I~OH OTBDPS

CH2CI2

4 D,,

FI~

OTBS

TBDPSO

OTBDPS 5

OH 6

Rychnovsky: Alkylation of cyanohydrin acetonides Me

Me

Me

O-~-O R

CN 7

Me

O,~-O + B r ~

Me R'

THF

Me Me

Me

oXo oXo

LiNEt2 "

R

~

8

FIGURE 2. Syntheticstrategies for 1,3-polyol chains.

R CN

' 9

SCOTT E. DENMARK AND SHINJI FUJIMORI Schreiber: Two-directional chain synthesis

M

e

O

1. RuCI2[(R)-BINAP ] 2. (EtO)2CHMe, H+ OMe 3. Li, NH3; then 0 3

~

~.~

oo

~ 1. RuCl2[(R)- BINAP] 2. (MeO)2CMe2, H+ 3. DibaI-H; VinylMgBr 4. (MeO)2CMe2, H+ =

10 O

M

e O ~ M e II II I II 0 0 O\ I0 0

T

Me 0 0 MUMe

II 0

11

0 0 MUMe

0 0 MUMe

12

Mori: Epoxide opening with lithiodithane

Ph Ph

Ph

Ph'J~-0

S/~

o

n-BuLi

Ls.J + 13

THF

v

v

v

-a

15

14

1. Hg(CIO4)2 Ph 2. NaBH4 Ph,,~ 3. (MeO)2CMe2 / O =- O ~

F" o, A . s x s j . "

Me

Me

O'/X~'O R

16 Evans: Aldol/carbonyl reduction

O R'~H

Bu2B0

+

O_Pg

~ R '

Et20

17 Me4NHB(OAc)3 MeCN, AcOH

OH 0 I II R~

OPg R'

18

OH OH OPg N

N' 19

FIGURE 2. (continued)

Stereoselective polyene synthesis has historically relied primarily on olefination reactions (Figure 3). Nicolaou and Moil employed conjugated phosphorus-based olefination reagents 20 and 21 to extend a polyolefin by three double bonds at a time. ~2,18 Rychnovsky employed Wollenberg's ~9 reagent 23 followed by elimination to install the dienal unit. Schreiber and

l

TOTAL SYNTHESIS OF RK-397

Nicolaou/Mori: O

R.,JQH

1.20 or 21 2. DibaI-H 3. MnO2 ,. R ~ C H O 22 20= (EtO)20P~-~/~.../~/COO Et 21 Ph3P~COOEt

Rychnovsky" O

1.23 2. MsCI, Et3N

H 23= B r M g ~ o E

t

24

Schreiber/Evans: [ ~

H .OAc Hg(OAc)2 ~ , O LiAIH4;then 02 = = O H C ~ c H O "

25

Ac

26

27

FIGURE 3. Synthetic strategies for polyene.

Evans took advantage of oxidative electrocyclic closure of cyclooctateraerie followed by reduction to provide triene dialdehyde 27. ~3 Another challenge arises from the macrocyclization in the presence of both polyene and polyol units. The most popular method to construct a macrocycle is the Yamaguchi method, which employs an activated anhydride derived from 2,4,6-trichlorobenzoyl chloride (Figure 4 ) . 20 The Yamaguchi macrolactonization works for a variety of large rings; however, the reaction sometimes requires vigorous heating to effect cyclization. An alternative method to close rings having more than 20 members is the intramolecular Horner-Wadsworth-Emmons olefination, originally developed by Stork and Nakamura. 21 The application of Roush-Masamune conditions to the macrolactonization makes this approach applicable to substrates with base-sensitive functional groups. 22 The first total synthesis of RK-397 was completed by Burova and McDonald in 2004. ~ Their approach was based on iterative coupling of

SCOTT E. DENMARK AND SHINJI FUJIMORI

Macrolactonization:

Me Me Me

1.2,4,6-trichlorobenzoylchloride 2. DMAP O

Me

--- "~

- "~

28 Me

Me./l....[/O M

O~_.J e

'

"

~

Intramolecular Olefination:

Me Me

Me

_

O LiCI, DBU

Me O

O

O O MUMe PO(OEt)2

O O MUMe 30

O O MUMe

OR

MeCN

H

Me _

A

M~

\

\

\

\

Me...J.... 6

\

~

,%~,OR

o. J

oFr

o~o

o

Me Me

Me Me

81

FIGURE 4. Macrolactone formation strategies.

six-carbon units by epoxide opening with an acetylide anion (Scheme 1). Chiral epoxide 33 was prepared by Jacobsen's hydrolytic kinetic resolution, and the attachment of acetylide anion derived from 32 furnished homopropargyl alcohol 34. The installation of 1,3-diol functionality was accomplished by hydration of the alkyne and subsequent stereoselective reduction. For the construction of polyene fragment, a Stille coupling of stannyltriene (C(9)-C(10)) was employed, and the closure of the macrocycle was accomplished by the intramolecular olefination (C(2)-C(3)) (Figure 5).

1

TOTALSYNTHESISOF RK-397

,/

intramolecular HWE ~ Stille coupling \

Me

"]

\

"

" 11

I'

Me.,P,,, 0

., OH

OH OH OH ~ O H epoxide opening

FIGURE 5. Key disconnections in McDonald's synthesis of RK-397.

H

O

O

32

Me'Me

~,~,.,,,.TMS

n-BuLi BF3-Et20

+

,.TMS

THF

0 0 Me.Me

o, -J

OPMB 33

34 1. HN(SiMe2H)2 2. Pt(DVDS) 3. H202, KF, KHCO3 4. Et2BOMe, NaBH4 ..TMS

0 0 Me'Me

OH OH OPMB 35

SCHEME 1

Another synthetic approach toward RK-397 has been reported by Schneider et al., ~ who employed a thermal oxy-Cope rearrangement of a silyloxy diene to construct a precursor for the polyol unit (Scheme 2). The diene 36 was readily prepared by auxiliary-controlled aldol addition in a stereoselective manner. The rearrangement was accomplished by heating the diene at 170 ~ and the subsequent acidic hydrolysis of silyl enol ether provided the aldehyde 37. After a few manipulations of 37, the silyl group was oxidized under conditions developed by Fleming to the corresponding diol 39. Using this method, Schneider constructed the

SCOTT E. DENMARK AND SHINJI FUJIMORI

polyol chain fragment of RK-397, however, the completed total synthesis has not yet been reported. 53 SO PhMe2Si~N 36

O O

~ t-Bu

OBn OH *~~[V[v~~OMe

1. 170~ 2. p-TsOH 70 - 75%

O

39

1. BF3(AcOH)2 ?" H202' KF ~ 72%

O SiMe2Ph O A ~~J'~~'~ N O 37

iI t-u--

OBn SiMe2Ph O O M e 38

SCHEME 2. Schneider's approach for the polyol fragment of RK-397.

The challenges of stereoselective construction of the 1,3-polyols as well as the all-E-polyene fragment in RK-397 could be efficiently addressed by ongoing methodological programs in these laboratories on the enantioselective, Lewis-base-catalyzed aldol addition 23 and siliconbased cross-coupling reactions. 24This chapter describes an efficient enantioselective total synthesis of RK-397 utilizing the methods mentioned above as key strategic steps. II. Retrosynthetic Analysis

To maximize synthetic convergence, the target was divided into four modules (Scheme 3). The disconnections at the lactone linkage and the C(10)-C(ll) bond provide a known polyene phosphonate fragment 40.13'14c'25 Obviously, the polyol fragment lends itself to myriad aldol/ reduction disconnections. 26 However, by carefully examining the pattern of stereogenic centers on the polyol chain, the most convergent synthesis might be achieved by the disconnections between the C(18)-C(19) and C(26)-C(27) bonds. This provides both the C(11)-C(18) and C(19)-C(26) fragments as an identical building block. In the forward sense, these disconnections require an aldol addition with 1,5-anti stereoinduction from the methyl ketone 41. The aldehyde functionality at C(11) and C(19) was masked as an alkenylsilane for better functional group compatibility. The remaining C(27)-C(31) fragment would be synthesized by use of Evans' chiral acyl oxazolidinone technology. 27

1

TOTAL SYNTHESIS OF RK-397

MeO 1 ~

~

~

10 ,~.~v.~,,,OH

~

OH OH OH OH OH OH

PO(OEt)2 0.~-~.~~~~~ EtO 40

10

BnMe2Si

..0

..Ph ,,',"10

41 Me Me.,,,J3,j/OPMB 26 41 Me'"~ H Me~SiMe2Bn II -42 0 0 O~v~O_

O 19

_

#h SCHEME 3

The polyene phosphonate 40 can be further subdivided into three units by sequential palladium-catalyzed cross-coupling reactions of bis-silyl diene 45 (Scheme 4). 28 The phosphonate 40 would be derived from hydroxy-ester 43 by functional group conversions at C(10), and the construction of the tetraene moiety in 43 would be accomplished through the sequential coupling of 45 with protected 3-iodo-allyl alcohol 46 and with iodoacrylate 44. The order of these two couplings can be changed depending on the reactivity of these iodides. However, ester functional groups are usually not compatible with TMSOK-promoted coupling conditions. Therefore, the coupling of 46 should be carried out with TMSOK activation, and the coupling of 44 should be performed with fluoride activation. The key building block 41 was envisioned to arise from vinylogous aldol addition of dienol ether 48. 29The aldol addition of 48 is catalyzed by a chiral dimeric phosphoramide, and the reaction should provide the 3~-addition product selectively. The conversion of 47 to 41 involves transformation of the C(25) ester to a methyl ketone and installation of the syn-diol. The disconnection at C(21)-C(22) in enoate 47 reveals dienol ether 48 and aldehyde 49 as substrates for the Lewis base-catalyzed vinylogous aldol

10

SCOTT E. DENMARK AND SHINJI FUJIMORI

addition. The fragment 42 would be derived from known aldehyde 5030 by olefination with methyl triphenylphosphanylideneacetate 51. The stereogenic centers at C(30) and C(31) would be established by Evans aldol addition with isobutyraldehyde.

PO(OEt)2 10

0~'~~~~~ EtO

OR

.

40

EtO

0.~-~ I EtO 44 1

43 OR SiMe20HI-~-./J 10

BnMe2Si~ 45

46

Ph

0 0"~0 19 Me'JL~[~~~~SiMe2Bn I 26 41

0 OH -21 19 EtO~'~SiMe2Bn 22

~

47

TBSO EtO~

0 22 H21~~SiMe2Bn

48

Me Me...J;~,.OPMB Ue,,,~~.~H i 0 42

PMBO

0

~ Me.~Me~Me 50

49

0 Ph3P~-~~OMe 51

SCHEME 4 III.

Synthesis of R K - 3 9 7

A. P R E P A R A T I O N OF F R A G M E N T 41

This synthesis plan was initiated with the preparation of key building block 41 (Scheme 5). Stereoselective Red-A1| reduction of 3-benzyldimethylsilylpropargyl alcohol 5231 provided allyl alcohol (E)-53, 32 which was oxidized to the 3-silyl-2-propenal 49 in good yield (61%, two steps).

1

11

TOTAL SYNTHESIS OF RK-397

BnMe2Si~'~-.~/OH

Red-AI Et20 " BnMe2Si~--"~~OH

52

74%

DMSO, (COCl)2 O Et3N 0H2012 ~" BnMe2SiA~.~H

53

83%

49

SCHEME 5

The vinylogous aldol addition was initially attempted with dioxanonederived dienol ether 54 (Scheme 6). The aldol addition of 54 was studied previously under catalysis with bisphosphoramide (R,R)-56, and the reaction proceeded smoothly with a broad range of aldehydes, including aliphatic aldehydes, and afforded good-to-excellent enantioselectivities. 29 The aldol addition of 54 to [3-trimethylsilyl acrolein 55 provided a good yield of the aldol product, however, the enantioselectivity was only modest. The additions to 3-alkoxypropanals, 58, were also considered as alternative approach to 41. However, additions to these aldehydes did not provide a synthetically useful enantioselectivity.

Me Me O,/~O

O + TMS ~ ' / " ' ~ ~ H 55

~OTBS 54 Me

(R,R)-56 (5 mol %) SiCl4

Me

O./~.O

OR O

Me Me OH O/~/~'O

0H2012

T

(R,R)-56(5 mol %) SiCI4, TBAI (20 mol %)

M

S ~ O 57: 63%, er 82/18 Me Me OH O/~/~'O

+

0H2012 54

R = Bn 58b: R = TBS 58c: R = TIPS

59a: 79%, er 64/36 (R = Bn) 59b: 58%, er 73/23 (R = TBS) 59e: 86%, er 85/15 (R = TIPS)

58a:

(R,R)-56 = ~k]~ ~ _ ~

... CH2 N' i 'MeMe / 2 SCHEME 6

Therefore, the crotonate-derived silyl ketene acetal was chosen as an alternative nucleophile (Scheme 7). Although the additions of 48 to 58

12

SCOTT E. DENMARK AND SHINJI FUJIMORI

or to 61 were low-yielding and unselective, the vinylogous aldol reaction with 49 (using chiral bisphosphoramide (R,R)-56) efficiently provided 47 in good yield with excellent ?-selectivity and enantioselectivity (75%, er 98/2). The beneficial effect of a silyl group at the [3-position was illustrated by the successful aldol additions of 48 to 55 and 63, whereas reactions with acrolein resulted in polymerization of the aldehyde. OTBS ~ o e t 48

OR O +

[''~

H

Bn 58b: R = TBS 58a: R =

OTBS ~ o e t 48

(R,R)-56 (5 mol %) SiCl4

O +

~"~"/~ H 61

OTBS ~ o e t 48

(R,R)-56 (5 mol %) OH O SiCI4,TBAI(20moI%) RO/~ -- v " ~ v / ~ -,4,,~ O- E t -" 0H2012 60a: 32%, er 85/15 (R = Bn) 60b: 44%, er 81/19 (R = TBS)

0H2012

O + BnMe2Si~ - ~ ' H 49

-"

(R,R)-56 (1.5 mol %) SiCI4 CH2C!2

"~'~~oet 48

O +

R3Si~ H 55: R = TMS 63: R = SiMe2Ph

SiCl4 0H2012

O

~~~~'~----------------------~~oet 62: not observed

OH

O

= BnMe2Si~'~'~~~----------------------~~oet 47, 75%, er 98/2

(R,R)-56 (5 mol %) OTBS

OH

OH

0

R 3 S i ~ O E t 64: 86%, er 97/3 (R = TMS) 65: 62%, er 98/2 (R = SiMe2Ph)

SCHEME 7

Although the additions to other 13-silyl enals also showed excellent enantioselectivities, we chose to proceed with compound 49 because the benzyldimethylsilyl group can be oxidized under very mild conditions. The vinylogous aldol addition could be performed on a 50-mmol (9.5 g of 49) scale without loss in yield or selectivity. The protected syn-diol functionality was installed using the tandem alkoxide addition/conjugate addition protocol developed by Evans and Gauchet-Prunet 33 (Scheme 8). This transformation proceeded with excellent diastereoselectivity to provide the benzylidene-protected diol 66 in good yield. The complete consumption of the starting material

1

13

TOTALSYNTHESIS OF RK-397

with minimal formation of the side products was achieved by maintaining the reaction temperature between 0 and 5 ~ Carrying out the reaction at lower temperature (-20 ~ yielded a significant amount of the aldol product between the resulting ester enolate and excess benzaldehyde. Ph OH

PhCHO KHMDS

O

BnMe2Si~

OEt

THF 74%

47

THF 87%

O

BnMe2Si~OEt 66: dr >19/1

Ph

HN(OMe)Me

/-PrMgCI

O~O

,.

MeMgBr

O~O O gnMe2Si~N..OMe 67

Me

Et20 90%

Ph 0~0

BnMe2Si~

0

Me 41 SCHEME 8

Finally, construction of the key building block 41 was completed by conversion of the ester functional group into a methyl ketone via Weinreb's amide in excellent yield. 34 First, the formation of the Weinreb amide from 66 was accomplished by using isopropylmagnesium chloride as a base. The controlled addition of methylmagnesium bromide at room temperature provided the desired methyl ketone without the formation of the over-addition product. B. PREPARATION OF THE F R A G M E N T 42

Several different routes employing Lewis base catalysis to construct the stereogenic centers in 42 were attempted (Scheme 9). The first approach involved the catalytic, asymmetric addition of a Z-crotylsilane to establish the two stereogenic centers in one transformation (path b, Scheme 9). The addition of (Z)-2-butenyltrichlorosilane catalyzed by

14

SCOTT E. DENMARK AND SHINJI FUJIMORI

chiral dimeric phosphoramides provides excellent diastereoselectivity and good enantioselectivity for aromatic aldehydes and alkenyl aldehydes. 35 Thus, it was envisioned that the addition of (Z)-71 to methacrolein would give the syn-crotylation product. Then, the selective reduction of the C(32)-C(33) double bond would provide 70, followed by the oxidative cleavage of the terminal methylene group to provide the desired aldehyde 50. However, the addition of (Z)-71 to methacrolein catalyzed by 74 only produced a trace amount of the crotylation product (Scheme 10). The remainder of the reaction mixture showed a small amount of the unreacted aldehyde and unidentifiable materials.

Me Me/J-..~.,,OPMB 42

a

Me Me@OH

O

68

PMBO O M e ~ H Me Me 50

Me

OTBS

Me~ O

~Ot-Bu Me 69

O

I

~ Me

Me Me 70

H Me/~~--SiCI3 33 (Z)-71

OH 0 Me Me 72

H

R (Z)-73

SCHEME 9

O Me ? H

74 (10 mol %) DIPEA

Me'~...

Me

SiCl3 CH2CI2,-78 ~

Me

Z-71 74=

H-~N //O

~

Ue/ SCHEME 10

OH - ~

75:trace

1

15

TOTALSYNTHESISOF RK-397

The second approach employed the aldol addition of a trichlorosilyl enolate to set the two stereogenic centers (path c, Scheme 9). The reaction of a Z-trichlorosilyl enolate under phosphoramide catalysis should provide the desired syn-aldol product. Three trichlorosilyl enolates 73 derived from propiophenone, 2,2-dimethyl-3-pentanone and propanal were prepared, and the aldol additions of these enolates to methacrolein were investigated (Scheme 11). The additions of the ketone-derived enolates were catalyzed by 76, and the addition of 73c was catalyzed by dimeric phosphoramide 56. Aldol products were obtained in modest yields, but the diastereoselectivities were very low for 77 and 79. The addition of 73b was syn (relative) selective, and the enantiomeric ratio of the corresponding benzoate was only 1.5/1.

Me~SiCI3~ Ph

(R,R)-76 (10 mol %)

O +

Me

H

CH2CI2

73a

Ph

-78 ~

OSiCI3 Me~t_B 73b

OH O

u

O +

-~ Me

77: 51%, syn/anti = 1/1.3

(R,R)-76 (10 mol %) H

CH2CI2 -78 ~

,-

OH O M e ~

t-Bu

78: 34%, syn/anti = 10/1 Me Ph,~1~ +O (R,R)-76 = ip. a..~....., Ph'

N,Me~

1. (S,S)-56 (10 mol %)

Me

OSiCI3 O 2. MeOH H Me? H CH2CI2 73c -78 ~

Me

OH OMe OMe

" Me 79: 54%, syn/anti = 1/2

SCHEME 11

The third option utilized the vinylogous aldol of dienol ether 69 to construct the whole carbon framework of 42 (path a, Scheme 9). The preliminary studies on aldol addition of 69 to benzaldehyde showed excellent site-, diastereo- and enantioselectivity. 29 The expected aldol product 68

16

SCOTT E. DENMARKAND SHINJI FUJIMORI

would have anti-configuration at the newly created stereogenic centers, therefore, the inversion at C(31) hydroxyl would be necessary. The aldol addition of 69 catalyzed by (R,R)-56, however, provided no desired aldol product (Scheme 12). It is most likely that methacrolein underwent the undesired side processes under these conditions. TM O Me~..H

+

M e ~ O T B

S

Ot-Bu

(S,S)-56 (5 mol % ) CH2CI2,-78~

69

OH O Me..~l.~.~Ot_Bu_ Me 80: not observed

SCHEME 12

After many unsuccessful attempts, it was decided that the C(27)- C(31) fragment would be prepared from the known aldehyde 50 (Scheme 13).3o The stereogenic centers in 50 were established by the aldol addition of chiral acyl oxazolidinone to isobutyraldehyde. From this aldehyde, Wittig olefination with methyl triphenylphosphanylideneacetate provided the conjugated ester 81. The reduction of the ester to the allylic alcohol with Dibal-H was followed by allylic oxidation with manganese(IV) oxide to provide the aldehyde 42 in good yield (61%, three steps, Scheme 13). PMBO

O

Me

H Me

Me 5O

Ph3PCHCO2Et " MeCN, reflux 90%

PMBO_ Me

I Me

_-Me 82

PMBO DibaI-H Me \ . / ~ . , , ~ / C O O E t ,] --_- Et20 Me Me 97% 81 MnO2 PMBO_ = M e ~ C H O OH 0H2CI2,reflux Me Me 70% 42 SCHEME 13

C. ALDOL ADDITION OF TRICHLOROSILYL ENOLATE DERIVED FROM 41

With the methyl ketone and the aldehyde in hand, the key aldol coupling was attempted. On the basis of previous studies, 36 the aldol addition

1

17

TOTALSYNTHESISOF RK-397

of the trichlorosilyl enolate derived from 41 was tested (Scheme 14). TMS enol ether 83 was prepared from methyl ketone 41 with diisopropylethylamine and TMSOTf ~3b(94%). The transsilylation of 83 to the corresponding trichlorosilyl enolate 84 was accomplished using SiC14 and mercury(II) acetate. 37 After 30 min, the volatile components were removed to provide the crude trichlorosilyl enolate 84, and this crude product was used directly for the Lewis-base-catalyzed aldol addition. With (R,R)-76 as the catalyst, aldol product 85 was obtained in good yield (81%), but the diastereoselectivity was only 2/1 favoring the desired diastereomer (27R)-85.

Ph

O~O

O

BnMe2Si~Me

CH2CI2

41

O/~O

OTMS

" BnMe2Si~ 83

94%

SiCI 4 I Hg(OAc)2 (2 mol % ) CH2Cl2

Ph

TMSOTf DIPEA

Ph 0~0

OSiCI3

/ B n M e 2 S i ~ 84

42, (R,R)-76 (10 mol %) CH2CI2 -78 ~ 81%, dr 2/1

Ph PMBO

M

e ~ s Me Me

OH 0

i

M

OH O

0~0

e

2

B

n

(27R)-85

,. ~

t

(27S)-85 SCHEME 14

Interestingly, by use of the enantiomeric catalyst (S,S)-76, diastereomer (27S)-85 was obtained in good yield (72%) with a 4/1 diastereomeric ratio. These results indicate that (27S)-85 forms in the matched case arising from 1,5-syn stereoinduction in the aldol addition. To improve the diastereoselectivity of the aldol addition, various chiral phosphoramides including dimeric species, chiral N-oxides, achiral Lewis bases such as phosphines, formamides, phosphine oxides and sulfides, and ureas were surveyed as the promoter, however, the diastereoselectivities were uniformly low (Chart 1 and Table 1).

18

SCOTT E. DENMARK AND SHINJI FUJIMORI

Ph0

Ph Ph%.IN, 0

IN'phN~ Ph

Me K'"-~N ,,0

N::N 0 H"'L [P--N/" ' ~N L}

Ph ~

86

87

MeMe / 2

~']N ,0

88

89

Ph" N~lel~e ] 2

(R,R)-56

90

~

Me

/2

74

Me

Me t-Bu n-BuO

91

~~/ t-Bu On-Bu

N'Me0 MeMe 92

0 L~/

0 93

CHART 1

The lack of selectivity in this aldol addition may be due to the coordination of benzylidene acetal oxygen to the trichlorosilyl moiety (Figure 6). The chelation by this additional oxygen will allow only one phosphoramide to bind, and the stereochemical information cannot be effectively transmitted from the chiral catalyst to the newly created stereogenic center. A dramatic difference in enantioselectivity between the one- and twophosphoramide pathways has been documented in the aldol addition of cyclohexanone-derived trichlorosilyl enolates where the one-phosphoramide pathway not only altered the diastereoselectivity but also led to reduction in enantioselectivity. 38 Another possibility is that the chelation may facilitate an uncatalyzed process by activating the trichlorosilyl moiety of the enolate. The product arising from the uncatalyzed pathway can compromise the observed diastereoselectivity of the aldol addition. This type of chelation was also proposed in analogous titanium(IV) enolates, and these enolates were found to give extremely low levels of stereoinduction. 39

TABLE 1 Aldol addition of trichlorosilyl enolate catalyzed by various Lewis bases Ph

1. SiCI4, Hg(OAc)2 OTMS 2.42, cat.

"""u ~ J BnMe2Si~

83

CH2012

---

Ph _

PMBO_

_OH 0

OH 0

0_~_0

Me

SiMe2Bn + " Me

Me

"

(27R)-85

(27S)-85

Entry

Lewis base a

Loading (tool%)

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

(R,R)-76 (S,S)-76 90 86 92 (R,R)-56 91 89 74 88 (R,R)-87 (S,S)-87 93 Ph3PO Ph3PS DMF Uncatalyzed

20 20 10 20 10 10 8 10 11 10 10 10 100 100 100 100

Yield (%)

No No

No No No

27R/27S

2/1 1/4 2/1 1/1

81 72 61 57 reaction 73 reaction 46 20 42 57 72 reaction reaction 33 reaction 41

2/1 1/1 1/1 2/1 2/1 1/3

1/1 1/1

aSee Chart 1. bReaction run at 0 ~

CI CI Ph si_Cl

oAo " b

Ph LB CI c'l

(R,R)-76

-

RoHo

BnMe2Si ~

] ci

e

,,,st< o'

0~0

0

\\

BnMe2Si 84 !

i background reaction !

85

' "one phosphoramide" , pathway |

85

FIGURE 6. Possible scenarios for the low diastereoselectivity.

20

SCOTT E. DENMARKAND SHINJI FUJIMORI

The aldol addition of TMS enol ether 83 catalyzed by bisphosphoramide 56 and SiC14 was also examined (Scheme 15). However, the reaction did not yield the desired product and only led to the elimination of the alkoxy group from aldehyde 42. Replacement of the PMBO group with other ether functions such as benzyloxy or naphthylmethoxy groups, or to noncoordinating functionality such as tert-butyldimethylsilyloxy or triisopropylsilyloxy groups did not prevent the elimination. In the aldol additions catalyzed by bisphosphoramide 56, simple olefinic aldehydes such as crotonaldehyde and acrolein were problematic due to the competitive conjugate addition of the ionized chloride. 23c In this case, however, the ionized chloride may act as a base to eliminate the alkoxide from activated 42 by deprotonation at 3,-position resulting in formation of 95. PMBO

Me_.x~x~CHO+Me 42

Me

~

0

Ph : ~ 0 OTMS

BnMe2Si

(R,R)-56

SiCI4,DIPEA . M e ~ C H O CH2CI2

83

Me

Me

95

SCHEME 15

Thus, both types of Lewis-base-catalyzed aldol additions were not able to provide the desired aldol product with high.diastereoselectivity. 23b The aldol addition of trichlorosilyl enolate 84 provided the aldol product in good yield, however, the diastereoselectivity is very low. Therefore, an alternative plan to connect these modules was considered. D. SYNTHESIS OF C(11)-C(33) FRAGMENT

On the basis of the studies by Evans et al. 39 and Paterson et al., 4~ the aldol additions of boron enolates derived from methyl ketones bearing a 13-oxygen stereogenic center provide high 1,5-anti stereoinduction. Indeed, better 1,5-anti stereoinduction could be achieved by substratecontrolled aldol addition using the dibutylboron enolate derived from 41 (Scheme 16). The boron-mediated aldol addition of 41 to 42 afforded 85 with excellent diastereoselectivity (85%, dr > 19/1). Although the origin of the 1,5-anti stereoinduction is not clear, the empirical analysis of these types of aldol additions indicated the nature of enolate metal, solvent and protecting group greatly affects the diastereoselectivity. Evans et al. 39

1

21

TOTALSYNTHESISOF RK-397

proposed the intermediacy of multiple sets of competing transition states based on the nonlinear relationship between the reaction temperature and diastereoselectivity. The transition state I leads to the observed 1,5-anti diastereomer. The conformation of the boron enolate can be rationalized by the avoidance of lone-pair repulsion between the [3-oxygen (benzylidene acetal) and the enolate oxygen. This repulsion would force the C(23)-O bond to be antiperiplanar to the C(24)-C(25) bond, and the two enolate faces would be sterically differentiated. The aldehyde will approach from the less sterically hindered side. A similar analysis has been provided for aldol addition of a boron enolate beating a [3-benzyloxy group. 4~ Also, Lee et al. 42 proposed an analogous model to explain the 1,5-anti stereoinduction in the addition of a lithium enolate bearing [3-alkoxy stereogenic center. Ph O~O

19

BnMe2Si ~

M

O

PMBO

e

+M e ~ C - H O Me Me 42 Ph

41 PMBO Me ~ Me

OH O

S

i

M

e

O1"1"O

2

B

n

Bu2BOTf DIPEA Et20,-78 ~ 85%, dr >19/1

H , 24

,H", 'u2

Me

85 SCHEME 16

The configuration at C(27) of 85 was confirmed to be R by Mosher ester analysis of the aldol product (Figure 7). 43 The Mosher ester was prepared by the reaction of 85 with 1-methoxy-l-trifluoromethylphenylacetyl (MTPA) chloride in pyridine/CDC13 and was analyzed by ~H NMR spectroscopy of the crude reaction mixture. Both diastereomers of the R- and S-MTPA esters (96 and 97, respectively) were analyzed, and their chemical shift differences at C(26)-methylene and C(28)-vinyl protons were measured. In the most relevant conformer of these esters, the CF 3 group, the ester carbonyl group and the C(27) hydrogen are all in the same plane (I and II, Figure 7). 44 The phenyl

22

SCOTT E. DENMARK AND SHINJI FUJIMORI

Ph

(R)-MTPA PMBo

H ~o

,~ o

o

o

PMBo

Me. ~ .,.~28~.,26,~ ~ .~/~ T ___v ~-.~ 27v v v ~ Me

Me H 96: 5H2C(26 ) = 2.70 ppm 5HC(28) = 5.48 ppm

(R)-MTPA:

Ph

(S)-MTPA \ .

,,,J,,., o

o

o

Me

o F~c~ ~-~__~,,,~ o OMe I

/~

-SiMe2Bn

Me H 97:~H2C(26 ) = 2.77 ppm ~HC(28) = 5.35 ppm

(S)-MTPA:

shielding

o

Me. A ..~28,~26~.,,~ ~ ~ -SiMe2Bn ~r~ ___v ~r~ 27. . . .

"

F~c~ ,/28

MeO'"~?

"~28

/,~, II shielding

FIGURE 7. Mosher ester analysis of the aldol product 85.

substituents of the MTPA ester can influence the chemical shifts of either the C(26) or the C(28) protons. In the R-MTPA ester, the C(26) protons of 96 are shielded (I) and therefore their chemical shift should be lower than that of the S-MTPA ester. On the other hand, the C(28) proton in 97 is shielded (II), and the chemical shift in S-MTPA ester should be lower than that of the R-MTPA ester. These shifts were observed and the assignment of configuration was made accordingly. The C(25) carbonyl group in 85 was reduced in the presence of diethylmethoxyborane and sodium borohydride, and the resulting syn-diol was protected as a benzylidene acetal to give 98 (86%, two steps, Scheme 17).45 The origin of the stereocontrol in the syn reduction arises from the formation of a boron chelate 100 with [~-hydroxy ketone and diethylmethoxyborane (Figure 8). The axial attack of hydride provides borate 101 leading to the observed syn-diol. The oxidative unmasking of the alkenylsilane to reveal the aldehyde functionality at C(19) was accomplished by the protocol developed by Tamao et al. 46 (73%). The benzyl group of the silane was cleaved by the action of fluoride, and the resulting silanol was oxidized by hydrogen peroxide under basic conditions. It is essential to completely remove the excess peroxide during the workup because the concentration of the crude 99 with trace amount of peroxide leads to decomposition. In addition, aldehyde 99 slowly decomposed even when it was stored at - 15 ~ and it was therefore necessary to use this aldehyde immediately after purification. 47

1

23

TOTALSYNTHESIS OF RK-397 o.~

Et.B/

,,

J

0

' '

H 100

ii

..m

~

Et

H

101

LJ hydride attack

FIGURE 8. syn Reduction of the aldol product 85.

!

PMBO Me. ~

/~..~

Me Me Me

Ph

OH 0 O/X'O ~ ~ ...~//~ 27 25 19 SiMe2Bn 85

Me~j .... ~.OPMB

M

e

98

,

"

OvO

Ph

15h

]

86%, dr >19/1

TBAF; Me2Bn H202'KHCO3D. Si THF/MeOH

~

O~/0

1. Et2BOMe, NaBH4 2. PhCH(OMe)2 CSA

73%

Me Me) . . .' (. OPMB ,,L / / ~ 2 7

Me'

25/~

---~ "/-

"1(

v

O\/0

OvO

Ph

Ph

]

99

/~19H

~

"-j

-11-" 0

SCHEME 17

With anticipation of 1,5-anti stereoinduction to establish the C(19) stereogenic center, the second iteration of the aldol addition was carried out using 41. The aldol addition product 102 was obtained in high yield and excellent selectivity (88%, dr > 19/1, Scheme 18). The aldol addition showed excellent 1,5-anti stereoinduction, and the configuration at C(19) was confirmed by Mosher ester analysis (Figure 9). 43 The analysis of the configuration was carried out in a similar manner. The chemical shifts of C(18)-methylene and C(21)-methyne protons in.R- and S-MTPA esters were measured. The results indicated that the configuration

24

SCOTT E. DENMARKAND SHINJIFUJIMORI

Me Me...,j. . . . OPMB ',~_.../ ~ 2 7 ~

Me" ~

y

y

O\/0 I Ph

99

Me t Me.,P. . . . OPMB '~ 27 M e ' " O\ O 102 II" Ph e Me....J....r/OPMB

~

~

~19

=Y -7-- Y OvO

0

15h

41, Bu2BOTf DIPEA Et20,-78 ~ 88%, dr > 19/1

BnMe2Siv~'~,,,O-,,,1,,,Ph 1. Me4NHB(OAc)3 11 I " 2. Me2C(OMe)2 " k.,. ,,O CSA 19 "1' ~ OvO OH O 87%,dr>19/1 Ph BnMe2Siv ~ , , . O . ~ l , , , Ph ~ I TBAF: "" L,~,,,O H202, KHCO3

Me" ~ 103

O\/O "1 Ph

OvO 15h

O .O Me'Me

THF/MeOH 86%

Me O.~--.~,.,O...],,, Ph / Me..P. . . . . OPMB H k.....,O DDQ ~..,./~27~ ~ ~ 1 9 ~ .,~ Me"' CH2CI2/H20 rt O~/O O.,.~fO 0 .0 90% 104 ][ _ Me'~Me Ph 15h iMe O ~ , , , O . . . ] , , , Ph Me..P.... r/OH H L..~,,.O I,,,. ~,,,,,,.27~ ~ ~ ~19~ .,,,,,I Me'" O\/O OvO O .O 105 | ae'~ Ph Ph Me SCHEME 18

at C(19) was S. The carbonyl group at C(17) was reduced using tetramethylammonium triacetoxyborohydride, and the resulting anti-diol was protected as an acetonide to give 103 (87%, two steps). 48 The stereochemical course of the anti reduction can be rationalized by the transition state illustrated in Figure 10. The formation of alkoxy borohydride III (Figure 10)

l

25

TOTAL SYNTHESIS OF RK-397

Me

BnMe2Si" v ' ~ ' " O " ' ] '''Ph

/

MeJ .... /OPMB [~ M

21

e

'

"

O\ / 0

t-~ ,,,O ']

~

OvO

][

-

Ph

Ph

106: (R)-MTPA 5H2C(18) = 2.86 ppm a HC(21) = 3.80 ppm

O, 0 Ror S-MTPA 107: (S)-MTPA 8H2C(18) = 2.90 ppm ,5HC(21)= 3.62 ppm i.

(S)-MTPA:

shielding

(R)-MTPA:

19

O

F ~_ 0 3C ~~/\

O

ncjL

0, a ~~'" l 8 ....

,.~

""

.H--~l8

j<,,' o-

~/o

'

MeO' ~1,~, J,/21"7-" OH ~.',,"

OMe_ .,.~21 """ H5""

II shielding FIGURE 9. Mosher ester analysis of aldol product 102.

| H H

m

OAc

,

H

OAc

JOH

I

'~-~~IB ~4

~OA c

~4

III FIGURE 10. anti Reduction of aldol product 102.

under the acidic conditions allows for an internal delivery of the hydride within the six-membered transition state and this model predicts the formation of the anti-diol. With the polyol chain installed, the stage was set to introduce the polyene fragment, which required unveiling of the alkenylsilane 103 to reveal aldehyde 104 as described previously. Before the installation of polyene fragment, the PMB group was removed by the action of DDQ to afford hydroxy aldehyde 105 (77%, two steps, Scheme 1 8 ) . 49

26

SCOTT E. DENMARK AND SHINJI FUJIMORI

E. P R E P A R A T I O N OF P O L Y E N E F R A G M E N T 40

Although there are several methods to prepare polyene phosphonate 40,13,14 the construction of the tetraenoate fragment was accomplished by the sequential palladium-catalyzed coupling of bis-silyl diene 45. 28 The coupling of 45 with various aromatic iodides under KOTMS activation had already been demonstrated; however, the coupling with vinyl iodides had not been fully explored. Therefore, the initial study involved a survey of various conditions for the coupling of 45 with THP protected 3-iodo2-propenol 465o (Table 2).

TABLE 2 Coupling of Silanol 45 with Iodide 46 Pd-cat. activator additive BnMe2Si/-,,.~",,.~'~SiMe20 H +

9" BnMe2Si

I/...,.~--.~/OTHP

45

/

/

solvent temp., time

46

Additive

Solvent

/

OTHP

108

Time/ Temperature (~

Yield ~

Entry

Catalyst

Activator

1

Pd(dba)e (5 mol%)

TMSOK (2 equiv.)

Dioxane

24h/~

NR

2

Pd(dba) 2 (5 tool%)

TMSOK (2 equiv.)

Dioxane

5 h/50

56% (5/1)

3b

Pd(dba)2 (5 mol%)

TMSOK (2.4 equiv.)

PPh 3 (5 mol%)

Dioxane

12 h/50

Trace

4b

Pd(dba) 2 (5 mol%)

TMSOK (2.4 equiv.)

AsPh 3 (5 mol%)

Dioxane

12 h/50

NR

5b

Pd2(dba) 3CHCI 3 (5 mol%)

TMSOK (2.4 equiv.)

Toluene

5 h/rt

60% (1/1)

6b

Pd2(dba)3CHCI 3 (5 mol%)

TMSOK (2.4 equiv.)

Toluene

5 h/rt

70% (2/1)

Pd2(dba)3CHCI 3 (5 tool%)

Nail (1.4 equiv.)

Toluene

6 h/rt

77% (3/1)

Cu(OAc) 2 (25 mol%)

aThe ratio of geometric isomer is shown in parenthesis. NR stands for no reaction, bl.2 equiv, of 45 was used. Cl.5 equiv, of 45 was used.

1

27

TOTAL SYNTHESIS OF RK-397

The standard conditions that were developed for the coupling with aromatic iodides did not yield the coupling product even after 24 h. However, the coupling took place when the reaction temperature was raised to 50 ~ Because homocoupling of the iodide was observed at elevated temperature, the effect of additives was investigated. Unfortunately, the addition of phosphine and arsine ligands suppressed the desired coupling under these conditions. In the coupling of heteroaromatic silanols, the use of the tris(dibenzylideneacetone)-dipalladium(0)-chloroform adduct showed superior results, 51a and therefore this palladium catalyst was tested for the coupling of 45. Indeed, this catalyst was more active than Pd(dba) 2, allowing the coupling to proceed at room temperature. Although the yield could be increased by addition of copper(II) acetate, the best yield was achieved for the reaction with in situ formation of the sodium silanolate. 5~bAddition of iodide 46 and palladium catalyst then initiated the coupling reaction. The olefin geometry of coupling product 108 was not been retained under the reaction conditions, however, the geometry could be corrected before the conversion to the phosphonate (vide infra). To complete the assembly of the tetraene unit, the second coupling of benzylsilane 108 was attempted (Scheme 19). Silane 108 was first activated with TBAF, and the resulting silanol was treated with ethyl 3-iodopropenoate 4752

Nail; Pd2(dba)3-CHCI3 246 HOMe2si~SiMe2

Bn

45

TBAF; Pd(dba)2, 47

THPO~~SiMe2Bn

toluene 77%, d r - 3/1

108 p-TsOH; 12

= THPOv'~.~~COOEt

EtOH

THF

88%

109

79%, dr-5/1

1. PBr3 2. P(OEt)3 HO

~

~

~

~

COOEt 93%

43 (EtO)2OP

~

~

~

~

COOEt

40 S C H E M E 19

28

SCOTT E. DENMARK AND SHINJI FUJIMORI

in the presence of Pd(dba)2 to afford tetraenoate 109 in good yield (79%). The fluoride-promoted cleavage of the benzyl group had to be carried out at low temperature (0 ~ because desilylation became competitive at higher temperature. The THP protecting group was removed by acidic hydrolysis to give hydroxy ester 43, and the treatment of this mixture of double bond isomers with a small amount of iodine during workup gave the all E-tetraenoate 43 in good yield (88%). ~4c The hydroxyl enoate 43 was readily converted to the phosphonate 40 following the literature procedure. TM The bromination of alcohol was achieved using phosphorus tribromide, and the resulting bromide was treated with triethylphosphite to afford phosphonate 40 in excellent yield. IV. C o m p l e t i o n of

the Synthesis

The final fragment coupling brought together phosphonate 40 and the polyol chain 105 using a Horner-Wadsworth-Emmons olefination protocol (Scheme 20). From this reaction to the final product, all reactions and purifications were performed with careful exclusion of light because of the photosensitivity of the polyolefin. To do this, the hood doors were covered with aluminum foil, and these doors were closed while the reactions were performed. Only during setting up of the reactions and purifications could the polyene containing intermediates be exposed to light. The olefination proceeded smoothly with the aid of LiHMDS to afford pentaenoate 110 in good yield (77%). The macrolactonization of the seco-hydroxy acid (prepared from 110 by saponification with LiOH) was efficiently effected via the mixed anhydride from 2,4,6-trichlorobenzoyl chloride. 2~Closure of the mixed anhydride in the presence of DMAP under high dilution successfully afforded a good yield of 111 with a minimum of larger oligomers (71% over three steps). Although the acetonide was easily removed from 111, deprotection of the more stable benzylidene acetals was initially problematic. A survey of acidic resins such as Dowex 50W | or Amberlyst-15 | in alcohol solvents including thiols revealed that full deprotection could not be achieved even after several days. It was also found that prolonged reaction times caused decomposition of the natural product. On the other hand, the use of p-toluenesulfonic acid showed significantly faster deprotection. However, the complete removal of the acid after workup and purification was difficult. Surprisingly, the use of concentrated HC1 (ca. 60 equiv.) in MeOH was found to be the most effective method for the global deprotection (93%).

1

29

TOTALSYNTHESISOFRK-397

Me

O ~ , , , O '. . . ] ' ,,Ph O

Me...j .... OH M e

"

'

O\/O ][Ph

H

~

OvO

~.. ,,O

O O MUMe

15h

PO(OEt)2 OEt 4O LiHMDS, THF -78 to 0~ 770

105

E t O O C ~ , , ' O " . ] , ,'Ph 1. LiOH 2. Et3N,2,4,6-trichloro/-Pr.... ..OH ~ ,,,O benzoylchloride 3. DMAP M e ' " ~ 71% 0\[I0 OvO 0 .0 Ph Ph Me'~Me 110 M2 ~ ~ ~ ~ ~ ,,,O.,..],,,Ph Me) . . . . ~ L,.. ,,o conc. HCI MeOH 93%

Me"

0\/0 [ Ph

OvO : 15h

0 0 MUMe

111 MeO 1

~

~

~

<1

Me),;,.[ ..0 M

e

,,,OH L,~ ,,,OH

' " ~ OH OH OH OH OH OH RK-397

SCHEME20

It was crucial to remove the acid at the end of the reaction with polymerbound piperdine to avoid decomposition of the final product during purification. The product was stable under silica gel chromatography, affording the natural product along with minor impurities presumably derived from decomposition of the product. To obtain the product in a state of high purity for thorough and unambiguous structural verification, further purification was carried out using preparative reverse-phase HPLC.

30

SCOTT E. DENMARK AND SHINJI FUJIMORI

Synthetic RK-397 exhibited identical spectroscopic and physical properties to those reported for the natural material (HNMR, CNMR, ORD, HRMS). 1 V.

Summary

RK-397 has been synthesized in a concise manner by employing a highly convergent synthetic strategy that features the use of an eightcarbon building block for 16 carbons in the polyol chain (Figure 11). The coupling of these modules was accomplished by substrate-controlled 1,5anti-aldol addition and subsequent stereoselective reduction. The synthesis highlights the highly stereoselective vinylogous aldol addition using chiral phosphoramide 56 for the construction of the key intermediate 41. The stereogenic center created by this reaction effectively established 8 out of 10 stereogenic centers in this molecule by substrate control. The use of an alkenylsilane as a masked aldehyde functionality is noteworthy for not only allowing high yields in the key aldol additions, but also for

HWE olefination macrolactonization/ kk~ O.~ ~

~

~

~

Me.,,.',,, [ . . . O . ,

...-~,_*..~ .

%~ ,,OH

OH OH'-,,,~ O~4/OH 1.5-anti aldol

carbon No.

reaction

,,OH

OH

selectivity

stereocontrol

C(13), C(21)

vinylogous aldol

49/1 er

catalyst control

C(15), C(23)

syn-benzylidene acetal

> 19/1 dr

substrate control

C(17)

anti-reduction

> 19/1 dr

substrate control

C(19), C(27)

1,5-anti boron aldol

> 19/1 dr

substrate control

C(25)

syn-reduction

> 19/1 dr

substrate control

C(30), C(31)

Evans aldol

> 19/1 dr

auxiliary control

formation

FIGURE 11. Derivation of stereogenic centers in RK-397.

1

TOTAL SYNTHESIS OF RK-397

31

its stability under numerous reaction conditions. The sequential crosscouplings with bis-silyl diene 45 to construct highly conjugated side chain 40 are also highlighted. The efficiency of the synthesis was significantly enhanced by identifying the building block 41. This approach minimized the steps for functional group interconversions and protection/deprotection of the polyol. The longest linear sequence is of 20 steps starting from 3-benzyldimethylsilylpropynol (52) with an overall yield of 4.3%. The previous synthesis by McDonald et al. also took advantage of a modular approach for the construction of the polyol, however, the preparation of the modules seems inefficient (Scheme 1). ~ The chiral epoxide module 33 is constructed through a hydrolytic kinetic resolution, and the yield for the resolution step is only 42%. In their synthesis, the overall yield for the longest linear sequence was 0.3% (30 steps from isobutyraldehyde). The two late-stage reactions are responsible for the low overall yield: the attachment of the triene fragment by Stille coupling (32%) and the macrocyclization by intramolecular olefination (29%). For the coupling of fragments 41 and 42, the Lewis-base-catalyzed aldol addition was employed and successfully provided the aldol product in good yield, however, only marginal 1,5-anti stereoinduction was attained. The compatibility of trichlorosilyl enolates with various functional groups present in 41 and 42 (e.g. benzylidene acetal and p-methoxybenzyloxy ether) was demonstrated. Moreover, the 1,5-syn diastereomer could be obtained in higher selectivity (dr 4/1), which is not possible by the addition of boron enolates. To obtain a better 1,5-anti stereoinduction by addition of the trichlorosilyl enolate, various Lewis bases were surveyed, but chiral phosphoramide 76 was the best catalyst. The selectivity might be improved by changing the protecting group on ]3-hydroxy group in 41 from benzylidene acetal to TBS group, though this would require additional steps in the synthesis. The selectivity obtained with a trichlorosilyl enolate bearing ]3-OTBS group is up to 5.5/1 using phosphoramide 76. 36 The key intermediate 41 can also be a suitable building block in the synthesis of other polyene macrolides and natural products that possess a polyacetate structural motif. Analogous intermediates can be prepared using a y-substituted silyl dienol ether that may allow preparation of a polypropionate-type structure. Although the alkenylsilane in 41 was oxidized to aldehyde in this synthesis, the benzyldimethylsilyl group can also be used in palladium-catalyzed cross-coupling reaction to form a C-C bond. The modular approach for the synthesis of RK-397 should facilitate preparation of various diastereomers of RK-397 through several modifications of the sequence: (1) by changing the catalyst configuration at the

32

SCOTT E. DENMARKAND SHINJI FUJIMORI

vinylogous aldol addition, (2) by employing the silicon-based aldol to establish 1,5-syn stereoinduction in the key aldol coupling and (3) by changing the reagents for the syn/anti carbonyl reduction. These variations may enable preparation of a library of RK-397 isomers to study biological activity of these polyene macrolides. Acknowledgements We are grateful to the National Science Foundation (CHE0414440) for generous financial support. We also thank Dr. H. Koshino (RIKEN) for providing 1H and 13C NMR spectra of the natural product.

References and Footnotes 1. (a) Kobinata, K., Koshino, H., Kudo, T., Isono, K., Osada, H., J. Antibiot. 1993, 46, 1616-1618. (b) Koshino, H., Kobinata, K., Isono, K., Osada, H., J. Antibiot. 1993, 46, 1619-1621. 2. Omura, S., Tanaka, H. In: Macrolide Antibiotics: Chemistry, Biology and Practice, Omura, S. (Ed.), Academic Press: Orlando, 1984, Chapter 9. 3. Bolard, J., Biochim. Biophys. Acta 1988, 864, 257-304. 4. Hartsell, S., Bolard, J., Trends Pharmacol. Sci. 1996, 17, 445-449. 5. Schaffner, C. P. In: Macrolide Antibiotics: Chemistry, Biology and Practice, Omura, S. (Ed.), Academic Press: Orlando, 1984, Chapter 12. 6. Gale, E. E In: Macrolide Antibiotics: Chemistry, Biology and Practice, Omura, S. (Ed.), Academic Press: Orlando, 1984, Chapter 11. 7. Mulks, M. H., Nair, M. G., Putnam, A. R., Antimicrob. Agents Chemother. 1990, 34, 1762-1765. 8. (a) Mechlinski, W., Schaffner, C. P., Ganis, P., Avitabile, G., Tetrahedron Lett. 1970, 11, 3873-3876. (b) Ganis, P., Avitabile, G., Mechlinski, W., Schaffner, C. P., J. Am. Chem. Soc. 1971, 93, 4560-4564. 9. (a) Rychnovsky, S. D., Rogers, B. N., Richardson, T. I., Acc. Chem. Res. 1998, 31, 9-17. (b) Rychnovsky, S. D., Richardson, T. I., Rogers, B. N., J. Org. Chem. 1997, 62, 2925-2934. 10. Rychnovsky, R. D., Chem. Rev. 1995, 95, 2021-2040. 11. For synthetic studies on RK-397, see: (a) Burova, S. A., McDonald, E E., J. Am. Chem. Soc. 2004, 126, 2495-2500. (b) Burova, S. A. McDonald, E E., J. Am. Chem. Soc. 2002, 124, 8188-8189. (c) Schneider, C., Tolksdorf, E, Rehfeuter, M., Synlett 2002, 2098-2100. 12. (a) Nicolaou, K. C., Chakraborty, T. K., Daines, R. A., Ogawa, Y., Simpkins, N. S., Furst, G. T.,J. Am. Chem. Soc. 1988, 110, 4660-4671. (b) Nicolaou, K. C., Daines, R.A., Uenishi, J., Li, W. S., Papahatjis, D. P., Chakraborty, T. K., J. Am. Chem. Soc. 1988, 110, 4672-4685. 13. (a) Poss, C. S., Rychnovsky, S. D., Schreiber, S. L., J. Am. Chem. Soc. 1993, 115, 3360-3361. (b) Dreher, S. D., Leighton, J. L., J. Am. Chem. Soc. 2001, 123, 341-342.

1

TOTALSYNTHESISOF RK-397

33

14. (a) Rychnnovsky, S. D., Hoye, R. C., J. Am. Chem. Soc. 1994, 116, 1753-1765. (b) Mori, Y., Asai, M., Okumura, A., Furukawa, H., Tetrahedron 1995, 51, 5299-5314. (c) Mori, Y., Asai, M., Kawabe, J., Furukawa, H., Tetrahedron 1995, 51, 5315-5330. (d) Evans, D. A., Connell, B. T., J. Am. Chem. Soc. 2003, 125, 10899-10905. 15. Richardson, T. I., Rychnovsky, S. D., Tetrahedron 1999, 55, 8977-8996. 16. Rychnovsky, S. D., Khire, U. R., Yang, G. J., J. Am. Chem. Soc. 1997, 119, 2058-2059. 17. Sinz, C. J., Rychnovsky, S. D., Angew. Chem. Int. Ed. 2001, 40, 3224-3227. 18. Vedejs, E., Bershas, J. P., Tetrahedron Lett. 1975, 16, 1359-1362. 19. Wollenberg, R. H., Tetrahedron Lett. 1978, 19, 717-720. 20. Inanaga, J., Hirata, K., Saeki, H., Katsuki, T., Yamaguchi, M., Bull. Chem. Soc. Jpn. 1979, 52, 1989-1993. 21. Stork, G., Nakamura, E., J. Org. Chem. 1979, 44, 4010-4011. 22. Blanchette, M. A., Choy, W., Davis, J. T., Essenfield, A. P., Masamune, S., Roush, W. R., Sakai, T., Tetrahedron Lett. 1984, 25, 2183-2186. 23. (a) Denmark, S. E., Stavenger, R. A. Acc. Chem. Res. 2000, 33, 432-440. (b) Denmark, S. E., Fujimori, S. In: Modern Aldol Reactions, Vol. 2, Mahrwald, R. (Ed.), Wiley-VCH: Weinheim, 2004, Chapter 7. (c) Denmark, S. E., Beutner, G. B., Wynn, T., Eastgate, M. D., J. Am. Chem. Soc. 2005, 127, 3774-3789. 24. Denmark, 5. E., Sweis, R. E In: Metal-Catalyzed Cross-Coupling Reactions, Vol. 1, 2nd ed., de Meijere, A., Diederich, E (Eds.), Wiley-VCH: Weinheim, 2004, Chapter 4. 25. Boschelli, D., Takemasa, T., Nishitani, Y., Masamune, S., Tetrahedron Lett. 1985, 26, 5239-5242. 26. Oishi, T., Nakata, T., Synthesis 1990, 635-645. 27. Evans, D. A., Gage, J. R., Org. Synth., Coll. Vol. VIII 1993, 339-343. 28. Denmark, S. E., Tymonko, S. A., J. Am. Chem. Soc. 2005, 127, 8004-8005. 29. (a) Denmark, S. E., Beutner, G. L., J. Am. Chem. Soc. 2003, 125, 7800-7801. (b) Denmark, S. E., Heemstra, J. R., Jr., Beutner, G. L., Angew. Chem. Int. Ed. 2005, 44, 4682-4698. 30. Evans, D. A., Dart, M. J., Duffy, J. L., Yang, M. G., J. Am. Chem. Soc. 1996, 118, 4322-4343. 31. Karaev, S. E, Kuliev, R. M., Guseinov, Sh. O., Askerov, M. E., Movsumzade, M. M., J. Gen. Chem. USSR 1982, 52, 1016-1019. 32. (a) Denmark, S. E., Jones, T. K., Org. Synth., Coll. Vol. VII. 1990, 524-527. (b) Ostwald, R., Chavant, P.-Y., Stadtmuller, H., Knochel, P., J. Org. Chem. 1994, 59, 4143-4153. 33. Evans, D. A., Gauchet-Prunet, J. A., J. Org. Chem. 1993, 58, 2446-2453. 34. Williams, J. M., Jobson, R. B., Yasuda, N., Marchesini, G., Dolling, U.-H., Grabowski, E. J. J., Tetrahedron Lett. 1995, 36, 5461-5464. 35. Denmark, S. E., Fu, J., J. Am. Chem. Soc. 2001, 123, 9488-9489. 36. (a) Denmark, S. E., Fujimori, S., Synlett 2000, 1024-1029. (b) Denmark, S. E., Fujimori, S., Org. Lett. 2002, 4, 3477-3480. (c) Denmark, S. E., Fujimori, S., Org. Lett. 2002, 4, 3473-3476. 37. Denmark, S. E., Stavenger, R. A., Winter, S. B. D., Wong, K.-T., Barsanti, P. A., J. Org. Chem. 1998, 63, 9517-9523. 38 Denmark, S. E., Su, X., Nishigaichi, Y., J. Am. Chem. Soc. 1998, 120, 12990-12991. 39. Evans, D. A., Cote, B., Coleman, P. J., Connell, B. T., J. Am. Chem. Soc. 2003, 125, 10893-10898.

34

SCOTT E. DENMARKAND SHINJI FUJIMORI

40. (a) Paterson, I., Gibson, K. R., Oballa, R. M., Tetrahedron Lett. 1996, 37, 8585-8588. (b) Paterson, I., Collett, L. A., Tetrahedron Lett. 2001, 42, 1187-1191. (c) Paterson, I., Oballa, R. M., Norcross, R. D., Tetrahedron Lett. 1996, 37, 8581-8584. (d) Evans, D. A., Coleman, P. J., Cote, B., J. Org. Chem. 1997, 62, 788-789. 41. Cowden, C. J., Paterson, I., Org. React. 1997, 51, 1-200. 42. Lee, C. B., Wu, Z., Zhang, E, Chappell, M. D., Stachel, S. J., Chou, T.-C., Guan, Y., Danishefsky, S. J., J. Am. Chem. Soc. 2001, 123, 5249-5259. 43. (a) Dale, J. A., Mosher, H. S., J. Am. Chem. Soc. 1973, 95, 512-519. (b) Seco, J. M., Quinoa, E., Riguera, R., Chem. Rev. 2004, 104, 17-117. 44. This is one of the three main conformers that have similar energy. For a detailed discussion of the Mosher ester method, see Ref. 43. 45. Chen, K.-M., Hardtmann, G. E., Prasad, K., Repic, O., Shapiro, M. J., Tetrahedron Lett. 1987, 28, 155-158. 46. (a) Tamao, K., Kumada, M., Maeda, K., Tetrahedron Lett. 1984, 25, 321-324. (b) Tamao, K., Ishida, N., J. Organomet. Chem. 1984, 269, C37-C39. (c) Miura, K., Hondo, T., Takahashi, T., Hosomi, A., Tetrahedron Lett. 2000, 41, 2129-2132. (d) Miura, K., Hondo, T., Nakagawa, T., Takahashi, T., Hosomi, A., Org. Lett. 2000, 2, 385-388. (e) Trost, B. M., Ball, Z. T., Joge, T., Angew. Chem. Int. Ed. 2003, 42, 3415-3418. 47. Aldehyde 99 seems to form a hydrate. The ~H NMR analysis of the hydrate showed the absence of aldehydic proton and the 0t-protons are shifted upfield. 48. (a) Evans, D. A., Chapman, K. T., Tetrahedron Lett. 1986, 27, 5939-5942. (b) Evans, D. A., Chapman, K. T. Carreira, E. M., J. Am. Chem. Soc. 1988, 110, 3560-3578. 49. Horita, K., Yoshioka, T., Tanaka, T., Oikawa, Y., Yonemitsu, O., Tetrahedron 1986, 42, 3021-3028. 50. Rossi, R., Bellina, E, Catanese, A., Mannina, L., Valensin, D., Tetrahedron 2000, 56, 479-487. 51. (a) Denmark, S. E., Baird, J. D., Org. Lett. 2004, 6, 3649-3652. (b) Denmark, S. E., Baird, J. D., Org. Lett. 2006, 8, 793-795. 52. (a) Takeuchi, R., Tanabe, K., Tanaka, S., J. Org. Chem. 2000, 65, 1558-1561. (b) Dixon, D. J., Ley, S. V., Longbottom, D. A., Org. Synth. 2003, 80, 129-132. 53. Note added in proof: Another total synthesis of RK-397 has recently been reported by Sammakia and co-workers. Mitton-Fry, M-J., Allen, A. J., Sammakia, T., Angew. Chem., Int. Ed. 2007, 46, 1066-1070.

STRATEGIES AND TACTICS IN ORGANIC SYNTHESIS, VOL. 7 9 2008 Elsevier Ltd. All rights reserved.

Chapter 2

A FORMAL TOTAL SYNTHESIS OF THE MARINE DITERPENOID, DIISOCYANOADOCIANE

Kelly A. Fairweather, Simon R. Crabtree, and Lewis N. Mander Research School of Chemistry Australian National University Canberra, Australia 0200

Introduction and Background Synthesis Plan Foundations and Proof of Concept Elaboration of the D-Ring Elaboration of the B-ring and Exploration of the Michael Reaction Improved Synthesis of an Advanced Tetracyclic Intermediate Deoxygenation, Curtius Rearrangement, and Completion of the Synthesis Conclusion Epilogue Acknowledgments References and Footnotes I~

II. III. IV. V. VI. VII. VIII. IX.

35 37 40 44 46 51 53 55 56 57 57

I. Introduction and Background

A fascinating variety of approximately 80 diterpenes possessing isonitrile and related functions have been isolated from marine sponges. 1 Three of these compounds are acyclic while the bulk of the remainder are based on one of four novel cyclic skeletons, namely kalihinane ( - preamphilectane) (1), amphilectane (2), cycloamphilectane (3), and isocycloamphilectane (4) 2 (Figure 1), examples of which are provided in Figure 2. Wells and coworkers isolated diisocyanoadociane (10) as well as six other isonitriles in 1976 from a marine sponge belonging to the genus Amphimedon (ex. Adocia) collected near Townsville, Australia on the Great Barrier Reef. 3,4 Subsequently, a further 12 related compounds were isolated from the marine sponge Cymbastela hooperi.5 Investigation of the biological activity of these compounds revealed significant antiplasmodial activity when tested in vitro against two clones of

36

1

KELLY A. FAIRWEATHER, SIMON R. CRABTREE, AND LEWIS N. MANDER

9

~

89

2 318 1 16 ~ 71 201~1, 5 15 lO~

2 1

110

1 '20

lO

,7

9

1 ~

2

3

'

FIGURE 1. Parent carbon skeletons for cyclic sponge diterpenes.

NC

0

_

. I. "~H

H

_

s H NC

I H

.

6 H ~C

7

NC H

-

8 Cl~

-

9

.....

n

NC

H

~c

10

H

-

NC

FIGURE 2. Representative isonitriles isolated from marine sponges.

the malaria parasite Plasmodium falciparum. The activity of the compounds towards the mammalian KB cell line was also tested in order to calculate an experimental selectivity index, which indicated whether the observed antiplasmodial activity was a specific or general toxic effect. Results showed that diisocyanoadociane (10), displays antiplasmodial potency and selectivity that rivals the in vitro results obtained with some clinically-used antimalarial drugs. 6 The combination of biological activity with the unusual perhydropyrene based structure makes 10 an enticing target for synthetic chemists and an elegant enantioselective synthesis (60% ee) based on sequential Diels-Alder cycloadditions has been reported by Corey and Magriotis (Scheme 1).7 In the final stages of their synthesis, however, the isonitrile groups were introduced without diastereochemical control. This intriguing molecule had also captured our attention 8 and in this chapter we describe our own efforts to construct this molecule.

2

37

TOTAL SYNTHESIS OF MARINE DITERPENOID, DIISOCYANOADOCIANE

0

MeO~

MeO2C"~ ......

enthy, ~

.

LDA,THF menthyl O -78~

TBDMSO 6 steps

....

"

threo/erythro-diastereoselection8 91 ca. 60% ee

PhMe =

T B D M S O ~ I H......

,. 5 steps BnO

= B n O ~ l 14 IH I '"'"4~0 H....

PhMe

150oc

185oc

~ ~

o H

,,,,"

5 steps

H

5 steps ~ O H

H

,,%

~. . . . ' . . . . . . ~

TMSCN =

'

,,,,,' H

TIC,,

I:1 -~OCOCF3

10 t~

NC

Plus three isomeric diisocyanides

SCHEME 1

II. Synthesis Plan

An outline of our strategy is provided in Scheme 3 and begins with the proposed construction of the phenanthrene-derived structure 13, which was based closely on our earlier synthesis of the phytoalexin, juncunol (11) (Scheme 2). 9

O~c~

Me U ~ ~ ~OMe L~

O~C~] l"alkylation 2. cyclization Me =

~L

OMe

HO

"%

OMe

1. Oxidative decarboxylation OMe 2. Demethylation

..% OH 11

SCHEME 2

1. Vilsmeier reactn 2. Wittig reactn

KELLYA. FAIRWEATHER,SIMONR. CRABTREE,ANDLEWISN. MANDER

38

After adding a suitable side chain to 13, we envisaged that the pyrene skeleton could be completed by means of an intramolecular Michael reaction (19 --->20) and that the two quaternary isonitrile centers could be constructed through a double Curtius rearrangement of the bis(acyl) azide derived from dicarboxylic acid 21 (Scheme 3). This last sequence had excellent precedent in the synthesis of 22 by Piers and co-workers. (Scheme 4). ~~

IO. ~~L. . ,

MeO2C'/~

I~ v

- -a. lkylation ......... MeO2C~

"-~ "OMe 0

OMe

.c.yclization ..........

)2C/~OMe

acylation

Me

-..................... 2. oxygenation

MeO

13

.....

1. reduction/protection

..........

O ~ C ~ ~

OMe

retro-aldol

O

............

OMe 14

HO

FG

O

~

~.....

elaboration .......... _~ at C-20

OMe

",,i '

H

RO2C" ~1:11

O

17

',,,,

-

LO

_ai_~p_~?j_~2~_i!i?_~_..

RO2C'"

1 ~ ~,~"CO~Me

H

H~'~'~

RO2C'"

O

'",,

"

Birch redn

OMe 16

2. Prepn of 4-ene

FG

...........

RO2

1.e l a b o r a t i o n at C-7 ...................

OMe

15

FG

,'""

a

_ 2.• ' "

!. Deo_xygena!io_n_.,,.

i~

Hydrolysis

2O I':1~O~: e H

1.Curtius H02

:~- Siel~s-- ~

..........

' ~

C

21 N~"CO2H

H 1

n

SCHEME3

1. (PhO)2PON3, Et3N 2. TMSEtOH, Et3N H02(~ --

T M s ~ O , . . . r ~ IXlHi I IO

SCHEME4

1. n-Bu4NF 2. AcOCHO 3. Ph3P, CCl4, Et3N

I{IC22

2

39

TOTAL SYNTHESIS OF MARINE DITERPENOID, DIISOCYANOADOCIANE

FG

FG

~2 0 H

pl I

"

a

RO2C"

RO2

OMe

O

v

H~.Co~IO Me

FIGURE 3. Stereocenters under thermodynamic control.

A vital part of our plan was the controlled elaboration of the l0 stereocenters. The use of chiral benzamides in the initial alkylation, as demonstrated by Schultz and co-workers, ~ could reasonably be expected to elaborate the first stereogenic center (C-11) with good enantioselectivity, but we elected to postpone this option and to return later, after establishing the viability of the overall synthesis by employing racemic intermediates. To address the question of diastereochemical control, we expected that the six centers at C-l, C-3, C-4, C-11, C-13, and C-15 (Figure 3) would be under thermodynamic control through enolization of the neighboring carbonyl groups in the intermediates, while equilibration of C-4 would be feasible, provided that the planned Michael reaction was reversible. It is well-established that metal-ammonia reductions of steroidal enones similar to 23 afford trans-fused products, ~2so there was every reason to expect that this process would afford the desired stereochemistry at C-8 (Scheme 5). FG

FG ttt

"

H

Li/NH3 •

"

H

RO2

RO2C'" O

O H~

23 SCHEME 5

There is also an extensive history of stereocontrolled elaboration of quaternary carboxylic groups. ~3As illustrated in Scheme 6, we expected simple alkylation at C-20 to proceed along the equatorial vector (steric control), whereas to achieve "axial" alkylation at C-7, we anticipated that the involvement of the C-6 carbonyl group (stereoelectronic control) would be necessary.

40

KELLY A. FAIRWEATHER, SIMON R. CRABTREE, AND LEWIS N. MANDER

I

I!

M OR

(a) alkylation (steric control)

(b) alkylation

(stereolelectronic control) SCHEME 6

III. Foundations and Proof of Concept

Assembly of the hydrophenanthrene 13 proceeded according to plan as did introduction of a formyl substituent at C-1 by means of the Vilsmeier reaction, provided that one was patient with this last step (7 days!). The next stage called for selective reduction of the styryl double bond, which was carried out by Li-NH 3 reduction (cf Scheme 7) to give 25, but in modest yield (44%). The corresponding methyl ester 26 (prepared by NaC102 oxidation ~4 followed by diazomethane treatment) afforded better results, but if the deep blue reaction mixture was quenched with NH4C1, over-reduction occurred, affording a mixture of diol and hydroxy ester. The addition of isoprene to remove any excess metal before the NHaC1 quench, however, resulted in a 75% yield of 27, accompanied by a moderate amount (14%) of lactone 28. As we have noted elsewhere, the protonation of enolate products by NH4C1 is faster than reaction with any remaining lithium, leading to over-reduction. ~5

Li, NH3 then NH4CI

MeO

,.

OMe

24

OMe

OH

25

I 1. NaClO2 2. CH2N2 Me

MeO2C~] )2C

~ /

0

Li, is~

then NH4CI ~' [

.

CO

~

21/"'~OMe

OMe OH SCHEME 7

27

OMe 28

2

TOTAL SYNTHESIS OF MARINE DITERPENOID, DIISOCYANOADOCIANE

41

HO~co2Me 27

~OMe

FIGURE 4. Determination of stereochemistry by NOE interaction.

The stereochemistry of 27 was determined by nuclear Overhauser effect (NOE) experiments that showed, inter alia, that H-9 was close to H-20, which is only possible with a cis-fusion of the C-and D-rings with the conformation illustrated in Figure 4; correlation spectroscopy (COSY) measurements established that the ester group was equatorial. We had hoped to obtain a product in which the C- and D-rings were trans-fused, corresponding to the configuration of the final target. Nevertheless, we confidently expected that the stereochemistry could be adjusted when we came to the deletion of the C-11 substituent by means of a retro-aldol or equivalent process. Similarly, we expected that the stereochemistry at C-1 could be inverted during the planned Michael addition. At this stage of the synthesis, we could choose to elaborate the D-ring further, or explore the construction of the A-ring. We chose the latter and to this end protected the hydroxyl function in 27 (methyl ether) before reducing the ester function with LiA1H 4 followed by oxidation, then addition of EtMgBr to afford 29. Further protection (MOM ether), Birch reduction and acid hydrolysis afforded enone 30, which unfortunately proved to be surprisingly resistant to conjugation. Basic catalysis was unproductive while care was required with acidic conditions to avoid loss of the MOM protecting group. Ultimately, we discovered that anhydrous HC1 in THF gave the best, albeit rather poor, outcome. The primary yield was modest, but recycling of recovered starting material produced 31 in a total yield of 63%. Given the ease with which steroidal substrates have been converted to conjugated enones, we were disappointed with this result, but pressed on. Fortunately, reductive acylation at C-7 by lithium ammonia reduction followed by in situ reaction with methyl cyanoformate 16 proceeded uneventfully to afford 32, as did C-alkylation of the resulting fl-keto ester (Scheme 8). In preparation for the planned Michael reaction (Scheme 9), we introduced the A4 alkene bond, 17 but when we removed the MOM protecting group, cyclization of the liberated hydroxy group took place to afford 35.

42

KELLY A. FAIRWEATHER, SIMON R. CRABTREE, AND LEWIS N. MANDER

1. KOH. DMSO, Mel 2. LiAIH4 ~; ~

1. MOMCI, iPrNEt2

3. PDC OMe 4. EtMgBr

2. Li, NH3, tBuOH MeOCH2'"'" OMe 29

H OMOM

H OMOM

H OMOM

.

MeOCH2,'""

MeOCH2, y

MeO

O

O

30

O

31

R

Nail, Mel ~ , 3 2

CO2Me

R= H 33 R Me

SCHEME 8

OMOM H

MeOCH2,'"'" 33

H_ OMOM H

1. LDA; PhSeCI

2. H202, Py O CO2Me 1. Me2BBr 2. LDA; PhSeCI

_H ~ / Me2BBr"

" MeOCH2,,,,

O

34

H

,H

MeO

co2 Me

H

35

O CO2Me

H OH -

1. S wer n MeOCH2,'"

K2CO3 ,

oxidn

.... SePh 2. H202, Py

M e O C H 2 , , , , ~ T ! ~']

CO2Me

-

37

,,,

",H

MeOH MeO 0

_

36

:

H c02 Me

0

38

ICt CO2Me

SCHEME 9

We therefore reordered the sequence as indicated, producing 37 in very modest yield, but sufficient to test the viability of the pivotal Michael process. In the event, treatment with DBU was ineffective, returning starting material, while t-BuOK destroyed our intermediate. However, reaction with K2CO 3 in MeOH at room temperature for only 2 h converted 37 into 38 in 71% yield. That cyclization had occurred was apparent from the appearance in NMR spectra of a secondary methyl group at C-3 in place of the ethyl group in the side chain. The stereochemistry of 38 was not rigorously proven but the observation of trans-diaxial couplings between H-3fl and H-4~ (11.4 Hz), and between H4~ and H-5/~ (12.9 Hz) is consistent with the assigned structure (Figure 5).

2

43

TOTAL SYNTHESIS OF MARINE DITERPENOID, DIISOCYANOADOCIANE

O

H3

MeO

4

5

~.fo

38

002 Me

FIGURE 5.

We were now faced with the choice of continuing the synthesis with diketone 38, or regarding the synthesis to date as a model for an alternative approach. To correct the stereochemistry at C-11, we envisaged introduction of a carbonyl group at C-20 and removal of the substituent at C- 11 by means of a retro-aldol reaction. We could then expect that diketone 42 would provide a thermodynamic sink for a series of base-catalyzed isomerizations (via enolates) as illustrated in Scheme 10. However, although the configuration at C-11 could presumably be inverted, it would require that the A-ring adopt a boat conformation as in 40. Alternatively, a D-ring boat (as in 41) would be enforced if inversion took place at C-1 (adjacent to the C-2 carbonyl group). Accordingly, to achieve the correct configuration at C-1 and C-11, it would be necessary to carry out a double isomerization on the 2,20-dione through a domino process. 18 Unfortunately, these two carbonyl groups would then end up in identical local environments and we could expect discrimination between them would be extremely difficult. We therefore chose to return to the earlier stages of the synthesis and focus on the elaboration of the D-ring, including the establishment of a trans-fusion between rings C and D.

O MeO

H3 4

38

base

O 5

H

~

[

'

H k,.,-,"O e

~

-~

O CO2Me

base I

9

O CO2Me

basel O O CO2Me

SCHEME 10

42

0 CO2Me

44

KELLY A. FAIRWEATHER, SIMON R. CRABTREE, AND LEWIS N. MANDER

IV. Elaboration of the D-Ring

Returning to intermediate 24, we extended the side chain and after protection (TBDMS ether) of the resulting alcohol 43, studied its hydroboration with a view to establishing a C-20 carbonyl group (Scheme 11). This reaction gave highly variable yields of 44, so we discontinued this approach. Instead, we oxidized 43 to ketone 45 with the Dess-Martin periodane 19 and reduced this product with lithium in liquid ammonia as for ester 26 to give hydroxy ketone 46, the stereochemistry of which was assumed to correspond to that of 27. Removal of the styryl alkene bond in this way gave us more freedom to introduce oxygen at C-20, which we proposed to carry out by epoxidation followed by Lewis acid mediated rearrangement. OTBDMS

o.

]

O~C~~~, 0 Me v ~ 24

EtMgBr 90% =- e O ~ C ~ ~ "OMe

M 43

1. tBDMSCI ..... 2. BH3DMS; = HO NaOAc,H202 Me OMe

~60% 10 O

~

Me v

45

44

OMe

0

1. Li, NH3,THF 2. isoprene 3. NH4CI "-.>- "OMe 75%

46

S C H E M E 11

Attempts to mask the ketone function of 46 as an acetal failed, so after protection of the primary hydroxyl (MOM ether), reduction with LiA1H4 afforded a single diastereomer that was converted into the benzoate 47. Epoxidation proceeded smoothly, but subsequent treatment with BF3-Et20 resulted in the formation of the cyclic ether 49. To avoid this complication, we replaced the MOM group with acetate and then Lewis acid treatment afforded the ketone 51 as a mixture of C-15 epimers. Finally, hydroxide treatment resulted in hydrolysis of the acetate function, a retro-aldol reaction and stereochemical equilibration at C-11 and C-15 to afford 52, the structure of which was determined by X-ray crystallography. 2~These results are summarized in Scheme 12.

2

0

~

45

TOTAL SYNTHESIS OF MARINE DITERPENOID, DIISOCYANOADOCIANE

OBz

, 2. O 1.IC O ChLi MOMCl m ,base ~ ~' ~ " AIH4 base

OBz

3

.A m C B P _ 60%D,- 0 NaHCO3,

OMe(90%)over3steps MOM;'" v 47-
M

R=MOM R Ac

BF3.Et20

m-CPBA (85%)

OBz

OBz i

OMe

48

H O ~ O ,,

~

O H,"

49 5O

OMe

OMe

BF3Et2075% 1 OBz

OBz KOH

O ,,

EtOH 70% 51

OMe

O v

v 52

"OMe

SCHEME 12

We were now in a position to elaborate the C-20 quaternary center. As outlined in Scheme 13, this was undertaken by means of a Wittig reaction with methoxymethylene triphenylphosphorane. After hydrolysis of the resulting enol ether 53 to aldehyde 54, we attempted C-methylation, but despite a number of precedents, 21 this reaction failed. We therefore returned to the enol ether 53 and subjected it to a Simmons-Smith cyclopropanation. 22 No matter how we tried with a variety of modifications, we could not push this reaction beyond a 40% yield. An improved yield (75%) was achieved with diazomethane-Pd(OAc)2, 23 but only after consuming large quantities of diazomethane. Nevertheless, acid treatment of the product 55 afforded the target aldehyde 56. 24 We had expected that methylenation would have taken place on the more exposed e x o face of 53, resulting in the stereochemistry assigned to C-20 in 55 and therefore to 56. Confirmation was obtained from NMR NOE correlation experiments that showed that the formyl group in 56 was s y n to the C-12 benzylic proton.

46

KELLY A. FAIRWEATHER, SIMON R. CRABTREE, AND LEWIS N. MANDER

After oxidation 14 of aldehyde 56 to the corresponding acid, we submitted this product to Birch reduction. We obtained a multitude of products, however, in which the benzoate group had undergone various stages of reduction, but not, as we had planned, reductive cleavage. So we removed the benzoate group and tried the Birch reduction again. This time over-reduction occurred as a result of intramolecular protonation 25 and the saturated ketone 58 was obtained. OBz

OBz H

O

NaHMDS, 76% OMe

OBz

.... H

H

~

THF

.... H

"

~"~'OMe

MeO

52

OMe

53

54

Zn/Cu, CH212 Et20, reflux (40%) or base; Mel t

CH2N2/Pd(OAc)4 (70%)

,,%,

H OBz -- , , , , H

OBz ,

,%,

H

.... H

-

ace one reflux, 60% =

Me

li....

OMe

OMe 56

55

NaOEt, EtOH

:

OH .... Li, NH 3

,

H

.....

:

OH .... H

HO2C 58

O

57

OMe

SCHEME 13

V. Elaboration of the B-ring and Exploration of the Michael Reaction

Given the modest yields in converting 52 into 56, and the difficulties with the Birch reduction, we elected to conserve material by postponing the elaboration of the C-20 center and press on with the synthesis of the main tetracyclic structure beginning with 52. This intermediate was reduced with L-selectride, 26 the resulting alcohol protected as a MOM ether, then the benzoate function replaced by a TBDMS group to afford 59 in

2

47

TOTAL SYNTHESIS OF MARINE DITERPENOID, DIISOCYANOADOCIANE

an overall yield of 55% for the four steps. Birch reduction (Scheme 14) proved to be stubborn, probably because of the bulky TBDMS group interfering with solvation of the radical anion intermediate. 27 However, a satisfactory yield of the expected dihydro product 60 could be obtained by using a massive amount of lithium metal (50 equivalents). Hydrolysis to the//,7-unsaturated ketone was straightforward, but conjugation to give 61, required a delicate touch to avoid partial loss of the protecting groups. This was achieved by using anhydrous HC1 in THE thereby affording a total yield of 71% over the three steps. OBz -

'",,

1. L-selectride

.....

H

2. MOMCI,/-Pr2NEt, DMAP

'",,

OTBDMS ..... H

-

3. NaOEt, EtOH O

Li, NH3, tBuOH

4. TBDMSOTf, 2,6-1utidine 52

OMe

MOMO'"

OMe

55%

59

OTBDMS i

1.AcOH-H20

2. NCCO2Me, Et20

OMe

61

OTBDMS

O

OTBDMS ii111

IIII 2

MOMO"

+ 121

1. Li, NH3, tBuOH

MOMO"

2. anhyd. HCI, THF 60

OTBDMS ,,,H

H

',,,

~

M O M O ' " ~

O

OTBDMS

li1

CO2Me +

MOMO '

Iq OH 63 (19%)

64 (13%)

121 _ O (14%) 62 CO2Me /

I

Nail, Mel, THF / (65%)

OT,O

KtOBu,Mel / (22% net)

S

MOMO"'

OT,O.S

O2Me v

- v

Iq

65

~

"O

MOMO' Iq

-

O

66 CO2Me

SCHEME 14

Reductive acylation, however, was really disappointing, affording a mixture of 62 (desired product), 63 and 64, all in poor yields. Moreover,

48

KELLY A. FAIRWEATHER, SIMON R. CRABTREE, AND LEWIS N. MANDER

although 63 behaved reasonably well when subjected to a practice Cmethylation (to give 65), C-methylation of 62 was disastrous (15% yield; 22% after recovery of starting material). We had honed this methodology on numerous similar substrates, 28 but were unable to identify where the problems lie, and so we regrouped, omitting the acylation step in the reduction of 61 and optimizing the yield of 64 into which we proposed to selectively introduce a A 4 alkene bond through Saegusa methodology, 29 thereby blocking C-5 enolate formation and effecting acylation at C-7 as outlined in Scheme 15. Enolization was not all that regioselective, but the "wrong" product (61) could be separated and recycled, while good yields were now obtained for both acylation and alkylation. The stereochemistry could again be checked by NOE difference spectra that confirmed that C-methylation had occurred on the upper face to afford 68 with the desired configuration at C-7. OTBDMS

OTBDMS

'

LDA

MOMO'"

,,,,

,

2

OTBDMS

.....

",,,

TMSCI

:_

....

+

MoMo" VHV 64

"O

v-

v

"OTMS

v-

H

1 92

2 steps 88% combined yield

OTBDMS

TBSO ,,,,

-

M )

~

N

MOMO" O

v - ~ ~O 121 68 CO2Me

"OTMS

Pd(OAc)2, p-BQ M e C N ,77%

H OTBDMS

....

= --~

MO

-_

v H

,,,,

1. LDA, HMPA; NCCO2Me 2. Nail, Mel 50%

=

MOMO'"

+ 61 v - v 67 IZI 7

"O

SCHEME 15

We were now close to checking out the Michael reaction and, anticipating problems with intramolecular addition of the side chain hydroxyl to the A 4 e n o n e if we removed the TBDMS group (cf 34~35), we reduced enone 68 to the alcohol 3~ and treated this product with TBAE Oxidation then afforded dione 69, which was treated with a variety of bases and acids under a range of conditions in an attempt to close the A-ring (Scheme 16). In sharp contrast to our earlier experience with dione 37, however, we were unable to induce cyclization. An inspection of

2

49

TOTALSYNTHESIS OF MARINE DITERPENOID, DIISOCYANOADOCIANE

molecular models indicated that the Michael reaction was unlikely to take place with 69 since it would require the D-ring to adopt a boat conformation to bring the participating functions within bonding distance, and even then orbital overlap between the reacting centers would be poor. It appeared that an e-configured side chain could assume the necessary geometry, however. When treatment of 69 with a variety of bases returned starting material, it seemed that the side chain was more stable in the/% configuration, despite its axial nature. Nevertheless, one could expect there to be a sufficient amount of the c~-epimer in equilibrium to undergo the desired cyclization. In desperation, we recalled an intramolecular Michael reaction described by Corey et al. during their synthesis of longifolene. 31 They employed Et3N-ethanediol in a sealed tube at 225 ~ for 24 h. By adopting this protocol, but at more moderate conditions (150 ~ we were relieved to see cyclization with 70, the structure being confirmed by X-ray crystallography. 32 Nevertheless, the yield of only 30% was obviously unacceptable. OTBDMS H_ .... H

.....

O

"'" ,'

1. NaBH4, CeCI 3 2. TBAF, 65 ~ "

MOMO'"

_ 68

O

3. DMP54%

H

MOMO''

_

121 CO2M e

69

I'llt (CH2OH)2, NEt3 150 ~ 20 h

H

O

121 CO2Me

O

H

",H

MOMO"'

30% -

-

O

70 H CO2Me SCHEME 16

There were several byproducts, among which was one that had lost the ester function and so we considered a reordering of the reaction steps that would introduce the ester group after the Michael reaction. The new approach is outlined in Scheme 17. Conversion of 67 into 72 followed the same routine as for the preparation of 69 from 68. Now, the simpler dione underwent cyclization to 73 in a yield of 68% employing the same conditions as for 69. The methylene groups flanking the C(6) carbonyl group

50

KELLY A. FAIRWEATHER, SIMON R. CRABTREE, AND LEWIS N. MANDER

in 73 are sterically equivalent except for the methyl group attached to C-3. Nevertheless, enolization with LDA followed by reaction with methyl cyanoformate 33 furnished a 9:1 mixture of the desired 75 with 74, albeit in modest yield. The two keto esters were inseparable, but on treatment with KzCO 3 in methanol, the ester substituent 74 epimerized, after which separation was possible. It was of interest that the axially substituted 77 was stable relative to 74 and the enol tautomer. Presumably, the C-3 methyl substituent destabilizes the latter two compounds because of the peri-interaction. This alternative approach came to a grinding halt, however, when once again what should have been a routine C-methylation procedure failed when applied to 75. OTBDMS 9

:

....

:

O

,,,,, 1 9 NaBH4. CeCl 3 ..

v67

OH

,,,H

v

DMP =

"0

v-

12t 7

v

"" '

-OH

H

v-

71 ~

72

v

"0

121

Et3N, (CH2OH)2 / 150 ~ O

o

O

,,,,

,,,,,

,,,, 9

,,,,' LDA, HMPA

',H

MOMO"'

+ v

- ~f_

75

H

CO2Me

MOMO"'

"O

/

-

CO2Me

.... i MOMO"

75:74, 9:1

~'O

0 73

74

0 ,,,,

NCCO2Me, 40%

0 ' ,,,,

9

MOMO'"

CO2Me

MOMO" v 76

: ~ "O H CO2Me

"O 77

S C H E M E 17

Such are the subtleties of organic synthesis that we could conjecture that the sequence beginning with 52 might be better behaved with a similar starting material, namely 57. And so, once more we altered direction with a final attempt to access a suitable tetracycle, this time returning almost to the beginning with a view to improving a number of the critical steps and streamlining some of the sequences.

2

51

TOTAL SYNTHESIS OF MARINE DITERPENOID, DIISOCYANOADOCIANE

VI. Improved Synthesis of an Advanced Tetracyclic Intermediate

The first part of the synthetic sequence to be improved was the preparation of ketone 45. We had in mind for some time to investigate the possibility of preparing this intermediate directly by means of a FriedelCrafts acylation (Scheme 18). Our first attempt using standard reagents at room temperature was a disaster. Nevertheless, it became clear that the acylation had taken place, but then cyclization had occurred to afford 78. The problem was easily rectified by conducting the reaction at - 7 8 ~ for 20 min, whereby ketone 45 was obtained in 75% y i e l d - a considerable improvement in time, effort, and yield over the original route via aldehyde 24.

EtC(O)CI, AICI3 25 ~ MeO2C 78

OMe

OMe 13 EtC(O)CI, AICI 3

-78 ~

MeO 45

OMe

SCHEME 18

The original sequence was then followed through to enol ether 53, but then we had to deal with the cyclopropanation reaction that had proven to be so sluggish. Fortunately, we discovered a report by Shi e t al. 34 who described the preparation of a modified Simmons-Smith reagent, CF3COzZnCHzI prepared from diiodomethane, diethylzinc, and trifluoroacetic acid. This reagent worked brilliantly and allowed the preparation of acid 79 in 80% yield over the three steps from enol ether 53. The application of the methodology that had previously been established using 59 as a model substrate, then took us through to the Michael stage without incident and with a few significant improvements, as outlined in

KELLYA.FAIRWEATHER,SIMONR. CRABTREE,ANDLEWISN. MANDER

52

Scheme 19. In particular, the desired regiochemistry of enol silyl ether formation (20:1) was considerably enhanced by using TMS triflate with 2,6-1utidine, while acylation of 82 to give 83 and its alkylation to give 84 afforded enhanced yields for both steps. OH 9,

:

H OTBDMS

.... .

1.2,6-1utidine, HO2C"'

TBDMSOTf = OMe 2. K2003, 57

.

.

.~,

.

H

:

1. Li, NH 3, t-BuOH

H2

A

3. anhyd. HCI

H

2. CH2N 2

00 8

3.2,6-1utidine TMSOTf

74%

OMe

79

1. Li, NH3, t-BuOH

HO2C""

2. AcOH-H20

HO2

MeOH 81%

OTBDMS . . . . . H ~H 9

',,,

.

.

.

.

H OTBDMS ~ .

Pd(OAc)2, p-BQ

H

79%

MeO2C,"

76%

81

OTMS

ILl

(20 9 1 ratio of regioisomers) OTBDMS _H ,,,H

9

, LDA, HMPA

MeO2C""

H '

OTBDMS .... H H

Nail, Mel

NCCO2Me MeO2

88%

83% O 82

9

-

12t

H

,,,

83

H

OTBDMS .... H

-

O

-,

CO2Me H O

..

H 'L ~ ~ ~ MeO2C,,

1. NaBH 4, CeCI 3 2. TBAF, 65 ~ 3. DMP

-

84

' MeO2C'"

H

54%

O

CO2Me

85

CO2M e

SCHEME19 We were pleasantly surprised to find that the Michael addition (Scheme 20) proceeded at a reasonable rate at a moderate temperature (70 ~ and in reasonable yield. Unfortunately, ester exchange with the ethanediol solvent resulted in a mixture of the methyl ester 86 (for which an X-ray crystal structure provided confirmation of structure) 32 and the hydroxyethyl ester 87. We also obtained dione 88 in which ester exchange had also occurred. Nevertheless, given that the ester functions

2

53

TOTAL SYNTHESIS OF MARINE DITERPENOID, DIISOCYANOADOCIANE

would soon be hydrolyzed, this outcome was more of an irritation than anything else. H

9

O

H

MeO2C" O 85 12t (~O2Me _

Et3 N, ethanediol 70 ~

f

0

0

H MeO2C'

'

0

H + C,,~A MeO2 ,"

',H

0 86, 30% A (~02Me

H " ,H

+

H C ~ H MeO2

- _ 0 H : 87, 28% 0~=.,.0...,,~ / HO

v - "-4/ ~-O I:t -: 88, 14% 0~=,,.0,.,..,~ /

HO

SCHEME 20

VII. Deoxygenation, Curtius Rearrangement, and Completion of the Synthesis

The penultimate phase of the synthesis required deoxygenation of the two ketone functions. We investigated the use of a modified Wolff-Kishner reaction, 35 but application to a model /%keto ester was very discouraging. Sodium cyanoborohydride reduction of the toluenesulfonyl hydrazone 36 of fl-keto ester 89 appeared promising (Scheme 21), but when applied to 86, we obtained a miserable 5% yield. H

H

1. TsNH-NH2 _-. [ ~ ~

- _ O 2. NaBH3CN H ZnCi 2 89 CO2 Me 66% SCHEME 21

90

121 CO2Me

54

KELLY A. FAIRWEATHER, SIMON R. CRABTREE, AND LEWIS N. MANDER

H

H

MeO2C"

Me02(~ 91

C02Me

92 O ~ , . . , . o ~ O M O M

FIGURE 6. Structures of deoxygenated esters.

Clearly, direct reduction did not appear to have much of a future, so we turned to the Barton-McCombie protocol. 37 A practice run on the model keto ester 89 (Scheme 22) was very encouraging and when applied to 86 and the MOM ether of 87, reasonable overall yields of the deoxygenated products 91 and 92 (Figure 6), respectively, were obtained. H ~ - _ 0 H 89 CO2Me

H I"NaBH4' = ~ cecI3 2. NaHMDS; CS2; Mel

.,EL S _ _ 0-- -s j H CO2Me

H n-Bu3SnH -~ ~ AIBN - 90 H CO2Me (70% over 3 steps)

SCHEME 22

The C-7 equatorial ester functions in both of these products was readily hydrolyzed by LiOH in aqueous methanol, but the axial ester group at C-20 was resistant to simple base-catalyzed hydrolysis. We therefore turned to the Johnson-Bartlett procedure (n-PrSLi/HMPA) 38 which afforded an excellent yield of the dicarboxylic acid 93. A standard Curtius protocol then afforded diamine 96 as illustrated in Scheme 23. The mass spectrum of 96 was the same as that obtained for the natural material, but when we ran the 1H-NMR spectrum of this material dissolved in CDC13 we found it to have significant differences with the spectrum of the material obtained from the natural product. 39 How could this be? Had we made a mistake over the assignment of stereochemistry along the line? This did not seem likely given that the X-ray crystal structure had confirmed the constitution of the diester 86. We therefore suspected that the culprit was a trace of HC1 in the CDC13 and so we compared the spectra of the TFA salts (using d4-methanol as the solvent). This time, the spectra were almost a match, but there was a small difference with just one multiplet.

2

55

TOTAL SYNTHESIS OF MARINE DITERPENOID, DIISOCYANOADOCIANE

Could we cross our hearts and say that they were identical? We were aware that the comparison spectrum had been run at a different concentration, however, and when this was adjusted to be the same as that of the synthetic sample, identical spectra were obtained. Phew! H

H Ill Ill

"

H

HO2(~ 931:'1

i l l II

IIii I

'

1 (COOl)2, DCM 9 ,,, 2. NaN 3, THF

',H

N3

:

-

CO2H

94

_

H O//",.. N3

toluene, reflux I

H

H

iiii I

to,uene, cone 4

H2N

,~o

i,'"

i

39% over 4 steps

N II

96

H

-

O

NH2

95

H

-

N-C=O

SCHEME 23

Since 96 had been previously converted back to the target diisocyanoadociane (10) by dehydration of the derived diformamide 97 (Scheme 24), 4~ the total synthesis of 10 was complete in a formal sense. H

H ,H

H2N'"

t-BuOCO2CH=O

Et3N 96121

:

NH2

H p-TsCl

= HC(O)I~H

pyr " _

97 ~ ~HCH=O

-

10 121 NC

SCHEME 24 VIII. Conclusion

With this synthesis we have demonstrated the considerable utility that benzenoid building blocks provide for the synthesis of polycyclic

56

KELLY A. FAIRWEATHER, SIMON R. CRABTREE, AND LEWIS N. MANDER

molecules. 4~ At the beginning of the synthesis we released an array of latent functionality that allowed the expeditious assembly of a hydrophenanthrene intermediate in which one alkene function could be used to attach the elements of the fourth ring (to complete the hydropyrene skeleton), while the other alkene group could be oxygenated to allow the elaboration of one of the quaternary centers. The second benzenoid unit was essentially inert while it was being carried through an extended sequence, until being ultimately transformed into a cyclohexenone moiety. However, we encountered repeated barriers to progress that leave us still puzzled, since most of the methodology was tested and found to work well on model substrates. In particular, we had difficulties with the conversion of Birch reduction products into ~,//-unsaturated cyclohexenones, the low yields from reductive acylations of enones, and the failure of fl-ketoesters to undergo simple alkylation reactions. Clearly, the equilibrium between the isomeric cyclohexenones is finely balanced and even the very small structural difference between our intermediates and steroids (which behave well) is enough to tip the balance in favor of the wrong isomer. Metal-ammonia reductions of cyclohexenones with capture of the enolates by electrophiles is often technically difficult and yields tend to be modest at best. 42 The sequence beginning with 83 in Scheme 19 is much more reliable and should have general applicability. Similarly, the cyclopropyl based method for the C-methylation of aldehydes (as in Scheme 13, but with the modified reagent CF3COzZnCHzI ) is a robust and effective alternative. 43 The failures with the classical Cmethylation of several /%ketoesters remain a puzzle. Finally, we had expected problems with the pivotal Michael reaction, since the orbital line up is not as good as we would have wished. The basic idea was fine, however, and we did obtain a good yield with the simplest of the examples, namely 72 --->73. IX. Epilogue

Organic chemists have made enormous progress over the past four decades in developing new methods and strategies for the construction of increasingly complicated molecules. However, the efficiency, reliability, and predictability of the methodology still leaves much to be desired as confirmed by the twists and turns in the present synthesis. The assembly of polycyclic molecules is especially challenging, since extended, linear routes tend to be necessary, resulting in an inexorable decline in the

2

TOTALSYNTHESIS OF MARINEDITERPENOID,DIISOCYANOADOCIANE

57

quantities of intermediates. It is therefore important that we continue to attempt difficult syntheses and in this way define the scope and limitations of our chosen procedures, as well as develop new strategies and methodology. Apart from some very notable exceptions, progress has been and will continue to be incremental, and relies on the collective efforts of a large number of synthesis groups beavering away and sharing information. This synthesis and the knowledge gained in its execution is offered as one such minor contribution to the common enterprise. Acknowledgments

We are indebted to Professor Mary Garson (University of Queensland) for the provi' sion of authentic samples of diisocyanoadociane (10) and diamine 96; to Dr Jamie Simpson (Monash University) for helpful advice; to Bruce Twitchin and Tony Herlt (ANU) for technical assistance; to Tony Willis and Alison Edwards (ANU) for X-ray studies; and to Chris Blake (ANU) for assistance with high field NMR spectra. References and Footnotes

1. Chang, W. C., Progress in the Chemistry of Natural Products, Vol. 80, Springer, Wien: New York, 2000. 2. The numbering system for "isocycloamphilectane" (4) derived from its presumed biosynthetic relationship to structures 2 and 3, and will be used for intermediates throughout this chapter. 3. Baker, J. T., Wells, R. J., Oberh~insli, W. E., Hawes, G. B., J. Am. Chem. Soc. 1976, 98, 4010. 4. Kazlauskas, R., Murphy, E T., Wells, R. J., Tetrahedron Lett. 1980, 21, 315. 5. K6nig, G. M., Wright, A. D., Angerhofer, C. K., J. Org. Chem. 1996, 61, 3259. 6. Wright, A. D., Wang, H., Gurrath, M., K6nig, G. M., Kocak, G., Newmann, G., Loria, E, Florey, M., Tilley, L., J. Med. Chem. 2001, 44, 873. 7. Corey, E. J., Magriotis, E A., J. Am. Chem. Soc. 1987, 109, 287. 8. Fairweather, K. A., Mander, L. N., Org. Lett. 2006, 8, 3395. 9. Cossey, A. L., Gunter, M. J., Mander, L. N., Tetrahedron Lett. 1980, 21, 3309. 10. Piers, E., Llinas-Brunet, M., J. Org. Chem. 1989, 54, 1483. 11. Schultz, A. G., Macielag, M., Podhorez, D. E., Suhadolnik, J. C., Kullnig, R. K., J. Org. Chem. 1988, 53, 2456. 12. Stork, G., Darling, S. D., J. Am. Chem. Soc. 1960, 82, 1512. 13. Evans, D. A., Stereoselective Alkylation Reactions of Chiral Metal Enolates. In: Asymmetric Synthesis, Vol. 3, Morrison, J. D. (Ed.), Academic Press: New York, 1984, p. 2. 14. Dalcanale, E., Montanari, F., J. Org. Chem. 1986, 51,567. 15. Hook, J. M., Mander, L. N., Woolias, M., Tetrahedron Lett. 1982, 23, 1095. 16. Crabtree, S. R., Chu, W. L. A., Mander, L. N., Synlett 1990, 169. 17. Sharpless, K. B., Lauer, R. E, Teranishi, A. Y., J. Am. Chem. Soc. 1973, 95, 6137. 18. Tietze, L. E, Chem. Rev. 1996, 96, 115.

58

KELLYA. FAIRWEATHER,SIMONR. CRABTREE,AND LEWIS N. MANDER

19. Dess, D. B., Martin, J. C., J. Org. Chem. 1983, 48, 4155. 20. The X-ray structure was solved by Dr Alison Edwards, whose expertise we gratefully acknowledge. 21. Ireland, R. E., Mander, L. N., J. Org. Chem. 1969, 34, 142. Ziegler, E E., Kloek, J. A., Tetrahedron 1977, 33, 373. 22. Charette, A. B., Beauchemin, A., Org. React. 2001, 58, 1. 23. Suda, M., Synthesis 1981, 714. 24. Wenkert, E., Acc. Chem. Res. 1980, 13, 27. 25. Cotsaris, E., Paddon-Row, M. N., J. Chem. Soc., Perkin Trans. 2 1984, 1487. 26. Greeves, N. In: Comprehensive Organic Synthesis, Vol. 8. Trost, B. M. (Ed.), Pergamon Press: Oxford, 1991, pp. 1-24. 27. Mander, L. N., Partial Reduction of Aromatic Rings by Dissolving Metals and by other Methods. In: Comprehensive Organic Synthesis, Vol. 8. Trost, B. M. (Ed.), Pergamon Press: Oxford, 1991, pp. 489-521. 28. Fairweather, K. A., Ph.D. Dissertation, Australian National University, 2006. 29. Ito, Y., Hirao, T., Saegusa, T., J. Org. Chem. 1978, 43, 1011. 30. Luche, J. L., J. Am. Chem. Soc. 1978, 100, 2226. 31. Corey, E. J., Ohno, M., Mitra, R. B., Vatakencherry, P. A., J. Am. Chem. Soc. 1964, 86, 478. 32. The X-ray structure was solved by Tony Willis, whose expertise we gratefully acknowledge. 33. Mander, L. N., Sethi, P. A., Tetrahedron Lett. 1983, 24, 5425. 34. Yang, Z., Lorenz, J. C., Shi, Y., Tetrahedron Lett. 1998, 39, 8621. 35. Furrow, M. E., Myers, A. G., J. Am. Chem. Soc. 2004, 126, 5436. 36. Hutchins, R. O., Maryanoff, B. E., Milewski, C. A., J. Am. Chem. Soc. 1971, 93, 1793. 37. Barton, D. H. R., McCombie, S. W., J. Chem. Soc., Perkin Trans. 1 1975, 1574. 38. Bartlett, P. A., Johnson, W. S., Tetrahedron Lett. 1970, 4459. 39. The authentic sample was provided by Professor Mary Garson. 40. Simpson, J. S., Garson, M. J., Org. Biomol. Chem. 2004, 2, 939. 41. Mander, L. N., Synlett, 1991, 134. 42. Stork, G., Rosen, P., Goldman, N. L., J. Am. Chem. Soc. 1961, 83, 2965. 43. A more direct strategy for elaborating a primary amino group at a quaternary center by adding tosyl nitrene to a methylene group has been reported by Wood (White, R. D., Keany, G. E, Slown, C. D., Wood, J. L., Org. Lett. 2004, 6, 1123). As applied to our substrate, however, we would have expected this method to afford the undesired C-20 epimer.

STRATEGIES AND TACTICS IN ORGANIC SYNTHESIS, VOL. 7 9 2008 Elsevier Ltd. All rights reserved.

Chapter 3 TOTAL SYNTHESES OF ZOAPATANOL Janine Cossy, Vdronique Bellosta, and Catherine Taillier Laboratoire de Chimie Organique, ESPCI, CNRS 10 rue Vauquelin, 75231 Paris Cedex 05, France

I. II. III. IV. V. VI. VII. VIII.

Introduction Nicolaou's Synthesis Chen's Synthesis Cookson's Synthesis Kocienski's Synthesis Kane's Synthesis Trost's Synthesis Our Approaches for the Total Synthesis of (+)-Zoapatanol A. Ring-Closing Metathesis Approach B. Homer-Wadsworth-Emmons Approach IX. Conclusion References and Footnotes

59 60 64 67 69 73 78 83 83 90 95 95

I. Introduction

(+)-Zoapatanol 1, montanol 2, tomentanol 3 and tomentol 4 are diterpenoid oxepanes isolated from the leaves of the Mexican zoapatle plant Montanoa tomentosa, which Mexican women have been using for centuries to prepare "tea" to induce menses, labor and to terminate early pregnancy. 1 Recent studies support the belief that zoapatanol and its metabolites might be responsible for the observed antifertility activity. 2 In 1979, the isolation and the structure of zoapatanol were described. 3 Due to its biological profile and its challenging structure, several synthetic approaches have been described 4 and seven total syntheses of zoapatanol have been reported 5-11 but only two of them were enantioselective, l~ Key issues for a successful synthesis of zoapatanol 1 are the stereocontrolled construction of the oxepane ring containing the two stereogenic centers, the introduction of the (E)-exocyclic double bond and the installation of the nonenyl side chain. Since (+)-zoapatanol was isolated as

60

JANINECOSSY,VI~RONIQUEBELLOSTA,AND CATHERINETAILLIER

4p

5p

OR'

HO,,.

O 1 2 3 4

R R R R

= = = =

(CH3)2C=CHCH 2(CH3)2CHC(CH3)=CHH2C=C(CH3)CH(CH3)CH 2(CH3)2C(OH)CH=CH-

R'= R'= R'= R'=

H H H Ac

Zoapatanol Montanol Tomentanol Tomentol

FIGURE 1. Oxepanederivatives isolated from M.

tomentosa.

a 1/1 mixture of epimers at C6, control of this stereocenter is not required (Figure 1).12 In 1980, the first two syntheses of (_+)-zoapatanol were disclosed, one by Nicolaou 6 and the other by Chen and Rowand, 5 and six other syntheses have been disclosed since then. 4,7-11 For all of them, the construction of the oxepane ring was achieved through the formation of the O1'-C7' or O1 '-C2' bonds except for one ~ in which the C4'-C5' bond was formed. II. Nicolaou's Synthesis

A convergent synthesis of (+)-zoapatanol was achieved. 6 In the retrosynthetic analysis, the oxepane ring would be obtained by nucleophilic attack of the tertiary alcohol on the epoxide present in compound I via a 7-exo-tet process that would allow the formation of the O 1 '-C7' bond. The syn-1,2-diol present in compound I would come from a chelation-controlled addition of a methyl Grignard reagent to the ~-benzyloxymethoxy ketone II according to the Cram-chelate model A (Figure 2). ~3 Ketone II would be synthesized from glycidol 5 as the starting material and bromide 10 as the precursor of the long side chain of zoapatanol (Scheme 1). Bromide 10, which would allow the introduction of the side chain of zoapatanol, was synthesized from 5-hydroxy-2-pentanone. By reaction with an excess of (methoxymethyl)triphenylphosphorane, ketone 6 was transformed in 75% yield to methoxy enol ether 7, which was directly and quantitatively converted to dithiane 8 [HS(CHz)3SH, HC1 gas, CHC13, 0-25 ~ Alkylation of the dianion derived from 8 (2.2 equiv, of n-BuLi, THF, - 7 8 ~ ---> - 1 5 ~ with 1-bromo-3-methyl-2-butene (1.1 equiv., - 7 8 ~ ---> - 1 5 ~ led to 9 (85% yield), which was converted to 10 in five steps, with a 60% overall yield, by using classical transformations: acetylation (Ac20, Pyr, DMAP, CHzC12, 0 ~ followed by removal of the

3

61

TOTALSYNTHESESOF ZOAPATANOL

Nu~% ~.O,,,,

Ph O.si...t_gu

O---M

Ph

CH2Ph A

FIGURE 2. Cram-chelate model.

CH2OH

O/-Ph k_O,,,

0 (+_)-Zoapatanol 1

OH O

I

0 ~~ ~ ~ 0

Br

/--Ph

OSi(Ph)2t-Bu

10

k~oH II

5

SCHEME 1. The Nicolaou strategy.

dithiane by using HgC12 (CaCO 3, MeCN/H20, reflux), formation of the dioxolane (HOCHzCHzOH , TsOH, benzene, reflux) then reduction of the acetate using LiA1H4 (ether, 0 ~ Finally, transformation of the resulting alcohol by using CBr 4 in the presence of PPh 3 ( - 4 0 ~ ~ 0 ~ produced the desired bromide 10 (Scheme 2). The construction of the oxepanic fragment began with the preparation of the protected aldehyde 11 from glycidol 5 employing classical chemistry (Scheme 3). Incorporation of the side chain of zoapatanol was accomplished by the coupling of the Grignard reagent derived from bromide 10 (Mg, THF, 25 ~ with aldehyde 11 at - 7 8 ~ providing a secondary alcohol (80% yield), which was oxidized to ketone 12 (Collins' reagent, 0 ~ CH2C12). This ketone was then treated with MeMgC1 (THE -105 ~ to afford the Cram-chelate-derived syn-1,2-diol 13, as the only detectable product (Scheme 3 and Figure 2).

62

JANINECOSSY,Vt~RONIQUEBELLOSTA,ANDCATHERINETAILLIER MeO

~

PPh3 M e O ~ O H

75%

7

6

quant.

I

.s/

sH

HCI (g)

r%

1. n-BuLi

S

~

O

H

85%

1. Ac20, Pyr, 4-DMAP 2. HgCI2, CaCO3, MeCN/H20 3. HOCH2CH2OH,TsOH

@

V

~

4. LiAIH4, Et20 5. CBr4, PPh3 60%

B

r

10

SCHEME 2. Synthesisof the side-chain precursor.

After deprotection of the silyl ether (n-Bu4NF), a chemoselective epoxidation of the allylic alcohol in the presence of the unactivated olefin was performed with tert-butyl hydroperoxide in the presence of vanadyl bisacetoacetate [VO(acac) 2] to produce the epoxy alcohol 14 (80%). ~4 This selectivity is due to the complexation of the vanadium derivative [VO(acac)2] by the free hydroxyl group of the allylic alcohol and by the oxidative reagent t-BuOOH, according to a six-membered ring pseudochair transition state of type B (Scheme 3). The construction of the oxepane ring by an internal epoxide ring-opening was realized with a high degree of regioselectivity by subjecting epoxide 14 to KCH2SOCH 3 in MezSO, leading to the desired oxepane 15 (75%). This latter compound was then transformed to the key ketone 16 by oxidative cleavage of the 1,2-diol using NaIO 4 via intermediate C (95%) (Scheme 3). The oxepanone 16 then underwent a condensation with the lithium salt of

3

63

TOTAL SYNTHESES OF ZOAPATANOL

1. DHP, H+

~ H

.OBn r/ l O"]""~'~

~

OH L__/

~~

3. TBDPSCl 4. PhCH2OCH2Cl,iPr2NEt 5. AcOH, THF/H20 6. Pyr.SO3 56%

5

|

r . . . o v Ph

~-~o., n-~u,i

OTBDPS

I 1.10, Ug, THF 2. CrO3, Pyr OBn

MeMgBr OTBDPS EHF,-100~ "

r" l~

i.

~

95%

~176 0

13

~

" 12

L I o~Vi ' ',L R~.~O-O-t-Bu H

1. n-Bu4NF 2. t-BuOOH VO(acac)2

80%

BnO

BnO k---O,,,

OH KCH2SOCH3 DMSO 75%

~J

14

OH

15

OH

l'-/4-~- 11ol" "~

95%

NalO4 EtOH/H20

C

| ~

BnO __ k--O'"~k/-~ --'J

~ O

L__/

CO2Me "1 (MeO,2P(O)CH2CO2Me| n-BuLi, THE - "

17

BnO ~ ~ - - O ' " ~ ~ :~O ~ ~ 0 ''J k_J

16

E / Z = 2.5/1 /

1. DIBAL-H 1 70% 2. Separation BnO ~--O . . . .

CH2OH 1. Li/NH3(liq) 2. AcOH/THF/H20

k~/

18

=

(+)-Zoapatanol1

80% SCHEME 3. Completion of the Nicolaou synthesis.

64

JANINE COSSY, VI~RONIQUE BELLOSTA, AND CATHERINE TAILLIER

trimethylphosphonoacetate (LDA, - 2 0 ~ to afford selectively the unsaturated methyl ester 17 as a 2.5/1 mixture of geometric isomers. After reduction of this mixture of isomers by DIBAL-H, the corresponding allylic alcohols were obtained as a mixture of E/Z-isomers (2.5/1) that were separated by chromatography to provide the desired stereoisomer 18 of E-configuration. Finally, deprotection of the secondary alcohol of the obtained oxepane 18 under Birch conditions [Li~H3(liq), - 7 8 ~ according to the mechanism presented in Scheme 4, followed by acidic treatment (AcOH, THF/H20) to release the carbonyl function, furnished ( +)-zoapatanol.

RCH2O_ C H 2 ~

f--"~e- ,. RCH2O~-CH2~--~,

1 RCH2OH =H20 RCH2O@ + H2C==~" eH3C

H20 SCHEME 4. Birch reduction of a benzyl group.

The Nicolaou total synthesis of racemic zoapatanol required 16 steps, and was accomplished in an overall yield of 12%. III.

Chen's Synthesis

Instead of using basic conditions to build the oxepane ring from an co-hydroxy epoxide, Chen and Rowand 5 chose to utilize acidic conditions to transform an e,~-epoxy alcohol to an oxepane by creating the O1'-C2' bond of (_)-zoapatanol. The synthesis of (_)-zoapatanol was envisioned starting from the ~,~-epoxy alcohol III. This compound would be prepared by selective epoxidation of 2-methyl-6-methylene-(E)-2,7-octadien-l-ol 19, which could be synthesized from myrcene (Scheme 5).

3

65

TOTALSYNTHESESOF ZOAPATANOL OH

~

(_+)-Zoapatanol1

RO;"~CH3LOH III

H

O

~

Myrcene

'

19

SCHEME5. The Chen strategy.

The regioselective oxidation of myrcene by SeO 2 (Scheme 6) produced the allylic alcohol 19, ~5 which after treatment with m-CPBA was transformed to the epoxy alcohol 20. Reaction of 20 with bromine gave an allylic dibromide (Scheme 7, via intermediate D), which was subjected to potassium acetate to give a diacetate intermediate. This latter was then tosylated to afford, after treatment under basic conditions (KzCO 3, CH3OH/H20 ), the pure diol 21. Reaction of 21 with the Grignard reagent 22 (3.3 equiv.) in the presence of a catalytic amount of LizCuC14 ( - 2 0 ~ ----> 0 ~ gave a complex mixture that was directly treated with acetic anhydride. After purification by chromatography followed by a basic hydrolysis, the key intermediate 23 was isolated in 32% yield. Next, treatment of 23 with trifluoroacetic acid (0.1 equiv.) in methylene chloride afforded

)••,•e=o

ene reaction= ~~e -OH

H~._.,/O

I [2,3]-sigmatropic

rearrangement

,, H20 OH

~/R O-Se-OH

SCHEME6. Allylicoxidationby seleniumdioxide.

1. i) SeO2

ii) NaBH4 He

2. m-CPBA 16%

Myrcene

//i Br2

/~~Br-

20

Br 14%

1. Br2, CH2012,0~ 2. AcOK, Acetone reflux 3. TsCI, Et3N 4. K2003, MeOH/H20

OH

OH TsO ~ 0::i

~k....

:

. THPO ~

23

(''"

O

H

O 22

MgBr

21

2.,,~:~O014

OTHP

1. CF3CO2H CH2CI2, 0 ~

3. K2CO3, MeOH/H20 32%

path b

iOAc

,~ THPO

17% 25

2. Ac20, Pyr

/iOAc

30%

THPO_ ~/

~

"-,,s-.---O 24

24%

1. Jones reagent (CrO3, H2SO4) 2. ~ L i

(+)-Zoapatanol 1 SCHEME 7. Chen's synthesis of (+__)-zoapatanol.

3

67

TOTALSYNTHESES OF ZOAPATANOL

the oxepane 24 as the major product (30% yield), resulting from nucleophilic attack of one of the primary hydroxy groups on the more substituted carbon of the epoxide (Scheme 7, path a). Under these conditions, tetrahydropyrane 25 was also formed in 17% yield, resulting from the nucleophilic attack of the hydroxy group on the less substituted carbon of the epoxide, according to a 6-exo-trig process (Scheme 7, path b). The final transformation of 24 to (_+)-zoapatanol was accomplished in two steps. After a deprotection-oxidation process of the primary alcohol using Jones reagent, the corresponding carboxylic acid was formed and its treatment with an excess of 3-methyl-2-butenyllithium led to (+_)-zoapatanol (24% yield, two steps). This synthesis of racemic zoapatanol spanned 13 steps from myrcene and gave an overall yield of 0.15%. A similar approach to Chen's strategy was used by Cookson et al. and Kocienski et al. to build up the oxepane ring of ( +)-zoapatanol as the O 1'-C2' bond of the oxepane was formed by treatment of an e,~-epoxy alcohol under acidic conditions. IV. Cookson's Synthesis

Cookson et al. have envisaged the synthesis of the key intermediate e,~epoxy alcohol IV from the homoallylic alcohol V, which would come from a methylalumination of the acetylenic compound VII, followed by functionalization of intermediate VI. The acetylenic compound VII would be obtained by alkylation of the enolate resulting from deprotonation of tert-butyl propionate (Scheme 8). //---x~

(+_)-Zoapatanol 1

", R O ~ c H 3 ~ j O H IV

VI

OH3

)

t-BuO2C VII SCHEME 8. The Cookson strategy.

oH

68

JANINE COSSY, VI~RONIQUE BELLOSTA, AND CATHERINE TAILLIER

The synthesis of (_+)-zoapatanol started with the alkylation of the enolate of tert-butyl propionate with 5-iodopent-1-yne. The straightforward reduction of the obtained acetylenic ester with LiA1H4 and subsequent benzylation of the resulting alcohol gave the terminal acetylenic compound 27 (68%). The transformation of the acetylenic compound 27 to the homoallylic alcohol 28 was achieved by methylalumination with trimethylaluminium in the presence of bis(cyclopentadienyl)zirconium dichloride (CpzZrCI2) (Scheme 9). 16The conversion of the alane intermediate of type VI into the more reactive aluminate with n-BuLi, followed by quenching with ethylene oxide allowed the formation of the desired homoallylic alcohol 28 with an overall yield of 62%. ~7 One proposed mechanism for the zirconium-catalyzed carboalumination involves the methylenation of CpzZrC12 with Me3A1 producing MeZrCpzCI and MezA1C1, followed by methylzirconation of the alkyne. After transmetalation of the resulting alkenylzirconium derivative with MezA1C1, the alkenyldimethylalane is formed and CpzZrC12 is regenerated. 18

Me3AI + Cp2ZrCI 2

_

-

Me2A I -

,,CI. ~ ZrCP2Me CI

R

R'

Me/~-~Z,rCP2 Cl ,Cl ~,[Me 2 R

1

IR R~-~R' = '+8 -5 Me--ZrCP2--CI- -AIMe2CI

m

R'

i

R'

+ CP2ZrCI 2 AIMe 2 SCHEME 9. Proposed mechanism for the Zr-catalyzed carboalumination.

Compound 28 was then transformed to the e,~-epoxy alcohol 31 in seven steps using classical chemistry. After mesylation of 28 (MsC1, Et3N), nucleophilic displacement of the mesylate using NaI in acetone, the corresponding iodide derivative was produced. Alkylation of the lithium anion of diethyl 2-triphenylphosphoranylidene butadienoate 29 by the previously synthesized iodide under basic conditions (LDA) afforded, after benzoic acid-catalyzed elimination of triphenylphosphine, the E-diester 30 in 63% yield. Reduction of 30 with DIBAL-H resulted in the formation of the corresponding unsaturated diol (97% yield). After acetylation of this diol

3

TOTAL SYNTHESES OF ZOAPATANOL

69

(Ac20, Et3N), epoxidation of the more electron-rich olefin with m-CPBA 19 and saponification with K2CO3 in MeOH, the desired e,~-epoxy alcohol 31 was isolated in 88% yield over four steps. This latter compound cyclized when treated with SnC14 in THF leading to the oxepane derivative 32 in 79% yield and to the corresponding tetrahydropyran derivative 33 in 7% yield. The epoxide ring-opening had presumably occurred stereospecifically with inversion of configuration. We have to point out that the use of SnC14 is crucial for the success of this cyclization and appears to be much better than CF3COzH, previously used by Chen et al., in terms of regioselectivity, as the oxepane derivative is formed preferentially to the pyrane derivative. The end of the synthesis of (_+)-zoapatanol from 32 was straightforward. After protection of the hydroxy groups as tert-butyldimethylsilyl ethers (TBSOTf, 2,6-1utidine), selective cleavage of the benzyl ether using Birch conditions [Li/NH3(liq), -78 ~ followed by oxidation of the resulting primary alcohol with PCC on alumina, and subsequent treatment of the corresponding aldehyde with silver(I) oxide, carboxylic acid 34 was isolated with an overall yield of 61%. Addition of prenyllithium to acid 34 then allowed the introduction of the desired [~,y-unsaturated ketone moiety. Finally, treatment with 25% HF in acetonitrile led to the cleavage of the silyl ethers, producing (+_)-zoapatanol (Scheme 10). Cookson's approach to (_+)-zoapatanol took 19 steps from tert-butyl propionate, and was achieved with an overall yield of 4.9%. V. Kocienski's Synthesis

Kocienski et al. envisaged the synthesis of zoapatanol by forming the O1'-C2' bond according to the same procedure developed by Cookson et al., i.e., the formation of the oxepane ring by SnC14-induced intramolecular cyclization of an ~,~-epoxy alcohol. The synthesis of the ~,~-epoxy ether VIII precursor of (_+)-zoapatanol was envisioned from diene IX, which would be obtained by the ring-opening of dihydrofuran 35. The stereochemistry of the exocyclic double bond would be controlled during the formation of the oxepane ring as the hydroxyl group at C 1' would be the only one to be able to open the oxirane ring (Scheme 11). Construction of the C3-C4 double bond of the key homoallylic ether IX was performed via a Ni(0)-catalyzed coupling between MeMgBr and 2-alkyl dihydrofuran 352o (Scheme 12). At first, the 5-1ithio-2,3dihydrofuran was alkylated with the appropriate alkyl iodide affording 35, which reacted with MeMgBr in the presence of (PPh3)zNiC12 to produce

1. LDA, THF,-78~ then ~ " " ' - ' " ~ I, HMPA

CO2t-Bu

.)

Bn

-78~ to rt 2. LiAIH4, Et20 3. Nail, DMF then PhCH2Br 68%

26

27

1. Me3AI, CP2ZrCI2 CICH2CH2CI, rt 2. n-BuLi then O

62%

/___&

I/OBn

/

CO2Et

1. MsCI, Et3N 2. Nal, Acetone

.OBn

3.C.O2Et 30

CO2Et

28

, ~ P P h 3 , LDA

29 CO2Et

88%

then PhCO2H Phil, reflux 63%

1. DIBAL-H, PhMe -78~ to -30~ 2. AcCI, NEt3 3. m-CPBA 4. K2003, MeOH

.OH

0 ~ ,

,--OH

OH

SnCI4

o:;'

33 (7%)

THF ,. BnO - - /

H O , , , ~

BnO

31

OH

oJ 32 (79%)

60%

1. ~ L i ~)-Zoapatanol 1

1. TBSOTf, 2,6-1utidine 2. Li, NH3(liq) THF, -78~ 3. PCC, AI203, rt 4. AgNO3, NaOH, MeOH OTBS

2. 25% HF, MeCN 44% 34 SCHEME 10. Cookson's synthesis of (_+)-zoapatanol.

3

71

TOTAL SYNTHESES OF ZOAPATANOL

(+)-Zoapatanol 1

>

1/--R OR "~2 1

, O',' VIII

2

35

IX SCHEME 11. The Kocienski strategy.

alcohol 36. The C2-C3 double bond of the key intermediate of type VIII was then introduced by using carbomagnesiation of butyn-l,4dio121 with Grignard reagent 37, which was obtained by standard methods from alcohol 36 (mesylation, transformation to the corresponding bromide by displacement of the mesylate by LiBr, and treatment with magnesium). The high trans-stereoselectivity (>95%) is probably due to an internal coordination of the alkenylmagnesium bromide 38. Protection of the hydroxy groups as acetates (Ac20, Et3N), and treatment of 39 with 1 equiv, of m-CPBA afforded epoxide 40. It is worth mentioning that the reaction was not totally chemoselective, as competitive epoxidation of the disubstituted terminal alkene also took place. After methanolysis of the acetate, opening of epoxy alcohol 41 occurred when this compound was subjected to SnC14 in THF at - 2 0 ~ affording the required oxepane 42. Conversion of the oxepane 42 to (_+)-zoapatanol started with the protection of the hydroxy groups as t-butyldimethylsilyl ethers (TBSOTf, 2,6-1utidine), followed by selective hydroboration of the terminal double bond, and Swern oxidation. Reaction of the aldehyde thus formed with dimethylsulfonium methylide 22 gave epoxide 43 via intermediate E. Nucleophilic cleavage of the resulting epoxide 43 was achieved with the homocuprate derived from 2,2-dimethylvinyllithium and CuI to afford the alcohol 44. Finally, Swern oxidation of the secondary alcohol and removal of the protecting groups with HF produced (+)-zoapatanol. The synthesis of (_+)-zoapatanol developed by Kocienski's group proceeded over 16 steps from dihydrofuran in 6.6% overall yield (Scheme 12).

1. n-BuLi

89%

i

rl sc"Et3

35 MeMgBr 92% (PPh3)2NiCI2 Et20/PhH

MgB ~

/

\

OMgBr Et20

36

OMgBr OMgBr

.--~~ ~ . . ~ / O A c

/

Ac20

OMgBr J

Et3N 50%

~ ~ / # M g B E 38

OH

2. LiBr 3. Mg, Et20 92%

~

OH

~OAc

39

I m-CPBA @ @ . . O R

SnCI4

64%

"ON

42

K2CO3 ~ 40 R = Ac MeOH 41 R=H quant.

L_..

1. TBSOTf, 2,6-1utidine 78% 2.9-BBN-H, H202, NaOH 3. Swern 4. Me2S=CH2

[R'~fH| OTBS

|174 .] ~ J'~S(Me)2

(~S(Me)2

/~2

R

vk_A

CuLi , Et20 -10oC 64%

43

..OTBS

~)-Zoapatanol 1

TBSO,,,/----~~

1. Swern 2. HF/MeCN ~ 73%

\

OH

J //O 44

SCHEME12. Kocienski'ssynthesisof (_+)-zoapatanol.

3

73

TOTALSYNTHESES OF ZOAPATANOL

VI. K a n e ' s

Synthesis

In their retrosynthetic analysis, Kane and coworkers envisaged access to (+_)-zoapatanol from the seven-membered ring lactone X, which would come from a regioselective oxidation of the cx,cx-disubstituted cyclohexanone XI under Baeyer-Villiger conditions. The synthesis of XI was envisaged from the Wieland-Miescher ketone 45 (Scheme 13).

(+)-Zoapatanol 1

) o x

OR'

45

XI SCHEME 13. The Kane strategy.

The synthesis of (_+)-zoapatanol was achieved from the Wieland-Miescher ketone (45), which was transformed to ketal 46 in five steps using classical transformations (Scheme 14). After acidic hydrolysis, the obtained cx,[3-unsaturated ketone 47 was treated with H202/NaOH to produce a mixture of diastereomeric epoxides 48 that were transformed to the monocyclic co-acetylenic ketone 50 via tosylhydrazone 49, by using an Eschenmoser fragmentation under acidic conditions via intermediate F (Scheme 14). 23 Ten steps were then necessary to transform ketone 50 to the key seven-membered ring lactone of type X (Scheme 13). After protection of the carbonyl group of 50 [HO-(CH2)2-OH, TsOH] followed by oxidative hydroboration of the acetylenic moiety using an excess of 9-BBN-H, alcohol 51 was produced as the major product. 24 This latter compound was then oxidized with Collins' reagent (CrO 3, Pyr) and the resulting aldehyde was treated with CH3Li to give a secondary alcohol. After a second oxidation with Collins' reagent, the resulting methyl ketone led to olefin 52 upon treatment with methyltriphenylphosphonium iodide/Nail in DMSO.

O

_OCH2Ph

1. NaBH4

2. Ac20, Pyr 45

I

3. HOCH2CH2OH, Ph, H+ 4. LiAIH4 5. BnBr, PhH/DMSO < 60%

46 I AcOH MeOH

; ?c.2Phl

_OCH2Ph H+

j

H

H

90% 1 _OCH2Ph

_OCH2Ph

H202, NaOH CH3OH 74%

48 I

H p-TsNHNH 2 CH3COOH CH2CI2

+s _OCH2Ph OCH2Ph

83%

Ts

5O

49

V--] 1. HO OH, H+ 80%

HO

2.9-BBN-H NaOH, H202 _OCH2Ph ~ :

1. CrO3, Pyr 2. CH3Li 3. CrO3, Pyr

O_CH2Ph

4. Ph3PCH31 68% 51

SCHEME 14. Preparation of the Kane intermediate.

52

3

75

TOTALSYNTHESESOF ZOAPATANOL

The alkene 52 was converted to a primary alcohol upon oxidative hydroboration (9-BBN-H/NaOH, H202), hydrolysis of the ketal function (H2SO4), and protection of the alcohol as its THP derivative to afford the ketone 53 in 81% yield (Scheme 15). This ~,~-disubstituted ketone was oxidized under Baeyer-Villiger conditions with m-CPBA, to produce the

1.9-BBN-H~ NaOH/H202

ttt

BnO

2. H2SO4, 0.002N 3. DHP, TsOH/EteO "O

O

L._/

52

OTHP

81%

O 53

m-CPBA NaOAc m

BnO ,,,/ T

H

P

O

~

o

\ ~

O

70%

OTHP

Ar ,(~.O O)

54

85%

o

LDA, (EtO)2POCI TMEDA, HMPA,THF ~BnO ,,,/~~ ~

THPO

Na/NH3(liq) HO,, O

r

II

O

O'~-OEt OEt

..•

~174

~"R

%" :

Et

H

55

57%1 H+

.j

THPO

57

~

1. NaH, BnBr 2. BH3then

O -

3. CrO3, Pyr 66%

HO,,,

OTHP 56

SCHEME 15. Kane's approach to the key oxepanone intermediate.

--~

76

JANINE COSSY, VI~RONIQUE BELLOSTA, AND CATHERINE TAILLIER

desired lactone 54. This reaction is highly regioselective as the more electron-rich bond preferentially migrates via a concerted mechanism (intermediate G, Scheme 15).25 In order to transform lactone 54 to the oxepanone 57, an 1,2-transposition of a carbonyl group was needed. This transposition was achieved in five steps. After formation of the lithiated anion of 54 using LDA and trapping of the enolate intermediate by diethyl chlorophosphate, the obtained enol phosphate 55 was reduced, via intermediate H, to an oxepene using an electron transfer process induced by Na in NH3(liq). Under these conditions, a Birch reduction took place and at the same time the benzyl ether was cleaved. The hydroxy group of 56 was then reprotected (Nail, BnBr) and the enol ether was regioselectively hydroborated to produce a secondary alcohol that was then oxidized using CrO 3 in pyridine. The desired oxepanone 57 was thus obtained in 66% yield (Scheme 15). The next stage of the synthesis focused on the elaboration of the side chain. Ketalization of 57 and concomitant deprotection of the primary hydroxy group followed by oxidation with Collins' reagent, produced an aldehyde intermediate. Next, addition of the Grignard reagent generated from 4-bromo-2-methyl-l-butene provided an alcohol that was acetylated to give acetate 59 (Scheme 16). Isomerization of the double bond of 59 under acidic conditions (TsOH, Phil), basic hydrolysis of the acetate (K2CO3,MeOH/H20), THP ether formation and reductive cleavage of the benzyl ether under Birch conditions [Na/NH3(liq), t-BuOH/THF] afforded alcohol 60. Acetylation of the secondary alcohol and nonselective hydrolysis of the ketal moiety (0.002N H2SO4) followed by reprotection of the secondary alcohol afforded the tetrahydropyran ether 61. Completion of the synthesis of ( +)-zoapatanol required the transformation of the ketone at C6' to the exocyclic (E)-2-hydroxyethylidene group and a functional group transformation on the side chain at C5. Thus, compound 61 was condensed onto triethylphosphonoacetate under basic conditions to provide the unsaturated ester 62 as an inseparable mixture of (E)- and (Z)-isomers (Scheme 16). Reduction of 62 with LiA1H4 gave diastereomeric diols (E/Z = 2/3 ratio) that were separated by chromatography to give the desired diol 63 (28% yield). Diol 63 was diacetylated, the tetrahydropyranyl protecting group removed (CH3CO2H/H20/THF) and the alcohol function oxidized (Collins' reagent) to give the corresponding ketone. Finally, treatment with an excess of tetrabutylammonium hydroxide was used to remove the acetate groups thus leading to (+_)-zoapatanol.

",

O

HO HO, TsOH P h H

L

87%

OH

OTHP

B~ nO

"~"--7~

57

58 1. CrO3, Pyr 77%

2 / ~ , - , MgBr

3. Ac20, Pyr

H O , , , / / - ~ O ~] THPO % ~ ~ ~ . -

O'~O. ZsOH,phH

6~0 / .

-2. Kf;flou31MeOH

1. Ac20, Pyr 63% 2. H2SO4(0.002N) 3. DHP,TsOH .

k k

.

BnO,,,~----~O ~

AcO~~~z ~

59

3. DHP/TsOH 4. Na/NH3(liq) t-BuOH 66%

"

CO2Et

',,

AcO,,, ~ ~ ~ - - - - O

61

o>O

(EtO)2F~C;H2CO2Et Nail = 98%

62 28% 1. LiAIH4 2. Separation

(+)-Zoapatanol1

1. Ac20, Pyr 2. CH3CO2H 3. CrO3, Pyr 4. n-Bu4NOH

HO,,, THPO

71%

SCHEME 16. Completion of the Kane synthesis.

,1OH "-~\~./

/-

/ 63

78

JANINE COSSY, VI~RONIQUE BELLOSTA, AND CATHERINE TAILLIER

Kane's racemic synthesis of zoapatanol required 40 steps from Wieland-Miescher ketone and proceeded in 0.1% overall yield. VII. Trost's Synthesis

The first enantioselective synthesis of zoapatanol was reported by Trost et al. in 1994. l~ The key step was an intramolecular Williamson reaction

to build up the oxepane ring. The control of the contiguous stereogenic centers at C2' and C3' was achieved via a Sharpless asymmetric epoxidation of an allylic alcohol using t-BuOOH, chiral diethyl tartrate (DET) and Ti(OiPr)4 .26 The configuration of the epoxide formed is dependent upon the enantiomer of DET used and can be predicted according to Scheme 17. The configuration of the (E)-exocyclic double bond was controlled by using a palladium-catalyzed 1,4-addition of triphenylsilanol to a vinyl epoxide. 27 (-)-DET ,,,-OH R 0 OH

.0. ,-OH R

(+I-BET SCHEME 17. Sharpless asymmetric epoxidation of allylic alcohols.

The synthesis of (+)-zoapatanol was envisioned from alcohol XII by using an intramolecular Williamson cyclization. Compound XII would be obtained from a nucleophilic attack of an alcohol on a ~-allylic palladium complex generated from the vinyl epoxide XIII. This compound would be synthesized from epoxide XIV, which would come from a Sharpless epoxidation applied to the allylic alcohol XV, which in turn would be obtained from methallyl alcohol 64 (Scheme 18). Methallyl alcohol 64 was transformed to allylic alcohol 66 in two steps (Scheme 19). After treatment of 64 with 2.2 equiv, of n-BuLi in the presence of tetramethylethylenediamine (TMEDA), the resulting dianion was quenched at first with 1-bromo-3-methylbut-2-ene and then with methoxymethyl chloride, allowing the isolation of dienic compound 65 (92%). The regioselective oxidation of one of the methyl groups in 65,

3

79

TOTAL SYNTHESES OF ZOAPATANOL

(+)-Zoapatanol 1

) "/

~

~ -~ "OH H3C

LG

XII

11 R'O/~/~OMOM "/ XlV

HO~

~

~ -~ "OH H3C XIII

"•OH

"

64

XV SCHEME 18. The Trost strategy.

achieved by using 4 mol% SeO 2 in the presence of t-BuOOH and 10% salicylic acid led to alcohol 66 in 53% yield. Use of L-(+)-diethyl tartrate [(+)-DET] as the chiral ligand gave epoxide 67 in 90% yield and with an ee greater than 95%. 2s The transformation of this latter epoxide 67 to the terminal epoxide 68 was accomplished sequentially, first by a titanium-promoted regioselective opening of the epoxide 29 with acetic acid as the nucleophile. A tosylation of the primary alcohol of the resulting 1,2-diol followed by treatment under basic conditions were used to build up epoxide 68. This sequence of reactions set the two contiguous stereogenic centers present on the oxepane ring of zoapatanol. In order to introduce the side chain present at C2', epoxide 68 was opened by using a cuprate derived from 3-butenylmagnesium bromide and the resulting product was transformed to diol 69 after acidic hydrolysis. The straightforward conversion of allylic alcohol 69 to the vinyl epoxide 70 was achieved in two steps. The first one was an oxidation with Dess-Martin periodinane 3~ (DMP) (Scheme 20), and the second one was Corey's epoxidation using the dimethylsulfonium methylide. 22 The key step, which is the vinyl epoxide ring-opening to stereoselectively produce the monoprotected diol 71, was achieved by treatment of

1. n-BuLi (2 eq) TMEDA 2. ~~-Br "~OH

~~A'OMOM

then MOMCI 92%

65

64

I SeO , t-BuOOH t-BuOOH L-(+)-DET Ti(O/-Pr4)

~OMOM

53% salicylic acid HO I

v ~

~OMOM

90% 67 (ee > 95%)

66

1. AcOH, Ti(O/-Pr4) 66% 2). TsCI 3. OH resin

A C O ~ o M O M ~'0 "ell

1. ~ M g B r Cul, THF

-/

k-)3I:- OH

/~OH

2. HCI 6N, MeOH 68

69 71% / 1. DMP 2. (CH3)2S=CH2

,.

OH

=

"/

~,-}3I: "OH 70

72

[Pd2(dba)3].CHCI3 67%

P(O/-Pr)3 Ph3SiOH, THF

1. Ac20 H 73

2. KF 79%

71

SCHEME19. PreparationoftheTrostintermediate.

3

81

TOTALSYNTHESES OF ZOAPATANOL

AcC~ OAc

OH

0

0

RI~"-R 2

RI-~R 2

.,. OAc .OAc

AcO~)I

AcO,

0 ~ H

OyR2 O

R1

0

/OAc +

O J] R1 / \ R 2

+

AcOH

0 SCHEME 20. Dess-Martin periodinane oxidation.

70 with [Pd2(dba)3].CHC13 in the presence of triphenylsilanol in THF at room temperature. 27 Under these conditions only the product resulting from cleavage of the distal bond of the epoxide was obtained, and exclusively with the required E-configuration. It is noteworthy that the formation of oxepane 72, which could arise from an intramolecular attack of the tertiary alcohol of 70 on the vinyl epoxide, was not observed. Acetylation of the primary allylic alcohol of 70, followed by the cleavage of the silyl ether using KF, led to the diol 73. The primary alcohol of 73 was then transformed to the corresponding primary triflate [(CF3SO2)20), 2,6-1utidine], and this latter spontaneously cyclized in situ under the reaction conditions, affording oxepane 74 (Scheme 21). At this stage the elaboration of the side chain was required to complete the synthesis. After oxidation of the terminal double bond under Wacker's conditions 31 (LiC1, PdC12, CuC1, DMF, H20, 0 2, rt) generating a methyl ketone (Scheme 22), the acetyl protecting groups were exchanged with tert-butyldimethylsilyl (TBS) groups in order to minimize problems of chemoselectivity at the end of the synthesis. The resulting ketone 75 was transformed to vinylstannane 76, via a vinyl triflate

82

JANINE COSSY, V]~RONIQUE BELLOSTA, AND CATHERINE TAILLIER

~ -/

OAc

(0F3SO2)20 ~ A~cO'"~/--~ -~~OAc 2'6"17~i~ ~ / / ~ v ~ - " O

\-)31~ un [ 73

74

1. LiCI,PdCI2, DMF,H20, 02, rt 71% 2. K2CO3,MeOH 3. TBSCI, imidazole

TBSO

1. KHMDS, -= CI~ | ' ) ; ~',.O T B S N(SO2CF3)2 TBSO

T ~ O ,,,.: ~ Me3Sn" ~3 "~-0 76

76%

2. [(CHs)sSn]2, Pd(PPh3)4, LiCI 67%

O 77

~'0 75

~CI 34 atm CO Pd2(dba)3.CHCI3 PPh3, Phil, 60~

~

TBSO,,, ~~~~Jl~~_

0~~

S O

/OTBS 11[Ph3PCuH]6,H20, Phil .~ (+)-Zoapatanol1 2 HF, H20, CH3CN,rt 85%

SCHEME 21. Completion of the Trost synthesis.

intermediate, in order to produce the 0~,[3-unsaturated ketone 77 by a carbonylative alkylation under 34 atm of CO in the presence of Pdz(dba) 3. CHC13 and prenyl chloride. The obtained 0~,[3-unsaturated ketone 77 was then reduced chemoselectively by conjugate addition of hydride by using the Strycker reagent [Ph3PCuH]632 and, after cleavage of the protecting groups using HF, (+)-zoapatanol was obtained. The Trost sequence led to the first enantioselective total synthesis of (+)-zoapatanol in 20 steps from methallyl alcohol in 1.6% overall yield.

3

83

TOTAL SYNTHESES OF ZOAPATANOL

R~

cat PdCI2, H20 CuCI 2, 02

~176

"f~

PdCl2 ~ (

Pd(0)

Hc, .'-m /

CI"

H-Pd-CI

Regeneration of Pd(ll) Pd(0) + 2CuCI 2 ~

~

OH I[,- _PdCI

PdCI2 + 2CuCI

Regeneration of Cu(ll) ..~.-R

H..O..H

fl-hydride ~ elimination f ~ OH

a'~ olefin "~complexation

c,,p

reductive ~ elimination \

O

O

_

-- R,,~,"

2CuCI + 1/202 + 2HCI-~,'- 2CuCI 2 + H20

nucleophilic attack

S C H E M E 22. Wacker oxidation.

VIII. Our Approaches for the Total Synthesis of (+)-Zoapatanol

Two synthetic strategies have been examined to construct the oxepane ring of (+)-zoapatanol, one utilizing a ring-closing metathesis (RCM) and the other one an intramolecular Horner-Wadsworth-Emmons (HWE) reaction. A. R I N G - C L O S I N G

METATHESIS APPROACH

The first strategy envisioned to synthesize (+)-zoapatanol relies on a RCM reaction to produce the seven-membered ring. When a diene is treated with a transition metal alkylidene complex, a metallacyclobutane intermediate is formed and a succession of [2+2]-cycloadditions and cycloreversions takes place (Scheme 23). Each step is reversible and the driving force is the elimination of a molecule of ethylene and the formation of a cyclic adduct. The retrosynthetic analysis revealed that an oxidation of oxepene XVII should lead to ketone XVI (Scheme 24), which is a precursor of zoapatanol according to Kane's synthesis. 70xepene XVII could be obtained by using a RCM applied to the unsaturated enol ether XVIII in which the two stereogenic centers could be controlled through application of a Sharpless asymmetric dihydroxylation 33 of the trisubstituted (Z)-olefin XX. The result of the enantioselective cis-dihydroxylation of olefins using an osmium catalyst [K2OsOz(OH)4 ] in the presence of K2CO 3, K3Fe(CN)6

84

JANINE COSSY, VI~RONIQUE BELLOSTA, AND CATHERINE TAILLIER

[M]--CH2 "~[M]

~[[M]

SCHEME 23

.f/--OH

HO,,,..//---',~ 0

9

{ R'01,.~O

1

R

'

O

XVI

,

RO~ O

,

,

.

~

~

"

XVIII

.<---

RO XVII

HOI,.~OR" HO~'~OR

OR"

XIX

XX

XXII

XXI

SCHEME 24. Our first retrosynthetic approach toward zoapatanol.

3

85

TOTALSYNTHESESOF ZOAPATANOL

(DHQD)2-PHAL

AD-mix-13

Rs " ~ " RE

K2OsO2(OH)4 K3Fe(CN)6, K2CO3 tBuOH/H20 'D" AD-mix-o~

HO

Rs (DHQ)2.PHAL

OH RM H RM

R L ~ H

HO

OH

(DHQD)2-PHAL = 1,4-bis(9-O-dihydroquinidine)phthalazine used in AD-mix-I[]

jo

.

o,; .

(DHQ)2-PHAL = 1,4-bis(9-O-dihydroquinine)phthalazine used in AD-mix-o~

N-N

SCHEME 25. Sharpless asymmetric dihydroxylation.

and cinchona alkaloids, (DHQD)2-PHAL or (DHQ)2-PHAL, can be predicted according to the model represented in Scheme 25. The olefin XX would be synthesized by using a Suzuki-Miyaura cross-coupling 34 between an organoborane derived from olefin XXI and (Z)-vinyl iodide XXII. This reaction follows the general mechanistic cycle represented in Scheme 26. After an oxidative addition of RX, the catalytic complex LzPd(0 ) is transformed to intermediate J. Complex J reacts with the organoborane partner by transmetalation and, after isomerization, complex K is formed. The final step, a reductive elimination, then produces the coupling product and regenerates the catalytic species, allowing propagation of the catalytic cycle.

JANINECOSSY,VI~RONIQUEBELLOSTA,ANDCATHERINETAILLIER

86

Pd(0)L2

~ R X

R-R'

Oxidative addition eliminationReduc \

R ~Pd j L

\

L_ ~

/L

L

R\Pd / \

\X J

R'M

~ ~ T r a

nsmetalation I$omerization

R' ~..

K

MX

SCHEME26. Catalyticcycleof the Pd cross-coupling.

It was desirable to adopt an orthogonal protecting group strategy to permit a selective deprotection of the hydroxy groups. The use of two silylated ethers that can be deprotected selectively, a tert-butyldiphenylsilyl (TBDPS) group for R and a TBS group for R", plus a benzyl group (Bn) for R' was thus envisaged (Scheme 24). To test our strategy an approach toward racemic zoapatanol was first performed. The synthesis of zoapatanol started with the preparation of the (Z)olefin 86 from vinyl iodide 80 (Scheme 27) and olefin 84 (Scheme 29). Vinyl iodide 80 was prepared in three steps from 1,4-butanediol 78 (Scheme 27). After monoprotection of 78 (TBSC1, Et3N, CH2C12), the resulting monosilylated ether was oxidized to aldehyde 79 by using a Swern oxidation [(COC1) 2, DMSO, Et3N, CH2C12] with an overall yield of 94%. Aldehyde 79 was then transformed to the (Z)-vinyl iodide 80 in

+ _ n-BuLi Ph3PCH2CH31 > Ph3P=CHCH3

12

NaHMDS + = Ph3PCHICH31

CH3 Ph3P=~ I 81

H O ~ o H 78

1. TBSCI, Et3N, CH2CI2 H ~ O T B 2. (0001)2, O DMSO, Et3N 79 94%

S

81 -20~ 47%

SCHEME27. Preparationof vinyliodide 80.

,, ~ O T B S rt /

"1 80

3

TOTAL SYNTHESES OF ZOAPATANOL

87

47% yield using a Wittig-type reaction with a-iodophosphonium ylide 81. 35 This latter was prepared from ethyltriphenylphosphonium iodide by deprotonation with n-BuLi followed by the addition of iodine and subsequent deprotonation of the resulting phosphonium salt using NaHMDS to form the ~-iodo ylide 81, which was added to aldehyde 79. The trisubstituted olefin (Z)-80 was obtained as a single isomer in accord with the literature data. 36 The modest yield can be explained by the formation of by-products. Under these conditions, the olefination can compete with a Darzens type reaction, which can produce a cis-epoxide XXIV or a methyl ketone of type XXV (Scheme 28). 35,36 This secondary reaction is likely responsible for the high (Z)-stereoselectivity of the olefination. Among the four adducts coming from the condensation between an aldehyde and the ~-iodo ylide 81, only two betaine intermediates can produce the (Z)-vinyl iodide XXIII via an oxaphosphetane. The two other betaines, which could potentially produce the (E)-vinyl iodide, are probably transformed preferentially to an ~,13-epoxy phosphonium salt, a precursor of the methyl ketone XXV or the cis-epoxide XXIV (Scheme 28). In order to prepare the (Z)-olefin 86 from iodide 80 using a Suzuki-Miyaura coupling, an organoborane derived from alkene 84 was prepared (Scheme 29). The alkene 84 was prepared in three steps from propionic acid 82. Thus, the dianion of 82 was alkylated with allyl bromide (Nail, then LDA, allyl bromide) leading to 2-methylpent-4-enoic acid 83 in 96% yield. 37 The reduction of 83 by LiA1H4 (THF, rt) followed by protection of the resulting alcohol (TBDPSC1, Imidazole, DMF, rt) led to the desired olefin 84 (86% yield, two steps). Olefin 84 was then transformed into the organoborane 85 (9-BBN-H, THF), which was not isolated but directly engaged in the Suzuki-Miyaura coupling with (Z)-vinyl iodide 80 in the presence of a catalytic amount of Pd(PPh3) 4 and K2CO 3 in dioxane at 85 ~ to afford the (Z)-olefin 86 in 82% yield. 34 It is worth noting that the (Z)-configuration was confirmed by differential ~H NMR-NOE experiments. To introduce the two contiguous hydroxyl groups at C2' and C3', olefin 86 was dihydroxylated [OsO 4, NMO, acetone/H20 (1/6), rt] to produce diol 87 in 61% yield, followed by transformation to benzyl ether 88. The best conditions for obtaining 88 in good yield were found to be the treatment of 87 with Nail (2.2 equiv.) in the presence of n-tetrabutylammonium iodide (40 mol%) and HMPA (3 equiv.) followed by the addition of benzyl bromide (1.1 equiv.). After 12 h at room temperature, compound 88 was isolated in 87% yield. To complete the construction of the oxepene ring precursor, compound 88 had to be transformed to diene 93. After selective cleavage of the tert-butyldimethylsilyl ether (TBAF, THF, rt, 1 h)

88

JANINE COSSY,VI~RONIQUEBELLOSTA,AND CATHERINETAILLIER

+

Ph3P

R' 81

1 Ph3P\ ~

+ Ph3P

..Me

R' I

R'

k._ Y

O-

O- + ~ ~ _ PPh3

Me

J

\

Me ~

O-

+ PPh3

R' I

R'

J Y

l O- Li +

O~

R ,,.,~~/pph 3

I

R

] ......

R'

I

R'

L,,~ ,,,,'+ '7 ..... ~'PPh3 R'

R'

I

O

i_

PPh3

XXIII

or

R,

0

R, ~

XXIV SCHEME 28. Intermediatesinvolved in the Wittig-type olefination.

0 XXV

3 TOTALSYNTHESESOFZOAPATANOL ~CO2H

82

1. Nail (1 eq) 2 LDA (1 eq) ~, 3. ~--.vBr 96%

89

CO2H 83

1. LiAIH4 ,. /~-~~OTBDPS 2. TBDPSCI, imidazole,DMF 84

I 9-BBN-H THF

"'/'~'OTBS / ~I 80

//"~/~'~OTBS ..~~~./OTBDPS

(~B~OTBDPS

Pd(PPh3)4,K3PO4, dioxane,85~ 82%

86

85

O/olOs04,NMO

61 Acetone/H20 Nail, n-Bu4NI, BnO,,,I//...~OTBS HMPAFFHF,BnBr,rt= ,~.~/~ 87% HO' OTBDPS

HO,,,[//"~~OTBS HO"~'v~ OTBDPS 87

88

1.TBAF

84o/o 2. PDC,DMF BnO,,,~

=-1

OTBDPS 91

B.nO,~~_

OTBDPS

BnO,,,[/'~~ H O " ~ ~/~OTBDPS

~OEt Hg(OCOCF3)2 ~ e t s50 N ~,

92

47%

Conditions see text ~.,

95

'"

OTBDPS 89

90

I(CF3COKO)H~-~~:t

68% MeP(Ph)3Br n-BuLi,THF

TBDPSO

_DIBAL-H

]

BnO,,,[../'V~ ~ O " ~

OTBDPS

93 I CI2(PCY3)2Ru=CHPh 70% Phil, 50~

~

x/

TBDPSO

SCHEME29. Oxepenesynthesisvia RCMreaction.

94

90

JANINE COSSY, VI~RONIQUEBELLOSTA, AND CATHERINE TAILLIER

and oxidation of the resulting primary alcohol by PDC (DMF, rt), lactone 89 was obtained in 84% yield (two steps). The transformation of lactone 89 to olefin 92 could be achieved via lactol 90 by treatment with a Wittig reagent as the lactol is in equilibrium with the corresponding hydroxy aldehyde 91. The reduction of 89 by DIBAL-H (THE - 7 8 ~ produced the corresponding lactol 90, which was treated directly with methyltriphenylphosphonium ylide [MeP(Ph)3Br, n-BuLi, THF] to afford olefin 92 (44% yield, two steps). 38 The tertiary alcohol of 92 was then etherified by treatment with Hg(OCOCF3) 2 and ethyl vinyl ether (Et3N, 50 ~ 39to afford the desired diene 93 in 47% yield via the mercuric intermediate K. Compound 93 was then converted to oxepene 94 in 70% yield by treatment with the "first generation" Grubbs catalyst [C12(PCy3)zRu=CHPh (30 mol%), Phil, 50 ~ 4~The resulting oxepene 94, by analogy to Kane's intermediate 56, was expected to lead to 95 via oxidative hydroboration (BH3-THF; H202, NaOH) followed by oxidation (CrO 3, Pyr, CH2C12).7 Unfortunately, when 94 was subjected to BH3.THF and then H202, NaOH, no reaction occurred. Moreover, when 94 was treated with various oxidizing agents (dimethyldioxirane or m-CPBA in MeOH) to produce a precursor of ketone 95, only degradation or recovery of 94 was observed. Due to this failure, a second route using a HWE reaction was envisioned to construct the oxepane ring of zoapatanol. B. H O R N E R - W A D S W O R T H - E M M O N S APPROACH

The retrosynthetic analysis of oxepinone XXVI revealed that the oxepene ring with the required stereochemistry could potentially be constructed through application of an intramolecular HWE reaction applied to the phosphono-aldehyde XXVII derived from anti-diol XXVIII. Control of the two contiguous stereogenic centers of XXVIII could be achieved by applying a Sharpless asymmetric dihydroxylation to the (Z)-trisubstituted olefin XXIX, using AD-mix-13. Olefin XXIX could be derived from a Suzuki-Miyaura coupling between vinyl iodide 96 and organoborane XXX. In this approach, an orthogonal protecting group strategy for R and R' was also required. Therefore, the protection of the secondary alcohol in XXVII by a benzyl group and the primary alcohol by a TBDPS group was planned (Scheme 30). The preparation of (+)-zoapatanol started with the synthesis of phosphono-aldehyde 101 from the silylated 2-methylpent-4-en-l-ol 84 and (Z)-vinyl iodide 96. This latter species was obtained by treatment of ethyl but-2-ynoic ester with NaI in acetic acid (70 ~ 95% yield). 4~ The high stereoselectivity of this reaction ruled out the possibility of direct HI

3

91

TOTALSYNTHESES OF ZOAPATANOL

HOt,,.~ 0

OH

Ro.r ~ XXVI

1

o~

~:

HOI~"[""~OR"

RO'v'"~,-~~:

.

R'O,,,./~

0H

P(OI(OMe)2

~-i~0

_

,....

XXVlll

XXVII 0

0

EtO1 ~ 96

XXlX XXX SCHEME 30. Retrosynthetic pathway involving a HWE reaction.

addition to the propynoate in AcOH, which has been reported to proceed with low stereoselectivity. 41aThe observed high level of stereocontrol may be explained by assuming that a trimolecular transition state of type L (Scheme 31), in which the halide and acetic acid lie on opposite sides of the carbon-carbon triple bond would be favored on steric grounds. Another explanation would be to consider that a vinyl anion intermediate is generated by nucleophilic addition of iodide to the electron-deficient alkyne. In this case, repulsion of the electron pairs on the halogen atom and the carbanion should destabilize intermediate M,, leading to the (E)-isomer, while the electron pairs are away from each other in intermediate M b, leading to the observed (Z)-isomer (Scheme 31). Alkene 84 was treated with 9-BBN-H (9-BBN dimer, THE rt) and the resulting organoborane product was then subjected to a Suzuki-Miyaura cross-coupling with (Z)-vinyl iodide 96 (Scheme 32) in the presence of Pd(PPh3) 4 and K3PO4 (dioxane, 85 ~ to afford the (Z)-~,~-unsaturated ester 97 in 74% y i e l d . 34,42 To transform 97 to diol 98, the enantioselective

JANINECOSSY.VI~RONIQUEBELLOSTA,ANDCATHERINETAILLIER

92 I--

' H---~ - -7--R ,,

I

=

H --OAc

R

~X~ H

L

H

Z-isomer

R = electron withdrawing group

% H

I

H+

H

R Ma

"I"

H

R

E-isomer

R

Mb

)=(

I

R

H

H

H+

Z-isomer

SCHEME31. Proposedexplanationsfor the mechanismof formationof 96.

Sharpless dihydroxylation was achieved with AD-mix-[3 (65% yield, 92% ee). 43 The phosphono-aldehyde 101 was then prepared from diol 98. After selective protection of the secondary alcohol of compound 98 using benzyl bromide in the presence of silver oxide and n-tetrabutylammonium iodide, 44 reduction of the carboxylic ester with LiA1H4 and protection of the resulting primary hydroxy group as a methoxymethyl ether, compound 99 was obtained in 66% yield. A rhodium-catalyzed insertion of ethyl diazoacetate (N2CHCOzEt, [Rh(OAc)2]2, toluene, 110 ~ followed by the condensation of an excess of the lithium salt of methyldimethylphosphonate with the resulting ester led to the [~-keto-phosphonate 100 (60% yield). 46 As an intramolecular HWE reaction was envisioned to construct the oxepane ring of zoapatanol, the methoxymethyl ether group had to be transformed into an aldehyde. Thus, after removal of the methoxymethyl ether protecting group with TMSBr, the corresponding hydroxyphophonate was obtained (84%) 47 and the oxidation of the primary alcohol was conveniently accomplished with PDC. The resulting crude aldehyde 101, which turned out to be unstable, was then directly subjected to Nail in THF to afford oxepinone 102 in 53% overall yield via an intramolecular HWE cyclization (Scheme 32). 48

3

93

TOTALSYNTHESES OF ZOAPATANOL

O Nal, CH3CO2H,70~

CO2Et

EtO~

95%

96 9-BBN dimer THF, rt

TBDPSO

]'BDPSO~

B~ 85

84

/

74%/ 96, Pd(PPh3)4 K3PO4 O

O

HO,,.r~OEt TBDPSO~~'@:

OH

AD-mix-fl H2NSO2Me

" ee = 92%

98

EtO~

/ TBDPSO...~..~o~

~ t-BuOH/H20

65% 97

1. BnBr, Ag20, n-Bu4NI 66% 2. LiAIH4 3. MOMCI,Nail, THF 1 [Rh(OAc)2]2 OMOM N2CHCO2Et, PhMe 2. MeP(O)(OMe)2 BnO,,/ n-BuLi, THF r" v 60% TBDPSO 99 TBDPSO 9

O MOMO p(OMe)2 ,.~ ~~:O II

100

O"

1. TMSBr 2. PDC --

(~

~ TBDPSO

0

NaH 45%

102

OTBDPS 101

SCHEME 32. Preparation of the oxepinone intermediate.

O II

--

P(OMe)2

94

JANINE COSSY, VI~RONIQUE BELLOSTA, AND CATHERINE TAILLIER

To complete the synthesis of (+)-zoapatanol, the unsaturated side chain as well as the [3,y-unsaturated ketone had to be introduced on oxepinone 102, but first this compound has to be reduced selectively in order to obtain the corresponding oxepanone. The chemoselective hydrogenation of 102 was performed in the presence of Pd/C (10%) in ethanol for 5 min. The benzyl protecting group was not affected and oxepanone 103 was isolated in 98% yield (Scheme 33). 49 The resulting oxepanone 103 was then treated with the lithium salt of triethylphosphonoacetate [EtO2CCHzP(O)(OEt)2, LiHMDS, THF, rt] to generate the corresponding 0t,[3-unsaturated esters (97%) as an inseparable mixture of E/Z-isomers (E/Z = 70/30 ratio by 1H NMR spectroscopy). 48",5~After reduction with LiA1H 4, the corresponding stereoisomeric allylic alcohols were separated by SiO 2 flash chromatography, and the desired (E)-allylic alcohol 5~ was obtained in 63% overall yield. This latter compound was then protected as a benzyl ether 104 (BnBr, Ag20, n-Bu4NI, CH2C12) in 98% yield. As a stable tetrahedral intermediate resulting from the addition of an organolithium to a Weinreb amide can serve as a carbonyl protecting group

~

O

OTBDPS

Pd/C, EtOH,98%5 min~ ~

102

0

OTBDPS

Ii. 103

1. EtO2CCH2P(O)(OEt)2 LiHMDS, THF 60% 2. LiAIH4 3. BnBr, Ag20, n-Bu4NI OBn

OBn MeO

BnO,,.~

Me ; N ~ / ~ _ _ : 0

0J 105

~ 1" n-Bu4NF

1 BnO,,.~

2. CrO3/H2SO4 r'" V "-.~-"O ~ 39HN(OMe)Me"HCl OTBDPS EDCI, iPr2NEt, DMAP 104 60%

Et20/THF [-

tl.

F-OBn-]

/

Li/NH3(liq)

t-BuOH, THF 66% L Me ""

106

= (+)-Zoapatanol 1

__J

SCHEME 33. Completion of the synthesis of (+)-zoapatanol.

3

TOTALSYNTHESESOF ZOAPATANOL

95

during the debenzylation of hydroxy groups under the Birch reduction conditions, 52 104 was transformed to the Weinreb amide 105 in order to elaborate the nonenyl side chain present in (+)-zoapatanol. Thus, after removal of the silyl protecting group in 104, the resulting primary hydroxy group was oxidized to the carboxylic acid (Jones' reagent) and this latter compound was directly converted to the Weinreb amide 105 [HN(OMe)Me, HC1, EDCI, iPrzNEt, DMAP, CHzCI2] with an overall yield of 60%. 53 Treatment of amide 105 with prenyllithium 54 (THF/Et20, - 7 8 ~ led to the stable intermediate 106, which was directly subjected to Birch reduction conditions 55 [Li/NH3(liq), t-BuOH/THE - 7 8 ~ to afford the desired (+)-zoapatanol in 66% yield. ~ In conclusion, this second approach involving four key steps, a Suzuki-Miyaura cross-coupling, a Sharpless asymmetric dihydroxylation, an intermolecular HWE cyclization and a chemoselective nucleophilic addition/Birch reduction process, allowed the total synthesis of (+)-zoapatanol in 19 steps from alkene 84 and 3% overall yield. IX. Conclusion

Up to now, seven total syntheses of zoapatanol have been reported and only two of them are enantioselective. Among the racemic ones, Nicolaou's synthesis is the shortest and the most efficient (16 steps, 12% overall yield). For the enantioselective syntheses, the synthesis using a Sharpless dihydroxylation to control the stereogenic centers at C2' and C3' and an intramolecular HWE cyclization, as the key steps, is the most efficient one. References and Footnotes

1. (a) Levine, S. D., Hahn, D. W., Cotter, M. L., Greenslade, E C., Kanijoa, R. M., Pasquale, S. A., Wacter, M. E, McGuire, J. L., J. Reprod. Chem. 1981, 524. (b) Quijano, L., Calderon, J. S., Fisher, N. K., Phytochemistry 1985, 24, 2337. (c) Oshima, Y., Cordial, G. A., Fong, H. S., Phytochemistry 1986, 25, 2567. 2. (a) Marcelle, G. B., Bunyapraphatsara, N., Cordell, G. A., Fong, H. S., Nicolaou, K. C., Zipkin, R. E., J. Nat. Prod. 1985, 48, 739. (b) Kanojia, R. M., Chin, E., Smith, C., Chen, R., Rowand, D., Levine, S. D., Wachter, M. E, Adams, R. E., Hahn, D. W., J. Med. Chem. 1985, 28, 796. 3. Lewine, S. D., Adams, R. E., Chen, R., Cotter, M. L., Hirsch, A. E, Kane, V. V., Kanijoa, R.M., Shaw, C. J., Wachter, M. E, Chin, E., Huetteman, R., Ostrowski, E, Mateos, J. L., Noriega, L., Guzm~n, A., Mijarez, A., Tovar, L., J. Am. Chem. Soc. 1979, 101, 3404.

96

JANINE COSSY,VI~RONIQUEBELLOSTA,AND CATHERINETAILLIER

4. (a) Davies, M. J., Heslin, J. C., Moody, C. J., J. Chem. Soc., Perkin Trans. 1989, 1, 2473. (b) Pain, G., DesmaEle, D., d'Angelo, J., Tetrahedron Lett. 1994, 35, 3085. (c) Shing, T. K. M., Wong, C.-H., Yip, T., Tetrahedron: Asymmetry 1996, 7, 1323. (d) Ovaa, H., van der Marel, G. A., van Boom, J. H., Tetrahedron Lett. 2001, 42, 5749. 5. Chen, R., Rowand, D. A., J. Am. Chem. Soc. 1980, 102, 6609. 6. Nicolaou, K. C., Claremon, D. A., Barnette, W. E., J. Am. Chem. Soc. 1980, 102, 6611. 7. Kane, V. V., Doyle, D. L., Tetrahedron Lett. 1981, 3027 and 3031. 8. Cookson, R. C., Liverton, N. J., J. Chem. Soc., Perkin Trans. 1985, 1, 1589. 9. (a)Kocienski, P., Love, P., Whitby, R., Tetrahedron Lett. 1988, 29, 2867. (b) Kocienski, P. J., Love, C. J., Whitby, R. J., Tetrahedron 1989, 45, 3839. 10. Trost, B. M., Greenspan, E D., Geissler, H., Kim, J. H., Greeves, N., Angew. Chem., Int. Ed. Engl. 1994, 33, 2182. 11. (a) Taillier, C., Bellosta, V., Cossy, J., Org. Lett. 2004, 6, 2149. (b) Taillier, C., Gille, B., Bellosta, V., Cossy, J., J. Org. Chem. 2005, 70, 2097. 12. Either the epimerization occurred during isolation or purification of the natural product, or the natural zoapatanol itself is a mixture of epimers at C6. Kanijoa, R. M., Wachter, M. P., Levine, S. D., Adams, R. E., Chen, R., Chin, E., Cotter, M. L., Hirsch, A. E, Huetteman, R., Kane, V. V., Ostrowski, P., Shaw, C. J., J. Org. Chem. 1982, 47, 1310. 13. (a) Still, W. C., McDonald, J. H., III, Tetrahedron Lett. 1980, 1031. (b) Collum, D. B., McDonald, J. H., III, Still, W. C., J. Am. Chem. Soc. 1980, 102, 2117. (c) Cram, D. J., Kopecky, K. R., J. Am. Chem. Soc. 1959, 81, 2748. 14. (a)Sharpless, K. B., Michaelson, R. C., J. Am. Chem. Soc. 1973, 95, 6136. (b) Sharpless, K. B., Verhoeven, T. R., Aldrichim. Acta 1979, 12, 63. 15. Btichi, G., Wriest, H., Helv. Chim. Acta 1967, 50, 2440. 16. VanHorn, D. E., Negishi, E., J. Am. Chem. Soc. 1978, 100, 2252. 17. Kobayashi, M., Valente, L. E, Negishi, E., Synthesis 1980, 1034. 18. Negishi, E., Kondakov, D. Y., Choueiry, D., Kasai, K., Takahashi, T., J. Am. Chem. Soc. 1996, 118, 9577. 19. A model study performed on (2E)-3-hydroxymethyl-7-methylocta-2,6-dien-l-ol demonstrated surprising resistance to epoxidation when treated with m-CPBA. Preliminary acetylation of hydroxyl groups circumvented this problem. 20. (a) Wenkert, E., Ferreira, V. E, Michelotti, E. L., Tingolini, M., J. Org. Chem. 1985, 50, 719 (and references therein). (b) Wadman, S., Whitby, R., Yeates, C., Kocienski, P., Cooper, K., J. Chem. Soc., Chem. Commun. 1987, 241. 21. (a) Ishino, Y., Wakamoto, K., Hirashima, T., Chem. Lett. 1984, 765. (b) Duboudin, J. -G., Jousseaume, B., J. Organomet. Chem. 1979, 168, 1. 22. Corey, E. J., Chaykovsky, M., J. Am. Chem. Soc. 1965, 87, 1353. 23. Schreiber, J., Felix, D., Eschenmoser, A., Winter, M., Gautschi, E, Schulte-Elte, K. H., Sundt, E., Ohloff, G., Kalvoda, J., Kaufmann, H., Wieland, P., Anner, G., Helv. Chim. Acta 1967, 50, 2101. 24. Hydroboration of 50 with 9-BBN followed by treatment with NaOH/H202 afforded a 85/15 mixture (after chromatography) of alcohol 51 and of the corresponding aldehyde. 25. (a) Criegee, R., Liebigs Ann. Chem. 1948, 560, 127. (b) House, H. O., Modern Synthetic Reactions, 2nd ed., W. A. Benjamin: Menlo Park, CA, 1972, p. 325.

3

TOTALSYNTHESESOF ZOAPATANOL

97

26. (a) Hanson, R. M., Sharpless, K. B., J. Org. Chem. 1986, 51, 1922. (b) Gao, Y., Hanson, R. M., Klunder, J. M., Ko, S. Y., Masamune, H., Sharpless, K. B., J. Am. Chem. Soc. 1987, 109, 5765. 27. Trost, B. M., Ito, N., Greenspan, E D., Tetrahedron Lett. 1993, 34, 1421. 28. The enantiomeric excess of 67 was determined by NMR spectroscopy of the Mosher esters: Dale, J. A., Mosher, H. S., J. Am. Chem. Soc. 1973, 95, 512. 29. Caron, M., Sharpless, K. B., J. Org. Chem. 1985, 50, 1557. 30. (a) Dess, D. B., Martin, J. C., J. Org. Chem. 1983, 48, 4155. (b) Dess, D. B., Martin, J. C., J. Am. Chem. Soc. 1991, 113, 7277. 31. Tsuji, J. In: Comprehensive Organic Synthesis, Trost, B. M., Fleming, I., Ley, S. V. (Eds.), Pergamon Press: Oxford, 199 l, pp. 449-468. 32. Mahoney, W. S., Brestensky, D. M., Stryker, J. M., J. Am. Chem. Soc. 1988, 110, 291. 33. Kolb, H. C., vanNieuwenhze, M. S., Sharpless, K. B., Chem. Rev. 1994, 94, 2483. 34. (a) Miyaura, N., Ishiyama, T., Sasaki, H., Ishikawa, M., Satoh, M., Suzuki, A., J. Am. Chem. Soc. 1989, 111,314. (b) Oh-e, T., Miyaura, N., Suzuki, A., Synlett 1990, 221. (c) Watanabe, T., Miyaura, N., Suzuki, A., Synlett 1992, 207. (d) Miyaura, N., Suzuki, A., Chem. Rev. 1995, 95, 2457. 35. Chert, J., Wang, T., Zhao, K., Tetrahedron Lett. 1994, 35, 2827. 36. Arimoto, H., Kaufman, M. D., Kobayashi, K., Qiu, Y., Smith, A. B., III, Synlett 1998, 765. 37. (a) Creger, E L., J. Am. Chem. Soc. 1970, 92, 1396. (b) Creger, E L., Org. Synth. 1970, 50, 58. 38. Compound 92 was isolated in an increased overall yield of 68% by performing the four consecutive steps from 88 without purification of intermediates. 39. (a) Germanas, J., Aubert, C., Volhardt, K. E C., J. Am. Chem. Soc. 1991, 113, 4006. (b) Nonoshita, K., Banno, H., Maruoka, K., Yamamoto, H., J. Am. Chem. Soc. 1990, 112, 316. (c) Tulshian, D. E, Tsang, R., Fraser-Reid, B., J. Org. Chem. 1984, 49, 2347. 40. For RCM reactions on vinyl ethers, in the presence of "first" or "second generation" Grubbs' catalysts see: Tuyen Nguyen, V., DeKimpe, N., Tetrahedron Lett. 2004, 45, 3443. (b) Rainier, J. D., Cox, J. M., Allwein, S. E, Tetrahedron Lett. 2001, 42, 179. (c) Arisawa, M., Theerladanon, C., Nishida, A., Nakagawa, M., Tetrahedron Lett. 2001, 42, 8029. (d) Okada, A., Ohshima, T., Shibasaki, M. Tetrahedron Lett. 2001, 42, 8023. (e) Chatterjee, A. K., Morgan, J. E, Scholl, M., Grubbs, R. H., J. Am. Chem. Soc. 2000, 122, 3783. 41. (a) Ma, S., Lu, X., Li, Z., J. Org. Chem. 1992, 57, 709 (and references cited). (b) Meyer, C., Marek, I., Normant, J. -E, Synlett 1993, 386. (c) Marek, I., Meyer, C., Normant, J.-E, Org. Synth. 1997, 74, 194. 42. (Z)-Configuration of the enoate 97 was confirmed by 1H NMR-NOE analysis. 43. The (2S,3S) absolute configuration of 98 was confirmed by the 1H NMR spectra of the two corresponding mandelates, following the procedure described by: Seco, J. M., Quifioa, E., Riguera, R., Tetrahedron: Asymmet~ 2001, 12, 2915. 44. Bouzide, A., Sauv6, G., Tetrahedron Lett. 1997, 38, 5945. 45. (a) Noels, A. E, Demonceau, A., Petiniot, N., Hubert, A. J., Teyssi6, P. H., Tetrahedron 1982, 38, 2733. (b) Jones, K., Toutounji, T., Tetrahedron 2001, 57, 2427. 46. Suemune, H., Akashi, A., Sakai, K., Chem. Pharm. Bull. 1985, 33, 1055. 47. Hu, X. E., Demuth, T. E, Jr., J. Org. Chem. 1998, 63, 1719.

98

JANINE COSSY,VI~RONIQUEBELLOSTA,AND CATHERINETAILLIER

48. (a) Maryanoff, B. E., Reitz, A. B., Chem. Rev. 1989, 89, 863. (b) Nicoll-Griffith, D. B., Weiler, L., Tetrahedron 1991, 47, 2733. 49. Boyer, F. -D., Lallemand, J. -Y., Tetrahedron 1994, 50, 10443. 50. Magnus, P., Miknis, G. E, Press, N. J., Grandjean, D., Taylor, G. M., Harling, J., J. Am. Chem. Soc. 1997, 119, 6739. 51. The (E)-configuration was confirmed by ~H NMR-NOE analysis. 52. Taillier, C., Bellosta, V., Meyer, C., Cossy, J., Org. Lett. 2004, 6, 2145. 53. Tashiro, T., Bando, M., Mori, K., Synthesis 2000, 13, 1852. 54. Prenyllithium was generated by reductive cleavage of phenyl prenyl ether with lithium in a mixture of Et20/THF (1/1); see: Birch, A. J., Corrie, J. E., Subba Rao, G. S. R., Aust. J. Chem. 1970, 23, 1811. 55. (a) Evans, D. E., Bender, S. L., Morris, J., J. Am. Chem. Soc 1988, 110, 2506. (b) Evans, D. E., Polniaszek, R. P., DeVries, K. M., Guinn, D. E., Mathre, D. J., J. Am. Chem. Soc. 1991, 113, 7613.

STRATEGIES AND TACTICS IN ORGANIC SYNTHESIS, VOL. 7 9 2008 Elsevier Ltd. All rights reserved.

Chapter 4 SYNTHESIS OF MOLECULAR TWEEZERS AND CLIPS BY THE USE OF A MOLECULAR LEGO SET AND THEIR SUPRAMOLECULAR FUNCTIONS

Frank-Gerrit Kliirner and Mireia Campa~6 Kuchenbrandt Institut fiir Organische Chemie Fachbereich Chemie Universitiit Duisburg-Essen Essen, Germany

I. II. III. IV. V.

Introduction and Background Retrosynthetic Analysis Preparation of the Molecular Building Blocks Synthesis of Molecular Tweezers and Clips Thermodynamic Parameters of Host-Guest Complex Formation with Molecular Tweezers and Clips VI. Host-Guest Structures and Dynamics VII. Synthesis of Water-Soluble Molecular Tweezers and Clips VIII. Binding of Biologically Interesting Substrates such as Enzyme Cofactors and Nucleosides in Aqueous Solution IX. Conclusion Acknowledgments References and Footnotes

99 102 103 107 117 123 138 141 149 150 150

I. Introduction and Background

The processes of molecular recognition and self-assembly/selforganization are of fundamental importance for the formation of organized chemical systems that result from the association of two or more chemical species. 1 These processes depend on weak but specific, mostly noncovalent intermolecular interactions, such as hydrogen bonding, 2 ion pairing, 3 and arene-arene interactions 4-6 in addition to the less specific van der Waals or dispersion forces. Furthermore, today coordinative metal-ligand bonds are frequently used for the programmed synthesis of supermolecules. 7 The solvent often plays an active role in these

100

FRANK-GERRIT KL,g,RNER AND MIREIA CAMPAlq/~ KUCHENBRANDT

processes by solvating or desolvating the interacting molecules during the receptor-substrate association. In particular, the hydrophobic effect in aqueous media can be very strong and determines the stability of the complexes to a substantial extent. 4 These noncovalent interactions play a key role in many biological processes such as protein folding, the bonding and catalytic transformation of substrates by enzymes, the formation of membranes and the transportation of neutral and ionic species through membranes, and the expression and transfer of genetic information. Multiple weak bonds are necessary to form supermolecules that are, on the one hand, sufficiently stable under normal conditions (e.g., room temperature in aqueous solution) and, on the other hand, sufficiently flexible to undergo conformational changes and partial or complete dissociation without changing these conditions dramatically. Today, the intermolecular interactions, and particularly the interplay between substrate and receptor via multiple noncovalent bonds are studied experimentally as well as theoretically by means of relatively simple synthetic receptors. These can act as models for the far more complicated biological systems. Besides the well-preorganized macrocycles, such as cyclodextrins, 8 cyclophanes, 9 carcerands, ~~ cryptophanes, l' cucurbit[n]urils CB[n], 12 and supramolecular capsules 13 (formed by self-assembly of suitable molecular building blocks), noncyclic compounds with cavities of flexible size, which are frequently termed as molecular tweezers ~4 and clips, ~5 proved to be effective as synthetic receptors. In this chapter, we focus on the synthesis and the supramolecular functions of molecular tweezers and clips 1-9 shown in Figure 1. These systems serve as host molecules binding guest molecules by multiple aromatic interactions (rt-rt, CH-rt and cation-Tt interactions). Aromatic interactions are important factors that determine the structures and properties of many higher organized chemical and biological systems. Examples are the base stacking in DNA and the protein folding caused by phenylalanine and other aromatic amino acid side chain interactions. ~6The preference of the edge-to-face over the face-toface orientation of two benzene rings, as it is found in the crystal structure of benzene ~7 or in the protein structures mentioned above, can be explained with a simple electrostatic model for benzene consisting of a positively charged cr framework sandwiched between two clouds of rt electron density. Quantum mechanical calculations predict a displaced face-to-face orientation slightly more stable than the edge-to-face configuration. 4,5 The tweezer and clip molecules 1-8 are substantially preorganized because of their belt-type structures. But bond angle distortions require little

4

101

SYNTHESIS OF MOLECULAR TWEEZERS AND CLIPS

RI

RI

/

R1

R~

a

b

e

d

e

f

R1

H

OAc

OH

OMe

OP(Me)O 2- M +

OPO32- 2M +

R2

H

OAc

OH

OMe

OP(Me)O 2- M +

OPO32- 2M +

g

h

i

j

k

!

R1

OCH2CO2Et

OCH2CO 2 M +

OH

OH

OH

OCH 3

R2

OCH2CO2Et

OCH2CO 2- M +

OAc

OCH2CO2Et

OCH2CO 2- M +

O(CH2)4-CH--CH 2

FIGURE 1. Tetra-, tri-, di-, and monomethylene-bridged tweezers and clips synthesized to date.

energy and, therefore, should induce a certain flexibility to these systems allowing the host "arms" to be expanded and compressed during the substrate complexation in a way comparable to the working principle of mechanical tweezers. Thus, a fit of the host geometry to the substrate topology induced by the complex formation can be expected to a certain extent.

102

FRANK-GERRIT KLARNER AND MIREIA CAMPAlqA KUCHENBRANDT

The size and shape of the host cavities can be systematically varied by the number and size of the spacer-units and the number of methylene bridges. The monomethylene-bridged dinaphthonorbornadiene (DNN) 9 can barely be considered to be a molecular clip, but, as will be pointed out later, 9 forms host-guest complexes comparable to the other host molecules 1-8, but in a 2:1 ratio. Finally, the parent compounds la, 2a, and 4 are simple hydrocarbons containing only nonconjugated benzene, and/or naphthalene tings arranged in a belt-shaped concave-convex topology, so that an aromatic guest molecule can be bound via multiple rt-rt and CH-rt interactions. Here one can address whether the magnitude of the complex stabilization is dependent on the orientation of the multiple arene-arene interactions, comparable to multiple hydrogen bonding, where, for example, the interaction between donor-donor (D-D) and acceptor-acceptor (A-A) is more stable than that between D-A and A-D. 18 II. Retrosynthetic Analysis

The skeleton of the molecular tweezers and clips 1-8 can be constructed by repetitive Diels-Alder reactions of dienes 10 or 13 and 15 or 16 with the bis-dienophiles 11, 12, or 14 (Scheme 1). At the outset of this work, it was already known that the Diels-Alder reaction between exo5,6-bismethylenenorbornene derivatives (e.g., 10) as dienes and norbornene and norbornadiene derivatives as dienophiles proceeded preferentially on the endo-face of the diene and the exo-face of the dienophile. ~9Thus, it could be expected, that the repetitive reactions of 10 with 11 or 12 and of 10 or 13 with 14 would lead to bis-adducts in which the methylene bridges were syn to each other on the same face of the bisadduct. This stereochemical arrangement of the methylene bridges is an essential prerequisite for the synthesis of the tweezers 1 and 2 or clips 3 and 4 by the use of these molecular building blocks. In the synthesis of the dimethylene-bridged systems 5-8, the clip topology is already present in the syn arrangement of the two methylene bridges in the bis-dienophile 11 or 12. Of course, the molecular building blocks, dienes 10, 13, 15, and 16 and the bis-dienophiles 11, 12, and 14, had to be prepared. The diene 102o was known and norbornadiene 14 was commercially available. The o-quinodimethane derivatives 15 21 and 16 22 a r e not stable. They only exist as transient intermediates in solution under normal conditions, but there are stable precursors known for both intermediates. The following section describes the preparation of these molecular building blocks.

4

103

SYNTHESIS OF MOLECULAR TWEEZERS AND CLIPS

1 (n=O) 2 (n=l)

>

10

3 (n=O) 4 (n=l)

11 (n=O) 12 (n=l)

10

.>

E

" 10 (n=O) 13 (n=l)

1o 10 (n=O) 13 (n=l) a1

5 (n=m=O) 6 (n=O,m= 1) 7 (n=l ,m=O) 15 (m=O) 16 (m=l)

11 (n=O) 12 (n=l)

15 (m=O) 16 (m=l)

SCHEME 1

III.

Preparation of the Molecular Building

Blocks

The preparation of the bis-dienophiles 11 and 1 2 23,24 a r e shown in Schemes 2 and 3 and that of dienes 10 and 13 in Schemes 4 and 5. The preparation of 11 started with a one-pot reaction producing the quinone 19. The Diels-Alder cycloaddition of 1,3-cyclopentadiene to p-benzoquinone led to adduct 17, which isomerized to hydroquinone 18 after addition of triethylamine. Compound 18 was subsequently oxidized with an excess of p-benzoquinone. Quinone 19 readily reacted almost quantitatively with 1,3-cyclopentadiene at - 7 8 ~ leading to a (60;40) mixture of the Diels-Alder adducts syn-20 and anti-20, which could be easily separated by recrystallization from toluene. Fifty-four grams of syn-20 and 33 g of anti-20 were prepared from 100 g of quinone 19 in one batch. Under basic conditions, syn-20 isomerized to the hydroquinone, which could be converted to the bis-dienophiles l l b , l l d , and l l g . In Scheme 2 the conversion of syn-20 to l l b is shown as one example. 23

104

FRANK-GERRIT KL,g,RNER AND MIREIA CAMPAlq,/~ KUCHENBRANDT

+ o

O MeOH

NEt3

o v

CHCI3, 82%

O 17

18

, v

-78~

O

98%

19

anti-20

syn-20 pyridine, Ac20, DMAP, 50~

78% AcO 11b

syn-20 SCHEME 2

~

+

toluene, 90~

O

Oo~~ 1. DBU 2. TMSCI

78 %

I MsI

THF 22 23

DDQ

Ac20 pyridine 95 %

toluene 48 %

OAc

24 12b

1. DBU 2. Mel 85 % MeO 12d

24

TosCI pyridine/ NEt3

NaBH4/ Ce(lll)CI3 methanol 95 %

5O% 12a 25 SCHEME 3

4

105

SYNTHESIS OF MOLECULAR TWEEZERS AND CLIPS

o oooc o O

AcCI

30 %

o

26

~

K2003 MeOH 98 %

94 %

= 4~CO2Me

o

O

CO2Me

cis-27 CO2Me

LiAIH4 THF

CO2M e

9

CH2OH

94 % 28

trans-27

"PPh3CI2" pyridine/ CH3CN 87 %

ID, ~ C H 2 C I 29

CH2OH

KOH/ [18]crown-6 THF 90 %

CH2CI

10

SCHEME 4

[~

]~ CHBr2 CHBr2

+

35

19 h, 180~

I~..CH2CI

62 %

Br

Nal / DMF 65~

31

[ ~

/ CH2CI

H2CI 30 CH2CI

_~CH2Cl 30 CH2CI ~"

66 %

DDQ 73 % =

15

H2C L

[18]Krone-6/ THF 74 % SCHEME 5

36

CH2C!

,

=

H2 CH2CI _1

C

30

KOH,85~ 36 h ~ C H 2 C I 34 CH2Cl

-]

33

~CH2CI CH2ClI4~~

-2 HBr

CIJ

I ~ c rggr L

32

2~176176 toluene 73 %

L.E"I "b,

_.J

r

/f----~--i~-F--/~

13

34

34

106

FRANK-GERRIT KLARNER AND MIREIA CAMPAIq/~ KUCHENBRANDT

The Diels-Alder reaction between 21 and 17 occurred stereospecifically on the endo-face of the diene 21 and exo-face of the dienophile 17. This led to adduct 22, in which both methylene bridges are on the same face of the molecule, as required for the synthesis of the bis-dienophile 12. Baseinduced keto--enol isomerization of 22 and trapping of the enol functions with trimethylchlorosilane gave the trimethylsilyl enol ether 23. Without further purification, 23 was oxidized with 2,3-dichloro-4,5-dicyano-l,4benzoquinone (DDQ) to give diketone 24, the precursor of the bisdienophile 12. Basic keto-enol isomerization of 24 and reaction with acetic anhydride or methyl iodide led to the acetoxy- and methoxy-substituted bisdienophiles 12b and 12d, 23 respectively. The parent bis-dienophile 12a was prepared by reduction of 24 with sodium borohydride in the presence of cerium(III) chloride and subsequent elimination of water from diol 25 by the use of tosyl chloride, pyridine, and triethylamine. 23,25 5,6-Bis-methylene-2,3-benzonorbornene 10 was prepared in six steps starting from indene according to a procedure published by Butler and Snow, which was modified by Kamieth and Wigger to obtain higher yields (Scheme 4). 2o At 200 ~ indene undergoes a sigmatropic 1,5-hydrogen shift to isoindene, which can be trapped by maleic anhydride through a Diels-Alder reaction leading to the cycloadduct 26. Esterification of the anhydride function in 26 with methanol and acetyl chloride and basic cis-trans isomerization gave the diester trans-27. Reduction of the ester groups of trans-27, conversion of the hydroxy groups of 28 into C1 groups and HC1 elimination from 29 gave the desired diene 10 in high yield (_>87%) for each step save the first one, which gave only a 30% yield under optimized conditions. The naphtho-substituted diene 13 could be prepared in two different ways. The first synthesis of 13 began with the in situ generation of dibromoo-quinodimethane 32 by 1,4-Br2 elimination from tetrabromo-o-xylene 31 using sodium iodide (Scheme 5). This reaction had already been described by Cava and coworkers in 1960.26 In the absence of a trapping reagent, the highly reactive o-quinodimethane derivative 32 can undergo an electrocyclic ring closure leading to 3,4-dibromo-l,2-benzocyclobutene, which can be also used as a precursor of 32 at a higher temperature (150 ~ In the presence of a dienophile such as maleic anhydride or N-phenyl maleic imide, 32 reacts to form the corresponding naphthalene derivatives after twofold HBr elimination of the primary Diels-Alder adducts under the conditions of reaction. Later in 1986, Paddon-Row and Patney 27 used this method to annulate naphthalene units to norbornene and norbornadiene systems.

4

SYNTHESIS OF MOLECULAR TWEEZERS AND CLIPS

107

The reaction of 31 with sodium iodide at 65 ~ in the presence of norbornene 3028 led to the corresponding naphtho-substituted norbornene 34 in 66% yield. 29 Compound 34 was also available via a second route starting from benzocyclobutene 35. 21 At 200 ~ 35 underwent an electrocyclic ring opening leading to o-quinodimethane 15, which could be intercepted by norbornene 30 to produce the Diels-Alder adduct 36. It was not necessary to isolate 36, which can be immediately dehydrogenated by DDQ to give 34. Twofold HC1 elimination of 34 with potassium hydroxide in the presence of crown ether ([ 18]crown-6) in THF led to the desired diene 13 in 74% yield. Although the overall yield of the second route (73%) was slightly higher than that of the first (66%), the first route to 34 was preferred due to the availability of 31 vis-~-vis 35. IV. Synthesis of Molecular Tweezers and Clips

The tweezer and clip molecules 1 - - 8 24,25,29-32 c a n be synthesized by the use of a molecular LEGO set consisting of the bis-dienophiles 11, 12, and 14 and the dienes 10, 13, 15, and 16, which were discussed in the previous section. The key steps in the synthesis of the tetra- and trimethylene-bridged tweezers 1, 2 and clips 3, 4 are repetitive DielsAlder reactions proceeding with a high degree of stereoselectivity on the exo-face of the bis-dienophile 11, 12, or 14 and on the endo-face of the diene 10 or 13, leading to the bis-adducts 37, 38, and 39, respectively (Scheme 6). The reaction of diene 10 with norbornadiene (14) and of 13 with 14 led to the Diels-Alder bis-adducts as shown in Scheme 6. In each adduct, all methylene bridges are syn to one another. Oxidative dehydrogenation of the cyclohexene rings in these bis-adducts by the use of DDQ produced 1-4 in reasonable overall yields (11-60%). The synthesis of the molecular clips of type 5 by repetitive Diels-Alder reactions of bis-dienophile 11 with o-quinodimethane 15 as suggested by the retrosynthetic analysis failed. From the reaction of l l b with 15 (generated in situ by thermolysis of benzocyclobutene 35 at 200 ~ the (2:1) Diels-Alder cycloadduct could be isolated in almost quantitative yield. However, the DDQ oxidation of the cycloadduct has failed to date. The successful one-pot synthesis of the molecular clips of type 5 and 6 started with the in situ generation of dibromo-o-quinodimethane 32 by 1,4-Br 2 elimination from tetrabromo-o-xylene 31 with sodium iodide and

108

FRANK-GERRITKLARNERANDMIREIACAMPAlq,/~KUCHENBRANDT R1

a1 160-170~ 3-5d toluene r

11

7

10

R1 o

R1

CI.~CN C~CNo~-

R2

yield

H H 33% b OAc OAc 59% d OMe OMe 30% g OCH2CO2Et OCH2CO2Et 28% a

. ~

160~ 5d toluene 12

38

10

o Ck.~CN

R1

R2

yield

H

30% 53% 29%

C~CN

O

a

b d

H

OAc OMe

OAc OMe

~+2 14

\ 170~ 3d 28 %

DDQ

I1120~

~

1 4 7 % ~

/=f 4

~

13

SCHEME6

repetitive Diels-Alder reactions of 32 with the bis-dienophile 11 or 12. The resulting cycloadducts were not stable and eliminated four molecules of HBr each under the conditions of reaction. This led to the desired clip molecules as shown in Scheme 7 for the diacetoxy-substituted systems 5b and 6b. 31

4

109

SYNTHESIS OF MOLECULAR TWEEZERS AND CLIPS

~ BF Br

BF

Br Nal, CaCO3 DMF, 55 ~ 100 mbar 82 %

~.

//....

11b

2

31Br

r

F

1

-4HBr ID,

gr

2

[~Br

Br

+ ~

Nal, NEt3 DMF, 65~ 100 mbar 6o %

Br

12b SCHEME 7

The preparation of the anthracene clip 7d was first tried starting from 2,3-bis(dibromomethyl)naphthalene and bis-dienophile l i d as building blocks analogously to the successful synthesis of the naphthalene clips 5. The one-pot reaction shown in Scheme 8, however, did not lead to clip 7d. The observation of dibromonaphthocyclobutene as the product suggested that the 1,4-Br 2 elimination of 2,3-bis(dibromomethyl)naphthalene by NaI proceeded to generate o-naphthoquinodimethane, but that the unimolecular electrocyclic ring closure to dibromonaphthocyclobutene was faster than the desired bimolecular Diels-Alder cycloaddition to bisdienophile l i d . Since the aromaticity of both rings has to be given up in o-naphthoquinodimethane, this intermediate is certainly less stable and hence more reactive than the corresponding o-quinodimethane derivative (generated from o-di(dibromomethyl)benzene under similar conditions). Evidently, in the case of o-naphthoquinodimethane, the bimolecular Diels-Alder reaction, which is limited in its rate by diffusion, cannot compete with the unimolecular electrocyclization.

110

FRANK-GERRIT KLARNER AND MIREIA CAMPAI~A KUCHENBRANDT

gr

~

gr

Br

Br

Nal, NEt3

Br

DMF

~B 1~[" e ~Br r3

SCHEME 8

The anthracene clips 7b-d could be prepared in four to six steps as shown in Scheme 9. 32 The tetrabromo-o-quinodimethane 42 (generated by 1,4-Br 2 elimination from hexabromo-o-xylene 41) reacted with bisdienophile l l b to afford the Diels-Alder adduct 43, which spontaneously eliminated HBr under the reaction conditions to give clip 44b. The acetoxy groups in 44b were converted into methoxy groups by basic ester hydrolysis and subsequent methylation without isolation of the intermediate hydroquinone 44e. The dimethoxy-substituted clip 44d could also be prepared directly in 34% yield starting from hexabromo-o-xylene 41 and bis-dienophile l l d . However, in this case, the isolation of pure 44d from the reaction mixture by liquid chromatographic separation turned out to be more difficult than in the case of 44b. Therefore, preparation of 44d via 44b was preferred. Debromination of 44d with n-butyllithium produced a formal bis-aryne, which was trapped with furan to afford a mixture of all three diastereomeric (2:1) cycloadducts syn, syn-, syn, anti-, and anti, anti-45d. The desired clip 7d could be obtained from this mixture of adducts without separation by deoxygenation with low valent titanium generated in situ from titanium tetrachloride and zinc powder. The dimethoxy-substituted clip 7d was converted into the hydroquinone clip 7e by reaction with boron tribromide 32,33and 7e was converted into the diacetoxy-substituted derivative 7b by esterification with acetic anhydride. Starting from the tetrabromonaphthalene clip 44d, the molecular clip 8d (with expanded benzo[k]fluoranthene sidewalls 34,35) could be prepared

4

111

SYNTHESIS OF MOLECULAR TWEEZERS AND CLIPS

Br Br"..~~Br

Br Nal, CaCO3,. B r . ~ . ~

11b

O100F'mbar LB r / " ~ ' ~ ' ~ 41 Br

65 %

42

Br B_~Ot"Ac

.

Br

Br" Br

~ Br ~BB r-

43

OM

1. PhNH-NH2, KOH, IPrOH, RT

-4 HBr

2. tBuOK, Mel, RT B( ur

90 %

i~r'Br

Br

B

44b

OM n-BuLi, C ~

TiCl4 / Zn _-THF, reflux 64 %

n-hexane / THF

-78oc - RT 40 %

O

O

syn, anti-45d anti, anti-45d

7d

OH

BBr3 CH2Cl2 -78~ - RT 98 %

Ac20 pyridine RT, 24 h 67 % 7b SCHEME 9

by a sequence of palladium-catalyzed Suzuki-Heck type couplings with 1-naphthaleneboronic acid analogously to the coupling between 1,2dibromobenzene and 1-naphthaleneboronic acid leading to fluoranthene. 36 The hydroquinone 8e and diacetoxy clip 8b could be prepared by conversion of the methoxy groups in the central spacer-unit of 8d into hydroxy and acetoxy groups (Scheme 10). Because the yield of the Suzuki-Heck type coupling was only 20%, an attempt was made to prepare clip 8b by the reaction of bis-dienophile l l b with bis(dibromomethyl)fluoranthene 46,

112

FRANK-GERRIT KLARNER AND MIREIA CAMPAlqA KUCHENBRANDT HO.B.OH

Pde(dba)3, P(Cy)3 DBU B~' "

44d

B~rkBr

DMF 155~ 48 h 20 %

BBr3

Ac20

0H2012 -78~ - RT 82 %

pyridine RT, 24 h 70 %

8r

8b

SCHEME 10

which was prepared in three steps as shown in Scheme 11. However, the reaction of l l b with 46 led only to the (1:1) adduct 47b, even with a large excess of 46. The adduct 47b could, however, be used to prepare the unsymmetrical clip 48b. Most natural receptors are chiral. To date, only few synthetic chiral receptors are known. 37 The molecular tweezers and clips discussed so far are achiral. There are two possibilities to obtain chiral derivatives. In the case of the naphthalene-spaced tweezers and clips of type 2 and 6, two different substituents in the central spacer-unit are sufficient to produce chiral systems (e.g., tweezers 2i-1 and clips 6i-1 are chiral). Chiral systems can also be obtained by substitution of the terminal tweezer or clip arene units. In Figure 2, chiral mono- and disubstituted clips are shown as representative examples. In the case of disubstituted clips, the formation of diastereomers is expected. Mixtures of the disubstituted clips meso- and rac-49, and rac-50 could be easily prepared by the reaction of bis-dienophile l l b with the methyl or (-)-menthyl 3,4-bis(dibromomethyl)benzoate ortho-quinodimethane shown in Scheme 12. The (-)-menthyl ester derivatives anti-50 and anti'50 are enantiomeric to each other only in the skeleton, but actually diastereomeric in a strict sense. Similarly, syn-50 is not a pure meso, but only a "mesoid" isomer with chiral substituents at the tips.

4

113

SYNTHESISOF MOLECULARTWEEZERSAND CLIPS

@

hydroquinone

_--

toluene, sealed tube 155~ 36 h 53% O Ci,,~CI gr

CI" "tf" "CI O

S,hoo,4

Xylene Reflux, 5 h 77 %

reflux, 36 h 99 %

Br 46

Nal, CaCO 3 DMF, 55~ 100 mbar

11b

gr Br Br -2 HBr

Nal, CaCO 3 ~

16%

.],

DMF, 55~ 100 mbar 47b

51%

48b

SCHEME 11

R2

R2

R2

R2

R2

FIGURE 2. Structures of chiral and meso-configurated dimethylene-bridgedmolecular clips.

114 0

FRANK-GERRIT KL,~RNER AND MIREIA CAMPAlq,/~ KUCHENBRANDT

Br

k~..~..Br Br

F

0

DMF, 55 ~ 80 mbar

L

COOR

ROOC

Br] - 4 HBr Brj

OA

ROOC

COOR

meso

COOR

rac

V (-)-Menthyl

R

49 OH3

50 (-)-Menthyl

yield

80 %

52 %

SCHEME 12

8.0

P3

6.0

> 4.0

.................

00x44

0 .rX ffl

"

2.0

c-

I

0

I

10

I

I

I

20

30

I

I

40

time/min FIGURE 3. HPLC data for a mixture of anti-50, anti'-50, and syn-50 detected by UV (bottom) and CD (upper) spectrometry at 254 nm.

4

115

SYNTHESIS OF MOLECULAR TWEEZERS AND CLIPS

These methyl- and (-)-menthyl ester derivatives could be separated by the use of HPLC on chiral columns as shown in Figure 3 for the (-)-menthyl esters anti-50, anti'-50, and syn-50. The structures of the separated stereoisomers were assigned by CD and NMR spectroscopy. Syn-49 as well as syn-50 were CD-silent and in their ~H NMR spectra the expected two signals were observed for the chemically nonequivalent acetoxymethyl groups, whereas in the spectrum of anti-49 and anti-50 or anti'-49 and anti'-50, only one signal was observed for these protons, as expected from the symmetry of these molecules. 38,39 Because the reaction of equimolar amounts of bis-dienophile l l b with 1,2-bis(dibromomethyl)benzene 31 led only to a mixture of 1:1 and 2:1 adducts (as well as l l b ) , the rational synthesis of the racemic carbmethoxy- and nitro-substituted clips 56b and 57b (Scheme 13) began with the preparation of the naphtho-substituted dienophile 55b.

BrBr

OAc

gr

~OAc AcO 51b

31= /ff

Br Nal,CaCO3,DMF . 100mbar, 55~ 5 h

OAc

~

83 %

NaOH, DDQ

CH2CI 2 70%

.~

52b

53

9

(Ac)20, DMAP r pyridine

70 %

toluene

-78 ~ RT 75~o 35

9

65

gr RI"~

Br Br 56 RI=CO2Me gr 57 RI=NO2

Nal,CaCO3, DMF 100 mbar,55~

55b

6h

65%(R1"CO2CH3) 54% (RI: NO2) 57b (RI=NO2) SCHEME 13

R1

116

FRANK-GERRIT KLARNER AND MIREIA CAMPAlq/~ KUCHENBRANDT

The reaction of diacetoxybenzonorbornadiene 51b with 31 in the presence of sodium iodide gave naphthonorbornadiene 52b in good yield. Compound 52b was converted to the corresponding quinone 53 by hydrolysis and oxidation of the acetoxy groups. This quinone reacted with 1,3-cyclopentadiene in a Diels-Alder reaction to give a (65:35) mixture of syn and anti adducts 54, which were easily separated by recrystallization and MPLC. After basic enolization and acetylation of the keto functions in syn-54, the reaction of 55b with tetrabromo-oxylene derivative 56 or 57 led to the desired racemate of mono-carbmethoxyor nitro-substituted clip 56b or 57b. The enantiomers of the carbmethoxy clip 56b could be again separated by HPLC on a chiral column. 39 The tweezer and clip derivatives c-I could be prepared by transformation of the acetate groups into other functional groups by standard methods. 24 The synthesis of the parent benzene tweezer la 25 and the chemically bonded stationary phases CBSP-benzo and CBSP-naphtho, 4~ which can be used for the HPLC analysis of substrate binding affinities (Scheme 14), illustrate the methods used for tweezer and clip functionalization. Most of the tweezers and clips discussed here are colorless and absorb light in the UV range because the arene subunits (e.g., benzene, naphthalene, or anthracene, which function as chromophores) are insulated from each other by sp3-hybridized carbon atoms of the norbornadiene subunits. Exceptions are the yellow clips 8b-d and 48b,. which contain benzo[k]fluoranthene sidewalls and absorb light in the visible region (/~max = 405 nm in CHC13). This band can be assigned to the expanded aromatic subunits of 8b-d and 4 8 b . 34'35 Therefore, it was quite surprising to find out that the quinones of the naphthalene, anthracene, and benzo[k]fluoranthene clips, which can be easily prepared by DDQ oxidation of the corresponding hydroquinone in almost quantitative yield, are intensely colored (Figure 4). Quantum chemical ab initio calculations of the UV/Vis spectra of the quinone clips containing benzene, naphthalene, or anthracene sidewalls suggest that the bathochromic shift of the absorption band of the longest wavelength observed for the quinones results from an increasing configuration interaction between the n to rt* excitation of the quinoid system and the rt to rt* excitation of the aromatic sidewalls (charge-transfer excitation). For the excited states, this configuration interaction can be explained to be a "through-bond" homoconjugation similar to that postulated for the quinhydrone of trypticene. 4~

4

117

SYNTHESIS OF MOLECULAR TWEEZERS AND CLIPS

1[}

LiAIH4, THF ~. 98%

Tf20, pyridine

98%

PdCI2(PPh3)2, HCO2H Ph2PCH2CH2CH2PPh2 NBu3, DMF 82%

lb (n=0) 2b (n=l)

NaOH,20~ 1,4-dioxane

_

1)HSiMe2Cl H2PtCI6(kat.) 2) ETCH,NEt3 92% (n = O) 87% (n = 1)

.

~ ~,~'"-- ~ C)/

Me~"Si~.Me ~

~

Me/Sif..Me l i I L

5 pm silica

H3

L'I-13

C A

CBSP-benzo (n = O) 0.07 mmol/g silica CBSP-naphtho (n = 1) 0.03 mmol/g silica

SCHEME 14

V. Thermodynamic Parameters of Host-Guest Complex Formation with Molecular Tweezers and Clips The magnetic anisotropy of the host arene units makes ~H NMR spectroscopy a very sensitive probe for uncovering the complexation of guest molecules inside the cavities of the tweezers and clips 1-8. The complex formation can be easily detected by pronounced upfield shifts of the guest signals in the 1H NMR spectrum of a mixture consisting of one of the host molecules 1-8 and one of the guest molecules. 42

118

FRANK-GERRIT KL,g,RNER AND MIREIA CAMPAN,/~ KUCHENBRANDT

t~

orange-red

_

dark blue

~max = 423 nm log e - 2 98

9 \

'...

\

.'"""i

~, -':;4~: _

"

.... ~,~

\

2 = 515 nm

:~::: ";~max= 470 nm

log

gll

\

'.

2.58

,og~.= 2.89

\

\

"" ~",~ '".

dark red-purple

A

-'"

\

\

;-max = 537 nm

"

~ : i log,; = 3.17

"

\

/

r

o

,v

..........-/ ,,

,

/

\ ....

'-.~

..-~

_

\,

-~,

-~

,

,

,

300

400

500

-....

\

\\

'...

\k\'~

"""i" 600

~'hX~"

,

700

2 [nm] FIGURE 4. Top: structures of the quinone clips and colors of their solids. Bottom: UV/Vis spectra of the anthracene clip (solid line), naphthalene clip (dashed line), and benzo[k]fluoranthene clip (dotted line) in CHC13 at 25 ~ (See color insert.)

4

SYNTHESIS OF MOLECULAR TWEEZERS AND CLIPS

119

The maximum complexation-induced ~H NMR shifts, A6max, of the guest protons, the association constants, Ka, and, hence, the Gibbs enthalpies of association, AG, were determined by ~H NMR titration experiments (Tables 1 and 2). In some cases, for example, the complexes between 1,2,4,5-tetracyanobenzene (TCNB) and the naphthalene tweezers 2a and 2b 24'25 or the trimethylene clip 4 29 (which are bright yellow due to charge-transfer bands at 420, 420, and 412 nm, respectively) were too stable for a determination of the association constants by NMR titration experiments. In these cases, the K~ values were obtained either by spectrofluorimetric and spectrophotometric titrations 43 or by isothermal titration calorimetry. 29 The molecular tweezers and clips 1-8 bind a variety of electrondeficient neutral and cationic substrates inside their cavities. Electronrich neutral or anionic guest molecules are not bound by these host molecules within the limits of NMR detection. Comparison of the naphthalene-spaced tweezer 2 with its smaller benzene-spaced analogue 1 (Tables 1 and 2) demonstrates that 2 is a better receptor for aromatic guest molecules than 1, whereas aliphatic guests such as acetonitrile, malononitrile, or di-n-butylammonium tetrafluoroborate are only bound inside the smaller cavity of 1. 25,44 The trimethylene clip 4 is, accordingly, a better receptor for aromatic guest molecules than the benzene-spaced tweezers 1, but usually a poorer receptor than the naphthalene tweezers 2 for guest molecules beating "small" substituents such as the linear sp-hybridized cyano group or nonbranched alkyl groups. 29An exception is TCNB, which forms a more stable complex with 4 than with 2b. Evidently, the rt stacking between the very electron-poor benzene ring of TCNB and the electron-rich naphthalene sidewalls of 4 is especially favorable. With sterically more demanding guest molecules, such as TNF (Table 1), 4 forms more stable complexes than 2. Similar receptor properties are observed for the dimethylene-bridged clips 5 and 7. 24'31'32 Because of the reduced number of methylene bridges from 4 to 3 to 2, the topology of the clip molecules 4-8 is more open than that of the tweezers 1 and 2, allowing sterically larger guest molecules to be included into the cavities of these clip molecules. But the reduction of the arene binding sites from five in tweezers 1, 2 to four and three in clips 3, 4 and 5-8, respectively, leads in most cases to a decrease in the complex stability. The finding that the anthracene hydroquinone clip 7e forms more stable complexes than the corresponding naphthalene clip 5e, can be explained by the larger van der Waals contact surface of the

120

FRANK-GERRIT KL)kRNER AND MIREIA CAMPAlq/i~ KUCHENBRANDT

TABLE 1 Association constants, K a [M-~], Gibbs enthalpies, AG [kcal/mol], and maximum complexation-induced 1H NMR shifts of the guest protons, A6max = 60 - 6complex,for the formation of host-guest complexes of molecular tweezers 1, 2 and clips 4--8 with neutral guests in CDC13 at room temperature 24,2529,32,35,42,45,46

Guest

Host

Ka

AG

A6ma x

la

l0

- 1.3

lb

40 110 110 43

-2.2 -2.8 -2.7 -2.2

3.5 2.2 4.3 4.1 2.9

> 105a 7.3 X 105

< -- 6.7 -8.0

4

1.4 X 10 7 (1"1) b

--9.8

5b 5e 7b 7e 7d 8b 8d

4.4 • 104 (2:1) b 140 2200 690 12,800 220 918 157 (11) 436 (2:1)

-6.3 -2.9 -4.6 - 3.9 - 5.6 -3.2 -4.0 -3.0 -3.6

4.7 3.5 3.6 4.1 4.7 4.0 3.8 3.5 2.4 2.9 3.6 3.3

2a

p-DCNB

2b 4 2a

2b NO

ON

NC

g

"CN

TCNB

la 2a

4

1100

-4.1

> 105~ 2600

<-6.7 -4.6

5.9

,b

NC TCNQ

5c

137

- 2.9

2.6

7b 7e 7d

130 640 40

-2.9 - 3.8 -2.2

2.4 3.4 1.3

8b 8d

584 117 (1"1) 229 (2:1)

-3.8 -2.8 -3.2

2.7 2.9 2.8

4

-2.9 -2.3 -3.8 -5.0

NC, CN l ~ e

5c

O 2 N?. y. ~N\ O / ~2 / - - - -t,- ~ \ \

7b 7c

130 50 570 4900

7d 8b

270 >2 X 104

-3.3 <--5.9

1.3(a), 1.1(b), 1.8(c), 1.5(d), 1.0(e) 1.6(a), 2.1(c), 1.9(e)

TNF

8d

6700

-5.2

1.4(a), 1.8(a), 1.8(c), 1.8(d), 2.1(e)

CH3CN

la

15

- 1.6

5.3

CH2(CN)2

la

35

- 2.1

4.5

a

b -'~

.

c'- d

NO2

aEstimated value, bDetermined by isothermal titration calorimetry.

1.1(a), 0.5(a), 1.0(a), 1.0(a),

1.5(b), 0.7(b), 0.6(b), 1.0(b),

4.6(c), 2.2(c), 2.7(c), 2.9(c),

4.0(d), 2.8(d), 3.3(d), 2.8(d),

2.1(e) 1.2(e) 1.7(e) 1.5(e)

TABLE 2 Association constants, Ka [M-l], Gibbs enthalpies, AG [kcal/mol], and maximum complexationinduced IH NMR shifts of the guest protons, A6max = g0 ~complex,for the formation of host-guest complexes of molecular tweezers 1, 2 and clips 4-8 with cationic or zwitterionic guests in CDC13 at room temperature 25,35,38'42,45,48 -

Guest

C02Me b , Et

I|

KS (Kosower salt)

tBu

Betaine

Host

-

Ka

AG

A6max

2a

1100

-4.1

4.1(a)

2b

3800

-4.8

4.0(a)

4

410

-3.6

3.4(a)

5b

140

-2.9

1.8(a), 2.4(b)

7b

360

-3.5

1.7(a), 2.5(b)

8b

475

-3.6

2.1(b)

2a

710

-3.9

2.0(a), 2.8(b), 1.6(c), 0.5(d), -0.1(e)

2b

950 a

-4.1

4.8(b)

4

32

-2.1

1.5(b)

la

568

-2.4

1.5(b)

la

130 c

-2.9

1.6(b)

4

990 c

-4.0

0.7(b)

(De)2B2+

7c

73 c

-2.5

2.3(a), 1.3(b)

G BE 4

la

25 c

-1.9

2.5

Tr +.BF 4

2b

6750 a

-5.2

3.2

n-Bu2NH~-BF4

la

30

-2.0

3.2(~-H)

anti'-50

13,600 e

- 5.6

anti'-50

3900 e

-4.9

R - Me

R I

(Me)2B2+

N b

9

R = (CH2)10OH

I

| R 2PF6

|

Ha,~CO2Me ~I~C I| N-~

H D-Trp-OMe-HC1

H3N| N ~ IcO2Me ~ ie H L-Trp-OMe-HC1 aIn CDC13/CD3OD (5"1). bin CDC13/acetone-d6 (1:2). Cln CDC13/acetone-d6 (1"1). JIn CDC13/CD3OD (11). eDetermined by CD titration experiments in THF/CH3OH/H20 (4:1:5).

122

FRANK-GERRIT KLARNER AND MIREIA CAMPAlqA KUCHENBRANDT

anthracene sidewalls. With small guests such as TCNB and TCNQ, dimethoxy clip 8d (with expanded benzofluoranthene sidewalls) forms a mixture of (1" 1) and (2:1) complexes, whereas the corresponding diacetoxy clip 8b binds these guests with 1"1 stoichiometry. With the larger guest TNF, both clips form highly stable 1"1 complexes. In the case of the complex TNF@8b, the guest binding quenches the long-wave emission of 8b at '~max -- 442 rim, which will be used for the determination of the complex stability. 47 The chiral recognition ability of chiral host molecules was examined with the optically active menthyl naphthalenecarboxylate clip a n t i ' - 5 0 by using the amino acid derivatives, L- and I)-tryptophan methyl esters (Trp-OMe.HC1), as chiral guests. The complex of a n t i ' - 5 0 with I)-TrpOMe.HC1 turned out to have an association constant KD - 13,600 M -1 and was substantially more stable than that with L-Trp-OMe-HC1 (KL = 3900 M -1, Table 2) Therefore, this chiral clip displays a relatively high enantioselectivity of K J K L - 3.5 compared to native and modified cyclodextrins (enantioselectivity - 1.3-3.6). 3s The binding of the core of dendrimers by the diacetoxynaphthalene tweezer 2b is particularly interesting. 49,5~Because of their large structures, dendrimers are extensively used as host molecules for a variety of metal ions or molecules. But in spite of their large structures, they can also be involved as guests in molecular recognition phenomena. In such cases, the host species does not interact with the whole dendritic structure, but only with specific component units. When the potential guest unit constitutes the core of a symmetric dendrimer, host-guest complex formation, which requires the threading of a ring-shaped host, cannot take place. In these cases, a tweezer can be used to clip the dendritic core and, hence, to influence the reactivity of the dendrimers. Tweezer 2b forms highly stable complexes with the core of the dendrimers shown in Figure 5. The fluorescence band of the tweezer 2b was quenched by addition of one of the dendrimers. This finding could be applied to determine the thermodynamic data listed in Table 3 by spectrofluorimetric titrations. In a few cases, these data could be independently determined by 1H NMR titrations. The NMR measurements also provide the information on the complexation-induced shifts, A6m~x, of the guest signals. The high values of A~Sma x a r e good evidence for the inclusion of the bipyridinium core inside the tweezer cavity. In the following section, these complex structures will be discussed in more detail.

4

123

SYNTHESIS OF MOLECULAR TWEEZERS AND CLIPS

0 ~o

Hm

+

(D0)2B2+- [ -

~m'~w' D1B2+

0

%

0

(D1)2B2+

0

+

O

D2B2+

~o~

h ~ , #+- ~ o ~ o~ T

(D3)2B2+

0

o ~ L~

FIGURE 5. Structures of dendritic guest molecules with bipyridinium cores.

VI. Host-Guest Structures and Dynamics

The structures of several complexes could be determined by singlecrystal structure analyses. 25,51 According to the structures shown in Figure 6, the naphthalene-spaced tweezer 2a has an almost ideal topology for the complexation of benzene derivatives, while the complexation of these substrates by the benzene-spaced tweezer 1 requires a substantial distortion of the receptor geometry. This distortion certainly explains why the complexes of aromatic or quinoid substrates with the benzene-spaced

124

FRANK-GERRIT KLARNER AND MIREIA CAMPAI~IA KUCHENBRANDT

TABLE 3 Association constants, K a [M-l], Gibbs enthalpies, AG [kcal/mol], and maximum complexationinduced 1H NMR shifts of the guest protons, A6max -- 60 - 6complex, for the formation of host-guest complexes of diacetoxynaphthalene tweezers 2b with the bipyridinium core of dendrimers as guests in CDzC12 at room temperature 49 Guest

D1B 2+ D2B 2+ D3B 2+ (D0)2B 2+ (D1)2B 2+ (D2)2B 2§ (D3)2 B2+

K a (• 103)

29 16 16b 8.4 27 b 18 b 9b

AG

-6.1 -5.7 -5.7 -5.4 -6.0 -5.8 -5.4

A6max

H~

H~

0.79 0.61

2.93 . .

0.91 1.35 c

Hm

.

2.41 . .

.

3.22 _a

H m'

N +-CH 2

N +-C H2CH 3

3.19

0.19 0.13

1.23 0.72

0.17 0.28 c

aNot detectable because of the signal broadening, bSpectrofluorimetric titration in CH2C12 at room temperature. CA6maxin CDzClz/acetone-d 6 (1:2).

p-DCNB @ 2a

TCN B @ 2a

TCNQ@ l a

(DO) 2B2+@ l h

TCNQ @ 2a

Cs+@ l k

FIGURE 6. Single-crystal structure analyses of complexes between molecular tweezers and various arenes and Cs + as guests.

receptor 1 are less stable than the corresponding complexes of the naphthalene-spaced receptor 2. The benzene-spaced tweezers lh and lk surprisingly form complexes with the cesium cation, Cs +. The formation of (Cs+)2 @ lh can be detected in its ~H NMR spectrum in CD3OD by the downfield shift of OCHzCOO- signal (A6=-0.15). According to the single-crystal structure (Figure 6) the Cs + cation interacts with four of the five benzene units inside the cavity of lk. No complexation is observed

4

125

SYNTHESIS OF MOLECULAR TWEEZERS AND CLIPS

d = 11.4A 5b

-"

d :8.3A KS@5b

i

d= 14.5A 7d

v

d:6.5A TCNB@7c

FIGURE 7. Single-crystal structure analyses of the naphthalene and anthracene clips 5b, 7d and the host-guest complexes KS@5b and TCNB@7c. 31,32

for the corresponding potassium salts. Cs + has obviously the optimum size (ionic radius 167 pm) whereas K + (133 pro) is too small for these multiple attractive cation-arene interactions, so that the stability usually observed for (1:1) alkali metal cation-arene complexes (Li + > Na + > K + > Cs +) is reversed in this case for K + and Cs +. Based on the single-crystal structures (Figure 7), the distance between the naphthalene sidewalls in the dimethylene-bridged clip 5b has to be compressed from 11.4 ,~ in empty 5b to 8.3 A in the complex of the Kosower salt KS@5b. 3~ The increase in steric strain resulting from this compression is certainly one reason why the complexes of 5 are usually less stable than those of the tetra- and trimethylene-bridged host molecules 2 and 4. A larger and even more impressive compression of this distance from 14.5 to 6.5 A is observed for the formation of the TCNB complex of anthracene clip 7b. 32According to force-field calculations, the expansion and compression of the sidewalls by angle distortion and outof-plane deformation of the aromatic sidewalls in 5a and 7a are lowenergy processes. For example, the compression from 10 (the global minimum) to 8 A in the parent naphthalene clip 5a and from 12.4 to 6.5

126

FRANK-GERRIT KL~i.RNER AND MIREIA CAMPAIq/~ KUCHENBRANDT

Kal + TCN B _.. -

Ka2 [TCNB@DNN]

21 ~

+ DN N _ -

[ T C N B @ 2 DNN]

CHCI3:

Kal = 360 M -1, Ka2 = 24 M -1 Kg = [ T C N B @ 2 D N N ] / [ T C N B ] = 8600 M -2

[DNN] 2

A(~max = 6.0 ppm T C N B @ 2 DNN

FIGURE 8. Formation and single-crystal structure of the (2:1) complex TCNB@2 DNN. 3

in the parent anthracene clip 7a, is calculated to require an energy of approximately 1.5 and 3.5 kcal/mol, respectively, which is, apparently, more than compensated by the noncovalent attractive host-guest interactions in the complexes of g and 7. DNN 9 ("the monomethylene-bridged clip") forms a 2:1 complex with TCNB in the crystal and in solution as well (Figure 8). In the crystalline state, the complex shows an optimal arrangement of the TCNB molecule between two DNN molecules without any distortion of the receptor geometry. There are attractive CH-rc and slipped face-to-face rt-rt interactions between TCNB and the naphthalene rings of DNN. Evidently, the gain in enthalpy resulting from the arene-arene interactions overcomes the unfavorable entropy term for the formation of a termolecular associate. Thus, the 2:1 complex is also stable in solution. 3~ Besides the single-crystal structures, the complexation-induced chemical 1H NMR shifts, A6max, of the guest protons provide important information on the complex structures, as has been recently shown for the complexes of p-DCNB, TCNB, and TCNQ as guest molecules with the parent naphthalene tweezer 2a as host. In the solid state NMR spectra of these complexes, the signals of the guest as well as host protons could be assigned by the use of ~H DQ MAS and 1H-~3C correlation spectra (Figure 9, bottom). 52,53 In the spectrum of complex p-DCNB@2a, two signals at & = 5.6 and 2.0 are observed for the nonequivalent p-DCNB protons H a and H b pointing either out of the tweezer's cavity or toward the benzene tings of the sidewalls (Figure 9, top, left), whereas in each spectrum of the two other complexes,

4

127

SYNTHESIS OF M O L E C U L A R TWEEZERS AND CLIPS

,

,

,

d~

#

...... ,

e ...

............

......

:+=.~ ..........~

~

. , , . ...... ,..

ji

2

4

i!!::~i:= i

I0

,. 150

140

....

...:=

.... ~ .

~ ......

~ ~'.~,o~

ppnl ............................................... 0

'

130

120 ppm

ppm

ppm i

-2 q

-2 ~I

01 2J

0 2

4~

4

6q

6

8q 1o -t

8 10 1150 140

130

1120 ppm

..... - - ~ ' ~ r * .... ~v-~'~---~'r9. . . . . t50 140 130 120 ppm

F I G U R E 9. Top and middle: single-crystal structure analyses of the host-guest complexes p - D C N B @2a (left), TCNB@2a (middle), T C N Q @ 2 a (right). Bottom: 1H rotor-synchronized 1H-13C R E P T - H S Q C

solid-state NMR spectra. (See color insert.)

only one signal is observed for the guest protons (in the case of TCNB @2a at 6 = 1.8 and of TCNQ@2a at 6 = 3.4). In addition, some of the signals assigned to the host protons, especially those that are close to the arene units of a neighbor tweezer molecule in the crystal lattice, show unusually large upfield shifts in the solid-state 1H NMR spectra. The 1H NMR chemical shifts of the guest and host protons were computed by quantum chemical ab initio methods for the monomeric complex as well as for larger segments of the crystal lattice (containing up to five complexes). These calculations were in good agreement with the solid-state ~H NMR data. The ~H NMR chemical shifts of the guest protons in the solution-state spectra of the complexes TCNB @2a and TCNQ@2a were very similar to those observed in the solid-state spectra (c5 = 2.0 and 3.9, respectively). Evidently, the complex structures in solution closely resembled those in the crystal. In the solution-state, however, the 1H NMR spectrum of

128

FRANK-GERRIT KLARNER AND MIREIA CAMPAlq/~ KUCHENBRANDT

p-DCNB @2a showed only one signal at 6 - 3.5 for the guest protons H a and H b e v e n at low temperature ( - 7 0 ~ This indicates that in solution the exchange of the nonequivalent protons H a and H b, resulting from mutual complex dissociation-association and/or rotation of the guest molecule p-DCNB inside the tweezer cavity, is fast with respect to the NMR timescale over a broad range of temperature (from + 21 to - 7 0 ~ This assumption is supported by a solid-state NMR study at different temperatures. In the solid-state 1H NMR spectrum of p-DCNB@2a, a broadening and finally a coalescence of the separated signals of H a and H b w e r e observed upon heating to 137 ~ From this observation, an activation barrier of AG ~ = 17.2 kcal/mol could be estimated for the exchange of H a and H b inside the tweezer cavity in the crystalline state. Two dynamic processes can be envisaged, which are consistent with the exchange of H a and H b, namely either a 180 ~ "rotation" around the long axis of the guest molecule or a 60 ~ "flip" between the two equivalent sites in the complex (Figure 10). According to quantum chemical and force-field calculations for the isolated complex in the gas-phase, the 60 ~ "flip" has a very low activation barrier (<-2 kcal/mol) and seems to be clearly favored over the 180 ~ "rotation" (calculated activation barrier c a . 8 kcal/mol) in view of the energy. These results certainly explain the missing temperature

H~ /~N

N

60~ " F l i p ' ~

N /Hb

N

' ~ ~ 8 0 ~ "Rotation"

H~ /~N

N

a) ab initio calculation (GIAO-HF/'IZP), b) force-field calculation (MMFF94) FIGURE 10. Left: possible conformational isomerization processes calculated for the exchange of the guest protons. Right: top view on the crystal lattice of the complex p-DCNB@2a. 52

4

129

SYNTHESIS OF MOLECULAR TWEEZERS AND CLIPS

dependence of the solution-state ~H NMR spectrum of p-DCNB@2a. However, if a larger fragment of the solid-state structure is considered (Figure 10), it becomes clear that the 60 ~ "flip" would move the guest CN groups in the transition state too close to the atoms of the neighboring complex. That prevents this motion and leads to the conclusion that the exchange of H a and H b in the solid state occurs via the 180 ~ "rotation" which does not affect the positions of the CN groups to each other. This steric restriction only exists in the crystalline state but not in the solution state where the complex molecules are separated from each other by solvent molecules. In solution, the low-energy "flip" can, therefore, certainly occur without steric restriction. In the TCNB complexes of the diacetoxy-substituted tweezer 2b and trimethylene-bridged clip 4, the TCNB protons are expected to be chemically nonequivalent. In the ~H NMR spectrum of each complex at 298 and 255 K, respectively, only one signal for both protons is observed, which is broadened by lowering the temperature and is finally split into two signals at 218 and 168 K, respectively. These exchange processes can be explained by a rotation of the TCNB molecule inside the tweezer or clip cavity. 42,46 From the line-shape analyses, the Gibbs activation enthalpies were determined to be AG ~ = 11.7 and 8.2 kcal/mol, respectively. The activation barriers calculated by force field (MMFF94) were of the same order of magnitude. These processes can be considered to be the dynamic equilibration of noncovalent conformers (Figures 11 and 12).

12 10

,o

"~ 6

"~6

~ u3 4

-~ 4 u3

0 .

0

.

.

50

.

.

1O0

e/[o]

0

150

.

0

50

.

.

1O0

.

150

e/[o1

FIGURE 11. The rotational barriers calculated by force field (MMFF94) 54 for the movement of TCNB inside the cavity of tweezer 2a (left) and clip 4 (right).

130

FRANK-GERRIT KLARNER AND MIREIA CAMPAlq,/~ KUCHENBRANDT

TCNB@2b

TCNB@2b

....

~--I ~'''~ ~

"

3.2

I

'2.8'

I

~

3.0

'

'

2.6'

2.i

l.lSocl i

......... i

~

3.2

3.0

I

i

2.8

-55~

~(H a, H b) = 2.4, 3.0 (slow exchange)

25~

6(H ~, H b) = 2.8 (fast exchange)

I ............ ~ .......... I

2.6

2.4

12s-cl

'3.2'

..... i~

......

,

3.0' '2.8'

2.6'

!

AG r = 11.7 k c a l / m o l

2.g

(ppm) -105~ H

-..

-

b

." C~H!TM N

NC

i

TCNB@4

•(H a, H b) = 2.4, 3.0 (slow exchange)

- 18~

~(H a, H b) = 2.8

(fast exchange) AG ~= = 8.2 k c a l / m o l

TCNB@4

FIGURE 12. Equilibration between noncovalent conformers of the TCNB complexes of tweezer 2b (top) and clip 4 (bottom) detected by the temperature dependence of the 1H NMR spectra of the host-guest complexes in toluene-d 6. The ~H NMR spectra of TCNB @2b at different temperatures are shown as representative example (middle). 42,46

In the ~H NMR spectra of 2:1 mixtures of tweezers 2a,b, or clip 4 with TCNB, tropylium tetrafluoroborate, Tr + BF4-, or bipyridinium salts (D0)2B 2+ 2PF6-, (D1)2B2+ 2PF6-, and D1B 2+ 2PF6-, which form stable host-guest complexes, separate signals of the 1:1 complexes and the excess free host were observed at low temperatures. In Figure 13, the temperature

4

131

SYNTHESIS OF MOLECULAR TWEEZERS AND CLIPS

TCNB@2a (H e)

TCNB@2a (H (H c)

\

(C|2CH__)

2a

2

(Ha) I

I

(H e)

j _ _ .;l__

21~ _

(C|2CH__)

2

e

e

61~

NC" ~ "CN TCNB (C12CH__) 2

101oC . . . . . . . . . . t

7,0

,

,

,

!

!

6,5

6,0

....

.

|

5,5

.... !

5,0

4,5

. . . .

i

,

4,0

6 [ppm] FIGURE 13. ~H NMR spectra (300 MHz, CDzCI2) of tweezer 2a and TCNB ([2a] 0 = 0.02 M, [TCNB] 0 = 0.01 M) at different temperatures. The superimposed red lines fit to the peaks result from the line-shape analysis. 46 (See color insert.)

dependence of the ~H NMR spectrum of a 2:1 mixture of tweezer 2a and TCNB is shown as a representative example. At low temperature, mutual complex formation and dissociation is slow with respect to the NMR timescale. An increase in temperature leads to a broadening and finally to a coalescence of these signals. From the line-shape analyses of the temperature-dependent spectra, the activation parameters for the dissociation of the complexes shown in Figure 14 can be determined. The comparison of the activation parameters shows that in the TCNB complex of clip 4, the Gibbs activation barriers of the guest rotation inside the host cavity and of the complex dissociation are smaller by 3-4 kcal/mol than those in the corresponding TCNB complex of tweezer 2b or 2a. This can be explained with the more open topology of the clip molecule 4. The finding of negative activation entropies for the dissociation processes seems to be surprising, but can be understood by the calculation of the transition states of the host-guest complex dissociation. Accordingly, in these transition states (Figure 15), the guest rotation inside the cavity has to be restricted and the guest molecule is still clipped between the tips of the host molecule. Both processes contribute negative terms to the entropy of activation.

132

FRANK-GERRIT KLARNER AND MIREIA CAMPAlq/i. KUCHENBRANDT

i+0 +01 9

%.

O+

O+

TCNB @2a

A~ AG~ T [~

Tr +@ 2b

= = = =

13.2 -10.6 16.8 60.9

8.8 -12.4 12.3 2.4

TCNB@ 4

9.5 -11.4 12.4 17.6

FIGURE 14. Activation parameters A H + [kcal/mol], AS ~ [cal/mol K], AG * [kcal/mol] and temperature of coalescence T [~ determined for the dissociation of the complexes TCNB @2a, Tr + @2b, and TCNB @4 from the temperature-dependent 1H NMR spectra of (2:1) mixtures of 2a and TCNB in (CDC12)2, 2b and Tr + BF 4- in CDC13/CD3OD (1:1), and 4 and TCNB in toluene-ds. 25,46,51(See color insert.)

TCNB .... 2a

TCNB .... 4

FIGURE 15. Transition states of the complex dissociation 46 calculated by force field (MMFF94).

From the thermodynamic and kinetic data already reported, the Gibbs # activation enthalpies of association, AGa~~, can be calculated (Figure 16). These data indicate that the binding of the guest by the molecular tweezers 1 and 2 requires a substantial activation barrier and does not occur via

4

133

SYNTHESIS OF MOLECULAR TWEEZERS AND CLIPS

AG

[kcal/mol]

l

[G . . . . . . . T] ~

T

AG~ass

1 A

diss

A G-ill ...... +

.................. T

GoT

Complex TCNB @ 2b Tr+@2b (D0)2B 2+@2b (D 1)2B 2+@ 2b D 1B 2+@2b (D0)2B 2+@ la

AG*~iss 15.7 12.3 12.6 12.0 10.5 12.0

AG,~, -7.9 -12.3 -3.9 -4.4 -4.3 -2.4

AGr 7.8 7.1 8.7 7.6 6.2 9.6

AG [kcal/mol] I

AG*diss = 12.4

1

. . . . . T. . . . . . AGas~ = - 9.5

.....l

.

.

.

.

I

*as~= 2.9

+ NC~ C N NC" v "CN

.

FIGURE 16. Gibbs enthalpy profiles of association and dissociation of the host-guest complexes with tweezers l a and 2b (top) 46'49 and clip 4 (bottom). 46

134

FRANK-GERRIT KL/~RNER AND MIREIA CAMPAlq/~ KUCHENBRANDT

diffusion-controlled process, whereas the activation barrier of association is much lower (close to diffusion control) for the complex formation between TCNB and clip 4. Also remarkable is the observation that the binding of the bipyridinium salt (D0)2B2+ (Figure 5) by the benzene-spaced tweezer la exhibits the largest activation barrier of the systems studied here even though this complex is thermodynamically less stable than the corresponding one of naphthalene-spaced tweezer 2b. All these findings 49 can be again explained with the different tweezer and clip topologies. The bond angles and the planes of the aromatic units of the molecular tweezers 1 and 2 have to be distorted during the guest insertion through the tips of the host molecule to a larger extent than those of clip 4 during this process. In addition, the observation that the activation barrier of binding of the bipyridinium salt D1B 2+ is smaller than that of (D 1)2B2+, indicates that in the case of D 1B2+, with only one dendron, complex formation can occur by threading the guest from the less substituted side through the open face of the tweezer. In the case of the host-guest complexes of the tweezers with the bipyridinium salts, the structures, dynamics, and reactivity are remarkable. According to the single-crystal structure of complex (D0)zB2+@l h (Figure 6) and force-field calculations of the other complexes, only one ring of the bipyridinium salt is positioned inside the tweezer cavity. Thus, different chemical ~H NMR shifts are expected for guest protons located inside and outside the host cavity. Contrary to this expectation, only one 1H NMR signal is observed for each type of guest protons (Table 3). Evidently, the complexes of the symmetrically substituted bipyridinium salts possess nonsymmetrical structures and the tweezer shuttles 49 rapidly from one to the other pyridinium ring, so that only averaged ~H NMR guest signals are observed. In the 1H NMR spectrum of (D1)2 B2+ @2b, a splitting of the signals assigned to the tweezer bridgehead protons was observed at low temperature ( - 1 0 5 ~ indicating the shuttling of the tweezer along the axis of the bipyridinium unit. From this splitting and the temperature of coalescence, the activation barrier was calculated for the shuttling process to be AG * = 9.2 kcal/mol, which is in good agreement with the barrier calculated by force-field methods (Figure 17). The different complexation-induced chemical 1H NMR shifts of the protons of the unsymmetrically substituted bipyridinium salt (D 1B 2+ and D2B 2+, Table 3) indicate that in these complexes, the N-ethyl-substituted pyridinium ring is preferentially included inside the tweezer cavity. The bipyridinium units of the dendrimers DnB 2+ and (Dn)zB2+ ( / 7 = 0 - - 3) are electroactive and undergo two successive, reversible, one-electron reduction processes at easily accessible potentials that correspond to the formation

4

135

SYNTHESIS OF MOLECULAR TWEEZERS AND CLIPS Ir

~

>

,::

,$

,,

> .........

.......

FIGURE 17. The shuttling process of tweezer 2a along the axis of the N,N-dimethyl-bipyridinium dication calculated by force field MMFF94 (AEc,,I # c -- 8 kcal/mol). 49

of a cation radical (B 2+ --->B +) and a neutral (B + ~ B) species. The cyclic voltammetric (CV) pattern for reduction of the dendritic cores is affected by the addition of tweezer 2b. In particular, both the cathodic and anodic peaks corresponding to the first one-electron reduction process of the bipyridinium core progressively move to more negative values upon addition of tweezer 2b (e.g., half-wave potential for the reduction of D1B 2+ and D1BZ+@2b: B 2+ ---->B+: - 0 . 2 9 or - 0 . 3 6 V), whereas the peaks corresponding to the second reduction process were practically unaffected (B + ~ B: - 0 . 7 7 or - 0 . 7 7 V). This finding indicates that the bipyridinium core is stabilized by complexation and that the complex of the radical cation resulting from the first reduction dissociates before the radical cation is further reduced. The behavior of the tweezer-bipyridinium complexes could also be determined in the gas phase by using mass spectroscopy (Figure 18). 55 The monoisotopic complex ions were isolated in a FT-ICR cell and subjected to collision with argon as collision gas. In the tweezer complex of (D0)2 B2+, the benzylic C - N bond is exclusively cleaved under these conditions leading to the tweezer-monocation complex, which subsequently dissociates the tweezer molecule, whereas the complex of the tweezer with the dendrimer of second generation (G2) loses the tweezer in the first step and then in a second step, the benzylic C - N bond in the bipyridinium core is cleaved. In the complex with dendrimer of first generation (G1), a competition between these two reactions - C - N bond cleavage and tweezer loss - is observed in the first step. Evidently, dendritic substituents of increasing generation at the bipyridinium core provide increasing stability to the naked dication in the gas phase so that the host-guest complex dissociation is favored over C - N bond cleavage with increasing dendrimer generation. This finding can be explained by force-field calculations (Figure 19). The G2

136

FRANK-GERRIT KL,g,RNER AND MIREIA CAMPAIqA KUCHENBRANDT

major pathway for

~ ~

(,N-2 R

///-- RCH?

- Tweezer"~

/~~-Tweezer

-RCH?/4R.,,.N..~ (~)

both pathways observed f o r- ~- ~ Ar LO Ar^O''~''N'~

G r~N~OvAr %

(D0)2B2+

| r~-N---R ~ R.,,.N/~~'~ Q

major pathway for Ar~ tAr O,1~O

=

~O

OAr N~-O'X3"O"Ar

r..O

Ar

Ar"~

"Ar

FIGURE 18. Pattern of fragmentation determined for tweezer-bipyridinium complexes by the use of electrospray ionization Fourier transform ion cyclotron resonance (ESI-FT-ICR) mass spectroscopy in the gas-phase: competition between tweezer loss and benzylic C-N bond cleavage in dependence of the size of the dendritic substituents. 55

FIGURE 19. Monte-Carlo simulation of bipyridinium structures (MMFF94). Left: (D0)2B2+, middle: (D1)(Me)B 2+, and right: (D2)(Me)B 2+. (See color insert.)

dendron is calculated to completely engulf the bipyridinium dication by back folding of the dendritic 'arms' leading to an 'intramolecular solvation' and hence to a stabilization of the bipyridinium core. The bipyridinium core is, however, calculated to be less efficiently contained by

4

137

SYNTHESIS OF MOLECULAR TWEEZERS AND CLIPS

folding of the G1 and GO dendron, respectively, resulting in the instability observed for naked dications in the gas phase. The decrease in the binding constants observed 49 for the complex formation between tweezer 2b and the dendrimers of increasing generation in solution (Table 3) may be the result of the stabilization of the bipyridinium core by back folding of the dendritic groups, which is evidently also efficient in CH2C12solution. The molecular tweezers and clips are able to influence chemical reactions as well. One example, which has been already discussed, is the reduction of bipyridinium salts in the presence of tweezer 2b. Another example is the methylation of pyridine derivatives such as 4-cyanopyridine. This reaction is highly accelerated in the presence of molecular tweezer 2b, but in chloroform equimolar amounts of the tweezer are required to run the reaction. 56 Thus, the tweezer does not function as a catalyst because the p r o d u c t - N-methyl-4-cyanopyridinium iodide - forms a highly stable host-guest complex with 2b that causes product inhibition. The product inhibition can be avoided by using a stirred two-phase solvent system consisting of water and chloroform. The starting material and the host-guest complex of the pyridinium salt with tweezer 2b resulting as primary product of the reaction are soluble in chloroform, whereas the decomplexed s a l t - N-methyl-4-cyanopyridiniumiodide - is only soluble in water. After the reaction, the free salt is transferred to the water phase, the empty tweezer remains in the chloroform phase and is ready to bind the starting material and accelerate its reaction with methyl iodide once again. Under the two-phase conditions, the tweezer catalyzes the reaction substantially (Figure 20). A complete kinetic and thermodynamic analysis shows that this reaction is accelerated by the tweezer by a factor of kcat/kunc~ t = 3010 and CN

CN 25 ~

100 h, H20/CHCI 3

+ CH3--1 @ I

CH3 in the presence of 2a (5 mol%) in the absence of 2a

I

yield: 95% <5%

FIGURE 20. Methylation of 4-cyanopyridine catalyzed by the molecular tweezer 2b in the twophase system (CHCI3/H20). ~6

138

FRANK-GERRIT KLARNER AND MIREIA CAMPAlqA KUCHENBRANDT

the transition state of the catalyzed reaction is stabilized by AGxs = - 6 . 4 kcal/mol compared to the uncatalyzed methylation of 4-cyanopyridine (Figure 21, top). A similar stabilization of the transition state resulting from the binding to a cyclophane receptor has been observed by Dougherty e t al. 57 for the reaction of N,N-dimethylaniline with methyl iodide (Figure 21, bottom). In this case, however, the factor of acceleration was determined to be only kcat/kunca t = 5, because the ground state of the starting N,N-dimethylaniline is already substantially stabilized by binding to the cyclophane receptor, contrary to the weak binding of 4-cyanopyridine to the tweezer 2b. The comparison of these two systems is an instructive example for the paradigm formulated by Linus Pauling more than 60 years ago that an efficient catalyst has to stabilize the transition state of a reaction rather than the ground states being involved. VII. Synthesis of Water-Soluble Molecular Tweezers and Clips

Efficient synthetic receptors with the capability for selective substrate binding in aqueous solution are important for the understanding and controlling of molecular recognition and self-assembly in chemical and biological systems. The molecular tweezers and clips discussed thus far are only soluble in organic solvents. Therefore, water-soluble derivatives substituted by hydrophilic ionic or neutral groups were prepared. In this review, only the tweezers le, 2e and clips 5e,f and 7e substituted by phosphonate or phosphate groups are discussed. The phosphonate-substituted tweezer le 58 and clips 5 e , 59 7 e 6~ could be prepared starting from the corresponding hydroquinones le, 5e, and 7e (Scheme 15). The deprotonated forms of these hydroquinones were treated with methanephosphonic acid dichloride followed either by methanolysis or hydrolysis of the remaining chloro substituents. The resulting methyl phosphonate or phosphonic acid was converted to the desired lithium phosphonates le, 5e, and 7e, by reaction with either LiBr or LiOH. The phosphate-substituted clip 5 f 61 w a s prepared by reaction of hydroquinone 5c with POC13 and Et3N, followed by hydrolysis with dilute HC1 and neutralization with 4 equiv, of LiOH (Scheme 16). For the synthesis of the tweezer 2e, a different route 6~had to be applied since the naphthohydroquinone tweezer 2e is highly sensitive to oxidation and has not been isolated to date. Therefore, hydroquinone bis-dienophile 12e (R = OH) 23 was converted to the corresponding methyl methanephosphonate 12n by reaction with CH3P(O)C12 in the presence of triethylamine

4

139

SYNTHESIS OF MOLECULAR TWEEZERS AND CLIPS

AG [kcal/mol] ~k

TS uncat

/l

i"......................................~T~--o,,

/ ~uncq

~_~ca,

/2~:4at~

/2i.i. . . . .1 /

i

CHa

OAc

"I

....................................

ON ~ass--~'~:,,~

/OAcI. \

I-

N+ I

CH3

I

AG [kcal/mol]

H3C,N,CH3

~,~-

FIGURE 21. Top: methylation of 4-cyanopyridine catalyzed by the molecular tweezer 2b in the twophase system (CHC13/H20)56; bottom: methylation of N,N-dimethylaniline accelerated by a cyclophane receptor. 57

140

FRANK-GERRIT KL.&RNERAND MIREIA CAMPAIq,/~KUCHENBRANDT O II | | Me--P--O Li I

OH

O 1) MePOCI2

R ~ , . ,

2) HClaq 3) 2LiOH

OH

R,R: ~

R

O

I | |

Me--P--O ki II O

yield:

~

~

5c

45 %

5e

7c

58 %

7e

1) MePOCI2 2) MeOH 3) LiBr yield: 45 %

le

SCHEME 15

OPO32- Li + 5c

1) POCI3 2) HClaq 3) 4LiOH OPO32- Li +

82%

5f

SCHEME 16

followed by methanolysis (yield: 30%). Compound 12n reacted stereoselectively with diene 10 via repetitive Diels-Alder cycloadditions, affording the bis-adduct 38n. At atmospheric pressure and high temperature (1 bar and 160 ~ only products of decomposition either of the cycloadduct or the starting materials were observed. An increase in pressure up to 12 kbar led to the expected acceleration of the bimolecular cycloaddition 62 and allowed the temperature of reaction to be lowered to 80 ~ Under these conditions, the desired cycloadduct 38n could be isolated in 65% yield. Subsequent DDQ oxidation of 38n led to tweezer 2o in 40% yield. Final dealkylation of the methyl phosphonate 2o was mildly

4

141

SYNTHESIS OF MOLECULAR TWEEZERS AND CLIPS

Me~P~OMe 1) MePOCI2

12kbar,80~

2) MeOH 30 %

I

Me~P~OMe

12c

II

65%

0

0

12n

o

II

Me--I~ll--OMe

e

|

Me--I~--OLi / 1) DDO, 120~ 2) LiBr 30 %

38n

v

2e

l./

SCHEME 17

effected by LiBr 63 and affords the desired naphthalene tweezer as lithium salt 2e in 75% yield (Scheme 17). VIII. Binding of Biologically Interesting Substrates such as Enzyme Cofactors and Nucleosides in Aqueous Solution

In aqueous solution, the phosphonate clip 5e forms a highly stable complex 59 with N-methylnicotinamide iodide (NMNA), which is frequently used as a model for the cofactor NAD § of many redox enzymes (Scheme 18). The large complexation-induced shifts (Afmax) of the signals assigned to the guest protons on the pyridinium ring suggest that this ring is positioned inside the clip cavity (Table 4). This assumption is confirmed by ab initio calculations of the chemical ~H NMR shifts for two different complex conformers, which were generated by a Monte-Carlo conformer search (MacroModel, force field: AMBER*/H20, Figure 22). A closer inspection of the experimental and calculated Almax values showed that the values calculated for one conformer (Figure 22, left) are larger and those calculated for the other conformer (Figure 22, right) smaller than the experimental values. This finding provided good evidence that the complex consists of both conformers equilibrating with each other

142

FRANK-GERRIT K L A R N E R AND MIREIA CAMPAlqA K U C H E N B R A N D T

F

R

O 5 [~~ 6\~ ~ 2

NH2 I

Me

R = OP(Me)O2 Li+: 5e R=OPO3 2 2Li + "5f

NMNA

O

NH2 5N NH2 6N ~1~-"2N

O

O "

o-,P-O-P,-O I'N I

i

O-

5

8A <~~~~N/~.,Jt 2A

6

9,

HO-P-O--L./o

~'A

OH

O- 4' &. I

OH OH

NAD§ [ ~

NH2

,

I

OH OH NMN NH2

NH2

5N

NH2

6N\N / 2N

O 0 , O-

I'N OH OH

O

8A NJIQN/..j 2A

-0 OH

9, HO-

1' I

I

OH OH AMP

OH OH

NH3+

NH2

N~N+~.. MeeN

Me 4' ~_~ \ TPP

N2

N

O-

I'A

NADH

8~

o

o

I

I

II II O--P--O--P--O O

O -

Me + !+ H 3 N ' ~ f " ~ S"-~

8~~~Y 2 N I

O~o -

4 l[~O'~r 1' OH OH TsO- SAM

S C H E M E 18

rapidly on the "NMR timescale" so that only averaged 1H NMR signals are observed for the guest protons. The enzyme cofactor itself, NAD +, is bound by clip 5e in water less efficiently by a factor of about 13 than the model compound NMNA.

4

143

SYNTHESIS OF MOLECULAR TWEEZERS AND CLIPS

TABLE 4 Association constants, Ka [M-~], Gibbs enthalpies, AG [kcal/mol], and maximum complexationinduced JH NMR shifts of the guest protons, Afm~x = 60 Ocomplex,for the formation of host-guest complexes of the phosphonate and phosphate clip 5e,f with N-methylnicotinamide iodide (NMNA) and enzyme cofactors in aqueous solution at room temperature 59,6j,64 -

Guest

NMNA

NAD +

NMN

AMP

NADH

TPP

SAM

Host

Ka

5e a 82,800 5e b 11,270 5f ~ 29,500

5e a 5e b 5f b

5e b 5f b

5e b 5f b

5e b 5f b

6200 4200 5630

550 1220

910 750

800 424

5e a,c 14,000

5e b 1200 5f b 5390

AG

-6.7 -5.5 -6.1

-5.2 -4.9 -5.1

-3.7 -4.2

A~max 2-H

4-H

5-H

6-H

N-CH 3

1.30 1.61 1.81

2.76 3.47 3.69

2.38 3.07 3.32

1.85 2.25 2.62

0.77 0.99 0.99

2N-H

4N-H

5N-H

6N-H

I'N-H

2A-H

8A-H

l'A-H

0.39 1.20 1.22

0.73 2.88 2.75

0.78 3.17 3.17

0.44 1.54 1.51

0.32 0.28 0.22

0.52 0.41

0.34 1.61 0.90

0.23 0.59 0.51

2-H

4-H

5-H

6-H

l'-H

4'-H

1.28 1.48

2.88 3.16

3.49 3.99

1.75 1.94

0.62 0.60

0.07 0.09

-4.0 -3.9

-4.0 -3.6

-5.6

-4.2 -5.1

2-H

8-H

1 '-H

3'-H

0.38 1.48

1.51 3.16

0.27 3.99

0.03 1.94

2N-H

4N-H

l'N-H

2A-H

8A-H

1 'A-H

0.22 0.12

0.20, 0.30

0.24 0.10

1.26 0.90

0.49 3.14

0.29 0.53

2-H

6-CH 3

N+CH2

4'-CH 3

5'-CH 2

5'-CH2CH 2

0.44

0.04

0.24

0.09

0.01

0.01

2-H

8-H

I'-H

4'-H

S+CH3

S+CH2CH

S ~-CH2CH2

1.0 1.l

0.9 0.9

0.5 0.5

0.7 0.8

1.7 1.8

0.6, 0.8 0.9

1.4, 4.7 0.8, 1.1

aln aqueous solution, bin aqueous buffer solution" pH 7.2, 1.06 mmol NaOH/1.33 mmol KH2PO 4 in 20 mL D20. C5e (R = OP(Me)O 2 +N(nBu)4 ).

FRANK-GERRITKLARNERANDMIREIACAMPAlq/~KUCHENBRANDT

144

!

2.6 ~ ~ ~135"34~~~

experimentalvalues calculatedby quantumchemicalmethods(HF/6-31G**) FIGURE 22. Comparison between A0max values obtained by quantum chemical calculations and IH NMR titration experiments in aqueous solution for the complex [email protected] (See color insert.)

In addition to the 1:1 complex (NAD+@5e), a 2:1 complex (2.NAD + @5e) was formed. The A6max values observed for the NAD § complexes in water were substantially smaller than those found for the NMNA complex and those calculated by ab initio methods for the two structures, one including the nicotinamide ring (Figure 23, left) and the other the adenine unit into the clip cavity (Figure 23, right). These findings exclude a major contribution of these two structures to the complex in aqueous solution. There are several reasons why the behavior of NAD § is different from that of NMNA. According to isothermal titration calorimetric (ITC) measurements, 59 NAD § forms a selfcomplex in water that has to be cleaved during the host-guest complex formation. This self-association, the sterically bulky ribose rings attached to the nicotinamide and adenine rings, and the negatively charged diphosphate unit may explain why the NAD § complex is less stable than the NMNA complex. In buffer at pH 7.2, which certainly mimics biological conditions better than pure water, NAD § forms a (1:1) complex with clip 5e displaying large A~max values for the nicotinamide

4

145

SYNTHESIS OF MOLECULAR TWEEZERS AND CLIPS

. ~:~ ,.. w~,,

C',tae ..,,"~ '

1.2

,

..4

c ~

.

,

.

.,

. ~

~

.. ,

, , ~

i

;i

~

(~

~

9

~

,,.

'

9

~

~iI

FIGURE 23. A6max values calculated by quantum chemical methods for two structures of complex NAD + @5e generated by a Monte-Carlo conformer search (MacroModel, AMBER*/H20). Left: the nicotinamide ring and right: the adenine-unit of NAD + is inserted into the clip cavity. 59 (See color insert.)

and adenine guest protons comparable to those of the NMNA complex of 5e. The comparison of the observed A~max values with those calculated for the two complex structures shown in Figure 23 allows the conclusion that in buffer solution, the nicotinamide ring, as well as the adenine unit, are inserted inside the clip cavity and a mutual interconversion proceeds rapidly on the "NMR timescale". The assumption that both units of NAD + are complexed inside the clip cavity has been confirmed with the finding that NMN and AMP, the two fragments formally obtained by the hydrolysis of the diphosphate group of NAD + both form complexes with clip 5e (Table 4). The question of why the behavior of NAD + is different in pure water and in buffer solution could be answered with the following observation. Commercially available NAD + as well as a 1:1 mixture of NAD + and clip 5e react in pure, acidic water solution (pH 3.3). Neutralization of an aqueous solution containing NAD + and 5e by the addition of dilute NaOH leads to an increase in the complexationinduced ~H NMR shifts of the guest signals to the values also observed

146

FRANK-GERRIT KLJ~RNER AND MIREIA CAMPAlq/~ KUCHENBRANDT

in buffer solution. Evidently, the change in the pH value and not the change in the salt concentration caused by the transition from pure water to buffer solution is responsible for the change in the complex structures and stoichiometry. In the case of the NMNA complex of 5e, the complex is less stable in buffer solution by a factor of 7.3 than in water. However, the differences in complexation-induced 1H NMR shifts are less dramatic than in the NAD + complex, indicating that the structural change here is small at the transition from water to buffer solution. A large solvent effect has also been observed for the complex formation of clip 5e with S-adenosylmethionine (SAM), which catalyzes a wide range of essential methylation reactions in nature. In pure water, the (1:1) complex precipitates by mixing solutions containing SAM and 5e. In buffer solution, a stable soluble (1:1) complex is formed exhibiting large A~ma x values for the methyl and methylene protons adjacent to the sulfonium group. Accordingly, the positively charged sulfonium group is located inside the clip cavity. The phosphate clip 51'65 shows the same tendencies as the phosphonate clip 5e for complex formation (Table 4) with the guest molecules shown in Scheme 18 in both water and buffer solution. Compared to 5e, the phosphate clip 5f shows a higher selectivity in the binding of the nucleosides containing a pyrimidine base (Figure 24). Whereas clips 5e and 5f preferentially bind flat guest molecules between their plane naphthalene sidewalls, the benzene-spaced tweezer le shows a high selectivity for the basic amino acids, lysine and arginine, which play a key role in numerous biological recognition processes 58 (Figure 25). Besides the amino acid derivatives (e.g., AcLysOMe, TsArgOEt, and AcHisOMe, Figure 25) oligopeptides such as KKLVFF (Ka = 38,000 M -~) are bound by tweezer le. In contrast to the benzene-spaced tweezer le and the naphthalene clip 5e, which do not show a significant self-assembly in water, the naphthalene-spaced tweezer 2e and the anthracene clip 7e form the highly stable self-assembled dimers (2e) 2 and (7e)2 in aqueous solution, whereas in methanol solution these systems exist as monomers. 6~The structures of the self-assembled systems could be elucidated by quantum chemical calculations of the ~H NMR shifts of the tweezer or clip signals, resulting from the formation of the dimers in water (Figure 26). The thermodynamic parameters were determined by 1H NMR dilution titrations at various temperatures for both equilibria (2.2e ~ (2e)2: AH =

4

147

SYNTHESIS OF MOLECULAR TWEEZERS AND CLIPS

NH2 2.15~ NI ~ N' ~ o,~,-oo8

0.1~'~0.06 i '

OH

2.20N

0,82 <~N I ~ N / . ~ N / / NH N

I

o-l o. l

O

-oo~,oo~ 0.60

0.07~'~ 0.3~ 0.48

[

r

l

OH

OH

Adenosine

OH

Guanosine

K a = 1 4 0 0 M "1

(K a = I 115

H2

I

K a = 1 070 M 1

M-1) O

(K a =

750 M-')

Nit 2.72 - 0.25

0.1,o.o

0.58 - 0.04 ]

HO---~O

[ 0.18

OH

N

2.63

I

0.76

OH

1

Uridine

r 0.49

OH OH Cytidine

K a = 5 2 4 0 M "1 K a = 9 6 9 0 M "1

(K a =

2 490 M-1)

(K a =

1 070 M-1)

FIGURE 24. Association constants, K,, and A~max values (in red) determined for the binding of the nucleosides to the phosphate clip 5f in buffer solution at pH 7.2. 65 In parenthesis the Ka values for the complex formation with phosphonate clip 5e are given. 64 (See color insert.)

[Ka [M-1] [

T

OH3,

,,O"

o i R~I. !"

4000

;,O

O

,

R"CH3

it

.oo

3000

2000 n=0-1

1000 m

m

m

m

~

m

m

m

/

Lys Arg His Asp Ser Thr Phe Leu Val Ala Gly FIGURE 25. Selectivity of the phosphonate-substituted tweezer le for N/C-protected amino acids' K~ values determined in buffered aqueous solution (c = 0.1 mM in 25 mM NaH2PO 4, pH 7.0). (See color insert.)

148

FRANK-GERRIT KLARNER AND MIREIA CAMPAlq,g, KUCHENBRANDT

A6ex p A'Scalc

0.7 3.1 b"

0.7 3.3

(H h, H h')

0.8

1.1

(H a)

2.3 1.0 0.4

3.4 1.1 0.4

(H b) (H c)

(H a' H a' )

b

2e

(2e)2 i

a,.

"

7e

(H d)

(7e)z

FIGURE 26. Comparison of the experimental and calculated ~H NMR shifts, A6 . . . . obtained from the ~H NMR spectra of 2e and 7e in CD3OD and D20 and from ~uantum chemical calculations of the monomers 2e, 7e and dimers (2e) 2, (7e)2.6~ (See color insert.)

-20.9 kcal/mol, TAS = -12.2 kcal/mol, AG = - 8 . 7 kcal/mol; 2.7e (7e)2: AH = - 13.8 kcal/mol, TAS = - 6 . 7 kcal/mol, AG = -7.1 kcal/ mol). In the presence of NMNA, each equilibrium is shifted toward the monomer by formation of a stable (1:1) host-guest complex. The thermodynamic parameters, particularly the negative enthalpy and entropy of association, provide good evidence that the unexpected selfassembly of tweezer 2e and clip 7e is the result of a nonclassical hydrophobic effect. Obviously, the enlargement of the tweezer's and clip's rt-face by additional benzene rings produces new host species with perfect self-complementarity. The spontaneous self-assembly of 2e and 7e is certainly further strongly supported by their rigid structure, leading to the highly ordered intertwined dimers.

4

SYNTHESIS OF MOLECULAR TWEEZERS AND CLIPS

IX.

Conclusion

149

The key steps in the synthesis of the molecular tweezers and clips of type 1-8 are repetitive stereocontrolled Diels-Alder reactions by using a molecular LEGO set consisting of bis-dienophiles and dienes. The tweezers and clips serve as selective hosts for electron-deficient aromatic and aliphatic guest molecules. The complex structures have been elucidated with X-ray, NMR, and computational methods. In most complexes of the tweezers 2, the plane of the aromatic guest is aligned parallel to the central naphthalene spacer-unit of 2, whereas in the complexes of the clips 4, 5, 7, and 8, the guest molecules are placed inside the host cavity with their plane of molecule nearly parallel to the naphthalene, anthracene, or benzo[k]fluoranthene sidewalls and orthogonal to the central spacer-unit. Evidently, attractive CH-rc as well as ~-rc interactions contribute to the complex stability. Quantum chemical calculation of the electrostatic potential surface (EPS) of host and guest molecules suggest that the attractive host-guest interactions observed here are predominantly electrostatic in nature. 66 But in the case of the host molecules with the expanded anthracene or benzo[k]fluoranthene sidewalls, dispersion provides an important contribution to the host-guest binding. In aqueous solution, hydrophobic interactions are predominant as demonstrated by the self-assembly of the phosphonate-substituted naphthalene tweezer 2e and anthracene clip 7e in water. The complex structures are not rigid, but quite flexible. The rotation of the noncovalently bound guest inside the host cavities could be observed by temperature-dependent 1H NMR measurements in solution and even in the solid state. This process is analogous to the rotation around a single C-C bond and hence can be considered to be an equilibration between noncovalent conformers. The molecular tweezer 2b was found to catalyze the methylation of 4-cyanopyridine; it shifted the redox potential of bipyridinium salts substituted by dendrimers to a more negative value by binding the bipyridinium core. The water-soluble clips 5e and 5t' form stable complexes with enzyme cofactors like NAD +, SAM, and TPP. The water-soluble tweezer le preferentially complexes with basic amino acid derivatives of lysine, argenine, and histidine. These findings are certainly a good starting point for a systematic investigation of the effect of water-soluble clips and tweezers on enzymatic processes involving cofactors that are bound by these host molecules.

150

FRANK-GERRITKLARNERAND MIREIACAMPAI~AKUCHENBRANDT Acknowledgments

E-G. K. thanks the Ph.D. students, postdocs, and colleagues cited in the references for their fine collaboration. This work has been supported by the Deutsche Forschungsgemeinschaft. References and Footnotes

1. (a) Lehn, J. M., Supramolecular Chemistry. Concepts and Perspectives, VCH: Weinheim, 1995. (b) Atwood, J. L., Davies, J. E. D., MacNicol, D. D., V6gtle, E, Suslick, K. S., Comprehensive Supramolecular Chemistry, Elsevier: Oxford, 1996. 2. (a) Prins, L. J., Reinhoudt, D. N., Timmerman, E, Angew. Chem. Int. Ed. 2001, 40, 2383. (b) Prins, L. J., Neuteboom, E. E., Paraschiv, V., Crego-Calama, M., Timmerman, E, Reinhoudt, D. N., J. Org. Chem. 2002, 67, 4808. 3. (a) Gallivan, J. E, Dougherty, D. A., J. Am. Chem. Soc. 2000, 122, 870. (b) Sakai, N., Mareda, J., Matile, S., Acc. Chem. Res. 2005, 38, 79. 4. Meyer, E. A., Castellano, R. K., Diederich, E, Angew. Chem. Int. Ed. 2003, 42, 1210. 5. (a) Hunter, C. A., Lawson, K. R., Perkins, J., Urch, C. J., J. Chem. Soc. Perkin Trans. 2 2001, 651. (b) Sinnokrot, M. O., Valeev, E. E, Sherrill, C. D., J. Am. Chem. Soc. 2002, 124, 10887. 6. (a)Kim, E., Paliwal, S., Wilcox, C. S., J. Am. Chem. Soc. 1998, 120, 11192. (b) Nakamura, K., Houk, K. N., Org. Lett. 1999, 1, 2049. 7. (a) Lehn, J. M., Angew. Chem. Int. Ed. 2002, 41, 3738. (b) Seidel, S. R., Stang, P. J., Acc. Chem. Res. 2002, 35, 972. (c) Waldmann, O., Ruben, M., Ziener, U., Mtiller, E, Lehn, J. M., Inorg. Chem. 2006, 45, 6535. (d) Yang, H. B., Ghosh, K., Das, N., Stang, P. J., Org. Lett. 2006, 8, 3991. 8. (a) Rekharsky, M. V., Inoue, Y., Chem. Rev. 1998, 98, 1875. (b) Tong, L. H., Lu, R. H., Inoue, Y., Prog. Chem. 2006, 18, 533. 9. (a)Diederich, E, Cyclophanes, Royal Society of Chemistry: Cambridge, 1991. (b) Ma, J. C., Dougherty, D. A., Chem. Rev. 1997, 97, 1303. (c) Philp, D., Stoddart, J. E, Angew. Chem. Int. Ed. EngL 1996, 35, 1155. 10. (a) Cram, D. J., Container Molecules and their Guests, Royal Society of Chemistry: Cambridge, 1994. (b) Warmuth, R., Yoon, J., Acc. Chem. Res. 2001, 34, 95. 11. Collet, A., Dutasta, J. P., Lozach, B., Canceill, J., Top. Curr. Chem. 1993, 165, 103. 12. Lee, J. W., Samal, S., Selvapalam, N., Kim, H.-J., Kim, K., Acc. Chem. Res. 2003, 36, 621. 13. (a) Hof, E, Craig, S. L., Nuckolls, C., Rebek, J., Angew. Chem. Int. Ed. 2002, 41, 1488. (b) Hooley, R. J., Van Anda, H. J., Rebek, J., J. Am. Chem. Soc. 2006, 128, 3894. 14. (a) Chen, C.-W., Whitlock, H. W., J. Am. Chem. Soc. 1978, 100, 4921. (b) Zimmerman, S. C., VanZyl, C. M., J. Am. Chem. Soc. 1987, 109, 7894. (c) Zimmerman, S. C., Saionz, K. W., J. Am. Chem. Soc. 1995, 117, 1175. (d) Fleischhauer, J., Harmata, M., Kahraman, M., Koslowski, A., Welch, C. J., Tetrahedron Lett. 1997, 38, 8655. (e) Kurebayashi, H., Haino, T., Usui, S., Fukazawa, Y., Tetrahedron 2001, 57, 8667. 15. Rowan, A. E., Elemans, J. A. A. W., Nolte, R. J. M., Acc. Chem. Res. 1999, 32, 995. 16. Burley, S. K., Petsko, G. A., Science 1985, 229, 23. 17. Cox, E. G., S., E R., Cruickshank, D. W. J., Smith, J. A. S., Proc. R. Soc. London 1958, A247, 1.

4

SYNTHESISOF MOLECULARTWEEZERSAND CLIPS

151

18. Zimmerman, S. C., Corbin, E S., Struct. Bond. 2000, 96, 63. 19. exo- and endo-addition of norbornene and norbornadiene derivatives are surveyed and discussed in: (a) Gleiter, R., B6hm, M. C., Pure Appl. Chem. 1983, 55, 237. (b) Brown, E K., Houk, K. N., J. Am. Chem. Soc. 1985, 107, 1971. (c) Wipff, C. G., Morokuma, K., Tetrahedron Lett. 1980, 21, 4445. (d) Houk, K. N., Paddon-Row, M. N., Caramella, P., Rondau, N. G. J., J. Am. Chem. Soc. 1981, 103, 2436. 20. Compound 10 was prepared in four steps starting from indene, see: (a) Butler, D. N., Snow, R. A., Can. J. Chem. 1975, 53, 256. (b) Alder, K., Pascher, E, Vagt, H., Chem. Bet. 1942, 75, 1501. The procedure was simplified by M. Kamieth: (c) Kamieth, M., Diplomarbeit, University GH Essen, 1995. 21. (a) Schiess, P., Rutschmann, S., Toan, V., Tetrahedron Lett. 1982, 23, 3665. (b) Luo, J., Hart, H., J. Org. Chem. 1987, 52, 4833. (c) Segura, J. L., Martin, N., Chem. Rev. 1999, 99, 3199. 22. Roth, W. R., Unger, C., Wasser, T., Liebigs Ann. 1996, 2155. 23. Benkhoff, J., Boese, R., Kl~irner, E-G., Liebigs Ann. Recl. 1997, 501. 24. Kl~irner, F.-G., Polkowska, J., Panitzky, J., Seelbach, U. P., Burkert, U., Kamieth, M., Baumann, M., Wigger, A. E., Boese, R., Bl~iser, D., Eur. J. Org. Chem. 2004, 1405. 25. Kl~irner,E-G., Burkert, U., Kamieth, M., Boese, R., Benet-Buchholz, J., Chem. Eur. J. 1999, 5, 1700. 26. (a) Cava, M. R, Napier, D. R., J. Am. Chem. Soc. 1957, 79, 1701. (b) Cava, M. R, Shirley, R. L., J. Am. Chem. Soc. 1960, 82, 654. 27. Paddon-Row, M. N., Patney, H. K., Harish, K., Synthesis 1986, 328. 28. Compound 30 was prepared by Diels-Alder reaction of 1,3-cyclopentadiene with 1,4-dichloro-2-butene: (a) Bowe, M. A. R, Miller, R. G. J., Rose, J. B., Wood, D. G. M., J. Org. Chem. 1960, 25, 1541. 1,4-Dichloro-2-butene was obtained by the reaction of 2-butene-l,4-diol with SOC12: (b) Brandsma, L., Preparative Acetylenic Chemistry, Elsevier: Amsterdam, 1971. (c) Gleiter, R., Merger, R., Nuber, B., J. Am. Chem. Soc. 1992, 114, 8921. 29. Kl~irner, E-G., Lobert, M., Naatz, U., Bandmann, H., Boese, R., Chem. Eur. J. 2003, 9, 5036. 30. Kl~irner, F.-G., Benkhoff, J., Boese, R., Burkert, U., Kamieth, M., Naatz, U., Angew. Chem. Int. Ecl. Engl. 1996, 35, 1130. 31. Kl~irner, E-G., Panitzky, J., Bl~iser, D., Boese, R., Tetrahedron 2001, 57, 3673. 32. Kl~irner,E-G., Kahlert, B., Boese, R., Bl~iser, D., Juris, A., Marchioni, E, Chem. Eur. J. 2005, 11, 3363. 33. The cleavage of aryl methyl ethers with boron tribromide was reported in Press, J. B. Synth. Commun. 1979, 9, 407. 34. Campafifi Kuchenbrandt, M., Kl~irner, F.-G., unpublished results. 35. Campafifi Kuchenbrandt, M., Dissertation, University of Duisburg-Essen, presumably 2007. 36. Wegner, H. A., Scott, L. T., de Meijere, A., J. Org. Chem. 2003, 68, 883. 37. For example, a chiral crown ether containing binaphthyl units reported by D. J. Cram: (a) Cram, D. J., Acc. Chem. Res. 1978, 11, 8. For a review on chiral tweezers, see: (b) Harmata, M., Acc. Chem. Res. 2004, 37, 862. 38. Fukuhara, G., Madenci, S., Polkowska, J., Kl~irner, E-G., Origane, Y., Mori, T., Wada, T., Inoue, Y., Chem. Eur. J. 2007, 13, 2473-2479.

152

FRANK-GERRITKL/~RNERAND MIREIACAMPA/qAKUCHENBRANDT

39. Madenci, S., Dissertation, University of Duisburg-Essen, 2006. 40. Kamieth, M., Burkert, U., Corbin, P. S., Dell, S. J., Zimmerman, S. C., Kl~irner, E-G., Eur. J. Org. Chem. 1999, 2741. 41. (a) Jansen, G., Kahlert, B., Kl~irner, E-G., Boese, R., Bl~iser, D., manuscript in preparation. (b) Iwamura, H., Makino, K., J. Chem. Soc. Chem. Commun. 1978, 720. 42. Kl~irner,E-G., Kahlert, B., Acc. Chem. Res. 2003, 36, 919. 43. Marchioni, E, Juris, A., Lobert, M., Seelbach, U. E, Kahlert, B., Kl~irner, E-G., New J. Chem. 2005, 29, 780. 44. Kamieth, M., Kl~irner, E-G., J. Prakt. Chem.~Chem. Ztg. 1999, 341,245. 45. Seelbach, U. E, Dissertation, University GH Essen, 2002. 46. Lobert,M., Bandmann, H., Burkert, U., Buchele, U. E, Podsadlowski, V., Kl~irner, E-G., Chem. Eur. J. 2006, 12, 1629. 47. Collaboration with Professor V. Balzani and Dr. E Ceroni, University of Bologna, Italy. 48. (a) Lobert, M., Dissertation, University of Duisburg-Essen, 2005. (b) Kahlert, B., Dissertation, University of Duisburg-Essen, 2005. 49. Balzani, V., Bandmann, H., Ceroni, E, Giansante, C., Hahn, U., Kl~irner, E-G., Miiller, U., Mtiller, W. M., Verhaelen, C., Vicinelli, V., V6gtle, E, J. Am. Chem. Soc. 2006, 128, 637. 50. Balzani, V., Ceroni, P., Giansante, C., Vicinelli, V., Kl~irner, E-G., Verhaelen, C., V6gtle, E, Hahn, U., Angew. Chem. Int. Ed. 2005, 44, 4574. 51. Kl~irner, E-G., Burkert, U., Kamieth, M., Boese, R., J. Phys. Org. Chem. 2000, 13, 604. 52. (a) Brown, S. E, Schaller, T., Seelbach, U. P., Koziol, E, Ochsenfeld, C., Kl~irner, E-G., Spiess, H. W., Angew. Chem. Int. Ed. 2001, 40, 717. (b) Ochsenfeld, C., Koziol, E, Brown, S. E, Schaller, T., Seelbach, U. E, Kl~irner, E-G., Solid State Nucl. Magn. Reson. 2002, 22, 128. 53. Schaller, T., Buechle, U. E, Kl~irner, E-G., Bl~iser, D., Boese, R., Brown, S. E, Spiess, H. W., Koziol, E, Kussmann, J., Ochsenfeld, C., J. Am. Chem. Soc. 2007, 129, 1293-1303. 54. (a) Halgren, T. A., J. Comput. Chem. 1996, 17, 490. (b) Spartan'04, Wavefunction, Inc.: Irvine, CA. 55. Schalley, C. A., Verhaelen, C., Kl~irner, E-G., Hahn, U., V6gtle, E, Angew. Chem. Int. Ed. 2005, 44, 477. 56. Verhaelen, C., Dissertation, University of Duisbug-Essen, presumably 2006. 57. McCurdy, A., Jimenez, L., Stauffer, D. A., Dougherty, D. A., J. Am. Chem. Soc. 1992, 114, 10314. 58. Fokkens, M., Schrader, T., Kl~irner, E-G., J. Am. Chem. Soc. 2005, 127, 14415. 59. Fokkens, M., Jasper, C., Schrader, T., Koziol, E, Ochsenfeld, C., Polkowska, J., Lobert, M., Kahlert, B., Kl~irner, E-G., Chem. Eur. J. 2005, I 1,477. 60. Kl~irner, E-G., Kahlert, B., Nellesen, A., Zienau, J., Ochsenfeld, C., Schrader, T., J. Am. Chem. Soc. 2006, 128, 4831. 61. Schrader, T., Fokkens, M., Kl~irner, E-G., Polkowska, J., Bastkowski, E, J. Org. Chem. 2005, 70, 10227. 62. van Eldik, R., Kl~irner, E-G., High Pressure Chemistry- Synthetic, Mechanistic and Supercritical Applications, Wiley-VCH: Weinheim, 2002.

4

SYNTHESISOF MOLECULARTWEEZERSAND CLIPS

153

63. (a) Rensing, S., Arendt, M., Springer, A., Grawe, T., Schrader, T., J. Org. Chem. 2001, 66, 5814. (b) Krawczyk, H., Synth. Commun. 1997, 27, 3151. 64. Polkowska, J., Bastkowski, E, Schrader, T., Kl~irner, F.-G., manuscript in preparation. 65. Bastkowski, E, Dissertation, University of Duisburg-Essen, presumably 2007. 66. Kamieth, M., Kl~irner, E-G., Diederich, F., Angew. Chem. Int. Ed. 1998, 37, 3303. Kl~irner, E-G., Panitzky, J., Preda, D., Scott, L. T., J. Mol. Model. 2000, 6, 318.

COLOR PLATE SECTION

orange-red

dark blue

/•

2max = 423 nm Inn ~ - 9 .qR '"~ . . . . . . -

=,..: .... ii~i~i /1, = 515 nm ~ log ~= 2.58

X

"-. 1 . " " i "-;4 .~" :

dark red-purple

......~"~

';{,max = 470 nm I o g ~ = 2.89

X

\

.""

\ \

....

.-: \

,Zmax = 537 nm log ~ = 3.17

.-

_

\\

co

o'1

Io .

'-/VIA

,,~ _

\.

.-

\

'-...---/ k " / ix

\ .-" .x

I ~'~'.-. - " ~

/ z /

'

\\\ ""-.,.. \

,

,

,

300

400

500

\

..

".

600

2 [nm] Please refer to Figure 4 of Chapter 4 in text for figure legend.

507

..~._q 700

508

COLOR PLATE SECTION

,

ppm -2 -

ppm ,

ppm -2

2 t o~

0

0-

2

2-

4

ii

6 8 l0 .........

[ .........

150

, .........

140

i .........

0

4-

tl!

61

8~

6810-

] .........

130

. . . . . . . . .

120 ppm

150

140

130

120

ppm

I

. . . . . . . . .

150

I

. . . . . . . . .

140

!

. . . . . . . . .

130

I

. . . . . . . . .

120

ppm

Please refer to Figure 9 of Chapter 4 in text for figure legend.

TCNB@2a (Ha)

TCNB@2a (H e)

2a (H e) ~j~

2a

(H a)

(HC) ~ 12n(HC)

(C|2CH__)2

21~ (CI2CH_)2 e

e

NC"~ "CN TCNB

(C12C~2 101~ .

7,0

.

.

.

!

6,5

.

.

.

.

.

.

6,0

.

.

!

.

5,5

| 5,0

| .... 4,5

A

!

4,0

6 [ppm] Please refer to Figure 13 of Chapter 4 in text for figure legend.

509

COLOR PLATE SECTION

J

f 3 +9

~

~

N

TCNB@ 2a

AS* AG ~ T [~

Tr+@2b

= = = =

13.2 -10.6 16.8 60.9

8.8 -12.4 12.3 2.4

TCNB@ 4

9.5 -11.4 12.4 17.6

Please refer to Figure 14 of Chapter 4 in text for figure legend.

~

R

Please refer to Figure 19 of Chapter 4 in text for figure legend.

510

COLOR PLATE SECTION

,%,,./ i ~ --

~

q Q

2.3 .7

2:.6

1..2 ~

~2.7 experimental values calculated by quantum chemical methods (HF/6-31G**) Please refer to Figure 22 of Chapter 4 in text for figure legend.

1.2 1.7

1.4

w~

Please refer to Figure 23 of Chapter 4 in text for figure legend.

511

COLOR PLATE SECTION

NH 2

-o.,~,-o.o~ HO-o

I

0.11 "~,0.06 ~ f " 1 r

OH

O

-o.o~,o.o~ HOo.jI 0.60

0.07 ~

OH

0.48

OH

Adenosine

OH

Guanosine

K a = 1 400 M -1

K a = 1 070 M -1

(K a = 1 115 M -1)

( K a = 750 M 4)

O

NH 2.72

N 2.63

I

- 0.25

HO~ ' - - - -JO--

~

-0.041

OH

H

0.58

2r 0.18

-0.211-0.20 O ~

o

I 0.76

OH

T--"--r

OH

0.49

OH

Cytidine

Uridine K a = 5 2 4 0 M "1

K a = 9 6 9 0 M "1

/'/(/&a2 490 =

M -~)

(K a = 1 070 M -1)

Please refer to Figure 24 of Chapter 4 in text for figure legend.

CH3 /O

T

n

o O'R',,. ,i "'.\ .." , ~ ~ P L - C H 3

5000

0 NH

4000

]-~

"~,~,<,,,u 6)

3000 2000

n=0-1

1000 i

l

~

H

H

H

i

~

H

l

l

Lys Arg His Asp Ser Thr Phe Leu Val Ala Gly

Please refer to Figure 25 of Chapter 4 in text for figure legend.

STRATEGIES AND TACTICS IN O R G A N I C SYNTHESIS, VOL. 7 9 2008 Elsevier Ltd. All rights reserved.

Chapter 5 APPLICATION OF C-GLYCOSIDES IN THE TOTAL SYNTHESIS OF (-)-GAMBIEROL Jon D. Rainier Department of Chemistry University of Utah Salt Lake City, UT 84112, USA

I. II. III. IV.

Introduction and Background Preliminary Studies Synthetic Strategies Total Synthesis of (-)-Gambierol A. A-C Tricycle B. F-H Subunit C. Subunit Coupling and Completion V. Summary Acknowledgements References and Footnotes

154 157 163 168 169 183 200 215 215 216

I. Introduction and Background

Dinoflagellates play a Jekyll and Hyde role in the marine environment; on the one hand they are an important part of the food chain, while on the other they have adverse effects on marine organisms, the environment, and human health owing to their ability to generate toxins. ~ These toxins have been implicated in a wide range of human health problems including food poisoning, neurotoxicity, and asthma-like symptoms arising from their ability to bind to ion channels. 2 A list of the most thoroughly studied dinoflagellates, as far as toxin production is concerned, would have to include Karenia brevis and Gambierdiscus toxicus. Toxins from G. toxicus have been implicated in ciguatera events and include ciguatoxin, gambieric acid, and gambierol (Figure 1).3 Toxins that come from K. brevis have been associated with red tide catastrophies and include the brevetoxins A and B, hemibrevetoxin B, and most recently, brevenal (Figure 2). 4

5

155

APPLICATION OF C-GLYCOSIDE IN THE SYNTHESIS OF (-)-GAMBIEROL Me

r,,~~

MeOH

H ~e~O~ H/I~'~OY

.~ " F f f ' ~ " ~

~

~O"H~'~O~o

HO~o~O/~V~g:mbieric HO,,. J " H J

""

.~-

acid A"

Me-H

/Me '~._

Me

_ HO- Me Me " O

HO

x__OH H

H

~H'~'L / ~ O ~ O y ~ \ H H O

--~

~-o-~,~,-o-~~-O~o.7..=~_)

j:;;,

[/ '"Me Ho

H

1

.... Me'O--(

o

gambierol

":"

----. : OH

H,-,

u,

Me

.

/

~ H

Me "oH

H ~

/-

H

0

HO" =

H

Me/ H U ~ H "

H

I

\-_O..?./-~'N~O\:

H~" OH ,OH

Me , OH

__ H

MeOH

ciguatoxin

I~le FIGURE l. Toxins from G. toxicus.

,oCHO H O-~-O H

.e, U ' - - J ~"~

"e,.e ~ ~:,, .-~-~.O:%o"H

HO. M e

;14 O

brevetoxinA

H',.Ojj)~-- " / / e

R"~ M~

Me ~

",)L--k I-I H

...+o_~

H~ HO ,

H

Me

H/----"'~--~OH

r_ ~

O "

"H

e'

brevetoxin B

H, Me

~

H

H

H

Me H/~x,,[/OH

~~176176

" ~o~~"

H OH brevenal: R = CHO brevenat dimethyl acetal: R = CH(OMe)2) FIGURE 2. Toxins from K. brevis.

OH hemibrevetoxin B

~ ~

156

JOND. RAINIER

Interestingly, some of the small molecules isolated from dinoflagellates have beneficial properties on the environment and potentially, on human health. Gambieric acid has been shown to be a potent antifungal agent, while brevenal and gambierol inhibit the binding of brevetoxin to VGSCs and may serve as antidotes to red tides. 5 In addition, brevenal has been shown to clear mucous from lung tissue at very low concentrations. As a consequence, it has been proposed that brevenal may lead to a therapy for cystic fibrosis. 6 We anticipate that other beneficial properties will be found as additional information is gathered on this fascinating family of natural products. Nakanishi has examined the biosynthesis of dinoflagellate toxins and has proposed that it is likely that they are derived from polyepoxide cascades (Scheme 1). 7 While the use of such a cascade to synthesize members of the dinoflagellate toxin family has not been successfully demonstrated thus far, impressive work has been accomplished in the McDonald and Jamison laboratories towards this goal. 8 HO

Me,~CHO Me

f..

Me

~

M.e M_e

k,,.O__.~ ~

o Ca::Me c6:

,~o,2...7,_;

....

/[0-"~

HO.

Me.~CHO O "" H ~ " H Hot',,//~ HO

Me Me Me Me H I--!_~ O _~,~,...--~,~.--~6 H ~H ~ : 0 : :,0 ~ n " l vLi " 10 ,,,~,, ~/H BREVETOXINB 0 Me i~e

SCHEME 1

The interesting structural features of the dinoflagellate toxins listed in Figures 1 and 2 have stimulated significant efforts from the synthetic community, including efforts from our laboratory. 9 From the outset, our strategy to these agents was to use an iterative approach where just a few reactions would be needed to generate their very symmetrical architectures. More explicitly, we wanted to couple a C-glycoside-forming reaction from a cyclic enol ether with an ring-closing metathesis (RCM)

5

157

APPLICATIONOF C-GLYCOSIDE IN THE SYNTHESIS OF (-)-GAMBIEROL

0 oxidation ~

" O

ring-closing

1. C-Nucleophile 2. ester formation

O" ~"Nu i

1

2

2. C-Nucleophile 3. ester formation

~'0

metathesis

i

4

3 Rp

1. oxidation

R'

Rp

~.~0.~

Nu

ring-closing

-0-1~ v

I~ "OH

metathesis

R'oR'

R'

6

5 SCHEME 2

reaction that would lead to a new cyclic enol ether (Scheme 2). Iteration of the sequence would quickly lead to more elaborate architectures. Shortly after beginning our efforts in this area, we decided that (+)hemibrevetoxin B would be an ideal and challenging target to showcase this approach. After completing this work, we chose to target gambierol, as it possessed both an interesting architecture and had an intriguing biological profile. Included in this chapter are a discussion of both our preliminary results and the application of the strategy outlined in Scheme 2 to the total synthesis of (-)-gambierol. II. Preliminary Studies

This work initiated in the summer and fall of 1996 at The University of Arizona. Fortunately, I was able to convince Shawn Allwein, a new firstyear graduate student to join my research group and to initiate our efforts. Subsequently, other students who contributed to this work included graduate students Jason Cox, Henry Johnson, and Scott Roberts. An undergraduate student, Brett Howard, also contributed, as did a very talented postodoc, Utpal Majumder. We continue these efforts and our current team consists of graduate students Scott Roberts, Jie Zhou, Clement Osei Akoto, and Yuan Zhang. Although both the C-glycoside and metathesis reactions that were central to the iterative approach outlined above were precedented in the literature, neither reaction sequence had been extensively studied. As far as C-glycoside synthesis was concerned, Kishi and Czernecki had independently used glycal anhydrides to generate C-glycosides. ~~In their studies, both groups had coupled cuprates with isolable Brigl anhydrides. ~ We were interested in an in situ preparation of C-glycosides utilizing dimethyl dioxirane (DMDO) as the oxidant in an analogous fashion to

158

JON D. RAINIER

Danishcfsky's work targeting the generation of glycosidic bonds. 12When we began our efforts, precedent for using metathesis chemistry to generate cyclic enol ethers from olefinic esters came from the work of the Nicolaou and Grubbs groups, where either the Tebbe reagent and/or the Petasis reagent had been used to convert olefinic esters into the corresponding cyclic enol ethers. ~3 Shawn first examined whether the iterative route in Scheme 2 could be carried out on the simplest system available, dihydropyran, and quickly discovered what others already knew" that oxygen substitution on the pyran rings helps to stabilize the corresponding glycosyl anhydrides. Following this realization, Shawn took a step back and examined the addition of C-nucleophiles to Danishefsky's anhydride 8, which comes from the reaction of tris-benzylidene D-glucal with DMDO. ~4 After some experimentation, he found that allyl magnesium chloride added to 8 with high stereoselectivity. ~5 Unknown to us at the time, Stick and Evans were independently examining nearly identical allylation reactions and had found similar results. ~6 Shawn and Jason Cox subsequently found that a wide variety of C-glycosides can be generated from glycosyl anhydrides and the appropriate nucleophile. ~7 Nucleophiles included propargyl Grignard, acetal Grignards, vinyl Grignard, methyl and aryl cuprates, and furyl zinc (Table 1). Our disappointments included our inability to stereoselectively couple homoallyl Grignard with 8 as the resulting product 16 would lead to the direct generation of oxepenes after RCM (entry 11). Having examined the C-glycoside-forming reaction, we next investigated the conversion of C-allyl glycoside 9 into the corresponding cyclic enol ether. Unfortunately, attempts at the single flask conversion mentioned above using olefinic ester 17 and the Tebbe or Petasis reagents were completely unsuccessful in our hands. Fortunately, however, a two-step Grubbs protocol that used an acyclic enol ether in an enol ether-olefin RCM reaction was more successful (Scheme 3). 18The conversion of an ester into the corresponding enol enol ether can be carried out using a variety of titanium reagents. In Grubbs' initial work and Clarke's subsequent studies targeting members of the ladder toxins, the Takai-Utimoto reagent was used to generate the acyclic enol ether and was therefore also employed by us. 19 We have subsequently found this reagent to be advantageous for a number of reasons, including its reactivity, which seems to lie somewhere between the Tebbe and Petasis reagents, that it is generated in situ, and, as discussed later in this chapter, that it is not limited to the generation of terminal enol ethers. In our hands, exposure of 17 to the Takai protocol resulted in the

5

159

APPLICATIONOF C-GLYCOSIDE IN THE SYNTHESIS OF (-)-GAMBIEROL TABLE 1 The generation of C-glycosides from 7

OBn ~,,,OBn

O~ ~ _

OBn , ~ ' " OH B nO , , , ~ , , ,Nu OBn, ~-

tZ"o~OBn

;',,k.o~OBn

7

,

OBn ,

HO,,,

Nu~'O ~OBn

8

OBn,,, , OBn

Nu....t " o ~ O B n O~

I] C-

Yield

Glycoside

(%)

(~

CHeCHCH2

9

82

0

9"1

CH2CH2CH(OMe)2

10

29

-40

1.7" 1

CH2CH2CH(OMe)z

10

74

-30

6:1

CH2CH2CH2CH(OEt) 2

11

65

0

>9:1

CHCH 2

12

-

0

5" 1

CHCH 2

12

57

-40

>95:5

CH2CCH

13

78

0

>95:5

PhMgC1

Ph

14

78

-60

1"1

Ph2CuLi

Ph

14

84

0

>95:5

15

78

16

85

Nucleophile

Entry

MgBr OMe

{ M: MgBr

M~/J"'OMe

M: CuMgBr OEt

Nu

Temperature

[~:a

BrMgCu " ~ ~ O E t MgBr ~'MgBr

~

/MgC1

10

MgBr

CH2CHCH2CH2

-60 to rt >95:5

0

1"1

generation of a mixture of cyclic enol ether 19 and acyclic enol ether 18 in 80% yield. 15,~7 Although we initially separated 18 and 19, we eventually found it best to carry a mixture of the two substrates into the subsequent ring-closing reaction. Enol ether-olefin RCM employing the Schrock catalyst 2020 converted the remaining acyclic enol ether 18 into cyclic enol ether 19. Subsequently, Shawn and Jason Cox found that the second-generation Grubbs catalyst could also be used to convert 18 into 19. 2~

160

JON D. RAINIER

OBn

OBn CH2Br2, Zn TiCI4, PbCIz

BnO~ ....~O..) .... Ac20 NEt3 DMAP (95%)

OBn BnO- ~ -O ..Me

.

BnO~..~OR

_

Bn:n~..O~,~ O "

Me

BnO~,

(80%)

g:R=H

19

18:19 = 65:15

17: R=Ac

i-Pr~'~kpr

RO No_____/C(CH3)2Ph RO"

20

R = (CF3)2CH3C SCHEME 3

TABLE 2

Optimization studies on the cyclization of 17

OBn

OBn OBn BnO- -i- -O Me BnO- ~ -0. /Me conditions ~- BnO~,,,,~?8~,,, ~]~ + ~ B n O.,,~O~,,. ~, 9 , 19

Bno~BnO~'r~OAc ....<.0/J..... . ~ 17 Entry

1 2 3 4 5 6

M (equiv.)

Time (h)

TiC14

TMEDA

TttF

Zn

PbCI 2

CH2Br 2

4 16 16 16 16 6

8 32 32 32 96 48

86 550 550 96 96 48

9 36 36 36 36 13

0.045 0.18 0.18 1.9 1.9 0.72

2.2 8.8 8.8 16 16 6

3 0.5 12 0.5 0.5 0.5

Yield (%) 18

19

0 15 15 30 50 30

0 65 0~ 0a 0~ 50

aMajor by-product was oligomer.

Interestingly, when Shawn employed 4 equiv, of the Takai reagent in the conversion, ester 17 was recovered intact. Subsequently, he found that 16 equiv, of the reagent did the trick and converted 17 into the aforementioned mixture of 18 and 19. Intrigued by the prospect of using the Takai reagent to carry out olefinic-ester cyclizations, Jason Cox examined the effect of TiC14, Zn, PbC12, THF, and TMEDA on the cyclization reaction of 17 and found that each of them was necessary (Table 2). Subsequently, he also examined the effect of different concentrations of each of the

5

161

APPLICATION OF C-GLYCOSIDE IN THE SYNTHESIS OF (-)-GAMBIEROL

critical reagents and found that the amount of 19 could be increased to 50% when the amount of PbC12 and TMEDA were increased relative to the amounts of the other reagents. However, he was never able to find conditions that resulted in the complete conversion of 17 into 19 without 18 also being present. As mentioned above, Nicolaou had carried out related cyclization reactions using the Tebbe reagent. ~3~,bMechanistically, he had proposed that cyclic enol ethers came from enol ether-olefin RCM reactions based upon experiments involving subjecting acyclic olefinic-enol ethers to the Tebbe reagent. Jason carried out an identical experiment using the Takai reagent and enol ether 18 and was not able to effect the cyclization reaction. If the reaction was pushed, decomposition of the starting material occurred. Based upon this, we proposed that cyclic enol ether 19 resulted from an olefin metathesis, carbonyl olefination mechanism analogous to the mechanism proposed by Grubbs for the generation of capnellene from norbomene (Scheme 4). TM OBn

OBn OBn

o

H

o

~

H 17 H

OBn OBn

"

o

,-

0

-

Ti/~ H 21H

~'~ "Ti=O"

J

Ho H 19

SCHEME 4

It occurred to us that we could easily test the hypothesis outlined in Scheme 4 through substrate modifications. That is, if the ester and olefin compete with one another for the presumed Ti methylidene intermediate, substrates that are more sterically congested about the ester relative to the olefin should give more cyclic enol ether, while substrates that are more congested about the olefin should give more acyclic enol ether. Utpal Majumder designed several monocyclic substrates to test this notion and found that substitution on the ester could be used to effect the amount of cyclic material (Table 3). 22 For example, although isobutyl ester and TBDPS ether 22 and 23, respectively, gave about the same results as methyl ester 17, the larger glycerate acetonide ester 24 resulted in exclusive cyclization. Substitution on the alcohol side of the ester also influenced the cyclization (Table 4). When a large TBDPS ether was incorporated at C(3) rather than the previously used benzyl ether, only cyclic product was obtained regardless of the substitution on the other side of the ester.

162

JON D. RAINIER TABLE 3 The influence of ester substitution on olefinic ester RCM

OBn o.H~H.OBn

R. OBn

TiCl4, Zn,PbCl2, CH2Br2, THF, TMEDA, CH2CI2 65 ~

o

SH H

' -oL I

H

H

OBn

C

OBn

R. O. OBn OHio ~l H

I:IA

Ester

R

Enol ether(s)

Yield a (%)

C:A b

1 2 3

17 22 23

CH 3 i-Pr CH2OTBDPS

19 25 26

80 88 86

5:3 2.4:1 2.1:l

4

24

27

71

>95:5

Entry

alsolated yields after chromatography, bFrom the crude ~H NMR of the reaction mixture.

TABLE 4 The influence of substrate substitution on olefinic ester RCM

OTBDPS OTBDPS H~Ht~O..si t-Bu TiCl4'Zn,PbCl2, R O H I H O ,t-Bu R..~O "t-Be CH2Br2'THF, = ~ T ~ "Si"t-Bu

Entry 1 2

Ester

R

Enol ether

Yield a (%)

C:A b

28 29

CH 3 i-Pr

30 31

84 89

>95:5 >95:5

aIsolated yields after chromatography, bFrom the crude ~H NMR of the reaction mixture.

Also consistent with the proposed mechanism was that sterically congested olefins gave more acyclic enol ether (Scheme 5). In contrast to the results with 28 and 29, when ketoside 32 was subjected to the Takai conditions, a mixture of products was obtained.

5

APPLICATION OF C-GLYCOSIDE IN THE SYNTHESIS OF (-)-GAMBIEROL

163

OTBDPS Ry O ' ~ O " s i ( t - B u ) 2 OTBDPS

HIH

TBSOH2CvO~O--sF II

I

I

13/~cHO.~O

,~-

t Bu~

J2

TiCI4, Zn,PbCI2, CH2Br2, THF,

-

TMEDA, CH2CI2 65 ~

32

(77%, 33:34 = 1.3:1)

OH3 H 33 OTBDPS

o,HIH, o "Si(t-Bu)2 ~

R"I~" ~

_

0

SCHEME 5

While our model studies clearly indicated that the C-glycoside, metathesis strategy could be used to generate polycyclic ethers, the question of whether this approach could be used to generate natural products of the dinoflagellate toxin family remained. As mentioned above, we initially explored this question by examining the synthesis of (+)hemibrevetoxin B. 23 Positive results in this work further encouraged us to tackle the more difficult ladder toxin natural product gambierol. 24,25 The remainder of this chapter discusses our gambierol total synthesis efforts. llI. Synthetic Strategies

We were initially attracted to three features of gambierol; the first was its interesting biological profile. Gambierol was reported to be a potent neurotoxin and, as an isolate from G. toxicus, there was some question as to whether gambierol's toxicity profile made it a ciguatera poison. 26 The second attractive feature was its octacyclic ring system. In contrast to our previous work on hemibrevetoxin B, the synthesis of gambierol would need to be convergent. We initially decided to split gambierol down the middle and to focus our attention on two tricyclic precursors that would come from the iterative C-glycoside chemistry outlined in Scheme 6: one would represent the A-C ring system (i.e. 37) and the other the F-H ring system (i.e. 38). As initially envisioned, we would couple 37 and 38 to give the D-ring via metathesis. The seven-membered E-ring would come from cyclization and reduction of the resulting ketal. If successful, we felt that this novel coupling strategy would be a powerful and potentially general method of combining subunits of polycyclic ethers as the subunit coupling would involve an esterification reaction and appeared ideal for complex molecule synthesis. Tied to this was that this approach would

164

JON D. RAINIER

OH - Me Me -

.Ov

H

O

H O

,/'-~-~ H O

--~u~v~~v~-o-J gambierol R"O

n v.

~

. v n 35

v

Me

Me 0 ~ 0

"~,,-"',~/

[] v

H

Me

~O+~ou

IVle 0 ..-~ . , L H" ~ : O X Me

H

v

-

//

Me

R"O -

H

F"P--,-" .

RO

36

o

-

~, ``%

M~

. H~ ~

R"O :-"e

H O ~

H

OY O

X

Me

o4...--..

37

~

M~ "~',.- 2" / H I .... 3M e O ~ 2 X

"OY

SCHEME 6

result in intramolecular C-C bond formation. Also advantageous was that there was some flexibility in the coupling strategy; we felt that we would be able to utilize metathesis chemistry to generate either the D- or E-rings. The reason for initially pursuing the D-ring with metathesis was that we felt that the D-ring would be easier to form than the seven-membered E-ring and there was no precedent for using metathesis in this manner. We were also drawn to the enol ether-olefin metathesis approach to coupling the subunits from our work with hemibrevetoxin B where we had coupled RCM with oxidation and cyclization/elimination chemistry to build hemibrevetoxin B's B- and C-rings in only five synthetic transformations from A-ring substrate 39 (Scheme 7). 27 The third feature of gambierol that attracted us was actually a presumed problem and involved gambierol's four angular methyl groups that are in a 1,3-diaxial orientation to one another about the gambierol B- and F-rings. These are positioned in such a way to make their incorporation difficult by any means and particularly for our C-glycoside approach. The problem essentially boiled down to a need to couple a C-nucleophile to

5

APPLICATIONOF C-GLYCOSIDEIN THE SYNTHESIS OF (-)-GAMBIEROL

HO2C/'--v,,'~ ,OMe 40 OMe DCC, DMAP, CH2CI2

H H R - .O~! ....~ 'OH

BnO

39

HoH,r//" R+u"~'" O

OMe

(93%)

BnO

R= BnO~/

1. TiCI4, TMEDA, CH2Br2, Zn, PbCl2 THF, CH2CI2

165

R. ~H.O . ~H~ ,

2.20 (15 mol%)

,.~Me

~.'" O-O;

41

,.

RHo H H H ""~u"~~O~Me

T H"O" ~)3 "OMe Me3AI,hexanes; BnO 42 -65~ to rt (75%)

(78%, 2 steps)

PPTS, 60~ pyridine, 135~ 9" (66%)

H H HO R- , . , ~ O ~ u - - ~ ~ , ~ ~ i H

" ~ H"OMe'~V)3 "OMe BnO 43 R = (CH2)3OBn

44

Me

BnO

SCHEME 7

the more substituted end of a glycal anhydride. We had previously encountered similar reactions: both in our model work and in the synthesis of the B-ring of hemibrevetoxin B. In the hemibrevetoxin B work, we found that allyl magnesium chloride coupled with the anhydride from 45 and DMDO to give the undesired 3 ~ alcohol 46 exclusively (Scheme 8). %,

H H R..,~ ~ . : . oT. I ~

/O-O; \"" Me

BnO

45

~MgBr,

~ 0~

R= BnO~~

H R -

H '

H

Y H "O'~ 'Me BnO 46

SCHEME 8

Presumably, 46 came from a pinacol rearrangement of an oxocarbenium ion from anhydride 47 (i.e. 48) followed by addition of allyl magnesium chloride to the intermediate ketone as outlined in Scheme 9. We attempted to overcome the rearrangement of 47 by modifying the conditions. We had previously found a direct relationship between temperature and the conversion of glycal anhydrides into oxocarbenium ions (i.e. the lower the temperature the less oxocarbenium ion formed) and set out to determine if temperature would affect the reaction of 47.

166 H

JON D. RAINIER

H

H

H

H

H

=-R- H ~ H0

~MgBr ,- 46

o

Bn~ H - Me 47 R= B

' 'Me

48 n O ~ ~ SCHEME 9

TABLE 5 The addition of allyl nucleophiles to 47

H R ~ MoH eH

":O conditions THF

H ~

BnO 50

BnO 47

BnO ~

Nu

H

H

"(H'O'~'Me

"Me "406 "~

BnO

49

R= B n O ~ # ~ Nucleophile

Temperature (~

R

Yield (%)

50:46:49

~MgBr

0

CH2CHCH 2

72

0:1:0

~MgBr

-60

CH2CHCH 2

50

2.3:1:0

- 6 0 to rt

CH2CH2CH2CH(OEt) 2

70

0:0:1

ClMgCu~)

To summarize a substantial amount of work, Shawn found that it does, although not in the desired sense. When 45 was subjected to allyl magnesium bromide at - 6 0 ~ he was able to generate 50 having the desired C-C bond connectivity but with undesired stereochemistry at C(11). A minor component of the reaction was 3 ~ alcohol 46. Thus, it turned out that for this reaction, temperature did not effect oxocarbenium ion formation but did effect its subsequent reactivity. We were also unable to use other allyl nucleophiles in the coupling with 47; for example, the use of allyl cuprate gave ketone 49 in 70% yield (Table 5). Presumably, the stereochemical outcome of the C-C bond forming reaction to 50 was a result of oxocarbenium ion formation and the addition of the nucleophile proceeding through chair transition state 51 (Figure 3). We decided to take advantage of the facility with which oxocarbenium ion 48 was formed by switching the order of C-C bond formation and

5

167

APPLICATION OF C-GLYCOSIDE IN THE SYNTHESIS OF (-)-GAMBIEROL

~O-~

H

I-~H ~ ~ e O H chair T.SNu 51 FIGURE 3. Rationale for the formation of 50 from 47.

TABLE 6 The use of Me3A1 to generate the hemibrevetoxin B C(11) stereocenter

RHo H o-o .,~u~',,,~, v~Me ~'~,'"' m H"O iv')3"OMe BnO 52 R=B n O ~ i H

H

BnO 54 Me

H

H

BnO 55

-1

R__'O H

~-

,.

H

';:o

H

1

Me

nucleophile

MeO

BnO 56

Entry

Nucleophile (equiv.)

Conditions

Yield (%)

54:55:56

1 2 3 4 5 6

MeMgBr (5) Me3AI (3) Me3AI (3) M%A1 (1) Me3A1 (20) MezA1C1(3)

THE - 6 5 ~ to rt "THE - 6 5 ~ to rt Hexanes, - 6 5 ~ to rt Hexanes, - 10 ~ to rt Hexanes, - 6 5 ~ to rt Hexanes, - 6 5 ~ to rt

65 70 50 20 75 78

1.6:0:1 5:2:0 1:0:0" 1:0:0 1:0:0 0:0:1

57

"The remainder of the material consisted of uncharacterizable by-products.

adding a methyl nucleophile to an anhydride that contained the atoms necessary for the generation of the C-ring. To this end, Shawn generated 52 and examined the addition of methyl nucleophiles to the corresponding anhydride, 53 (Table 6). The addition of methyl magnesium bromide to 53 resulted in the generation of a 1.6:1 mixture of the desired addition product 54 and the hydride migration product 56 in 65% yield. Encouraged by this, Shawn felt that he could enhance the yield of 56 and suppress products from hydride migration if the addition of the methyl nucleophile were somehow forced to occur in an intramolecular fashion. That is, a methyl nucleophile

168

JOND. RAINIER

capable of coordinating to the anhydride was needed and he was immediately drawn to the use of Me3A1. When 53 was subjected to 3 equiv, of Me3A1 in THF at - 6 5 ~ Shawn isolated a 5:2 mixture of the desired 3 ~ alcohol 54 and the undesired alcohol 55. The key to success in the reaction turned out to be the use of a large excess of Me3A1 and a non-polar solvent (entry 5). Presumably, the Me3A1 addition reaction proceeded through charged intermediate 57. Subsequent to these studies, Jason Cox examined the ability of aluminum reagents to deliver nucleophiles to glycosyl anhydrides and found it to be a general method of generating a-C-glycosides} 8 IV. Total Synthesis of ( - ) - G a m b i e r o l

Although we had not been able to add Grignard reagents to ~-substituted anhydrides, our success with A1Me3 emboldened us to approach the challenges associated with the synthesis of gambierol optimistically. Our initial analysis of the gambierol A-C and F-H rings is outlined in Schemes 10 and 11, respectively. As envisioned, central to the generation of 37 would be our ability to generate the B-ring from 59 and the requisite angular methyl groups that were oriented in a 1,3-disposition to one another. Although precedent suggested that the A-ring precursor would come from a hetero-Diels-Alder chemistry and 61, this compound had not been synthesized previously. R"O

R'O~

R"O

- Me

Me

H,-.,H

R'O~]~

~ OMe~ Me -

R'O..

OR

M ~ O . ~ Me ~-_-.o.-- v

37

~

58

59

OR

>

O

~..

> 1H

H H 60

SCHEME10 H O ~ O

H O H

R'OxH H

Me ~ Z / , '& Me- 0 ~ 0

H

H

RO H OR'

H I

ou x

Me'OR" H 62

SCHEME11

63

5

APPLICATION OF C-GLYCOSIDE IN THE SYNTHESIS OF (-)-GAMBIEROL

169

Our planned synthesis of the F-H subunit took a slightly different tact. While still employing C-glycoside chemistry and RCM, we would begin with the G-ring intact, build in the F-ring, and then turn back to the H-ring. What made these efforts most attractive was that the starting point for these studies would be I>glucal. As with our plans to 58, we were most concerned with the generation of the 1,3-angular methyl groups and the synthesis of the F-ring where an unprecedented metathesis reaction to a tetrasubstituted enol ether was required. One advantage to the approach outlined in Scheme 11 was that we would address the problematic issues early in the synthesis and therefore would not invest an undue amount of time following a terminally flawed plan. A. A - C T R I C Y C L E

Fresh from his successes with hemibrevetoxin B, Jason Cox decided to initiate our studies to gambierol by examining the generation of the A-ring. As alluded to above, our approach to the gambierol A-ring required that we be able to carry out a hetero-Diels-Alder cycloaddition reaction between the methylated version of the Danishefsky diene 67 and 7-alkoxy aldehyde 66 (Table 7). Although 67 had been employed in related reactions, these had not been enantioselective. 29 Thus, Jason explored the use of asymmetric catalysts for the generation of 71 and initially examined Keck's BINOL system due to its demonstrated ability to carry out the cycloaddition between aldehyde 66 and Danishefsky's diene in 65% yield and 95% ee. 3~ In spite of numerous attempts, when 66 was exposed to 77 no reaction occurred using the standard Keck conditions [Ti(Oi-Pr) 4, (S)-BINOL (20 mol%), 4 ,~ MS]. Jason also examined, albeit to a lesser extent, other Lewis acids. The use of a mixture of Me3A1 and (S)-BINOL resulted in some product being formed, but in a disappointing 10% yield and with only 8% enantioselectivity. Although the yield for the reaction increased considerably when the catalyst was B(OPh) 3 and (S)-BINOL, the enantioselectivity was still negligible. After these disappointing results, Jason turned to other catalysts capable of carrying out the desired transformations and settled upon Jacobsen's chiral Cr(III) adamantane catalyst 73. 3~ This catalyst had previously been shown to be capable of catalyzing enantioselective hetero-Diels-Alder reactions and seemed to have a large substrate scope. In contrast to his work with B INOL, Jason found that only 3 mol% of 73 was required to catalyze the reaction between 66 and 67. In addition, pyranone 71 was isolated in 90% yield and with 94% ee. Henry Johnson also examined the reaction between 67

170

JON D. RAINIER

TABLE 7 The influence of adamantane catalyst 73 and binol (74)-derived catalysts on the hetero-Diels-Alder cycloaddition of 67 and 68 with aldehydes 64--66

O

OTMS BnO.(..~CHO + 64: n = 1

M e O @ R 67: R = Me

65:n=2

catalyst conditions ,. B n O ~ ~ ~ . 1 R=Me

68: R = H

66:n=3

69: n = 1

Me

C•OH

~

H Entry

73 (5 mol%)

R R=H

72: n = 3

70: n = 2 71:n=3

...-OH

74 (20 mol%)

RCHO

Diene

Catalyst

Conditions

Dihydropyran

Yield (%)

% ee

1

66

67

73

71

90

94

2 3 4 5

66 66 66 66

67 67 67 68

74 74 74 73

71 71 71 72

0 69 20 74

0 8 93

6

65

67

73

70

78

93

7

64

67

73

4 ,~ MS, rt, 70 h; TFA, 0 ~ CH2C12, 1 h Ti(Oi-Pr)4; TFA B(OMe) 3 A1Me3 4 ,~ MS, rt, 70 h; TFA, 0 ~ CH2CI 2, 1 h 4 A MS, rt, 70 h; TFA, 0 ~ CH2C12, 1 h 4 ,~ MS, rt, 70 h; TFA, 0 ~ CH2C12, 1 h

69

68

69

and aldehydes 64 and 65 (entries 6 and 7) to give 69 and 70, respectively, and Danishefsky's diene 68 with aldehyde 66 to give 72 (entry 5). Interestingly, enantioselectivity was dependent upon the length of the tether between the aldehyde and the benzyl ether (compare entries 1, 6, and 7). The enantioselectivity of these reactions was established using HPLC; the absolute stereochemistry of 71 was established using Kitagura's modification of the Mosher analysis. 32

5

171

APPLICATION OF C-GLYCOSIDE IN THE SYNTHESIS OF (-)-GAMBIEROL

With a method of generating 71 in hand, we next turned to its conversion into the corresponding C-glycosides 81 and 82, but first needed to establish the C(6) stereocenter. To this end, we applied the Luche reduction, which afforded 73 (Scheme 12). Although 73 is epimeric to the desired C(6) stereocenter, we both hoped to be able to epimerize it in a subsequent step but, more importantly, we were hopeful that this stereocenter would help direct the formation of the C(7) and C(8) stereocenters in the C-glycoside forming reaction. Thus, conversion of the C(6) alcohol to benzyl and PMB ethers 75 and 76, respectively, using standard conditions followed by their exposure to DMDO and propenyl magnesium chloride, resulted in the generation of a mixture of C-glycoside diastereomers favoring the desired isomers 81A and 82A, respectively (Table 8).

O

OP R

B

n

O

~ 4 -~

H 71 (R = CH3) 72 (R = H)

NaBH4, CeCI3.7H20 MeOH, 0

--~

i~.~

~ -

NaH, PMBCI ( 7 4 : P = H , R = H

BnO...~---.....~._ O..~8 (95%, 2 steps)~'77: p H 73: P = H R = CH 3 Nail, BnBr, Bu4NI, THF (92%, 2 steps) - 75: P = Bn, R = CH 3 Nail, PMBCI, DMF (95%, 2 steps) ~ 76: P = PMB, R = CH 3 TMSCI, imidazole, DMF

78: P = TMS, R = CH 3

TBDMSCI, imidazole, DMAP, DMF

,. 79: P = TBDMS, R = CH 3

TBDPSCI, NEt 3, DMAP, DMF

o - 0 , -65 ~ to rt; ~MgCI,-65

R

~C

~ to rt

~

,- 80: P = TBDPS, R = OH3

,,,OH B n O ~ o

_.....~,~..~ H

OH gnO..v..........~_. O...~....

H

H

8 1 A : P = Bn, R = CH 3 8 2 A : P = PMB, R = OH 3 8 3 A : P = PMB, R = H 8 4 A : P = TMS, R = O H 3 85A: P = TBDMS, R = CH 3

81B: 82B: 83B: 84B: 85B:

OPMe

BnO

.... ~ H

H

P = Bn, R = CH 3

P = PMB, R = CH 3 P = PMB, R = H P = TMS, R = C H 3 P = TBDMS, R = CH 3

OPMe

BnO

,,

H 86

SCHEME 12

87

PMB, R : H

172

joN D. RAINIER TABLE 8 Conversion of enol ethers 75-80 into C-glycosides 81-85

Entry

Enol ether

C-Glycoside

A:B a

Yield (%)

1 2 3 4 5

75 76 77 78 79

81 82 83 84 85

7.5:1 b 7.5:1 b >95:5 2:1 c 2:1 c

90 78 95 20 46

6

80

-

-

<5

aRatio was determined from ~H NMR of crude reaction mixture, bAlso isolated 5% of diastereomeric glycoside 86. CMajor by-product was acetone adduct 87.

Although 81A and 82A were isolated in useful quantities, it was surprising that their reactions were not more selective; PMB ether 77 lacking a C(7) alkyl substituent gave 13-C-glycoside 83A in 95% yield and with >95:5 diastereoselectivity. Interestingly, the choice of C(6) protecting group influenced the effectiveness of the C-glycoside forming chemistry. Glycals containing C(6) silyl ethers reacted much more sluggishly and with even lower diastereoselectivity than the corresponding C(6) benzyl ethers (entries 4-6, Table 8). In an effort to better understand the results with 75 and 77, Scott Roberts, in collaboration with Anita Orendt from our computational facility, calculated transition states for the oxidation of models of 75 and 77 and found similarities and notable differences between the transition structures for the respective oxidations. 33 Consistent with our experimental results was that the difference in energy between attack from the side of the benzyl ether and from the side opposite the benzyl ether was smaller for [3-methyl substrate 75 than it had been for the corresponding [3-unsubstituted substrate 77. Similarities included a twist boat transition state conformation for both substrates. The most significant differences centered on the bond lengths of the C-O bonds in the transition states. Due to the opposing charge stabilization of the C(2) methyl group and the pyranyl oxygen, the C(1) and C(2) C-O bond lengths were more synchronous by 0.26 * in the transition state for the oxidation of the 13-methyl substrate than they were for the substrate lacking substitution at the 13-position. Based upon our analysis of the transition structures, 75 shows lower selectivity than 77 because the interactions between DMDO and the C(3) benzyl ether are diminished in 75, largely

5

173

APPLICATION OF C-GLYCOSIDE IN THE SYNTHESIS OF (-)-GAMBIEROL TABLE 9

Conversion of A-ring substrates 81 and 82 into cyclic enol ethers 88 and 89, respectively RO

RO ,,,OH

Me

1. Ac20, i-Pr2NEt, CH2CI2 2. TiCI4, CH2Br2, PbCI2, TMEDA, Zn, THF, CH2CI2 R'

R'

82: R = PMB 3. RCM (See Table)

88: R = PMB

81" R = Bn

89: R = Bn

R'= B n O ~ % L

MsN. NMs Ci "]/ F3C, .Ru_-~ M /~" ] CI/I Ph e O ...."Mo~C(Me)2Ph PCy3 Od 90

Me+CF3

20

CF3

Entry 1 2 3

ROH

R

Yield (acetate) (%)

RCM catalyst, conditions

81 81 82

Bn Bn PMB

77 77 85

90 (20%), Phil, rt 20 (20%), hexanes, 65~ 90 (10%), Phil, rt

Yield (RCM)

(O~)a,b

74 70 80

aTakai protocol gave a l'l mixture of cyclic and acyclic enol ethers, bTwo steps.

due to DMDO being shifted away from the benzyl ether in the transition structure from 75. Having found a reasonable route to the A-ring, Jason turned his attention to the B-ring. After acylation of 81 and 82, he subjected the resulting esters to the Takai-Utimoto reagent to give a 1:1 mixture of cyclic and acyclic enol ethers (Table 9). As with our earlier results, the formation of mixtures from these reactions was not a problem, the mixtures were simply subjected to the Schrock Mo catalyst 20 or the second-generation Grubbs catalyst 90 to cyclize the remaining acyclic material. Having generated the requisite B-ring enol ether, we turned our attention to the formation of the B-ring C-ketoside. As mentioned above, we had viewed the generation of this ketoside to be a major challenge as its synthesis required the stereoselective addition of a carbon nucleophile to the more substituted end of the anhydride. When Jason subjected 89 to DMDO and propenyl magnesium chloride, he isolated a modest yield of C-ketoside 92 (entry 1, Table 10). Surprising

174

JON D. RAINIER TABLE 10 The coupling of propenyl nucleophiles with the anhydride from 89

RO Me ~eO.~Me

R'-'~o~v 89

~,"" O-O; ,,.

~M

Bn?Me Me ~ I ~

R'"~o~v ""OH

1 2 3

R'

91

R'= B n O ~ " h ~

Entry

BnlOMe Me + ~ O: ..~ 92

M

91:92

Yield (%)

MgC1 B(allyl) 2 Al(allyl) 2

<5:95 1"1.5 1.5"1

50 45 48

BnO ~Me

Me tt

BnO H

H

,',O H

93 FIGURE 4. The addition of nucleophiles to 93 proceeds via a trans-diaxial attack.

to us was that the coupling reaction had occurred in the undesired sense (i.e. from the same face as the angular methyl group). This implied a direct addition of the nucleophile to the anhydride rather than via the intermediacy of an oxocarbenium ion as had been anticipated. In an attempt to force the reaction to proceed through the desired oxocarbenium intermediate, we examined the coupling of 93 with triallylaluminum and triallylborane. Unfortunately, these reactions resulted in the generation of gross mixtures of stereoisomeric C-ketosides. Clearly, our concerns about the use of anhydrides to generate the gambierol B-ring had proven themselves to be well founded. The relatively low yields observed in the reactions of 89 were probably a consequence of the instability of the intermediate anhydride (i.e. 93). Ring-opening of the presumed ground-state conformer proceeds through a chair transition state giving trans-diaxial addition products (Figure 4). Consequently, relatively weak nucleophiles (i.e. acetone) would be capable of decomposing this substrate. 34

5

APPLICATIONOF C-GLYCOSIDEIN THE SYNTHESIS OF (-)-GAMBIEROL

175

In light of the results of the direct addition of propenyl magnesium chloride to 93, Jason felt that the simple reversal of the order of the C(11) C-C bond formation might solve the problem of establishing the C(1 1) center (Scheme 13). That is, in a similar fashion to our hemibrevetoxin B work, the incorporation of a C-ring precursor into the B-ring anhydride and the coupling of this species with methyl magnesium chloride would, in theory, result in the desired adduct. Unfortunately, this strategy was also unsuccessful and instead gave ketone 97 from a stereoselective hydride migration in 75% yield. The enhanced yield in this reaction is probably the result of our not concentrating the intermediate anhydride as a result of using Messeguer's "acetone-free" DMDO that can be generated as an c a . 0.2 M solution in CH2C12.35

BnOMe ,,,OH

DCC, DMAP, CH2CI2 _~ Ho~.OMe 94 OMe

81

MeO BnO i ~ Me ... /----._/ OMe ,,,O H

(64%)

R'= B n O ~ ~ h . 1. TiCI4, PbCI2, CH2Br2, TMEDA, Zn, THF, CH2CI2 2.90 (20 mol%), Phil, rt (62%, 2 steps)

~,,, , O-O; . MeMgCI, THF,-60~ (75%)

o

95

BnO ...[. M~eo, 1/~

BnO~~_.

H

0 __~~ H

H

OMe /[..

v

"OMe

96

BnO OMe wMe H / ~ ' - ~ O ~ O M e BnO.~ / A " 4 ..4...O..4......L...O H " "H ~ / "97

SCHEME 13

Following these disappointing results, it was apparent that a reassessment of our synthetic plans to the C(11) ketoside was needed. Among the various possibilities, Jason became intrigued with the possibility of exchanging an intramolecular C-C bond forming reaction for the intermolecular variant that he had been attempting. More specifically, he became interested in employing a C(10) allyl vinyl ether in a Claisen rearrangement to generate the C(11) ketoside (Scheme 14). Although related rearrangements had been utilized to generate C-glycosides, all previous examples that we were aware of had come from precursors having

176

JOND. RAINIER

R'O

R'O Me

Me

__base..

~ R

._

p_r

R'~Me O Me R'O O 0 "< ~ __[3,3]__., or.,~o --Me ~Me R SCHEME 14

the allylic component as part of the pyranyl ring system. In these cases, the control of the C-glycoside center was predetermined by the stereochemistry of the allylic center. 36 In our substrate, we hoped that subtler influences would control the outcome of the reaction. Namely, we envisioned that the C(7) angular methyl and/or the trans-pyranyl ring system would direct the generation of the new C(11) stereocenter. That the proposed reaction would lead to a C(10) ketone was an added benefit as it would enable us to avoid a subsequent epimerization reaction; reduction of the ketone from the axial face would result in the desired C(10) alcohol. Jason initially examined the conversion of ketone 98 into the corresponding allyl vinyl ether and envisioned the trapping of the more stable enolate from 99 with allyl bromide. Although his initial attempts to employ this sequence were unsuccessful, this approach was not studied in sufficient detail to determine whether it could be successful. Ultimately, a more interesting and productive approach presented itself to him. This alternate approach avoided the generation of the ketone and involved the in situ generation of the enol ether instead of the allyl ether and is illustrated in Scheme 15. The execution of the strategy began with the epoxidation of 88 and 89 using m-CPBA in methanol to give ketals 102 and 103, respectively, as a 2:1 mixture of anomers in high yield. Jason then converted the ketals into the corresponding allyl ethers 104 and 105 using standard conditions. Jason and I were extremely pleased to isolate C-ketosides 108 and 109, each as an 8:1 mixture of diastereomers and in 97% yield, when 104 and 105 were subjected to PPTS and pyridine at 100 ~ Not only had the PPTS conditions generated the enol ether but they had also induced the desired rearrangement. The relative stereochemistry at the newly formed C ( l l ) center in 108A and 109A was established through the observance of the expected NOE enhancements between the angular methyl groups.

5

APPLICATION OF C-GLYCOSIDE IN THE SYNTHESIS OF (-)-GAMBIEROL

177

R'O Me R'O M" OMe ~eo i~Me m_CPBA,MeOH ~ e o ~ , ' Me ~ B r RvH'O'~ v

R- BnO~~"~

88: R'= Bn 89: R'= PMB

Rv~'O'~ v ~"OH Nail, THF 102: R ' = Bn (90%)

103: R'= PMB (92%)

R'O Me OMe ,,[~O....~Me .. ,.

PPTS, pyridine [

R'wOMe Me ~ ..~]

oc

104: R'= Bn (75%) 105: R'= PMB (78%) R'O Me Me " Rv,--,"O',--,v l o ~ O H

H A

o

L

106: R'= Bn -J 107: R'= PMB

R'O wMe Me , , , 1 0 8 : R' = Bn (97%yield) R~"~O'~"~-"B O 109: R'= PMB (97%yield) A:B = 8:1 SCHEME 15

i

Me Me 101

[ 111

FIGURE 5. Proposed transition state for the generation of 108.

Presumably, the major product resulted from a chair-like transition state to give the trans-pyranyl system as indicated by 110 (Figure 5). Rearrangement to the opposite face would suffer from steric interactions between the angular methyl group and the side chain as indicated by 111. Having finally solved the C(ll) problem, Jason next turned to the inversion of the C(6) stereocenter. This was accomplished on 109 using standard conditions; namely, oxidative removal of the PMB ether, Mitsunobu inversion, and silyl ether formation (Scheme 16). We examined [3-PMB ether 109, ~-TMS ether 112, and ~-TIPS ether 113 in the subsequent chemistry. We decided to initially model our chemistry to the C-ring to determine the feasibility of the RCM and oxidation chemistry. To this end, Jason generated model substrate 116 from the NaBH 4 reduction of 113 and the coupling of the resulting alcohol 114 with acid 115 (Scheme 17).

178

JON D. RAINIER

PMBO wMe

R = TIPS 1. DDQ, CH2CI2, H20 (97%) 2. DEAD, PPh3, p-NO2C6H4CO2H; NaOH, THF (80%) 3. TIPSOTf, 2,6-1utidine, DMAP, DMF (100%) R = TMS

Me

R'H'O'H v

1. DDQ, CH2CI2, H20 (97%) 2. DEAD, PPh3, p-NO2C6H4CO2H; NaOH, THF (80%) 3. TMSOTf,NEt3, DMF (97%)

109

R = BnO~-"$_

RO : Me Me

"O

112: R = TIPS 113: R = TMS

SCHEME 16

O TMS__OMe Me

NaBH4 MeOH

R",--,"O" ,-, v \'O 113 ii

ii

R = BnO~'~.

TMSO r - Me M e | ~..~U~A , . ~ j.L 0 R~,-,-~O---,-,v,-,~O-- v

O-O; = DIBAL-H (69%)

HO~/'~'f~OMe 115 OMe

~ ~t~ ~ R " H ' O " H V H"OH 114

DCC, DMAP (91%, 2 Steps)

TMSO 1.TiCI4, Zn, - Me Me PbCi2, CH2Br2 ,,, ~ . ~ O ~ ~ . . . --...~OMe 2.20, hexanes 60 ~ R

-~~v~

116

•/•"

TMSO r - Me M e |

OMe

(77%, 2 Steps)

,5 Me H " ~ ~-.,~,~O. , ..T. /~[.,_...~l,

.

' ~ / H ' ~ / ~IRH

117

R"H'O'H v H-O- H ~L 119

TMSO " Me Me

. y-

<"CH(OMe)2

TMSO Me Me .~~ ~-~,OH 0

118

PPTs, PhCI; pyridine, 135 ~ (95%)

l

o1

o

R'H'O'H v H'O" H v 121)

~CH(OMe)2

F3C.

Me,~

II

O...."Mo~C(Me)2Ph ol

Me: ~ C F 3 CF3 SCHEME 17

20

5

APPLICATION OF C-GLYCOSIDE IN THE SYNTHESIS OF (-)-GAMBIEROL

179

The two-step RCM chemistry sequentially employing the Takai protocol and either the Schrock or the Grubbs catalysts gave 117; the reagent of choice turned out to be the Schrock catalyst 20 as it gave 117 in a significantly higher yield. From 117, the stereocenters on the C-ring were established using an oxidation/reduction protocol that presumably proceeded through transition state 118 and resulted in the same product that one would expect to see from a hydroboration, oxidation sequence. 37 The chemistry to model substrate 120 was completed by subjecting 119 to PPTS, PhC1, and pyridine. 3s The gambierol coupling strategy that is outlined in Scheme 6 required a slightly modified protocol to the requisite C-ring coupling precursor. In contrast to our work with the model, we believed that the most direct route to the C-ring and the A-C coupling precursor from 109, 112, or 113 would involve the generation of the corresponding unsubstituted enol ether (e.g. 122) and its subsequent conversion to an olefinic alcohol. Henry Johnson followed up Jason Cox' work to carry out these studies. From alcohol 114, vinyl ether formation gave metathesis precursor 121 (Scheme 18). 39 Enol ether-olefin RCM using the secondgeneration Grubbs catalyst 90 provided 122 as the precursor to the gambierol C-ring. TMSO -- Me

~ Me|

-

,,,

R

0

1. TsOH, ethylvinylether -60~176

(95%)

OH 114 R= B n O ~ %

TMSO__Me

-_

2. TMSOTf, NEt3, 0 ~ (98%)

90 (20 mol%) Phil, rt (95%)

TMSO - Me

i/~O~,,,,

Me|r~

ii

..,2....A.,..A,~ .p R"H'O'H v H- O 121

Me

R 122

SCHEME 18

In contrast to our ability to reduce the anhydride from model substrate 117, all of Henry's attempts to couple the anhydride from 122 (i.e. 123) with allyl nucleophiles failed to deliver the desired allyl C-glycoside (Scheme 19). Instead he isolated a considerable amount of ketone 126 resulting from a 1,2-hydride migration or 3 ~ alcohol 125 from allyl addition to 126.

180

JON D. RAINIER

TMSO " Me Me

TMSO Me Me

1

M

0-0

122

TMSyn Me Me

H

IIM = MgCi, MgBr, MgCI/Bu3SnOTf,

~O.~~,OH~

I

R,-H-O- H v ~-O- H v

-.~

Cu(allyl)l, BusSn/BusSnOZf, Al(allyl) 2, a(allyl)2

124 TMSO : Me Me

TMSO - Me Me

OH ....

125

126 SCHEME 19

It is not clear why this reaction failed to give C-glycoside products, as we have successfully coupled other tri- and tetracyclic anhydrides with allyl nucleophiles. For example, in his G-ring of gambieric acid efforts, Utpal Majumder was able to generate 128 in 73% yield from tetracyclic dihydropyran 127 (Scheme 20).

M

OBn |H

H

127

H

~, ~'~"

(73%)

OBn H Me| H

H

H

128

SCHEME 20

Our failure to couple allyl nucleophiles with 123 forced us to examine a somewhat circuitous route to the C-ring and to reconsider the successful DIBAL-H reduction sequence. In contrast to our results to model 120, the A-C coupling precursor required the use of the propanoic acid homologue of 115. To this end, esterification of the alcohol from 109 with 129 provided ester 130 in 90% yield (Scheme 21). In light of our success with 116, Henry found it surprising that the subsequent Takai reaction of 130 only delivered a 13% yield of acyclic enol ether 131. In addition to recovering starting material, the remainder of the material from this reaction consisted of diene 132 (50%) and unsaturated enol ether 133 (7%); both

5

181

APPLICATIONOF C-GLYCOSIDEIN THE SYNTHESISOF (-)-GAMBIEROL 1. NaBH4,EtOH (100%) 2. CH2cI2DCC' DMAP

PMBO ~,~Me Me O ~ BnO /-v v~'o~v

\'0

I.L,,..~

MeO

TiCI4, PbCI2,CH2Br2,Zn~ "

TMEDA, THF, CH2CI2

(90%)

.,~

-"

OMe 0

109

OPMB /Me Me O

BnO,,L 129

OH

130

Me MeO

PMBO /Me Me ,.,.-,._/.O~.~,j~...~ BnO~ 0 " ~ ' ~ ' " O. ~

+22% recovered 130

131 <"/-- OMe (13%) MeO PMBO Me Me O ~

PMBO /Ueo

B

BnO

n

O

(50'/0) Me(~

Me

~

0

(7%) Me(~

SCHEME 21

R I Me Me ~ O ~ ' ~ R,-H-O- ~ v q

"0

R /Me 1. NaBH4, MeOH 2: DOC, DMAP, 138

R'

~

//" MeS O .... O .~ 0

H

112: R = ~-OTIPS 109: R = I~-OPMB R' = B n O ~ ' ~ 5 _ R = o~-OTIPS: 1. CH2Br2, TiCI4, Zn, PbCI2 2.90 (20 mol%), Phil, rt R = 13-OPMB 1. CH2Br2, TiCI4, Zn, PbCt2 2.20 (20 mol%), hexanes, 60 ~

~OTIPS

134: R = o~-OTIPS(86%, 2 steps)

135: R = 13-OPMB(88%, 2 steps)

O /R Me

HO.JJ'..v-'~OTIPS

Me

R''H'O'H V H ' o "

v

"OTIPS

136: R = oc-OTIPS(70%, 2 steps)

137: R = 13-OPMB(80%, 2 steps)

SCHEME 22

of these resulted from a competitive [3-elimination reaction of the acetal under the reaction conditions. The C-ring problem was ultimately solved using a synthon of acetal 129, namely silyl ether 138 (Scheme 22). Beginning with ketones 112

182

JON D. RAINIER

and 189, reduction and esterification with 138 provided 134 and 135, respectively. Two-step enol ether-olefin RCM gave the cyclic enol ethers 136 and 137. Note that the majority of the material isolated from the Takai reaction of 134 and 135 was acyclic enol ether (we observed c a . 8% cyclic enol ether). Presumably this was the result of the increased steric environment about the olefin relative to the ester as was discussed above. With []-PMB substrate 137, DMDO oxidation and DIBAL-H reduction gave alcohol 148 as a single diastereomer in 80% yield (Scheme 23). With 0~-TIPS substrate 136, Henry found it more efficient to carry out the conversion using hydroboration and oxidation to give alcohol 139 in 91% yield. 4~ Presumably, the C(11) angular methyl group directed the oxidation reaction. As was discussed previously for the generation of model substrate 128, we believe that the anhydride reduction sequence involves aluminum ate complex 141.

R ~Me

Me

R'

/ oy-i,o R /Me

~'"' O-O; OTIPS

DIBAL-H

R'"H'O'HVH'O"

H

Hv

"OTIPS

139: R = e-OTIPS (75%) 140: R = ~-OPMB (80%)

R'= B n O ~ k 136: R = (z-OTIPS (70%, 2 steps) 137: R = ~}-OPMB (80%, 2 steps)

Me

BH3*THF; H20 2, NaHCO3 (91%) rrrr Me H Lr, ~,"O'A~__ L

L_&r-"A 141 "

"OTIPS

SCHEME 23

Our initial subunit coupling reactions were carried out on 13-OPMB derivative 142; its synthesis from 140 is illustrated in Scheme 24. Generation of the C(13) pivaloyl ester was followed by TIPS ether hydrolysis using TBAE Oxidation of the resulting 1o alcohol and hydrolysis of the C(13) ester gave the corresponding hydroxy aldehyde. Wittig olefination completed our synthesis of A-C coupling precursor 142.

5

APPLICATION OF C-GLYCOSIDE IN THE SYNTHESIS OF (-)-GAMBIEROL

BO PMiMe~.. u M _ H. e r""~O~ '''O' ' J&..A....A....2.~ /~ R'" -_-H-O-H v : - . - O -H v "OTIPS 140

183

PMBO IMe Me H ~..~O~,,,OH,,,

1. PivCI, DMAP, pyridine(90%) 2. TBAF ( 9 3 % ) 3. (COCI)2, DMSO, NEt3 (98%) 4. LiOH, MeOa; SiO2 (90%) 5. Ph3P=CH2 (92%)

R,,r O ~ O _-"'~"~../"~ H H H H 142

R'= g n O / ~ / SCHEME 24

B. F - H S U B U N I T

We next set our sights on the generation of the gambierol F-H subunit. At the outset of this work, we anticipated that the biggest challenge would be the F- and G-tings where we were again faced with the generation of a C-ketoside. In spite of the fact that we had been largely unsuccessful in our previous attempts to couple Grignard reagents with ~-substituted anhydrides as direct precursors to ketosides, the potential efficiency of the anhydride to ketoside approach convinced us that it deserved further examination. We selected bis-silyl D-glucal derivative 144 as a precursor to the G-ring (Scheme 25). Not only would the C(25) stereocenter serve as a handle for the introduction of the C(23) and C(24) centers, but also the C(26) and C(27) centers would come directly from D-glucal. Jason Cox initiated this work and chose to utilize a TBDPS and a cyclic silylene as the alcohol protecting groups. He made these choices to insure orthogonality and because these groups had been reported to be robust under the conditions required to incorporate the ~-methyl group. As shown in Scheme 25, this OH

24

~OH 143

',,,,~ O-O; ClMg~ (93%)

OTBDPS

~

1. t-BuSi(OTf)2, 2,6-1utidine (76%) O"si(t-Bu)2 2. TBDPSCI, imidazole (100%) Me./K--O__~-../O 3. t-BuLi; Mel (95%) H 144 OTBDPS OTBDPS H o

0'" ~ " ~

'7-o

Me

"Si(t-Bu)2 I

~

H145

|

]

SCHEME 25

_~

HOH..~O.si(t_gu) 2 Me

H

146

184

JOND. RAINIER

choice turned out to be fortunate and much more important than was initially envisioned. The synthesis of 144 from D-glucal began with the sequential generation of the cyclic silylene and the TBDPS ethers followed by the incorporation of the C(23) methyl group. 4~ Following its synthesis, Jason subjected 144 to DMDO and propenyl magnesium chloride and, to both his and my delight, isolated [3-ketoside 146 in 93% yield. As far as we are aware, this was the first time that a Grignard reagent had been coupled in a stereoselective fashion with an ~-substituted anhydride to give the corresponding ketoside, where the newly formed C-O and C-C bonds were trans to one another. As our gambierol F-ring efforts involved an RCM reaction to a tetrasubstituted enol ether (vide infra), we required a 2-methylallyl nucleophile. Surprising to us was that the use of the same conditions for the generation of 146 (i.e. formation of the anhydride in CH2C12,concentration of the reaction mixture, solvation of the resulting residue, and addition of the nucleophile) but using 2-methylpropenyl magnesium chloride as the nucleophile failed to give the corresponding C-ketoside and instead resulted in the generation of ketone 148 in 78% yield (entries 1-3, Table 5). This was disappointing and forced us to consider alternative methods of generating the requisite disubstituted alkene. However, this was about the time that Utpal Majumder joined the gambierol team (Jason Cox went off to a postdoctoral position in Amos B. Smith III's group). Utpal decided to examine this reaction more closely and found that the conditions used to generate the intermediate anhydride were important. Namely, Utpal found that the use of Messeguer's "acetone-free" conditions of generating DMDO were critical. 35 Messeguer's conditions involve using CHzC12to extract DMDO from the water and acetone used in its generation. By using these conditions Utpal found that he could directly couple the intermediate anhydride 145 with 2-methylpropenyl magnesium chloride without concentrating 145 to give 147 in 65% yield (entry 4, Table 11). Originally, I was very skeptical of this reaction because the CHzC12 solution of DMDO from the Messeguer's conditions contains a considerable amount of acetone. Fortunately, Utpal did not share my skepticism and ran the experiment in spite of my concerns. Further optimization showed that the Grignard counterion also played a role in the reaction; when a bromide instead of a chloride was used, the yield of C-ketoside 147 improved to 92%. Significant to our gambierol efforts was that Utpal found the reaction to be scalable (ca. 8 g) and highly diastereoselective (>95:5). In addition to the preparation of the anhydride and the choice of Grignard counterion, Utpal found that the choice of the C(25) substituent

5

APPLICATION OF C-GLYCOSIDE IN THE SYNTHESIS OF (-)-GAMBIEROL

185

TABLE 11 The addition of 2-methylpropenyl nucleophiles to the anhydride from 144

OTBDPS OTBDPS OTBDPS ~O"si(t-Bu)20~O_ cond,tlons H'H O ~ -~O'si( H .. 9 | H o"~~~O'si(t-Bu)2 t-Bu)2 Me/~"O-~14 O ~MgX "~~2"~ O H- ~./O additive 147 148 Entry

X

Conditions a

Additive

147:148

Yield (%)

1 2 3 4 5

C1 C1 C1 C1 Br

A A A B B

None ZnC12 CuI None None

<5:95 <5:95 <5:95 >95:5 >95:5

78 78 81 65 92

aA: concentration of intermediate anhydride; residue was dissolved in THF; nucleophile added. B" Messeguer's conditions (DMDO added to the enol ether as a c a . 0.2 M solution in CH2C12); nucleophile added directly to the anhydride without concentration.

TABLE 12 The generation of C-ketosides 152-154 from cyclic enol ethers 149-151

R ~

R H o "Si(t-Bu)2

~O-O; ""

R

H| H R'H-O~O"si(t-Bu)2,

= Me~'~O'~ O "~'-- MgBE-~'~eO ~'~O R' 149: R = H 6 150:R= OTBDMS 151"R= OTIPS R H OTBDMS OTIPS

R' CH 3 CH 3 H

13:~ 2:1 4:1 >95:5

a | H R'H-O"~'~O"si(t-Bu)2,

+ ~~MeO ~ ' ~ O a

Ketoside

Yield (%)

152 153 154

60 80 91

was key to the success of the ketoside forming reaction (Table 12). The use of C(25)-deoxy substrate 149 gave a 60% yield of 152 as a 2:1 ]3:a mixture and the C(25) TBDMS ether 150 gave 153 as a 4:1 ]3:~ mixture in 80% yield. The difference in selectivity between the TBDPS and TBS groups appears to be steric in nature as C(3) TIPS-protected glycal 151 also underwent a stereoselective coupling reaction to give 154 in 91% yield. When a mixture was formed, it was most interesting to us that the

186

JON D. RAINIER

mixture did not lie at the C(24) hydroxyl group but instead at the newly formed C(23) C-C bond. Thus, the C(25) substituent was not only influencing the oxidation reaction but, to our surprise, was also playing a role in the subsequent formation of the C(23) C-C bond. We currently believe that the reaction requires a group at C(25) that is of sufficient size (i.e. >OTBDMS) to serve as a protecting group for the adjacent anhydride enabling it to avoid decomposition via oxocarbenium chemistry prior to C-C bond formation. Having demonstrated that ~-methylglycals having the appropriate C(3) substituent can be used to stereoselectively synthesize C-ketosides, we became interested in determining whether glycals of higher synthetic utility might also undergo these transformations and settled upon ~-benzyloxyglycal 155. Scott Roberts examined this and was extremely pleased to find that the anhydride from the oxidation of 155 with DMDO (i.e. 156) was efficiently converted into [3-C-ketosides 157 and 158 when subjected to propenyl and propynylmagnesium chloride, respectively (Table 13). In contrast to related substrates, anhydride 156 was surprisingly robust, as exemplified by its recovery following aqueous workup and its stability to silica gel chromatography. Next, Scott examined the efficiency with which non-allylic and propargylic Grignard nucleophiles coupled with the anhydrides from 144 and TABLE 13 The coupling of A-benzyloxy anhydride 156 with carbon nucleophiles

TBDPSO [],~O H . Si(t_Bu)2 B n O v ~ o -~""/O H 155

TBDPSO ~"" H' s i ( .... ~ ~ / O O-O=_ u'~J"'c~-/~d t-Bu)2 BnO--/ "" 1:4

156 TBDPSO Nucleophile _-- HO'"~"O"si(tgu)2

R+o BnO /

Entry

Nucleophile

~ H

Ketoside

R

Yield (%)

13:~

~MgCI

157

CH2CHCH 2

91

>95:5

~ ""~ MgBr

158

CH2CCH

99

>95:5

5

APPLICATION OF C-GLYCOSIDE IN THE SYNTHESIS OF (-)-GAMBIEROL

187

TABLE 14

159-161 PO HO, H O. " _ ~ _ H O~si(tBu)2 O

The generation of C-ketosides

R Me

PO H r~O"si(tBu)2 ae'/J'LO-'~/OH

~,'"' O-O; ~ Nucleophile

144: P = TBDPS 151 P = TIPS

Entry

Glycal

P

1

144

TBDPS

2 3

144 151

TBDPS TIPS

~MgBr PhMgBr PhMgBr

Me'/~O~ O H 144

[3 PO HO,,,[..~O..si(tBu)2 Me...~---o.~O

Nucleophile

TBDPSO ~ H O..Si(tBu)2

H

C-Ketoside

R

Yield (%)

[3:0~

159

CHCH 2

45

1"13.5

160 161

Ph Ph

75 72

<5:95 <5:95

TBDPSO TBSO ~.. H.O S, tBu . '"r T ",/ MeO2C O OTBS ~2-~"

~'~,'"' o-o; MeO

162

163

TBDMSOTf(2.2eq), -7~ CH2Cl2 (72%,>95:5) SCHEME 26

151 (Table 14). Interestingly, vinyl magnesium bromide and phenylmagnesium chloride both coupled with the anhydrides to give the 0~-anomers 159-161, respectively. Scott also employed the anhydride from 144 in a Mukaiyama-type addition reaction with ketene acetal 162 (Scheme 26). The addition of 162 to the anhydride from 144 in the presence of TBSOTf gave ~-C-ketoside 163 in 77% yield as a single diastereomer. To the best of our knowledge, the generation of 163 is novel in that it represents the first example of the addition of any ketene acetal to a glycosyl anhydride. The disparate stereochemical outcomes described above can be rationalized (Figures 6-8). For propargyl and allyl nucleophiles, coordination of

188

JON D. RAINIER

BoPso ., 6 ] L

MgXR j

FIGURE 6. Rationale for the addition of allyl and propargyl nucleophiles to 145.

I

imXMgO--" o so.--.r"--(~/~''%" \ x

[--IM OTBDPS eO~~(~)

,-Bu)2

FIGURE 7. Rationale for the addition of non-allyl and propargyl Grignards to 145.

Nu:

OTBDPS

Me TBDMSO

t-Bu)2 H 167

u

FIGURE 8. Rationale for the addition of ketene acetal 162 to 145.

the Mg counterion to the axial lone pair of the pyranyl oxygen and ligand transfer via six-membered transition structure 164 would lead to the observed 13-addition products. As mentioned above, the importance of C(3) substitution on these reactions is presumably tied to the ability of the substituent to sterically protect the epoxide and inhibit the competitive formation of oxocarbenium ion intermediates. Alternatively, this group could simply serve to lock the glycosyl anhydride into the conformation shown having the pyran ring in a bow tie with the TBDPS or TIPS ethers in pseudo-equatorial positions. Nucleophiles (phenyl, vinyl, and silyl ketene acetals) that are incapable of forming a six-membered transition structure presumably react through

5

APPLICATION OF C-GLYCOSIDE IN THE SYNTHESIS OF (-)-GAMBIEROL

189

oxocarbenium ion intermediates. Curiously, while ketene acetal 162 adds to the si-face of the oxocarbenium ion intermediate, Grignard nucleophiles add to the re-face. The Grignard additions can be rationalized either by invoking the conformation having all groups equatorial and a chair transition state as depicted for 165 or through a directed addition and a boat conformer as illustrated for 1 6 6 . 42 In an effort to explain si-face addition for 162, Scott calculated the low energy conformer for the oxocarbenium resulting from the interaction of anhydride 145 with TBSOTf and found it to exist in a boat as depicted by 167 (Figure 8). 43 Assuming 167 to also be the reactive conformer, approach of the nucleophile to the oxocarbenium from the face opposite the adjacent pseudo-axial TBDMS ether would give the observed product. Presumably, this conformation minimizes relatively severe gauche interactions between the C(2) TBDMS and C(3) TBDPS ethers. Interestingly, as evidenced by its C(3) and C(4) 1H NMR coupling constants (Ju(2),H(3) = 1.0 Hz, JH(3~,H(4) = 4.2 Hz), the pyranyl ring in 163 exists in a boat conformation. Having explored the scope of our C-ketosides efforts, we returned to gambierol. Utpal found both the esterification and acyclic enol ether forming chemistry of 147 to be sluggish. Esterification of 147 with 168 required a large excess of acid and prolonged reaction times to deliver 169 in 75% yield (Scheme 27). The conversion of 169 into the corresponding acyclic enol ether 170 was also relatively slow and gave a 35% yield of 170 (47% recovered starting material). Although clearly concerned about the yield of 170, Utpal was equally anxious about the generation of the F-ring so he decided to examine its formation prior to attempting to optimize the formation of 170. In light of

OTBDPS H H|H O~O"si(t-Bu)2 e

I~

O HO~j.....~......~/O p M B 168

= RO~"~'~

DCC, DMAP (75%)

OTBDPS 0 H"~O"si(t-Bu)2 e

TiCI 4, Zn, PbCI2, CH2Br2 ("- 169: X = O TMEDA, THF, 65oC k,~ 170: X = CH2 (35%) SCHEME 27

190

JON D. RAINIER TABLE 15 Tetrasubstituted enol ether 171 from RCM

OTBDPS

OTBDPS

PMBO"~'~]/0,, O"si(t-Bu)2 catalyst ,. P M B O I r~r 2 ~ . . . 0 conditions .~'~'Me "170

MsN.

MeO..~O..si(t_Bu) 2 H _ ~0 171

NMs

CI "]/ F3C~ I CI/ au_-~. I Ph Me/~" 0 ....,Mo~C(Me)2Ph PCY3 0d 90 Me__~CF3 20 OF3

Entry 1 2 3 4 5

Catalyst (mol%)

Conditions

Yield (%)

20 (20) 90 (20) 90 (20) 90 (45) "~ 90 (45) a

Hexanes, 65 ~ Phil, rt Phil, rt Phil, 65 ~ Phil, 80 ~

0 0 0 5 82

a90 was added in three portions (15% + 15% + 15%).

the fact that the F-ring required the generation of a tetrasubstituted enol ether it was not surprising that these reactions were also challenging. Subjecting 170 to either the Schrock Mo alkylidene catalyst 20 at 65 ~ or the second-generation Grubbs catalyst 90 at room temperature resulted in the complete recovery of starting material (Table 15, entries 1-3). The stability of the Grubbs catalyst at elevated temperatures turned out to be critical. 44 When 170 was subjected to 90 (45 mol%, added in three portions) at 65 ~ a small amount (ca. 5%) of tetrasubstituted enol ether 171 was isolated (entry 4, Table 9). When the temperature of the reaction of 170 was increased to 80 ~ Utpal isolated 171 in 82% yield (entry 5). In contrast to the enol ether RCM reactions of the substrates that have been described previously in this chapter, we believe that the reactions of the more sterically encumbered substrate 170 proceed through a less reactive Fischer carbene intermediate (i.e. 172); thus, the need for elevated temperatures (Scheme 28). 45 To the best of our knowledge, this reaction represented the first use of enol ether-olefin RCM to build a tetrasubstituted olefin.

5

191

APPI,ICATION OF C-GLYCOSIDE IN THE SYNTHESIS OF (-)-GAMBIEROL

OTBDPS H H TO'~O"si(t-Bu)2

90

~eeO~ e H"~0

[OTBDPS R' H H I' - ~ ~ 1 7 6

7

LI Ru, ~JM~:'e- -l l0H~ 0

]/

170

I

172

OTBDPS R'

O

M

2 /.~.O..si(t_Bu) | _

~...0 H

171

R'= P M B O ~ r r r SCHEME 28

Although clearly pleased with the generation of 171, the described sequence and especially the Takai-Utimoto reaction would need to be much more efficient for it to be of use in our gambierol work. In light of the sensitivity of the Takai-Utimoto protocol to the steric environment of the substrate and that the C(25) alcohol needed to be removed for the generation of gambierol, Utpal decided to examine whether the C(25) deoxy substrate might undergo the Takai procedure more efficiently than had 169 (Scheme 29). Unfortunately, the most direct route to the deoxy substrate (i.e. removal of the TBDPS group from ester 169 and deoxygenation) proved unworkable as it required forcing conditions that resulted in competitive removal of the silylene and/or the decomposition of 169. Removing the TBDPS group prior to ester formation circumvented this problem.

OTBDPS HO'-.~O'- Si(t-Bu)2

OTMS ~.H f .H 1. Nail, HMPA(92%) ~ H"'~/~-.[..'O-si(t_Bu)2 2. TMSOTf,i-Pr2NEt(90%) J'-- i :.~....O ~M2"H Me 172 OH H

-~'O"si(t-Bu)2

1. 168, DCC,DMAP(90%) 2. HOAc,H20 (98%)

o/~..o.:<.~o 173

SCHEME 29

192

JON D. RAINIER

Key to the success of this route was the selective generation of C(25) TMS ether 172. Incorporation of the ester was followed by TMS ether hydrolysis to give alcohol 173. Utpal used the Barton-McCombie protocol to deoxygenate 173 (Scheme 30). In the initial xanthate formation, the reaction temperature was critical; elevated temperatures resulted in a substantial quantity of 175 from ester migration and C(24) xanthate formation. S o./LL..sMe ,..~0.,~ H 0 R "~ "Si(t-Bu)2 OH

O

...•H

R..~O .~O..si(t_Bu) 2 O ~...__O1__~.....~O CS2, Nail" Mel 173

Conditions

Me 174 + R

conditions MeS.

a = PMBO~rF

O1

j

o'H/~'H'o

S, tBu

e 12t

174:175

yield

0~ to rt

1:1

85%

-40~ to -10~

>95:5

89%

SCHEME 30

Deoxygenation of 174 using free-radical conditions (Bu3SnH/AIBN, 80 ~ gave 176 (Table 16). If run at relatively high concentration (0.125 M), the desired product was generated in 90% yield. At lower concentrations, tricycle 177, from a competitive 6-endo cyclization of the intermediate radical, became the dominant, or at very low concentrations, the only isolated product. 46 As predicted, Utpal found that the yield of enol ether improved when the steric environment about the ester was decreased (Scheme 31). TMS ether 172 gave acyclic enol ether 178 in 50% yield with 33% recovered starting material when exposed to the Takai-Utimoto reagent. The conversion problem was solved by turning to the C(25) deoxy substrate 176. When subjected to the Takai-Utimoto protocol it gave an 83% yield of acyclic enol ether 179. Utpal also found enol ether 179 to perform identically to 170 when subjected to the second-generation Grubbs catalyst 90 to give an 83% yield of tetrasubstituted enol ether 180 (Scheme 32).

5

193

APPLICATION OF C-GLYCOSIDE IN THE SYNTHESIS OF (-)-GAMBIEROL TABLE 16 Deoxygenation of xanthate 174 H

H

p MB O . " ' - v . " ' ~ O ~ - ~ O - -

Si(t_Bu)2

S e

O..'~L-.SMe Bu3SnH, AIBN ,Phil, 80~

R ...~O H..~..O.. Si(t_gu)2

PMBO L.

I/t 176

O~MMeeO~173e H 4

LT T "s~l,-Bul~ . . ~ o ~f-..~ ~ Me H 177 Concentration (M) 0.003 0.014 0.125

176:177 a

Yield (%)

<5:95 1:1.4 >95:5

50 60 90

2177 was isolated as a 3:1 mixture of diastereomers.

R R~]/O H . ~ O . .

Si(t_Bu)2

TiCI4, Zn, PbCI2, CH2Br2 TMEDA, THE 65 ~

e 1:4 172: R = OTMS

R= PMBO/~~s

176: R = H

~

R ..]]./O

O..Si(t_Bu)2

Me 178: R = OTMS (50%, 33% recovered 172) 179: R = H (83%)

SCHEME 31

MsN

OH

HO -silt-,u/

NMs

Ci/l~u--Nph PCy3 O H

H o

90 ~ PMBO/"v'~ ~ "Si(t-Bu)2 Phil,(83o/0 80~) Me-"L ~ v ~ -~eO"~" -~v O 180

SCHEME 32

194

JON D. RAINIER

As with 170, the Schrock catalyst was unable to convert 179 and only resulted in recovered starting material. Remaining to be formed on the F-ring were the C(20) and C(21) stereocenters. From tricyclic enol ether 180, DMDO oxidation and DIBAL-H reduction of the intermediate anhydride 181 provided the requisite stereocenters and 182 in a highly efficient fashion (Scheme 33). As discussed previously, we believe that the generation of the C(20) stereocenter comes from a directed reduction. In contrast to the reaction of 136 (Scheme 23), the DMDO/DIBAL-H sequence of 181 was superior to hydroboration and oxidation, which resulted in the generation of a 5:1 mixture of diastereomers. R

y

OH H ~"'~ O'si(t'Bu)2

0-0;

R~,71/O~/'~O" Si(t-Bu)2

6

180

181

R= PMBO~~ DIBAL-H

1

H H H R" , b O ~ " ~ O " Si(t_Bu)2 nJ;L .,,L . . L . . o HO'M~'JMeO" H v 182

(93%) >80:1 diastereomeric ratio

SCHEME 33

Having completed the F-ring, Utpal moved to the seven-membered H-ring. Required was not only the incorporation of the 3 ~ alcohol and the olefin but also a handle to attach the side chain. He initially opted to employ an acid-mediated cyclization and elimination approach to this problem with the notion that the side chain would be incorporated in a subsequent C-glycoside-forming sequence. The olefin and 3 ~ alcohol would come from the corresponding ketone and a Saegusa oxidation followed by an axial attack of a methyl nucleophile on the ketone. To examine this, the cyclic silylene from 182 was removed using HF.pyridine and the resulting triol was transformed into the corresponding 1o triflate and 2 ~ TBS ether (183, Scheme 34). To incorporate the acetal required for the hydroxy acetal cyclization reaction, Utpal examined the coupling of the triflate with acetal magnesium cuprate 187. Unfortunately, this reaction was completely unsuccessful, resulting in either recovered starting material or decomposition of the triflate. This result forced him to examine a more circuitous route to the same compound. In contrast to 187, allyl

5

APPLICATION OF C-GLYCOSIDE IN THE SYNTHESIS OF (-)-GAMBIEROL

H H H R2~O1 ,~.~O..si(t_Bu)2 1. HF.pyridine (100%) HO,.MU _ _ , ~ o ~ O Me H

195

=_ R, ,H, ~ O H . ~ ~ OHT B S

2. Tf20, 2,6-1utidine(62%) 3. TBSOTf, 2,6-1utidine(92%)

HO,,.~../~. O..~-.,..jOTf Me Me H

182

183

R= PMBO~#( ~ C u M2g C l

H H H T BS R,,,~O,...~~O '

~_

(88%)

H O ~ e O ~ 184

H H H R~O..~~OTBS

1. MeOH, HCI .

o q. "H

185

60 ~ pyridine,

"]

CHO

135 ~

1. TMSCI,i-Pr2NEt (100%)

2.BH3~ H202,KOH (62%) 3. (COCI)2, DMSO, NEt3 (85%) H H H,, R]~O~--~u~k /

T SO- e eO186

(37%,2 steps)

OMe BrMgCu'~..~eMe)2 187 SCHEME 34

cuprate underwent an efficient coupling with triflate 183 to give adduct 184. 47 Hydroboration and oxidation of the olefin gave aldehyde 185 after Swern oxidation. Treatment of the aldehyde with methanol and HC1 resulted in the formation of the methyl acetal along with the removal of both the TMS and TBS ethers. Treatment of this compound with PPTS, pyridine, and heat gave the H-ring enol ether in 37% yield for the last two transformations. Although we were pleased to have generated the H-ring using the acetal cyclization chemistry shown in Scheme 34, we were not satisfied with the low yield in the 185 to 186 conversion; nor was it clear that 186 was ideal for the incorporation of the H-ring side chain. After much deliberation, we came to the conclusion that the incorporation of the H-ring side chain could best be accomplished using an enol ether olefin RCM reaction as this would enable us to incorporate a precursor to the side chain into the H-ring precursor (Scheme 35). The H-ring alkene would come from an elimination reaction of the appropriately substituted C(30) ketone. Subsequent to coupling, the addition of MeMgBr to the ketone

196

JON D. RAINIER

X

..."5

,o,

POlO

_...,~,,,h/OR, ". . . . . . . . . . . . . . . . . .

IVle IVle- H L 187

PO"MUMeO'H'~t" 188 OR'

14

II X

FI H

H

X

H.-. H )

H

H

H.-.H/~

ba, e

o

PO

189

OR'

MeMgBr ..............

X H H H~H) R,, ~ , , , . O ~ U ~ ' . OH 191

SCHEME 35

would give the H-ring. Coupling with the A-C ring would then be followed by the attachment of the side chain. Because he was concerned with our ability to incorporate the requisite H-ring functionality, Utpal decided to examine the route in a model substrate, namely vinyl C-glycoside, 192 (Scheme 36). 17 He oxidatively cleaved the alkene using ozone to give aldehyde 193 after reductive workup with DMS. The coupling of 193 with allyl magnesium bromide gave 2 ~ alcohol 194 as a 3.8:1 mixture of diastereomers in a 75% unoptimized yield. Since he felt that both isomers would be amenable to elimination and enone formation, he carried both diastereomers forward into the RCM chemistry after conversion of the alcohol into the corresponding methyl ether. Removal of the TBS ether and esterification gave the metathesis precursor 195. By sequentially subjecting 195 to the Takai protocol and then the second-generation Grubbs catalyst 90, he was able to form the oxepene 196 in an unoptimized 50% yield for the two steps. With 196 in hand, Utpal examined its conversion into the corresponding enone. Although not optimized, he found that he could convert the intermediate anhydride into ketone 197 in 40% yield when 196 was treated with DMDO followed by MgC12 in THE Unfortunately, he was unable to find conditions to eliminate the methoxy group in 197 to give 198 without decomposing the remainder of the substrate.

5

197

APPLICATIONOF C-GLYCOSIDE IN THE SYNTHESIS OF (-)-GAMBIEROL BnO_ BRO~OTBS

BnO

03~DMS

BRO~.~OTBS

BnO~ ....L'O'~'"Jl 192 '~

BnO~ ....L...O.,P,.,CHO 193 BnO BnO%..,,L..,y~OTBS

1. ~ M g B r 2. Nail, Mel (75%, 3 steps)

1. TBAF

BnO~ .... I 1 ~ . . JO ....~

,,

0

194 OMe

2. HO OTBS,...~/ (80%, 2 steps)

/ -OTBS BnO '~:O BnO@..~O

BnO H

1. TiCI4, Zn, CH2Br2, PbCI2, TMEDA

DCC

TBSO ~

.

BnO~ ....t.,,O.J ....~ 195

2.90 (20 mol%)

OMe

(50%, 2 steps)

BnO~ ....~176HJ~oMe

TBSO BnO H ) BnO - : 0

DMDO; MgCI2 =BnO~, (40% yield)

TBSO BnO H ) BnO : : 0 0

197

//

OMe

"

0

BnO~, 198

H

SCHEME 36

While we probably could have overcome our inability to generate the enone in model substrate 198, we decided to use Yamamoto's Saegusa oxidation approach to the H-ring enone. 2sb In addition to its precedent, this approach would allow us to avoid carrying a mixture of diastereomers through much of the H-ring sequence as we had done in our model work to 197. To carry out this plan Utpal returned to the real gambierol F-H system and 182 (Scheme 37). Removal of the TBS group, esterification, and C(21) TMS ether formation gave olefinic-ester 201. As had our model substrate 195, 201 underwent a successful RCM reaction when exposed to the Takai-Utimoto conditions and Schrock's molybdenum catalyst 20 (Scheme 38). In contrast to our previous oxepene syntheses, ~7 the second-generation Grubbs catalyst 90 was less successful than the Schrock catalyst 20, giving a 39% yield of 202 over the two steps. With the H-ring skeleton in place, we were now prepared to examine the formation of the C(30) and C(31) stereocenters. To our delight the use of the DMDO oxidation, DIBAL-H reduction sequence worked nicely

198

JON D. RAINIER H H H R'"~2g~O"si(t-Bu)2

2/LL ..L ..L .o

1. HF,pyridine (100%) 2. Tf20, 2,6-1utidine -40 ~ TBSOTf, 2,6-1utidine 0 ~ (79%)

HO'~e~eO'H~ 182

R : PMBO~/ H

H

H ~

H

H H

.

,,

-40~

Me Me 199

H

(88%)

200

1. TBAF (85%)

3. TMSCI, i-Pr2NEt (100%)

O'~/'~OTBS H ~ O

R H H ,,, O

2. TBSOCH2CO2H, DCC, DMAP (95%)

TMSO 201

SCHEME 37

O~OTB

9,F -t- -1~ 1

1

1

~

TBSO

s 1. TiCl4, Zn, PbCI2, CH2Br2 T EO ,THF, 6 ~ (71% (10% recovered s.m.))

H

H H DDH~ . . y / OO~ '~

/ .

TMSO~eOH~~

2.20, hexanes, 65~ (88o/o)

201

TMSO -

or 90, Phil, rt (50-55%)

H 202

R = PMBO~/ _ H _ H

i-Bu2AI" (92%)

H ,.., H / ~ O T B S

T M S O " M U MeO" H ~ 203

H

SCHEME 38

and resulted in the generation of 203 in 92% yield as a single diastereomer (Scheme 38). Interestingly, substitution at the ~-position on the oxepene is critical to achieving high diastereoselectivity. When unsubstituted oxepene 204 was subjected to DMDO, we isolated a mixture of diastereomers in spite of the presence of an angular methyl group at the junction between the six- and seven-membered rings (Scheme 39). Scott Roberts and Anita Orendt calculated transition states for the oxidation of both substituted and unsubstituted oxepenes and found that

5

199

APPLICATION OF C-GLYCOSIDE IN THE SYNTHESIS OF (-)-GAMBIEROL

H

Me Me

H

O.A..:_zO \ - - ~ : _ z O--~

t-Bu , ~ "Si

,

T

I

Me Me

o.,--,~_o ~ o - - - ~ o

O-O

t-Bu, , ~ 'Si ,

I

I

V'

205 2:1 mixture of diastereomers by ~H NMN

204

SCHEME 39

substitution helps to make the bond formation more asynchronous. 33 The result of this is that interactions between the axial hydrogen on the allylic carbon and DMDO are accentuated in the disfavored transition state. In the oxidation of the unsubstituted substrate, the reaction is not as asynchronous and the interaction between DMDO and the axial hydrogen is not as important, leading to lower levels of diastereoselectivity. From 203, Utpal found Yamamoto's TPAP and Saegusa oxidation conditions to result in the incorporation of the requisite enone and to give 206 (Scheme 40). Addition of methylmagnesium bromide gave 3 ~ ether 207 following silyl ether formation. The stereoselectivity in this transformation is interesting; we believe that axial attack of methylmagnesium bromide is dictated by developing eclipsing interactions between the C-O bond and the adjacent C(30) silyloxymethyl substituent during the transition state that would lead to the undesired axial alcohol. 48 The completion

R H A H

TPAP, NMO (90%) H ,,..,H / ~ O T B S 1. 2. LiHMDS, NEt3,

TMSO"MU 1~12" H ~ 203 R = PMBO~rF

H H H ~ H/"~OTBS R~ / O ~ ' - ' - ~

TMSCI ,. ~ 3. Pd(OAc)2, CH3CN TMSO

H

2

(90%, 2 steps)

~

O

206

r R H A H

H ,.,, H/~OTBS

1. MeMgBr,-70 ~ (94%) 2. TBSOTf, CH2CI2, rt (96%)

~ ~ -/~'OTBS T M S O ' M U MeO" H ~ " Me 207

1. DDQ, CH2CI2, H20 (98%) 2. TPAP, NMO, CH2CI2 3. NaCIO2, 2-methyl-2-butene NaH2PO4, H20, t-BuOH (90%, 2 steps)

HO2C,.

H O....~~O.~

~ . . ~ ~ TM~ MeO" H~ 208

SCHEME 40

OTBS F~. OTBS Me

200

JON D. RAINIER

of the synthesis of the F-H coupling precursor 208 involved oxidative hydrolysis of the PMB group, TPAP oxidation of the resulting 1~ alcohol, and sodium chlorite oxidation to the corresponding carboxylic acid. C. SUBUNIT COUPLING AND COMPLETION

Having completed the A-C and F-H precursors, we were finally prepared to examine their coupling chemistry as was outlined in Scheme 6. Henry Johnson took on this task. His initial efforts were wildly successful. He began with the generation of 209 from the esterification of 208 with 142 (Scheme 41). The two-step enol ether-olefin RCM reaction gave dihydropyran 210. Oxidation of the cyclic enol ether using DMDO and directed reduction using DIBAL-H provided the corresponding 2 ~ alcohol as a 3:1 mixture of diastereomers. Of note is that the use of hydroboration/ oxidation on 210 resulted in the competitive reduction of the H-ring olefin. That the reaction gave a mixture of C(17) diastereomers favoring the undesired ~-isomer was not a problem; we took advantage of the thermodynamic stability of the desired C(17) [3-stereochemistry by oxidizing the C(16) alcohol and equilibrating the C(17) stereocenter to give 211. 49 Equilibration resulted in a 4:1 mixture of isomers that could be separated and recycled. With 211 in hand, it remained to form the E-ring. As was mentioned above, prior to carrying out this work, we were much more confident about our ability to generate the E-ring using acidic conditions than we had been about generating the D-ring using metathesis. As often happens in situations like this, our analysis was off the mark. All attempts to effect the cyclization of 211 were unsuccessful. Included were attempts to generate the corresponding mixed thioketal through the use of EtSH and various acids and the generation of the cyclic ether directly through the use of BiBr 3 and Et3SiH or TMSOTf and PhzMeSiH. 5~ Based upon the lack of olefinic protons in the ~H NMR spectra of recovered samples, we believe that the H-ring olefin was undergoing competitive decomposition under the reaction conditions. In an attempt to avoid the olefin decomposition problem, Henry examined the corresponding C(28)-C(29) saturated substrate (Scheme 42). Although somewhat less than ideal in that the use of this substrate would require that the olefin be introduced post-coupling, at the very least these experiments would enable us to determine the overall feasibility of the approach. Subunit coupling, metathesis, and oxidation/reduction were carried out as described previously for the generation of 211 to give 215. Unfortunately, all of the conditions that were attempted to convert (211) into the corresponding

5

APPLICATION OF C-GLYCOSIDE IN THE SYNTHESIS OF (-)-GAMBIEROL

201

PMBO

H O ~ HO ~H ~

O

I

I

u ~OTBS H On/

~(,

I

142

~IOTBS

DCC, DMAP, CH2CI2, rt (90%)

""

208 PMBO ~Me Me

H

H _ H

u ~OTBS H On /

1. TiCI 4, TMEDA, THF CH2CI2, Zn, PbCI2, CH2Br2

%OTIS 2.20 (20 mol %) (75%, 2 steps) 121 " i-21 PMBO M

t11 " 121 I TMSO" I~e

I~e" I11 ~ 209

e

M

e

H

H

B n O ~ o ~ O ~ T M S O

H

"'

H fOTBS

Hon./

%OTIS

-~"v~: O~"Me e H

(80%, 3:1 0~:13mix) 2. TPAP, NMO (80%)

210 PMBO

wMe

Me

H

H

B n O ~ o ~ o ~ . . . t-21

H

H

1. O-O (acetone free); DIBAI-H, -65 ~

,, f O T B S

H On /

%OTIS

conditions~/Z

HO- - * ~ " ' ~ : O ~ " M e e H

conditions DBU

f

211 C(17) [3-isomer

EtSH, acid

Phil, 80 ~ ~ 212: C(17) or-isomer (4:1 mixture)

BiBr3, MeCN, Et3SiH, rt Ph2MeSiH, TMSOTf, CH2CI2, 0 ~

PMBO

- O O o. OTOS 213: R=SEt

Me "0 ~--~2 8 H ~9~ ~

214: R = H

Me/~.

11

O ...., M o ~ C ( M e ) 2 P h

ol M e + C F 3 20 CF3

SCHEME 41

O,S-ketal failed. The main products were either the acyclic dithiane 217 or decomposition when attempts were made to push the reaction. Also unsuccessful were our attempts to convert hydroxy ketone 215 directly into the octacycle using BiBr 3, Et3SiH or TMSOTf, Et3SiH. Clearly, the use of the C(21) 3 ~ alcohol as a nucleophile to generate gambierol's E-ring was problematic in our hands.

202

JON D. RAINIER PMBO ~Me

Me

H

H

H

u ~OTBS H O~. /

H

00o. o

EtSH, H+

OTiS

215 HO ~Me

O

Me

H

H

O

O

u ~OTBS H O n./

H

BnO

TBS !:1

e PMBO ~Me

co

ns=_

Me

H

H 216 H

B n O \ ~ . / ~ - ~ _ ,, _..-~../~_ ,~_.~,,./1L._ _/

Et21;-

conditions BF3*Et20, PhCl

~

H

H

Me- " O ~ o T B

S

H

AgNO 3, NCS, 2,6-1utidine, 3A MS, SiO2 AgCIO4, NaHCO3, 3 A MS, SiO2, CH3NO2

SCHEME 42

From the efforts described above, it was clear that an alternate coupling protocol was needed. Because of our continued belief that it could become a highly efficient means of generating polycyclic ethers, we opted to continue to pursue an enol ether-olefin RCM strategy. However, instead of employing metathesis to generate the "easier" D-ring, we would use it to generate the seven-membered E-ring (Scheme 43). Subsequently, a ketal cyclization and reduction sequence similar to that employed by Sasaki in his gambierol synthesis would be employed to generate the D-ring. TM To examine the new approach, our syntheses of both the A-C and the F-H subunits required modification. The generation of the A-C substrate 222 was carried out from 140 according to the sequence of reactions illustrated in Scheme 44. To simplify the removal of the alcohol protecting groups subsequent to the coupling of the A-C and F-H precursors, we opted to incorporate silyl ethers and exchanged the C(1) benzyl ether for a TBDPS ether. Acid-catalyzed hydrolysis of the 1~ TIPS ether in the presence of the 2 ~ TIPS ether and 1o TBDPS ether gave 221 after bis-TES ether formation. Selective hydrolysis of the 1o TES ether and oxidation provided coupling precursor 222. Henry constructed the new gambierol F-H precursor 225 according to the sequence illustrated in Scheme 45. Oxidative hydrolysis of the PMB group

5

APPLICATION OF C-GLYCOSIDE IN THE SYNTHESIS OF (-)-GAMBIEROL

HO -

Me Me

H

203

H

/:-q,. o...I..,--..~ o.,F.'--- ~ a HO

0 H

r, '-'..

'-'.

r,

~;~-',_/,. __7-~ a~

_--~=/--

O.-7~ / H" -'--- : OH Me

PO Me Me OH -

eo

O

:-

H

oH

H-

H0 H ROe "

PO

Me Me

H

p

H

Me H

H

H R

<....:q..oq...-q,,,o o POv~~

O ~ O

~'/~O

H

HO ~ e ' ~ \ 2 / ~ ~

219

"Me

220

SCHEME 43

eo. o.

TIPSO

1. LiDBB, THF,-78~ to -40~ (80%) 2. TBDPSCI, NEt3, DMAP, 0H2CI2,-78~ (95%) 3. CSA, CH2CI2, rt (92%) 4. TESCI, NEt3, DMAP, CH2CI2 (100%)

. eo. OTES

TIPSO

1. AcOH, H20/MeOH(97%) 2. TPAP, NMO, CH2CI2 (97%) 3. NaCIO2,2-methyl-2-butene, NaHPO4, H20, t-BuOH (95%)

221

TIPSO ~/eBO'~~OTeS

222

SCHEME 44

was followed by TPAP oxidation of the resulting 1~ alcohol and incorporation of the olefin to give 224. Subsequent to olefin formation, hydrolysis of the C(21) 3 ~ TMS ether also resulted in the hydrolysis of the C(32) TBS ether. Reincorporation of the TBS ether gave coupling precursor 225.

204

JON D. RAINIER

OPMB OHC...] H A H H ,-, H/~OTBS H H ... H/~OTBS 1. DDQ, H20 O O . . ~ ~ u ~ u OTBS CH2CI2 (98%) "~ OTBS TMSO"MUI~leO" ~ ~ Me 2. TPAP, NMO TMSO'MU MeO" H ~ Me 223 207 H 32 H O H H O -i/~OTBS Ph3PCH2 = (85%, 2 steps) TMSO~O : ~ IVle Me Me H 224 L~

1. CSA, MeOH (90%) _ ~ 2. TBSCI, NEt3, DMAP (94%)-

U

H A H H ,..,H/~OTBS ~ o TBS 225(94%)

SCHEME 45

With the precursors in hand, Henry examined their unification. Not surprisingly, based upon his earlier work that attempted to use it to generate an E-ring ketal, esterification of the C(21) 3 ~ alcohol 225 with acid 222 proved challenging. In contrast to the ketal-forming chemistry, however, Henry was able to find conditions to effect the transformation. As outlined in Table 17, he ultimately settled on a modified Yamaguchi protocol to give a 92% yield of ester 226. 52 Important were elevated temperatures and, because the intermediate anhydride was not stable for indefinite periods of time, that the formation of the anhydride be monitored by ~H NMR. Henry was now prepared to examine the metathesis chemistry to the E-ring. Unfortunately, his attempts to generate the acyclic enol ether corresponding to ester 226 were completely ineffective (Scheme 46). In spite of carrying out the experiment a number of times, Henry never observed the formation of any acyclic enol ether but instead isolated a very small amount of cyclic enol ether 229 along with a mixture of products all lacking the terminal olefin. Not surprising to us based upon previous experience, the use of the Tebbe and Petasis reagents were equally unsuccessful. Frustrated by his difficulties with the generation of the acyclic enol ether, Henry attempted several other methods to the desired E-ring. Included were two strategies that were closely related to the enol ether-olefin RCM chemistry. One was an intramolecular cyclization reaction using

5

205

APPLICATIONOF C-GLYCOSIDEIN THE SYNTHESIS OF (-)-GAMBIEROL TABLE 17 Subunit coupling of 225 with 222 H

HA H ~'..~~U~~~U

H ,.., '~ ~ O T B S

HO/---"'~_-_"O / - ~ Me Me H

B

225

PMBO IMeA B

n

O

conditions

OTBS I~le

PMBO r~eo n

u

~

Conditions

~

OTES \ O o

~

H

H

oH

O

~H

.... M~ O.-~"k

OTBS

-/~--nTBS

H~M~

Yield (%)

DCC, DMAP (COC1)2; Nail (COC1)2; Zn 227 Condition A, 12 h Condition B, 0.5 h Condition B, 4 h Condition B, 1.3 h

o 222

Me

226

Entry

O

Me H,,OT~)S

Me O

0 0 0 0 30-60 35 0 92a

O

NO2

2 CI

CI

0

227

228

Condition A: 228 (6 equiv.), NEt3 (7.5 equiv.), DMAP (7.5 equiv.), CH2C12, -20 to -5~ Condition B: 228 (6 equiv.), NEt3 (7.5 equiv.), DMAP (7.5 equiv.), CH2C12,40 ~ "Progress of anhydride formation was monitored by ~H NMR.

PM ~ e o

~.~.......OTES~--.-., H u O H B n O ~ o ~ o ~ ' ~ o ~ O . ~ O T B S CH2Br2,0H2012,Zn0' PbCI2, TMEDA,THF, TiCI4 PMBO.,

Me

226 Md O ~ O T B S

H OTES/----N H

BnO~

)\ H

H

H

O oTOS (3-5%)

+ productsfrom terminalolefin decomposition SCHEME 46

206

JON D. RAINIER

dichloroester 230 and a Takai cyclization protocol (Scheme 47). This resulted in complete decomposition of the starting material and no indication of the formation of 229. The other strategy was similar to one that had been demonstrated by Hirama and co-workers 53 during their work towards ciguatoxin where they had employed the Takeda protocol to cyclize an ester having a pendant bis-thioacetal. When 231 was subjected to the Takeda reagent (CpzTi[P(OEt)3]2), Henry once again only observed decomposition of the starting material (Scheme 47).

PMBO . BnO_

Me

H OTES X - ~ , ,

~ ' ~ ( ) ' ~

~ o ~ O ' ~

O...~ O\H -

H H

T~

M~ Mg O ~ g T B S

230:X=CI 231"X=SPh X = CI | X = SPh Takai + Cp2Ti[P(OEt)3]2

PMBO ..

Me

HOTESF--~ H O H

O H

H H

Me

SCHEME 47

In search of a method of carrying out the desired coupling without having to completely rework our strategy to these molecules, we became intrigued by the possibility of utilizing a Ramberg-B~icklund reaction in the coupling. As precedent, Franck had described the use of a modified Ramberg-B~icklund reaction to generate C-glycosides where the requisite bromide was generated and reacted in situ (Scheme 48). 54 OBn ~ O O BnO BnO---~r----v,-~v R 232 O

OBn C2F4Br2 ~ O - BnO KOH/AI203 - BnO~R (85%) 233 SCHEME 48

OBn Et3SiH ~_ = BnO O _ TFA (60%) BnO-234 ~ " f - "H~ \ R

5

207

APPLICATION OF C-GLYCOSIDE IN THE SYNTHESIS OF (-)-GAMBIEROL

Henry modeled the desired coupling using isobutyl thioether derivative 235. When he subjected 235 to Franck's conditions and CzF4Br 2, he saw complete decomposition of the starting material and no indication that the desired alkene had formed (Scheme 49). He also attempted to utilize a more classical Ramberg-B~icklund approach wherein the intermediate halide was isolated, but was unsuccessful in these efforts as well. 55

"

S _,L_ /

~

\ {'---0~0

C2F4Br2 KOH/AI203

]'- ~-tO,~OTBS ~ 3 T B

S

decomposition

SCHEME 49

Having failed in our intramolecular C - C bond forming reactions, we went through what only can be described as a period of desperation during which we considered a number of different intermolecular C - C bond couplings. Among these was a Mukaiyama aldol coupling between an E-H silyl enol ether and an A-C aldehyde. 56 Once again the C(21) carbon atom proved to be too sterically hindered to react; Henry was unable to induce coupling even when an excess of isobutanal was used as the coupling partner (Scheme 50).

TBDPSO~

l'v TMSO'~/-~O\~

OTIPS - Me Me H~-r=e

v ,/..or 237

H H

CHO 236 M~ Me " O ~ 3 T B S

N.R.

Lewisacids SCHEME 50

The difficulty in carrying out intermolecular substitution reactions on the E - H rings can best be illustrated by our inability to not only alkylate the enolate from 238 with 240 but also to methylate it using methyl iodide (Scheme 51). We were also unsuccessful in our attempts to utilize the metalloenamine from 239 in a coupling reaction with 240. The problem here was not enolate formation because quenching of the enolate from 238 and LDA with D20 resulted in deuterium incorporation at C(21).

208

JON D. RAINIER

H

x<'--"C2~

H H or,40

Me

M~ O ~ O T B S I-I Me

238: X = 0 239: X = NNMe2

T.O SO.

OTIPS : Me Me

N.R. Hcvr,--~ . . . .

] L.,v~o~o

_.~OT f

240

SCHEME 51

After having spent some time considering a number of alternate strategies and having failed in all of them, we came to the conclusion that the metathesis approach might not be so bad after all (we had at least observed some cyclic product) and that we needed to go back and try to solve its problems. Thus, we retraced our steps and went back to the reaction that gave a 5% yield of cyclic material (Scheme 46). In light of the steric environment about the ester in 226 and our previous experiences in model substrates (see Tables 3 and 4 and Scheme 5), it was not surprising that acyclic enol ether formation was problematic and that the titanium methylidene from the Takai-Utimoto reagent was preferentially reacting with the olefin. At this stage of the project we were focused on generating the acyclic enol ether and reasoned that we would have a higher likelihood of the reagent reacting with the ester if an internal instead of a terminal olefin were used. To this goal, Henry synthesized internal olefin substrate 243 and subjected it to the Takai-Utimoto reaction conditions (Scheme 52). Although more promising than any other experiment that we had run to this point, this reaction was also problematic; preferential decomposition of the olefin was again observed with the only identifiable products being cyclic enol ether 244 and terminal olefin 226. Following this result, it finally became clear to us that it was going to be difficult to generate the desired acyclic enol ether and we decided to change our focus to the generation of cyclic enol ether 244. The clue as to how to solve the decomposition problem came from the isolation of terminal olefin 226 from the reaction of 243 (Scheme 52). Mechanistically, the interaction of the Ti methylidene with the alkene in 243 produces one of two intermediate titanacyclobutanes (Scheme 53). When using CH2Br 2 as the carbon source in the Takai protocol (R = H), the intermediate that leads to cyclic enol ether 244 has Ti oriented at the more hindered, "internal" position of the alkene (i.e. 247). The alternative "undesired" orientation

5

APPLICATIONOF C-GLYCOSIDEIN THE SYNTHESISOF (-)-GAMBIEROL

209

H o.c. o. HoC~176

Ph3PCHCH3~- I~le . . . . ~[,21 ~._...~. _ .-//~'.OTBS ~ ~ F~. OTBS (85%, 2 steps) ' - TMSO'~UMeO"~ ~ Me TMSO"_:_v.:.'O" _ : _ ~ Me Me Me H 223 241 H _ H

H ,-, H / ~ O T B S

1. CSA, MeOH (90%) 2. TBSCI, NEt3, DMAP (94~ .

"'-

HO'MUI~I~O" ~ ~

Me

242 TIPSO [~eo.~.H'oT~)

S

CI

~ CI /

222 (93%)

O

~

'J~CI, DCC, DMAP "CI

TIPSO Me Me M e ' - " ~ .,~ ME_O~I~,-,.~.OTES ~ H ~.0 '~ oH

.~ .~-o-~-~-o-~.~ " ~ ,- --"4~ H

T.D~SO..-

~4~

228

. ~o~

OT.S

.... ~e O ~ O T ~ S

CH2Br2' CH2CI2,Zn~ PbCl2, TMEDA, THF, TiCl4 TIPSO.

Me

H OTES/---N J

~ 6 " ~ O . . ~ ,k.,,.4--./?~o ~

[/"H-O ill ..]

H

"

244

H

. O \+. . ~ ~HO' UHO T B S Me

.~

,t. Me OTES~~ H "1" "1 o '~oH

/14 U-H ~ / M " 2 T B S

(15-30%)

226 (10-20%)

OTBDPS

SCHEME 52

proceeds through titanocyclobutane 245 having Ti proximal to the methyl group and leads to terminal olefin 226. Based upon the poor conversions observed when 226 was independently subjected to the reaction conditions (Scheme 46), we believe that the reasons for the low and capricious conversions in the reaction of 243 was related to the instability of terminal olefin 226 to the reaction conditions; it was being siphoned out of the

JOND. RAINIER

210 "Ti"

R "Ti"~/ R

\P-Y"t

Me 0

H OTES"~~H ,-, 0 u\

,4-0

243 R ][ R=H, Me

"Ti"~/

R

R "Ti"

Me _ H .OT~S~'~H 0

Me _ H . O ~ H r , 243: FI = Me

"~ H

H

Me J 247: R = H 248: R = Me

1LiTi" ~

~

0

\

,.k,/~4 u H

~

"Ti"=O Me.

_ H'l' OTES L'~HH 0

Me 244

"~ H

t:4 249

6

,.,.,.r

SCHEME53 metathesis pathway via a competitive non-productive decomposition reaction. To overcome this and to improve the overall efficiency of the process, we proposed to employ a Ti alkylidene (R 4: H) rather than a methylidene (R = H). Reaction of the alkylidene in the "undesired direction" would simply result in the re-generation of the more stable internal olefin starting material (i.e. 243) probably as a mixture of olefin isomers. Ultimately, the alkylidene would react to give 248 and cyclic product 244 after the decomposition of 248 to 249 and cyclization. Critical to this proposal was that the Takai-Utimoto protocol had been shown to be amenable to the generation of a variety of alkylidenes through the use of different dibromoalkanes in its in situ preparation. 57 It was one of those very wonderful days in the lab that one occasionally has when this hypothesis proved to be accurate (Scheme 54). By subjecting

5

211

APPLICATIONOF C-GLYCOSIDEIN THE SYNTHESISOF (-)-GAMBIEROL

TBDPSO

TIPSO Me Me M e " / " ~ /:-... I ..O.'i':--... OTES ~ H 0 oH A ~ "~-~ "~ '~ ' ~ Xr-~" H H ,,,,

OTBS

OOTBS

....

CH3CH2Br2,CH2CI2,Zn~ PbCI2, TMEDA, THF, TiCI4 TIPSO.. : M e TBDPSO~ ~

Me H 0TES/---k J H 0.0\~..~...OUOTBSt0 ~ .

~.~//~'-_0 ' "H - H H

n

244 ....

TIPSO.. t,A., H OTES -Me "'~ ~ | / ,,._ : u

H H

/1~ U-H ~.~I~-2TBS

H

\F -O

(60%)

.

H

H H

T o o

O T B S 2 5 0 Me

(30%)

SCHEME 54

~-TIPS substrate 243 to the Takai-Utimoto ethylidene reagent that was generated from 1,1-dibromoethane instead of dibromomethane, Henry was able to isolate oxepene 244 in 60% yield. Interesting was that for the first time the reaction also led to a substantial quantity of acyclic enol ether 250. Qualitatively, the titanium ethylidene that comes from the use of 1,1dibromoethane was much more reactive than the corresponding methylidene reagent. We plan to examine this phenomenon more closely in the near future. Henry spent a considerable amount of time on the enol ether-olefin RCM reaction of 250 (Table 18). He found that the reaction only worked when 250 was subjected to both the second-generation Grubbs catalyst 90 and ethylene; it was important that 250 be converted into the corresponding terminal olefin prior to it undergoing cyclization. Elevated reaction temperatures during the cyclization reaction helped to minimize the generation of dihydropyran 251 from isomerization of the olefin and cyclization (compare entries 4 and 5). Interestingly, the Schrock catalyst 20 was not effective at all in the 250 to 244 transformation. With the ability to transform 250 into 244, our overall yield for the conversion of ester 243 into heptacycle 244 increased to a respectable 80%.

212

JON D. RAINIER TABLE 18 The conversion of acyclic enol ether 250 into cyclic enol ether 244

cata, st ~-0 i:.I

Entry 1 2 3 4 5

Me ~F 250

conditions

/ }~.0. H -

Me-'rP s 244

Conditions

Yield (244)

90 (20 mol%), Phil, rt to 80 ~ 90 (20 mol%), PhCH 3, 110 ~ 20 (20 mol%), hexanes, 60 ~ 90 (20 mol%), Phil, 80 ~ ethylene (1 atm); N 2 purge, 90 (20 mol%), 40 ~ 90 (20 mol%), Phil, 80 ~ ethylene (1 atm); N 2 purge, 90 (20 mol%), 80 ~

0% (60% recovered 250) 0% (60% 251) 0% (80% recovered 250) 30% (40% 251)

F3 Me 004"M~ MeICF3~_ 20 OF3

65% (20% 251)

MsN_ NMs H H cl~Scl/iRu__~ph r r F ~ %" 90PCY3 / 251

With the E-ring finally in hand, Henry turned his attention to the reductive cyclization chemistry to the D-ring. Toward this goal, selective oxidation of the cyclic enol ether with DMDO followed by reduction of the intermediate epoxide with DIBAL-H gave 2 ~ alcohol 253 as a 10:1 mixture of diastereomers (Scheme 55). Neither of us anticipated the facial selectivity in the dioxirane reaction as the generation of 253 implied that DMDO approached 244 from the side of the C(21) angular methyl group. As described earlier with respect to the gambierol H-ring, this phenomenon seems to be a general feature of fused oxepenes having
5

213

APPLICATION OF C-GLYCOSIDE IN THE SYNTHESIS OF (-)-GAMBIEROL

TIPSO. Me ~Me 0

' I ' H T E S ~ i "c~ H 4/ ~ 'f~'-',. n'

TBDPSO,

H H

I-:t 244M~ Me O ~ e O T B S H

,,,,,/ O-O

,,

O\

o;,,

/

DIBAL-H TIPSO.. MeTESOHo ~ HO H ~Me 0 ; '" H H TBDPSO-~ ~ 0 ~ ' ~ 0 . ~ u ' ~ - ~ 0 " ' ~ " OTBS 253 SCHEME 55

TIPSO.. MeTESOHo ~ ;Me O'"~ ~ . ' ..... ( TBDPSO-...,

~

0

~

0

oUoT s

',,f. -0\ H

~

H H

,,le u-H ~.~M"eOTBS TPAP, NMO(93%) 254:255 = 10:1

TIPSO.. MeTES-OO- ~ ;Me -_

,. 254: c(16) ~-H I imidazole ( 40~ \ 255: C(16) a-H

HO H

H H

1. CSA, MeOH (90%) 2. Zn(OTf)2, EtSa (86%) 3. PhaSnH, AIBN (97%)

TIPSO..

~Me 0

Me

H H H o ' i / - - - - X ~'O H

TBDPSO'~o~O'~-" H

H

H

Hl~e M ~ O ~ e O T B 256

SCHEME 56

S

214

JON D. RAINIER

effective in this equilibration. Following its formation, ketone 254 was treated with CSA to remove the C(13) TES group. As we had seen previously in our synthesis of the F-H coupling precursor (see Scheme 45), it was fortuitous that these conditions also removed the C(32) TBS group. Gambierol's octacyclic core was completed by subjecting the hydroxy ketone from 254 to Zn(OTf) 2 and EtSH. In contrast to our attempts to generate the E-ring (see Schemes 41 and 42), this reaction generated a single O,S-ketal diastereomer without any degradation of the H-ring olefin. Undoubtedly, the more facile cyclization to form the D-ring is responsible for this result. Reduction of the O,S-ketal using Ph3SnH then gave the gambierol octacycle in 97% yield. It remained to attach the skipped triene side chain and remove the remaining protecting groups. To accomplish this, Henry borrowed heavily from the work of Yamamoto and Sasaki. 2s,s8 Oxidation of the 1~ alcohol was followed by the conversion of the resulting aldehyde into the corresponding diiodoalkene using a modified Corey-Fuchs addition reaction (Scheme 57). 59 Stereoselective reduction using

H H TIPSO.. Me I-I O ' i ~ J'O H \Me O TBDPSO~ ~ 0 ~~-.~ ~ - _ C)~ u ~

H H O " ~ " OH

O OT0S

257

1. TPAP, NMO (100%)

2. CHI3, KOt-Bu, PPh3 (94%)

T~

H

TIPSO,. Me -Me O

H O'i~J'O

H

H

"

H H O

H

I I

258

1. Zn-Cu (couple) 1. SiF4 (95%) 2. Pd2(dba)3,P(furyl)3,Cul, DMSO Bu3Sn - ~ : , J ~ , HO.. :Me -

H

Me

H H H O'{/----~j' O H

O

259

H H O

gambierol SCHEME 57

(85%) _~

5

APPLICATION OF C-GLYCOSIDE IN THE SYNTHESIS OF (-)-GAMBIEROL

215

Zn(Cu) amalgam gave the Z-iodide. 6~ Sasaki and Yamamoto had previously observed that gambierol's skipped triene was unstable to the conditions required for removal of the silyl groups. Thus, deprotection was carried out prior to skipped triene generation; global deprotection worked best when SiF 4 was used. 61 HF-pyridine also worked but Henry found that it was more difficult to purify the resulting triol. The main advantage with SiF 4 was that the by-products were volatile. Not surprising based on its basicity was that TBAF was completely ineffective and led to decomposition. Finally, Stille coupling using Sasaki's protocol with dienyl stannane 25927 provided (-)-gambierol. The spectroscopic and physical data for synthetic gambierol was identical to that reported previously. Impressive in this work is that only 12 post-coupling transformations were required to complete the synthesis from A - C and F-H subunits 222 and 242, respectively, thus illustrating the effectiveness of the enol ether-olefin RCM approach to these substrates. Overall, this synthesis required 44 steps (longest linear sequence from o-glucal), gave a 1.5% overall yield of gambierol, and provided 7.5 mg of gambierol that has been used in a number of ion channel binding experiments. 62,63

V. Summary

To conclude, this chapter has not only described our total synthesis of the marine ladder toxin gambierol, but it has also described some of the detours that always come with total synthesis endeavors. This work has described a new subunit coupling strategy to polycyclic ethers, led to novel syntheses of C-ketosides, and has helped to better understand the DMDO oxidation of cyclic enol ethers. Current investigations include the exploration of the biological properties of synthetic gambierol and analogues as well as the application of this coupling strategy to other marine polycyclic ethers.

Acknowledgements

The work described in this review was carried out by a dedicated, enthusiastic, and thoughtful group of graduate students and postdoctoral associates. They are Dr. Shawn Allwein, Dr. Jason Cox, Dr. Henry Johnson, Dr. Scott Roberts, and Dr. Utpal Majumder. I am indebted to them and to the National Institutes of Health, General Medical Sciences (GM56677) for their support of this work.

216

joN D. RAINIER

References and Footnotes

1. See: (a) Baden, D. G., Bourdelais, A. J., Jacocks, H., Michelliza, S., Naar, J., Environ. Health Perspect. 2005, 113, 621. (b) Pearn, J., J. Neurol. Neurosurg. Psychiatry 2001, 70, 4. 2. Trainer, V. L., Baden, D. B., Catterall, W. A., J. Biol. Chem. 1994, 31, 19904. 3. Brody, J. E., New York Times September 8, 1993. (a) Lehane, L., Lewis, R. J., Int. J. Food Microbiol. 2000, 61, 91. (b) Schnorf, H., Taurarii, M., Cundy, T., Neurology 2002, 58, 873. 4. (a) Flewelling, L. J., Naar, J. P., Abbott, J. E, Baden, D. G., Barros, N. B., Bossart, G. D., Bottein, M. -Y. D., Hammond, D. G., Haubold, E. M., Heil, C. A., Henry, M. S., Jacocks, H. M., Leighfield, T. A., Pierce, R. H., Pitchford, T. D., Rommel, S. A., Scott, P. S., Steidinger, K. A., Truby, E. W., Van Dolah, E M., Nature 2005, 435, 755. (b) Goodnough, A., Aguayo, T., New York Times October 8, 2005. 5. (a) Nagai, H., Murata, M., Torigoe, K., Satake, M., Yasumoto, T., J. Org. Chem. 1992, 57, 5448. (b) Nagai, H., Koichiro, T., Satake, M., Murata, M., Yasumoto, T., Hirota, H., J. Am. Chem. Soc. 1992, 114, 1102. (c) Nagai, H., Mikami, Y., Yazawa, K., Gonoi, T., Yasumoto, T., J. Antibiot. 1993, 46, 520. (d) Bourdelais, A. J., Campbell, S., Jacocks, H., Naar, J., Wright, J. L. C., Carsi, J., Baden, D. G., Cell. Mol. Neurobiol. 2004, 24, 553. 6. Abraham, W. M., Boudelais, A. J., Sabater, J. R., Ahmed, A., Lee, T. A., Serebriakov, I., Baden D. G., Am. J. Respir. Crit. Care Med. 2005, 171, 26. 7. (a) Lee, M. S., Qin, G., Nakanishi, K., Zagorski, M. G., J. Am. Chem. Soc. 1989, 111, 6234. (b) Chou, H. N., Shimizu, Y., J. Am. Chem. Soc. 1987, 109, 2184. 8. Valentine, J. C., McDonald, E E., Neiwet, W. A., Hardcastle, K. I., J. Am. Chem. Soc. 2005, 127, 4586. (b) Simpson, G. L., Heffron, T. E, Merino, E., Jamison, T. E, J. Am. Chem. Soc. 2006, 128, 1056. 9. For reviews discussing synthetic work to dinoflagellate toxins see: (a) Marms~iter, E E, West, E G., Chem. Eur. J. 2002, 8, 4346. (b) Nakata, T., Chem. Rev. 2005, 105, 4314. (c) Inoue, M., Chem. Rev. 2005, 105, 4379. 10. (a) Klein, L. L., McWhorter, W. W., Jr., Ko, S. S., Pfaff, K.-P., Kishi, Y., Uemura, D., Hirata, Y., J. Am. Chem. Soc. 1982, 104, 7362. (b) Bellosta, V., Czernecki, S. J. Chem. Soc., Chem. Commun. 1989, 199. 11. Brigl, E, Z. Physiol. Chem. 1922, 122, 245. 12. Halcomb, R. L., Danishefsky, S. J., J. Am. Chem. Soc. 1989, 111,6661. 13. (a) Nicolaou, K. C., Postema, M. H. D., Claiborne, C. E, J. Am. Chem. Soc. 1996, 118, 1565. (b) Nicolaou, K. C., Postema, M. H. D., Yue, E. W., Nadin, A., J. Am. Chem. Soc. 1996, 118, 10335. (c) Stille, J. R., Grubbs, R. H.,J. Am. Chem. Soc. 1986, 108, 855. (d) Stille, J. R., Santarsiero, B. D., Grubbs, R. H., J. Org. Chem. 1990, 55, 843. 14. Adam, W., Bialas, J., Hadjiarapoglou, L., Chem. Ber. 1991, 124, 2377. 15. Rainier, J. D., Allwein, S. E, J. Org. Chem. 1998, 63, 5310. 16. (a) Best, W. M., Ferro, V., Harle, J., Stick, R. V., Tilbrook, D. M. G., Aust. J. Chem. 1997, 50, 463. (b) Evans, D. A., Trotter, B. W., C6t6, B., Dias, L. C., Rajapakse, H. A., Tyler, A. N., Tetrahedron 1999, 55, 8671. 17. Allwein, S. P., Cox, J. M., Howard, B. E., Johnson, H. W. B., Rainier, J. D., Tetrahedron 2002, 58, 1997.

5

APPLICATIONOF C-GLYCOSIDEIN THE SYNTHESISOF (-)-GAMBIEROL

217

18. (a) Fujimura, O., Fu, G. C., Grubbs, R. H., J. Org. Chem. 1994, 59, 4029. (b) Clark, J. S., Kettle, J. G., Tetrahedron Lett. 1997, 38, 123. (c) Clark, J. S., Kettle, J. G., Tetrahedron Lett. 1997, 38, 127. 19. Takai, K., Kakiuchi, T., Kataoka, Y., Utimoto, K., J. Org. Chem. 1994, 59, 2668. 20. Schrock, R. R., Murdzek, J. S., Bazan, G. C., Robbins, J., DiMare, M., O'Regan, M., J. Am. Chem. Soc. 1990, 112, 3875. 21. Rainier, J. D., Cox, J. M., Allwein, S. P., Tetrahedron Lett. 2001, 42, 179. 22. Majumder, U., Rainier, J. D., Tetrahedron Lett. 2005, 46, 7209. 23. Rainier, J. D., Allwein, S. P., Cox, J. M., J. Org. Chem. 2001, 66, 1380. 24. (a) Satake, M., Murata, M., Yasumoto, T., J. Am. Chem. Soc. 1993, 115, 361. (b) Morohashi, A., Satake, M., Yasumoto, T., Tetrahedron Lett. 1998, 39, 97. 25. For the total synthesis of gambierol from other labs see: (a) Fuwa, H., Kainuma, N., Tachibana, K., Sasaki, M., J. Am. Chem. Soc. 2002, 124, 14983. (b) Kadota, I., Takamura, H., Sata, K., Ohno, A., Matsuda, K., Satake, M., Yamamoto, Y., J. Am. Chem. Soc. 2003, 125, 11893. 26. Ito, E., Suzuki-Toyota, E, Toshimori, K., Fuwa, H., Tachibana, K., Satake, M., Sasaki, M., Toxicon 2003, 42, 733. 27. Rainier, J. D., Allwein, S. P., Cox, J. M., Org. Lett. 2000, 2, 231. 28. Rainier, J. D., Cox, J. M., Org. Lett. 2000, 2, 2707. 29. (a) Miyashita, M., Yamasaki, T., Shiratani, T., Hatakeyama, S., Miyazawa, M., Irie, H., Chem. Commun. 1997, 1787. (b) Barrett, A. G. M., Carr, R. A. E., Attwood, S. V., Richardson, G., Walshe, N. D. A., J. Org. Chem. 1986, 51, 4840. 30. (a) Keck, G. E., Li, X.-Y., Krishnamurthy, D., J. Org. Chem. 1995, 60, 5998. (b) McDonald, E E., Vadapally, P., Tetrahedron Lett. 1999, 40, 2235. 31. Dossetter, A. G., Jamison, T. E, Jacobsen, E. N., Angew. Chem. Int. Ed. 1999, 38, 2398. 32. Ohtani, I., Kusumi, T., Kasman, Y., Kakisawa, H., J. Am. Chem. Soc. 1991, 113, 4092. 33. Orendt, A. M., Roberts, S. W., Rainier, J. D., J. Org. Chem. 2006, 71, 5565. 34. For another report of acetone adducts from anhydrides see: Lohman, G. J. S., Seeberger, P. H., J. Org. Chem. 2003, 68, 7541. 35. Ferrer, M., Gibert, M., S~inchez-Baeza, F., Messeguer, A., Tetrahedron Lett. 1996, 37, 3585. 36. See: (a) Ireland, R. E., Wilcox, C. S., Thaisrivongs, S., Vanier, N. R., Can. J. Chem. 1979, 57, 1743. (b) Fraser-Reid, B., Dawe, R. D., Tulshian, D. B., Can. J. Chem. 1979, 57, 1746. (c) Vidal, T., Haudrechy, A., Langlois, Y., Tetrahedron Lett. 1999, 40, 5677. (d) Wallace, G. A., Scott, R. W., Heathcock, C. H., J. Org. Chem. 2000, 65, 4145. 37. For another example of this reduction see: Fujiwara, K., Awakura, D., Tsunashima, M., Nakamura, A., Honma, T., Murai, A., J. Org. Chem. 1999, 64, 2616. 38. Cox, J. M., Rainier, J. D., Org. Lett. 2001, 3, 2919. 39. Dujardin, G., Rossignol, S., Brown, E., Tetrahedron 1995, 36, 1653. 40. Tsushima, K., Araki, K., Murai, A., Chem. Lett. 1989, 1313. 41. Lesimple, P., Beau, J.-M., Jaurand, G., Sinay, P., Tetrahedron Lett. 1986, 27, 6201. 42. For an excellent discussion of the influence of substituents on the conformations of oxocarbenium ions see: Ayala, L., Lucero, C. G., Romero, J. A. C., Tabacco, S. A., Woerpel, K. A., J. Am. Chem. Soc. 2003, 125, 15521. 43. The structure optimization was completed using Gaussian03, using the B3LYP hybrid functional and the D95"* basis set. See: (a) M. J. Frisch et al., Gaussian 03,

218

44. 45.

46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56.

57. 58. 59. 60. 61. 62.

63.

JON D. RAINIER

Revision C02, Gaussian, Inc.: Wallingford, CT, 2004. (b) Lee, C., Yang, W., Parr, R. G., Phys. Rev. B: Condens. Matter 1988, 37, 785 and Becke, A. D., J. Phys. Chem. 1993, 98, 5648. (c) Dunning, T. H., Jr., J. Chem. Phys. 1989, 90, 1007. Sanford, M. S., Love, J. A., Grubbs, R. H., J. Am. Chem. Soc. 2001, 123, 6543. For studies on the formation and reactivity of Fischer carbenes from enol ethers and 90 see: (a) Louie, J., Grubbs, R. H., Organometallics 2002, 21, 2153. (b) Liu, Z., Rainier, J. D., Org. Lett. 2005, 7, 141. For a related cyclization from an anomeric radical onto a pendant alkene see: Groeninger, K. S., Jaeger, K. F., Giese, B., Liebigs Ann. Chem. 1987, 731. Winkler, J. D., Rouse, M. B., Greaney, M. E, Harrison, S. J., Jeon, Y. T., J. Am. Chem. Soc. 2002, 124, 9726. See: Eliel, E. L., Wilen, S. H., Stereochemistry of Organic Compounds, Wiley: New York, 1994, pp. 880-886. Nicolaou, K. C., Reddy, K. R., Skokotas, G., Sato, F., Xiao, X.-Y., Hwang, C.-K., J. Am. Chem. Soc. 1993, 111, 6476. For an example of the successful use of BiBr 3, Et3SiH see: Evans, P. A., Cui, J., Gharpure, S. J., Hinkle, R. J., J. Am. Chem. Soc. 2003, 125, 11456. For an example of the successful use of Et3SiH, TMSOTf see: Suzuki, K., Nakata, T., Org. Lett. 2002, 4, 3943. Inanaga, J., Hirata, K., Saeki, H., Katsuki, T., Yamaguchi, M., Bull. Chem. Soc. Jpn. 1979, 52, 1989. Uehara, H., Oishi, T., Inoue, M., Shoji, M., Nagumo, Y., Kosaka, M., Le Brazidec, J.-Y., Hirama, M., Tetrahedron 2002, 58, 6493. Yang, G., Franck, R. W., Byun, H.-S., Bittman, R., Samadder, P., Arthur, G., Org. Lett. 1999, 1, 2149. Boeckman, R. K., Jr., Yoon, S. K., Heckendorn, D. K., J. Am. Chem. Soc. 1994; 116, 7459. For an example of the use of a Mukaiyama aldol reaction to couple subunits see: Kobayashi, S., Furuta, T., Hayashi, T., Nishijima, M., Hanada, K., J. Am. Chem. Soc. 1998, 120, 908. Okazoe, T., Takai, K., Oshima, K., Utimoto, K., J. Org. Chem. 1987, 52, 4410. Kadota, I., Ohno, A., Matsukawa, Y., Yamamoto, Y., Tetrahedron Lett. 1998, 39, 6373. Michel, P., Rassat, A., Tetrahedron Lett. 1999, 40, 8579. Kadota, I., Ueno, H., Ohno, A., Yamamoto, Y., Tetrahedron Lett. 2003, 44, 8645. Corey, E. J., Yi, K. Y., Tetrahedron Lett. 1992, 33, 2289. In comparison, Sasaki's synthesis involved 71 steps (longest linear sequence) and 0.71% overall yield while Yamamoto's synthesis required 66 steps (longest linear sequence) and 1.2% overall yield. (a) Johnson, H. W. B., Majumder, U., Rainier, J. D., J. Am. Chem. Soc. 2005, 127, 848. (b) Johnson, H. W. B., Majumder, U., Rainier, J. D., Chem. Eur. J. 2006, 12, 1747. (c) Majumder, U., Cox, J. M., Johnson, H. W. B., Rainier, J. D., Chem. Eur. J. 2006, 12, 1736.

STRATEGIES AND TACTICS IN ORGANIC SYNTHESIS, VOL. 7 9 2008 Elsevier Ltd. All rights reserved.

Chapter 6 A BIOMIMETIC APPROACH TO THE ROCAGLAMIDES EMPLOYING PHOTOGENERATION OF OXIDOPYRYLIUMS DERIVED FROM 3-HYDROXYFLAVONES John A. Porco, Jr. and Baudouin Gerard Department of Chemistry Boston University and Center]br Chemical Methodology and Library Development Boston, MA 02215, USA

I. II.

Introductionand Background SyntheticStrategies: A BiomimeticApproach via Photogeneration of Oxidopyryliums III. Synthesisof (-)-Methyl Rocaglate and Related Natural Products A. Model Studies Using 3-Hydroxyflavone B. Synthesis of ( +)-Methyl Rocaglate IV. EnantioselectiveSynthesis of Rocaglamide Using Chiral BrCnstedAcid V. Summary Acknowledgments References and Footnotes

219 222 227 227 230 232 239 239 239

I. Introduction and Background

The plant genus Aglaia (Meliaceae) from the tropical rain forests of Indonesia and Malaysia is the source of natural products featuring unique and densely functionalized ring systems (Figure 1). ~ One class of compounds from this genus possess the cyclopenta[b]tetrahydrobenzofuran ring system and includes rocaglamide 1, 2 aglaroxin C 2, 3 cyclorocaglamide 3, 4 and the recently isolated dioxanyloxy-modified derivative silvestrol 4. 5 The structurally related aglains containing a cyclopenta[bc]benzopyran ring system (e.g., thapsakon 5 and aglaforbesin 6) have also been isolated from Aglaia. Finally, forbaglin derivatives (e.g., 7 and 8, Figure 1) are benzo[b]oxepines derived from formal oxidative cleavage of the aglain core. The rocaglamides have been shown to display potent insecticidal 6

JOHNA. PORCO,JR. ANDBAUDOUINGERARD

220

HO~- ,.lJ---N O Me- MeO MeO HO .L

~,/

,.,L

.~/i-~/O z ........ ~HQ o~ .

MeO~.i~:[/.O ) OMeMeO;!"~ :~'~OMe 1 rocaglamide

o Meu" " HO~- 11_..

N~ .~

2 aglaroxinC

O

HO~ MeO HO

-

"

~. ~......... NMe20~OH~~., .... 1~.,,~. _./L/;O2Me()~~

........ /

(~:('"

'i

OMe

!O~oM e'~H

i

"'~ ii'-:i::i

3 cyclorocaglamide

OMe

4 silvestrol

Rocaglamide derivatives

MeO O H ~ " N / ~ ,,., I .J- ......'--..~,.,

MeMeO 0 H ~ ~ M ~ ~-~/t-J'k

o

......

MeO O ~ N ' ~ M e M e o 1 ~.~ ,,.~. Me ........

...........

O~~O~~ MeO~~~oI~HN"~-MeO.......... . O/i~ M?OMe 5 thapsakonA

M M e e ~ M e

Me02C~~ O M e

6 aglaforbesinA

7 forbaglinB

Aglain derivatives

O ~N~Me ~.~ f. I I...Me .......

.......

0~.......O/.. Me02/ ~ M

e

8 thapoxepineA

Forbaglin derivatives

FIGURE 1. Rocaglamidesand relatedcompoundsfromAglaia. and anticancer activities 2 (IC50 = 1.0 ng/mL against P388 human cancer cell lines). Subsequent studies have shown that the rocaglamides exhibit potent antiproliferative and antileukemic activities 7 and also inhibit TNF-a and PMA-induced NF-~:B activity in the nanomolar range in human T cell lines. 8 The most recently isolated rocaglate, silvestrol 4, displays cytotoxic activity against human cancer cell lines comparable to the anticancer drug paclitaxel (Taxol). 5 According to proposals by Proksch ~and Bacher, 9 the rocaglamides may be derived from a biogenetic pathway wherein 3-hydroxyflavone (3-HF) 9 reacts with cinnamic amides 10 in a Michael reaction to afford 11 which in turn forms cyclopenta[bc]benzopyran 12 after intramolecular aldol reaction (Figure 2). The resulting ~-hydroxy ketone may participate in three distinct pathways: (i) oxidative cleavage at the methylene bridge to afford the benzo[b]oxepine (forbagline) core 13, (ii) reduction of the ketone to aglains 14, or (iii) skeletal rearrangement to cyclopenta[b]tetrahydrobenzofuran ketone 15 en route to rocaglamides 16. Examination of the structures of the aglains (e.g., 5 and 6, Figure 1) shows that the overall addition of the cinnamic amide moiety occurs in different regiochemical orientations. The bisamide-derived side chains (cf odorine l~ 17, Figure 2, inset) frequently occur in aglains, aglaforbesins, and forbaglines but have not yet been encountered in the rocaglamides.

6

221

A BIOMIMETIC APPROACH TO THE ROCAGLAMIDES

O MeO O

H

...~

MeO/~.,,~,~...,. O2,.~~ = 9 l ~'J'OMe (

H

NR2

MeO O . ~ - ~

NR2

Me?H~coNR2

"- MeOQ.~.~O,..~__~x~ 11 ~ ~

O Me Mc

iinset

e

[ ~,,. N~',,Me

MeO 0

L odorine17

13Me02C~ X

.CONR2 P

oxidative

rearrangement OMe

HO MeO

0

MeO ~ UQ)LQ,~

Me?-

MeO ~ u~tQ,~

OMe

reductionX,~

/H ?~ CONR 2

M

H OH MeO" ~ "O~'--"

OMe

OMe

OMe

FIGURE 2. Proposed biosynthesis of rocaglamide derivatives.

H I

OeOO

O

oeO

"e2"5,

MeO

":/ "OMe

19

MeO v 20

18 Q'---~--,N~~: Me

I~.~ ~_~L-~.. + ~-~ ~ ~e ~ ~ 21

O- '

O Me

MeO (b) J~- ~- .OH MeO

..... -]~ ]

]~oMO"

OMe

---2-~-~ 6

'-~/ i ~ 17

FIGURE 3. Alternative biosynthetic proposals.

Alternative biosyntheses previously proposed for the rocaglamides are illustrated in Figure 3. Stanton and coworkers ~ proposed photocycloaddition of flavone 18 with cinnamic acid dimethylamide 19 to afford the benzo[b]cyclobutapyran-8-one adduct 20, which may be followed by addition of water and further rearrangement to 1. Dumontet and coworkers proposed that the aglains and aglaforbesins (cf Figure 1) may be

222

JOHN A. PORCO, JR. AND BAUDOUIN GERARD

derived from cycloaddition of flavylium 21, the tridesmethyl derivative of a compound found in the Aglaia forbesii bark extract, 12 and odorine 17 (Figure 3). Due to their broad range of biological activities, the rocaglamides have attracted the attention of organic chemists. A number of synthetic approaches to the tricyclic core structure have been developed. Trost and coworkers established the first synthesis and stereochemical assignment of (-)-rocaglamide based on an oxidative cyclization to create the dihydrobenzofuran ring. 13 Davey's approach to the racemic tricyclic rocaglamide skeleton involved an efficient keto-aldehyde pinacolic coupling. This pathway has recently been improved in studies reported by Dobler and coworkers. 14 The key step of their approach involves acyloin ring closure of a keto-aldehyde followed by Stiles carboxylation. To the best of our knowledge, no syntheses of the related aglain, aglaforbesin, or forbaglines have been reported in the literature thus far. Moreover, a unified synthetic approach to the aglains-forbaglines-rocaglamides based on biogenetic considerations still remained to be developed. At the outset of our studies, it was our specific aim to develop a biomimetic approach toward the total synthesis of rocaglamides and the related aglain and forbagline derivatives from Aglaia, including both natural products and analogues. II. Synthetic Strategies: A Biomimetic Approach via Photogeneration of Oxidopyryliums

Our first approach toward a biomimetic synthesis of rocaglamide and related natural products was based on a modification of Stanton's ~1 proposal where 3-HF 9 and cinnamate derivatives 22 could participate in a metal-assisted [2+2] cycloaddition. We were interested to access the benzo[b]cyclobutapyran-8-one 23 in order to manipulate the skeletal framework toward the aglain 24 or rocaglamide 25 core structures using an c,-ketol rearrangement (Figure 4). However, after extensive metal and solvents evaluation, no cycloaddition products were observed. An alternative would be to develop a [2 + 2] photocycloaddition using copper (I) catalysis following the work of Salomon and coworkers. ~5 In order to evaluate the prospects for photocycloaddition chemistry, we sought further information about the UV absorption and photochemistry of 3-HFs. To our surprise, the photobehavior of 3-HF is well documented and is largely focused on an interesting phenomenon called excited state intramolecular proton transfer (ESIPT).

6

223

A BIOMIMETIC APPROACH TO THE ROCAGLAMIDES

MeO 0

MeO "~ -0-9 ~']~

1""~0 ()'~ ~ H~~j R R2~"R3 I'~0 H

0

OMe

22 0

~ cycloaddltlon'" aeO 0 Q [~~_~.,'~" R MeO"~"~O~~-~ 23 ~'~I /~ _"~

o HO I/0

MeO ~ O J ~

'"7/

O"e

// -OMe

~'~'~

Ketolashdt

0 0

MeO - "~~-~" 25

OMe

FIGURE 4. Proposed [2+ 2] cycloaddition toward rocaglamide and aglain core structures.

Literature reports have documented ESIPT of 3-HF derivatives ~6 as the phenomenon being responsible for the red shift fluorescence of 3-HF at -500-600 nm after excitation in the UV range (290-350 nm). The overall ESIPT process involves excitation of 3-HF (N) at the first excitation state (N*) and then generation of a putative tautomeric form of 3-HF (T*) wherein the proton of the hydroxyl group at C3 position migrates to the ketone at C4 to yield an oxidopyrylium species (Figure 5). This particular intermediate decays via fluorescence emission to yield the phototautomeric form at the ground state (T), which undergoes a back proton transfer (BPT) to regenerate the normal form of 3-HF (N). According to recent reports, it has been established that the acidity or basicity of certain structural features in molecules are enhanced in the excited state. ~7 If two of these groups (one acidic and one basic) are in close proximity to each other, it is possible for a proton to migrate from an acidic to a basic site. The pK~ values for the hydroxyl and ketone groups at C3 and C4 have been found to be 9.6 and -2.88, respectively, in the ground state. Upon excitation, these values change to - 0 . 1 6 and 1.48, respectively, in the first excited state. ~8 The overall change in pK~ values provides a driving force for ESIPT in which 3-HF is transformed to

224

JOHN A. PORCO, JR. AND BAUDOUINGERARD

~IF OH

q*

First excited state (IV*) [ PKa (hydroxy) = -0.16 / pKa (carbonyl) = 1.48

N*

....

.... - .....

H

Normal form (IV)

I

lifetime of T*: ns scale

J

Xem: 500-575 nm ,.,~.... .~

s~

...... I"

PKa (hydr~ = 9"6 / pKa (carbonyl) = -2.88|

T* ::.

Xem" 370-425 nm

;L~x: 290-350 nm O

First excited state phototautomeric form (T*)

ESIPT

TF

(~o'H |

o'H

l

N Ground state phototautomeric form (7")

I

lifetime ~ T: gs scale

1

ESIPT" Excited State Intramolecular Proton Transfer BPT" Back Proton Transfer

FIGURE 5. ESIPT and fluorescence emission of 3-HFs.

a photoexcited tautomeric form (oxidopyrylium) T* (Figure 5). The photophysical behavior of 3-HF and fluorescence emission of T* (510-570 nm) ~9 have been described in the literature. However, it has been difficult to measure the rate of the proton transfer that is dependent on various external factors including the nature of the solvent. In non-polar solvents, the hydroxyl proton forms an intramolecular hydrogen bond with the carbonyl at C4 and the 3-HF undergoes ESIPT with a rate constant in the picosecond range. ~6Lifetimes of the excited state T* and the ground state T have been estimated to be in the nanosecond ~6 and microsecond 2~ timescales. The lifetimes of T* and T suggests that the oxidopyrylium intermediate may react in the ground state (microsecond timescale). For 3-HF, the rate of ESIPT has been shown to increase in polar protic solvents (e.g., methanol) ~6in which case 3-HF likely forms a monosolvated complex with methanol involving possible "double proton transfer" (Figure 6). 2~ Inspired by the recent work of Wender and coworkers involving the use of oxidopyrylium 26 as a key intermediate for the construction of complex natural products such as (+)-resiniferatoxin and phorbol 27 (Figure 7a), 22 w e recognized that the photogeneration of oxidopyrylium (pyrylium betaine) species 28 via ESIPT would provide an efficient and likely biomimetic approach to the rocaglamides 29 and related derivatives (Figure 7b).

6

225

A BIOMIMETIC APPROACH TO THE ROCAGLAMIDES

Me,

Me

..o

Me

..o-.

i :

!

O-H

@dH/

o|

OH

Me

,,o-,.

,,,.,-

d

;

v.-

FIGURE 6. A possible mechanism for double proton transfer.

T B S O ~ AcO

-IT

O Me

c I

OAc

DBU, CH2CI2, 25~ oxidopyrylium formation

"~

Me

1,3-dipolar cycloaddition

26 HO

OH -~

Me M e / , , , ~ ~ ~ Me

OAc

~oMe~

H

0

(a)

OH

) HO

OTBS

phorbol 27 O

O

MeO OH.R ~ ~ MeO O [~~22 R R OR ~L~T~OH = NR2, . ..~.,.~,,~~ h'o.ESIPT-------Meu - u ll~ ~'L oxidopyrylium formation MeO 9 v OMe 2~)8 v OMe I i I

t

MeOH(~(~ ,,',~-"-O

OM~ ..<

~

MeO

(b)

29

OMe

....

. . . .

MeO ~

"0~'--' 24 ~

OMe

FIGURE 7. (a) Wender's oxidopyrylium cycloaddition, formation, and application toward the synthesis of phorbol. (b) Proposed photogeneration of an oxidopyrylium species and dipolar cycloaddition toward the synthesis of the aglain and rocaglamide core structures.

226

JOHN A. PORCO, JR. AND BAUDOUINGERARD HQ 0 ,eOO

OH

...... 22____R_>..

Me

R = NR2, OR 9

" Me (~ keto124 [

"'~OMe

shift

HO O MeO HO" ''~'R

M ..<..[H]__ ~

__[H].......

O

-

QO

,,"/ ......... [O1 )1(" "'~.

O "I~R

Me

ONe agtain core structure

0

MeO

O ...-'~R

M e O @ 31 O OMe rocaglamide core structure

X-~-~,OMe

OMe forbaglin core structure

SCHEME 1. Unified biomimetic approach to the aglains-forbaglines-rocaglamides.

However, this approach introduces several new challenges including regioand stereoselectivity of the photocycloaddition process. Our biomimetic proposal toward the rocaglamides is summarized in Scheme 1. The approach involves generation of an oxidopyrylium 28 from a 3-HF derivative 9 and subsequent dipolar cycloaddition with cinnamate derivatives 22 to afford aglain skeleton 24, which may be further reduced to aglain 30. There is significant literature concerning cycloaddition of oxidopyryliums with various dipolarophiles. 23 Oxidative cleavage of the aglain to the forbagline (benzo[b]oxepine) core 31 may be conducted using Pb(OAc) 4 in benzene/MeOH. 24 The aglain skeleton 24 may alternatively be converted to the dehydrorocaglamide skeleton 25 by a-ketol (acyloin) rearrangement. 25 Such rearrangements may be conducted using acidic or basic conditions or with metal catalysts and have been used with success in a number of natural product syntheses, including K252a, 26 taxusin, 27 and other taxanes. 28 Two requirements are necessary for an ~-ketol shift: (i) the hydroxyl and carbonyl group should be in the same plane, 26 (ii) the migrating bond should roughly be parallel to the p orbitals of the C = 0 29 or more preferably the rt* orbital. 3~According to molecular models (Figure 8, with R = OMe), the migration of bond C5-C6 in 24 should be favored by orbital alignment considerations (dihedral angle H O - C 5 - C 7 = O = 18 ~ and C 6 - C 5 - C 7 - 0 = 97 ~ to afford the dehydrorocaglamide core 25 which also occurs in nature. 3~ A molecular model of the product (25, Figure 8, b) illustrates that the hydroxyl and carbonyl may not be coplanar (dihedral angle O-C5-C6=O = 53 ~

6

227

A BIOMIMETIC APPROACH TO THE ROCAGLAMIDES

O

7//

HO ~

MeO

,,C02Me

1

2"

OMe

MeO HO

5

b)

0 5 ,,,,C02M e .,,~~~

MeO

25 OMe FIGURE 8. Ketol shift of the aglain to the rocaglamide core.

suggesting that the cz-ketol shift in this instance may be irreversible. 26 Hydroxyl-directed reduction of 25 should afford rocaglamides 29. I I I . Synthesis of (-)-Methyl Rocaglate and Related Natural Products A. M O D E L S T U D I E S U S I N G 3 - H Y D R O X Y F L A V O N E

Our initial efforts toward understanding the cycloaddition reactivity of the oxidopyrylium species T* (Scheme 2) initially focused on model studies with commercially available 3-HE The appropriate wavelength for the reaction was selected after studying absorption spectra of both starting materials. Irradiation at 350 nm (uranium filter) resulted in good wavelength-dependent absorptivity coefficient differential of 3-HF 32 vs. methyl cinnamate 33. In initial studies, we employed 5 equiv, of dipolarophile 33 and screened different solvents (MeOH, MeCN, octane, and CH2C12). After irradiation for 2 h at room temperature, the 3-HF was consumed and a mixture of products was observed resulting from presumed e n d o and e x o [3 + 2] cycloaddition (Scheme 2). According to spectroscopic data

228

JOHN A. PORCO, JR. AND BAUDOUIN GERARD

0

0 Ph J-~CO2M e

0

HO..

",.

,',

33 MeCN, 350 nm 2h, rt

36

~

14%

34

//~/

&

SiO 2

56% O

CC OH"CO2Me

X-ray of 34

35

SCHEME 2. Photochemical [3+2] cycloaddition.

and X-ray analysis of a crystalline derivative, the major compound (56%) was confirmed to be endo cycloadduct 34 in which the phenyl ring of the dipolarophile is anti to the oxido bridge. TM Interestingly, an equilibrium between 34 and the benzo[b]cyclobutapyran-8-one 35 was observed during silica gel purification likely resulting from an acid-mediated ketol shift. 32 This equilibrium between the two core structures was found to be controlled by heating the mixture in EtOAc to afford 34. Monitoring of the photocycloaddition by ~H NMR (CD3CN) also confirmed formation of 34 as the major product. Compound 36 (14%) was identified by ~H NMR and X-ray crystal structure analysis of a derivative obtained by reduction and acylation with 4-bromobenzoyl chloride. The formation of the regioisomer 36 featuring an aglaforbesin-type ring system (cf 6, Figure 1) suggests the possibility of a stepwise cycloaddition 33 process (Scheme 3) involving Michael addition, proton transfer, ~-ketol shift, and C-C bond formation. Further studies will be needed to probe the details of this potential stepwise mechanism. Conversion of aglain 34 to the forbagline framework was conducted using Pb(OAc)4 in benzene/MeOH 24 at room temperature affording benzo[b]oxepines 37:38 as a mixture of keto-enol tautomers (2:1 by ~H NMR, 85% yield) (Scheme 4). However, attempted thermal ~-ketol rearrangement 34 of 34 did not afford any observable ketol shift product. Alternatively, treatment of 34

6

229

A BIOMIMETIC APPROACH TO THE ROCAGLAMIDES

Ph

I ~~

Ph

Ph

C 9O M 2e proton ~-- %

~

O2Me

CO2Me

transfer o~-Ketol shift

1) LiAIH 4

~---'v~

-[

1. proton transfer

2Me

2. C-C bond formation DMAP, THF

36

RO HO "

R

0~

~ ~.

R: Br X-Ray

SCHEME 3. Proposed mechanism tbr the production of 36.

O

,H O

Pb(OAc) 4

30% MeOH in toluene 85%

r

+

MeO2C 37

MeO2C 38

SCHEME 4. Conversion of the aglain to the forbagline core.

with protic or Lewis acidic conditions resulted in decomposition. However, treatment of 34 with NaOMe/MeOH 32 afforded a 1:1 mixture of keto-enol tautomers 39:40. The success of such basic conditions for ~-ketol rearrangement may be explained by formation of the enolate of 41, which may drive the ketol shift equilibrium 35 toward the rocaglamide core. Further support for this assumption was provided by treatment of 34

230

JOHN A. PORCO, JR. AND BAUDOUIN GERARD

O

0

HO ,,\

MeONa,MeOH 65~ 90%

+

40

Me4NBH(OAc)3 MeCN/AcOH 95%

%~

inset:

o

HO

H041~~

Me 42

43 SCHEME 5. Conversion of the aglain to the rocaglate framework.

with Nail (2.1 equiv., THF, rt) and quenching the reaction mixture with thionyl chloride to afford the stable 1,3,2-dioxathiolane 42 (48%). 36 Hydroxyl-directed reduction ~3 of 39/40 afforded rocaglate 43 (95%) (Scheme 5). B. SYNTHESIS OF (_+)-METHYL ROCAGLATE

We next proceeded to study flavanoids with trimethoxy substitution suitable for the synthesis of the rocaglamides and related compounds. Attempted photochemical cycloaddition of kaempferol derivative 3v 9 with methyl cinnamate 33 failed to afford a [3 + 2] cycloadduct using acetonitrile as solvent. Several experimental conditions were screened at room temperature including excitation wavelength (254 and 350 nm), solvent, dipolarophile concentration, and additives (e.g., metals) without success. One major product obtained during this process was the photorearrangement product 3-aryl-3-hydroxy-l,2-indandione 4438 (Scheme 6). However, when 3-HF 9 and methyl cinnamate 33 were irradiated in MeOH at 0 ~ aglain 45, as well as benzo[b]cyclobutapyran-8-one 46 (33 and 17%, respectively), were obtained after silica gel purification. The use of methanol as solvent may help ESIPT to occur via double proton transfer (cf Figure 6). Basic conditions (NaOMe) were used to effect

6

231

A BIOMIMETIC APPROACH TO THE ROCAGLAMIDES

Me0

0

MeO

ph../'~/CO2Me

0H.~~

0

33 0

Me0

h v (350 nm) rt, MeCN

9

v

MeO HO 44

0Me

~

"0Me

SCHEME 6. Photolysis of a trisubstituted 3-HF 9.

Me00 0 MeO v - O 9

"1~"~ ~0ae

hv(350 nm) ~0oC, MeOH Me0 33 (10 equiv)

+ v

.0.....~ - -

(33%)

HO

0Me

Me

[

, %_, (1 7%) T

OMe

NaOMe (2.1 equiv) MeOH, 95%

HO +

Me0 (endo, 51%)

j Me000 ,e

0

...... Me0 4~

0Me

u~_.~k

(eXv,,_7%)

Me4NBH(OAc)3 '~ MeCN, AoOH Me0 47 " ~ o - ' a

"0Me

e

SCHEME 7. Synthesis of (+_)-methyl rocaglate.

a-ketol-rearrangement of both 45 and 46 to afford a mixture of e n d o and cycloadducts 47. Reduction of 47 afforded (_+)-methyl rocaglate 4839 (51%) and the corresponding exo stereoisomer 49 (27%) (Scheme 7). The overall yield of 48 was 25% over three steps from readily available materials. Regarding the stereoselectivity of the photocycloaddition process, molecular models of the e n d o transition state (Figure 9) show the possibility for re-stacking interactions of the two aromatic tings of the 3-HF and methyl cinnamate. This arrangement also suggests likely steric interactions between the methyl ester of the dipolarophile and the methoxy group at C5 of the 3-HE In order to avoid this steric constraint, the methyl cinnamate may also approach via an exo transition state where the methyl ester and methoxy groups at C5 are away from each other. The favorable re-stacking interaction present in the e n d o transition state may explain the product ratio (2/1) obtained for methyl rocaglate stereoisomers. exo

232

JOHN A. PORCO, JR. AND BAUDOUIN GERARD

MeO MeO

OH

HO HO- ,,,,C02Me

MeO@

|

endo

MeO

transitionstate

OMe m

28

|

~..~L,,,.O a e

(

O

MeO~

~

HO

C

02Me

%

exo

transitionstate

FIGURE 9. e n d o and exo transition states in the photocycloaddition process.

IV. Enantioselective Synthesis of Rocaglamide Using Chiral BrOnsted Acid

According to the biogenetic hypothesis for aglain derivatives, chiral cinnamides such as odorine 17 have been proposed as likely precursors and could be utilized in nature as chiral dipolarophiles. In preliminary experiments, we evaluated use of readily available chiral auxiliaries including a chiral N-acyloxazolidinones 50, 4o menthol esters 51, 41 and N-acyl sulfinamides 5242 derived from cinnamic acid to determine if they will participate in photochemical cycloaddition. This process involved evaluation of various solvent systems and reaction conditions to determine the reactivity of dipolarophiles with kaempferol derivative 9 (Scheme 8). Photocycloadditions were tested in a wide range of solvents, micellar 43 and reverse-micellar 44 media. Micelles and reverse-micelles have been used to catalyze and control photochemical reactions 45 and have been shown to successfully mimic biological membranes. 46 However, thus far we have not observed photocycloaddition products under any conditions using chiral dipolarophiles, including 50-52. We rationalize these preliminary results by steric congestion of the chiral auxiliaries, which may block approach of the dipolarophile to the oxidopyrylium dipole for the production of 53. During our studies toward the synthesis of (_+)-methyl rocaglate, 4v we found that the cycloaddition required polar protic solvents such as methanol in order to proceed. As described previously, it has been proposed that

6

233

A BIOMIMETIC APPROACH TO THE ROCAGLAMIDES

O MeO

O

O

MeO u,-, ~ u~\

MeO OM 9

" ~

MeO

53

O

"~

50-52 OMe Me , e O O .

j

R*" Ph

Me

50

Me

Me

51

Me

52

H

0 (~'uN.

SCHEME 8. Evaluation of chiral auxiliaries in the [3 +2] dipolar cycloaddition.

ESIPT may be enhanced in such solvents due to the formation of solvated complexes involving "double proton transfer. ''48 In order to access chiral, non-racemic methyl rocaglate, we therefore investigated use of chiral BrOnsted acids 49 in aprotic solvents as host-guest 5~ complexes to mediate photochemical cycloaddition. Activation of carbonyl compounds via hydrogen bonding has shown significant utility in asymmetric catalysis 51 in various organic reactions including Diels-Alder cycloaddition, 52 aldol condensation, 53 and the Morita-Baylis-Hillman reaction. 54 We thus evaluated various chiral BrCnsted acid catalysts in an attempt to activate the 3-HF substrate and potentially enhance the rate of ESIPT via double proton transfer (Scheme 9). The electron-rich carbonyl of the 3-HF may form intermolecular hydrogen bonds with the protic, chiral additive as well as a potential intramolecular hydrogen bond with the hydroxyl group at C3 of the 3-HE There is literature precedent for stabilization of dipoles via hydrogen bonding. 55 Under photoirradiation, intramolecular proton transfer should occur and the resulting oxidopyrylium should be stabilized via hydrogen bonding with either the hydroxyl group at C4 or the phenoxide at C3. We evaluated various chiral protic additives including TADDOL 52b-d 54, B INOL 54 derivatives 55, diyne-dio156 56, cinchonine 57 in aprotic solvents (toluene, - 1 0 ~ 350 nm). The enantioselectivity of the asymmetric photocycloaddition

234

JOHN A. PORCO, JR. AND BAUDOUIN GERARD

0 MeO 0 Ph/..,,.~-,,.C02M e MeO HOJ' . . . . 0 I Jl. nil 33 b equiv j . ~/~jt.;U2Me MeOHo/~,,,,CO,~.,2Me i~-'~ y " " "Chiral additive1 equivD, i i ~ i / ~ = - ~ MeONa, MeOH M ..~-.~~toluene, -10 oC, 10 h . ~ ~ k . _ ~ . v47Y~O 2:0.03 M ; ~A'OMe hv(350 nm) MeO 60~ MeO 47 +

Me

OMe

46 after purification H M ~ c ~ ..... MeO v u~ } 49

e

H M ~

=

MeO "~ O ( %

~~~OMe

Chiral additive i

e

48

Me><.Me

Me4NBH(OAc)~ MeCN, AcOH, rt

~OMe R: 0

54

from L-tartrate

HO HO

Br OH OH Br

55a

55b racemic

54a ee: 7 / 5 (endo/ exo) 54b ee: 25 / 23 (endo / exo)

0o. 6 5o. 56

57

mixture obtained

SCHEME 9. Primary screening of chiral BrCnsted acids.

was evaluated on methyl rocaglate obtained after ketol shift and hydroxyldirected reduction. Initial experiments identified T A D D O L 49b'e reagents as suitable chiral mediators (Table 1). For example, photochemical cycloaddition of 9 with methyl cinnamate 33 (5 equiv.) using 1-phenyl TADDOL 54a (1 equiv.) in toluene at 0 ~ afforded 24% overall yield and 7% ee of (-)-methyl rocaglate 48 after ketol shift and reduction and 31% yield and 25% ee using 1-naphthyl TADDOL 54b. Encouraged by these results, we proceeded to evaluate additional TADDOL derivatives in the photochemical cycloaddition. Based on optical rotation data, use of TADDOL derivatives derived from L-tartrate were shown to favor the natural (-)-enantiomer of 48. According to the previous work of Rawal and c o w o r k e r s 49b involving use of TADDOLs as chiral BrCnsted acids for Diels-Alder cycloaddition, we decided to evaluate the effect of temperature as well as the nature of both the aryl substituents and ketal side chain of the TADDOFs. Investigation of reaction temperature showed a noticeable effect on the

6

235

A BIOMIMETIC APPROACH TO THE ROCAGLAMIDES TABLE 1 Development of enantioselective photochemical cycloaddition a HO M eO O

TADDOL,I" ht~ > 350no_70 ~

M e2OC

toluene/CH2CI 2 MeO

v

u 9

R

v 'v~ - ~ - ~ O M~e] 33

R

HO

MeO ",..~/'~"-/~'COMe HO ~ / . L _ ~

2.3.NaOMe/MeOHMe4NBH(OAc)aMeO "~

O ~_..

M~COr~2Me +

OMe

48

R Me, Ar" phenyl, 54a

Ar Ar

~ - k / ~.. MeO

, v

u "O~Me 49

Ar Ar

R Me, Ar phenanthren-9-yl,54c HO OH A r pyren-l-yl, 54h ~ _ X '~' XO .,-[ R H, Ar phenanthren-9-yl,54d Ar-"(~H H~"Ar R phenyl,Ar 9phenanthr-9-yl,54e HOAr-r-~A~u 'Ar'~AOH cyclohexyl Ar: phenanthr-9-yl,54f R cyclooctyl,Ar pyren-l-yl,54g OXO Ar~/ ~ A r

Entry 1 2e 3e 4 5 6 7 8 9 10f 11 12

Additive 54a 54b 54a 54b 54c 54d 54e 54f 54f 54g 54h

Yield of 45 32 60 61 51 92 90 70 54 79 73 58 22

(%)b

Yield of 48/49 (%)c

Enantiomeric excess of 48/49 (%)d

45/19 41/15 49/7 35/4 52/9 71/14 69/14 72/22 67/19 47/9 61/16 52/7

Racemic 7/5 25/18 15/7 40/36 60/58 25/22 Racemic 71/51 53/30 82/68 89/78

~Reactions conducted with 1 equiv, of 3-HF, 1 equiv, of additive, and 5 equiv, of methyl cinnamate in toluene/CHzC12 (2/1) at - 7 0 ~ for 12 h. bIsolated yield. CIsolated yield for the a-ketol rearrangement/reduction sequence, dDetermined by chiral HPLC analysis, eReaction conducted at 0 ~ in toluene, fReaction conducted in the presence of anhydrous CH3OH (5 equiv.).

enantioselectivity of the cycloaddition. Indeed, we found that by decreasing the temperature to - 7 0 ~ and using a mixed solvent system (2:1, PhCH3:CHzC12 to avoid low viscosity and poor substrate solubility), an improvement in the enantioselectivity of the cycloaddition could be achieved. For example, when additive 54b (naphthyl TADDOL) was used at - 7 0 ~ a 40% ee could be obtained in comparison to 25% at 25 ~ In order to facilitate photochemistry experiments at low temperature, we designed a photochemical apparatus (Figure 10) wherein a Hanovia 450 W medium pressure mercury lamp was housed in a quartz immersion

236

JOHN A. PORCO, JR. AND BAUDOUIN GERARD

External cooling system -70~

Uranium filter (ht) > 350 nm)

Hanovia 450 W medium pressure mercury lamp Merry go-round

Acetone Cooling bath

FIGURE 10. Photochemical reaction setup for low-temperature experiments. (See color insert.)

wells cooled with a Thermo Neslab ULT 80 system circulator (external cooling system). Pyrex test tubes (16 x 100 mm) were mounted on a support approximately 0.5 cm from the immersion well lamp (merry-goround). Encouraged by these results, we next investigated the effect of the aryl substituents on the TADDOLs as well as the nature of the ketal group. We observed that the size of the aryl group improved the enantioselectivity of the cycloaddition likely due to increased shielding provided by the larger rt surface of the aromatic tings. For example, use of a TADDOL additive bearing a 9-phenanthrenyl substituent provided an enantiomeric excess of 60% in comparison to 40% for TADDOL with 1-naphthyl substituent. Using 9-phenanthrenyl substituent, we evaluated different TADDOL featuring various ketal side chains. Interestingly, when either methylene (54d, entry 5) or diphenyl (54e, entry 6) TADDOL acetals were employed as additives, rocaglates 48 and 49 were obtained either in racemic form or in low enantioselectivity. Infrared spectroscopy and X-ray crystal structure analysis of 54d and 54e s7 (Figure 11) suggested the possibility that the loss of enantioselectivity in these photocycloadditions may be due to the presence of TADDOL conformers involving intramolecular H-bonding between the hydroxyl groups and the rt system of the phenanthrene ring for 53e and the

6

237

A BIOMIMETIC APPROACH TO THE ROCAGLAMIDES

~

,

~,.~:.

< ..

j

p ....q 0

~

A~Ar

.o--", 0

OH O ~ O Ar~Ar Ar Ar Ar: phenanthren-9-yl

FIGURE 11. X-ray analysis of TADDOLs 54d and 54e. (See color insert.)

acetal ether oxygens for 54d rather than well-defined intramolecular, hydrogen bonding between the two hydroxyl groups. 58 Gratifyingly, when dimethyl acetal was changed to a cyclohexyl acetal, an increase of enantiomeric excess up to 71% was observed (53% overall yield). The highest enantioselectivity was achieved using a dimeric TADDOL (89%, entry 12) but with low conversion. Fortunately, recrystallization of 48 obtained from TADDOL 54g led to the formation of centrosymmetric racemate 59 crystals and the isolation of 48 (94% ee, 86% recovery) from the mother liquor (Figure 12). The TADDOL complexing agent could be recovered in high yield by precipitation from methanol. 6~A control experiment involving addition of 54g and 5 equiv, of methanol (entry 10) led to a loss of enantioselectivity presumably due to achiral background reactions promoted by the protic cosolvent. 52d In order to explain the enantioselectivity observed in the [3 + 2] photocycloaddition, we propose an assembly involving oxidopyrylium 28 and TADDOL 54g (Figure 13). The well-defined arrangement of TADDOL may form a hydrogen bond with the oxidopyrylium ylide via its free hydroxyl group which may stabilize the dipole. 55 A computational study (B3LYP/6-31 +G*) of the oxidopyrylium intermediate indicates a high

OMe

od

M e O ~

..

OMe

~:~'

,,,0H{O l~igO / H-. "0

MeOu,., 0

L

MeO

e OMe

FIGURE 12. X-ray analysis of methyl rocaglate 48 as centrosymmetric racemate crystal. (See color insert.)

MeO OH Q 0

MeO OMe Electrostatic potential map (bond density) for oxidopyrilium 28

approach of the dipolarophile

:~.~

_/~

FIGURE 13. Proposed arrangement for enantioselective photocycloaddition. (See color insert.)

6

A BIOMIMETICAPPROACHTOTHEROCAGLAMIDES

239

degree of electron density on the phenoxide oxygen suggesting this site as a strong point of interaction for hydrogen bonding. The stereofacial approach of the dipolarophile may be controlled by shielding of the aryl group at the p s e u d o equatorial position of the seven-membered ring formed by an intramolecular H-bond between the two hydroxyl groups. 49b V. Summary

In conclusion, we have described a photocycloaddition approach toward the rocaglamide natural products. The key strategy involves dipolar cycloaddition of oxidopyrylium ylides derived from ESIPT of 3-HF derivatives and a-ketol rearrangement of the initially formed again to the rocaglamide core. An asymmetric version of the [3 + 2] dipolar cycloaddition using specifically functionalized TADDOL derivatives as chiral BrCnsted acids has been developed successfully for the synthesis of (-)-methyl rocaglate. A model for enantioselectivity has been proposed involving hydrogen bonding of the TADDOL additive with the oxidopyrylium intermediate. Our synthetic approach, while not "biomimetic" in the absolute sense, sheds light on the probable biosynthesis of the rocaglamides and related natural products. Further studies on related photocycloaddition processes as well as applications toward complex natural product synthesis are currently in progress. Acknowledgments

We are grateful to the National Institutes of Health (GM-073855), Bristol-Myers Squibb, and Merck Research Laboratories for generous financial support. We thank Professor Viresh Rawal (The University of Chicago) and Professor Guilford Jones (Boston University) for helpful discussions, Professor Daniel O'Leary (Pomona College) for collaborative studies on TADDOL derivatives, and Dr. Emil Lobkovsky (Cornell) for X-ray crystal structure analyses. References and Footnotes

1. (a) Proksch, R, Edrada, R., Ebel, R., Bohnenstengel, E I., Nugroho, B. W., Curr. Org. Chem. 2001, 5, 923-938. (b) Kim, S., Salim, A. A., Swanson, S. M., Kinghorn, A. D., Anticancer Agents Med. Chem. 2006, 6, 319-345. 2. King, M. L., Chiang, C.-C., Ling, H.-C., Fugita, E., Ochiai, M., McPhail, A. T., J. Chem. Soc. Chem. Commun. 1982, 20, 1150-1151. 3. Ohse, T., Ohba, S., Yamamoto, T., Koyano, T., Umezawa, K., J. Nat. Prod. 1996, 59, 650-653. 4. Bringmann, G., Muhlbacher, J., Messer, K., Dreyer, M., Ebel, R., Nugroho, B. W., Wray, V., Proksch, E, J. Nat. Prod. 2003, 66, 80-85.

240

JOHN A. PORCO, JR. AND BAUDOUINGERARD

5. Hwang, B. Y., Su, B.-N., Chai, H., Mi, Q., Kardono, L. B. S., Afriastini, J. J., Riswan, S., Santarsiero, B. D., Mesecar, A. D., Wild, R., Fairchild, C. R., Vite, G. D., Rose, W. C., Farnsworth, N. R., Cordell, G. A., Pezzuto, J. M., Swanson, S. M., Kinghorn, A. D., J. Org. Chem. 2004, 69, 3350-3358. 6. (a) Nugroho, B. W., Gussregen, B., Wray, V., Witte, L., Bringmann, G., Proksch, E, Phytochemistry 1997, 45, 1579-1585. (b) Chaidir, J., Hiort, J., Nugroho, B. W., Bohnenstengel, E I., Wray, V., Witte, L., Hung, E D., Kiet, L. C., Sumaryono, W., Proksch, E, Phytochemistry 1999, 52, 837-842. (c) Nugroho, B. W., Gussregen, B., Edrada, R. A., Wray, V., Witte, L., Proksch, P., Phytochemistry 1997, 44, 1455-1461. 7. (a) Cui, B., Chai, H., Santisuk, T., Reutrakul, V., Farnsworth, N., Cordell, G. A., Pezzuto, J. A., Kinghorn, A. D., Tetrahedron 1997, 53, 17625-17632. (b) Lee, S. K., Cui, B., Mehta, R. R., Kinghorn, A. D., Pezzuto, J. M., Chem. Biol. Interact. 1998, 115, 215-228. (c) Bohnenstengel, E I., Steube, K. G., Meyer, C., Quentmeier, H., Nugroho, B. W., Proksch, P., Z. Naturforsch. 1999, 54, 1075-1083. 8. Baumann, B., Bohnenstengel, E, Siegmund, D., Wajant, H., Weber, C., Herr, I., Debatin, K.-M., Proksch, E Z., Wirth, T., J. Biol. Chem. 2002, 277, 44791-44800. 9. Bacher, M., Hofer, O., Brader, G., Vajrodaya, S., Greger, H., Phytochemistry 1999, 52, 253-263. 10. (a) Babidge, E J., Massy-Westropp, R. A., Pyne, S. G., Shiengthong, D., Ungphakorn, A., Veerachat, G., Aust. J. Chem. 1980, 33, 1841-1845. (b) Nagasaka, T., Yamamoto, H., Hamaguchi, E, Heterocycles 1988, 27, 2219-2224. (c) Purushothaman, K. K., Sarada, A., Connolly, J. D., Akinniyi, J. A., J. Chem. Soc. Perkin Trans. 1 1979, 3171-3174. (d) Inada, A., Nishino, H., Kuchide, M., Takayasu, J., Mukainaka, T., Nobukini, Y., Okuda, M., Tokuda, H., Biol. Pharm. Bull. 2001, 24, 1282-1285. 11. Hailes, H. C., Rapahel, R. A., Staunton, J., Tetrahedron Lett. 1993, 34, 5313. 12. Dumontet, V., Thoison, O., Omobuwajo, O. R., Martin, M.-T., Perromat, G., Chiaroni, A., Riche, C., Pa'/s, M., S6venet, T., Tetrahedron 1996, 52, 6931-6942. 13. Trost, B. M., Greenspan, E D., Yang, B. V., Saulnier, M. G., J. Am. Chem. Soc. 1990, 112, 9022-9024. 14. Dobler, M. R., Bruce, I., Cederbaum, E, Cooke, N. G., Diorazio, L. J., Hall, R. G., Irving, E., Tetrahedron Lett. 2001, 42, 8281-8284. 15. (a) Salomon, R. G., Kochi, J. K.,J. Am. Chem. Soc. 1974, 96, 1137-1144. (b) Ghosh, S., Raychaudhuri, S. R., Salomon, R. G., J. Org. Chem. 1987, 52, 83-90. 16. (a) Chou, P.-T., J. Chin. Chem. Soc. 2001, 48, 651-682. (b) Ormson, S. M., Brown, R. G., Prog. React. Kinet. 1994, 19, 45-91 and references therein. 17. (a) Ireland, J. E, Wyatt, P. A. H., Adv. Phys. Org. Chem. 1976, 12, 131-221. (b) Weller, A., Prog. React. Kinet. 1931, 1, 188-214. 18. Wolfbeis, O. S., Knierzinger, A., Schipfer, R., J. Photochem. 1993, 21, 67-79. 19. Bader, A., Ariese, E, Gooijer, C., J. Phys. Chem. A 2002, 106, 2844-2849 and references herein. 20. Itoh, M., Fujiwara, Y., Sumitani, M., Yoshihara, K., J. Phys. Chem. 1986, 90, 5672-5678. 21. (a) McMorrow, D., Kasha, M., J. Phys. Chem. 1984, 88, 2235-2243. (b) Brucker, G. A., Kelley, D. E, J. Phys. Chem. 1987, 91, 2856-2861. 22. (a) Wender, E A., Jesudason, C. D., Nakahira, H., J. Am. Chem. Soc. 1997, 119, 12976-12977. (b) Wender, P. A., Lee, H. Y., Wihelm, R. S., Williams, P. D., J. Am. Chem. Soc. 1989, 111, 8954. 23. (a)Potts, K. T., Elliot, A., Sorm, M., J. Org. Chem. 1972, 37, 3838-3845. (b) Hendrickson, J., Farina, J. S., J. Org. Chem. 1980, 45, 3359-3361. (c) Sammes, P. G.,

6

24. 25. 26.

27. 28. 29. 30. 31.

32. 33.

34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44.

45.

A BIOMIMETICAPPROACHTO THE ROCAGLAMIDES

241

Street, L. J., J. Chem. Soc. Perkin. Trans. 1 1983, 1261-1265. (d) Ohmori, N., Miyazaki, T., Kojima, S., Ohkata, K., Chem. Lett. 2001, 30, 906-907. For the use of oxidopyrylium species in natural product, see: (e) Sammes, E G., Street, L. J., J. Chem. Soc. Chem. Commun. 1983, 12,666-668. (f) Wender, P. A., Mascarenas, J. L., J. Org. Chem. 1991, 56, 6267-6269. (g) Wender, E A., Mascarenas, J. L., Tetrahedron Lett. 1992, 33, 2115-2118. (h) Baldwin, J. E., Mayweg, A. A. W., Pritchard, G. J., Adlington, R. M., Tetrahedron Lett. 2003, 44, 4543-4545. Baer. E., J. Am Chem. Soc. 1940, 62, 1597-1606. Paquette, L. A., Hofferberth, J. E., Org. React. 2003, 62, 477-567. Wood, J. L., Stoltz, B. M., Tetrahedron. Lett. 1996, 37, 3929-3930. (b) Wood, J. L., Stoltz, B. M., Goodman, S. N., J. Am. Chem. Soc. 1996, 118, 10656-10657. (c) Tamaki, K., Huntsman, E. W. D., Petsch, D. J., Wood, J. L., Tetrahedron. Lett. 2002, 43, 379-382. Holton, R. A., Williams, A. D., J. Org. Chem. 1998, 53, 5983-5986. Paquette, L., Hofferberth, J. E., J. Org. Chem. 2t)tl3, 68, 2266-2275. (b) Paquette, L. A., Lo, H. Y., Hofferberth, J. E., Gallucci, J. C., J. Org. Chem. 2003, 68, 2276-2281. Nickon, A., Nishida, T., Lin, Y-I., J. Am. Chem. Soc. 1969, 34, 6860-6861. (a) Priyakumar, U. D., Sastry, G. N., Mehta, G., Tetrahedron 2004, 60, 3465-3472. (b) Tomoda, S., Senju, T., Tetrahedron 1999, 55, 5303-5318. Dreyer, M., Nugroho, B. W., Bohnenstengel, E I., Ebel, R., Wray, V., Witte, L., Bringmann, G., Mtihlbacher, J., Herold, M., Hung, E D., Kiet, L. C., Proksch, P., J. Nat. Prod. 2001, 64, 415-420. Creary, X., Inocencio, E A., Underiner, T. L., Kostromin, R., J. Org. Chem. 1985, 50, 1932-1938. (a) Padwa, A., Carlsen, P. H. J., J. Am. Chem. Soc. 1977, 99, 1514-1523. (b) Di Valentin, C., Freccero, M., Gandolfi, R., Rastelli, A., J. Org. Chem. 2000, 65, 6112-6120. Lui, J., Mander, L. N., Willis, A. C., Tetrahedron 1998, 54, 11637-11650. Piers, E., Romero, M. A., Walker, S. D., Synlett 1999, 7, 1082-1084. Shipman, M., Thorpe, H., Clemens, I., Tetrahedron 1999, 55, 10845-10850. Sim, K. Y., Phytochemistry 1969, 8, 1597-600. Waiis, A., Corse, J., J. Am. Chem. Soc. 1965, 87, 2068-2069. (b) Matsuura, T., Takemoto, T., Tetrahedron 1973, 29, 3334-3337. Ishibashi, E, Satasook, C., Isman, M. B., Towers, G. H. N., Phytochemistry 1993, 32, 307-310. Thom, C., Kocienski, E, Synthesis 1992, 6, 582-586. Buschmann, H., Scharf, H.-D., Hoffmann, N., Esser, E, Angew. Chem. Int. Ed. Engl. 1991, 30, 477-515. Backes, B. J., Dragoli, D. R., Ellman, J. A., J. Org. Chem. 1999, 64, 5472-5478. (a) Turro, N. J., Proc. Natl. Acad. Sci. USA 2002, 99, 4805. (b) Tascioglu, S., Tetrahedron 1996, 52, 11113-11152. (a) Andrade, S. M., Costa, S. M. B., Pansu, R., Photochem. Photobiol. 2000, 71, 405-412. (b) Biasutti, M. A., Correa, N. M., Silber, J. J., Curr. Opin. Colloid Interface Sci. 1999, 3, 35-51. (a) Ramnath, N., Ramesh, V., Ramamurthy, V., J. Photochem. 1985, 31, 75-95. (b) Hailes, H. C., Raphael, R. A., Staunton, J., Tetrahedron Lett. 1993, 34, 5313-5316. (c) Tung, C.-H., Guan, J.-Q., J. Org. Chem. 1998, 63, 5857-5862. (d) Wolff, T., J. Photochem. 1981, 16, 343-346. (e) Singh, V., Raju, B. N. S., Indian J. Chem., Sect. B, 1996, 35B, 303-311. (f) Mayer, H., Schuster, F., Sauer, J., Tetrahedron Lett. 1986,

242

46.

47. 48. 49.

50.

51. 52.

53. 54. 55.

56. 57. 58.

59. 60.

JOHN A. PORCO,JR. AND BAUDOUINGERARD 27, 1289-1292. (g) Berenjian, N., De Mayo, E, Sturgeon, M. E., Sydnes, L. K., Weedon, A. C., Can. J. Chem. 1982, 60, 425-436. (a) Rodgers, M. A. J., Lee, P. C., J. Phys. Chem. 1984, 88, 3480-3484. (b) Gandini, S. C. M., Yushmanov, V. E., Borissevitch, I. E., Tabak, M., Langmuir 1999, 15, 6233-6243. (c) Scalise, I., Durantini, E. N., J. Photochem. Photobiol., A 2004, 162, 105-113. (d) Yu, J. H., Weng, Y. X., Wang, X. S., Zhang, L., Zhang, B. W., Cao, Y., J. Photochem. Photobiol., A 2003, 156, 139-144. (e) Chauhan, S. M. S., Sahoo, B. B., Bioorg. Med. Chem. 1999, 7, 2629-2634. Gerard, B., Jones, G., II, Porco, J. A., Jr., J. Am. Chem. Soc. 2004, 126, 13620-13621. Le Gourrierec, D., Ormson, S. M., Brown, R. G., Prog. React. Kinet. 1994, 19, 211-275. (a) McDougal, N. T., Schaus, S. E., J. Am. Chem. Soc. 2003, 125, 12094-12095. (b) Thadani, A. N., Stankovic, A. R., Rawal, V. H., Proc. Natl. Acad. Sci. USA 2004, 101, 5846-5850. (c) Nugent, B. N., Yoder, R. A., Johnston, J. N., J. Am. Chem. Soc. 2004, 125, 3418-3419. (d) Yamamoto, H., Momiyama, N., J. Am. Chem. Soc. 2005, 127, 1080-1081. (e) Bhasker, V., Gravel, M., Rawal, V., Org. Lett. 2005, 7, 5657-5660. (f) Yoon, T. E, Jacobsen, E. N., Angew. Chem. Int. Ed. 2005, 44, 466-468. (g) Taylor, M. S., Jacobsen, E. N., Angew. Chem. Int. Ed. 2006, 45, 1520-1543. (a) Grosch, B., Orlebar, C. N., Hertdweck, E., Massa, W., Bach, T., Angew. Chem. Int. Ed. 2003, 42, 3693-3696. (b) Legrand, S., Luukinen, H., Isaksso, R., Kipel~iinen, I., Lindstr6m, M., Nicholls, I. A., Unelius, C. R., Tetrahedron: Asymmetry 2005, 16, 635-640. (c) Bauer, A., Westkaemper, F., Grimme, S., Bach, T., Nature 2005, 436, 1139-1140. (d) Tanaka, K., Fujiwara, T., Org. Lett. 2005, 7, 1501-1503. (e) Wessig, E, Angew. Chem. Int. Ed. 2006, 45, 2168-2171. (a) Pihko, E M., Angew. Chem. Int. Ed. 2004, 43, 2062-2064. (b) Schreiner, P. R., Wittkopp, A., Org. Lett. 2002, 4, 217-221. (a) Braddock, C. D., MacGilp, I. D., Perry, B. G., Synlett, 2003, 8, 1121-1124. (b) Du, H., Zhao, D., Ding, K., Chem. Eur. J. 2004, 10, 5964-5970. (c) Huang, Y., Unni, A. K., Thadani, A. N., Rawal, V. H., Nature, 2003, 424, 146. Tang, Z., Jiang, E, Yu, L.-T., Cui, X., Gong, L.-Z., Mi, A.-Q., Jiang, Y.-Z., Wu, Y.-D., J. Am. Chem. Soc. 2003, 125, 5262-5263. McDougal, N. T., Schaus, S. E., J. Am. Chem. Soc. 2003, 125, 12094-12095. (a) Quadrelli, P., Fassardi, V., Cardarelli, A., Caramella, E, Eur. J. Org. Chem. 2002, 13, 2058-2065. (b) Adembri, G., Paoli, M. L., Sega, A., J. Chem. Res. 2003, 3, 126-127. Narasimha Moorthy, J., Venkatesan, K., J. Org. Chem. 1991, 56, 6957-6960. Irurre, J., Alonso-Alija, C., Piniella, J. F., Alvarez-Larena, A., Tetrahedron: Asymmetry 1992, 3, 1591. (a) Beck, A. K., Bastani, B., Plattner, D. A., Petter, W., Seebach, D., Braunschweiger, H., Gysi, E, LaVeccia, L., Chimia, 1991, 45, 238-244. (b) Seebach, D., Daninden, R., Marti, R. E., Beck, A. K., Plattner, D. A., Kuhnle, E N. M., J. Org. Chem. 1995, 60, 1788-1799. Lei, X., Johnson, R. E, Porco, J. A., Jr., Angew. Chem. Int. Ed. 2003, 42, 2913-3917 and references therein. Gerard, B., Sangji, S., O'Leary, D. J., Porco, J. A., Jr., J. Am. Chem. Soc. 2006, 128, 7754-7755.

512

COLOR PLATE SECTION

A6ex p A6calc

b'

(H a, H a' )

0.7

0.7

3.1

3.3

(H b, H b')

0.8 2.3 1.0 0.4

1.1 3.4 1.1

(H b) (H c)

0.4

(H d)

b

2e

(2e)2 i

7e

(H a)

(7e)2

Please refer to Figure 26 of Chapter 4 in text for figure legend.

External cooling system -70~

Uranium filter (ha9 > 350 nm)

Hanovia 450 W medium pressure mercury lamp Merry go-round

Acetone Cooling bath

Please refer to Figure 10 of Chapter 6 in text for figure legend.

513

COLOR PLATE SECTION

~ J

X-rayof54d

~' .... o

0 ~ Ar~Ar

"~ ~ J

F",

X-rayo 54e

0

0

~

Ar"~~~---~Ar Ar : phenanthren-9-yi Ar Ar

Please refer to Figure 11 of Chapter 6 in text for figure legend.

OMe

o6 ,~'

o H-O ,Tg'c; H.( M e O HO ~

"0 ,,,,"

e

MeO~ OMe Please refer to Figure 12 of Chapter 6 in text for figure legend.

OMe

514

COLOR PLATE SECTION

MeO

OH

0 0

MeO OMe

Electrostatic potential map (bond density) for oxidopyrilium 28

approach of the dipolarophile

Please refer to Figure 13 of Chapter 6 in text for figure legend.

STRATEGIES AND TACTICS IN ORGANIC SYNTHESIS, VOL. 7 9 2008 Elsevier Ltd. All rights reserved.

Chapter 7 SYNTHESIS STUDIES OF DOLABELLANES AND TRANSANNULAR PROCESSES LEADING TO RELATED DITERPENES David R. Williams Department of Chemistry Indiana University Bloomington, IN 47405, USA

I. II. III.

Introduction and Background Biosynthesis Preliminary Studies A. Synthesis of Neodolabellenol B. Dolabellanes via the Julia Condensation IV. Synthesis of a Dolastane via a Dolabellane Precursor A. Transannular Cyclizations of Dolabelladienones V. Fusicoccane Synthesis via a Dolabellane Progenitor: Total Synthesis of (+)-Fusicoauritone VI. Summary Acknowledgments References and Footnotes

243 245 248 248 250 252 255 258 264 265 265

I. Introduction and Background

The dolabellanes are a large family of diterpenoid natural products that are chiefly characterized by a common [9.3.0] cyclotetradecane skeleton. Dolabellanes first gained attention as antibacterial metabolites isolated from the digestive glands of the soft-bodied sea hare, Dolabella californica. ~ Although the germacrenes had previously been well recognized as tenmembered terpenes isolated from higher plants, 2 the initial appearance of the eleven-membered carbocyclic system of the dolabellanes spurred widespread interest. In the years following the initial discoveries, many reports confirmed the distribution of these diterpenes in marine ecosystems, s Dolabellanes have been described as important secondary metabolites of mollusks, brown algae, soft corals, and gorgonians, but this common carbon

244

DAVID R. WILLIAMS

framework has also been identified among substances from fungi, liverwort, and higher plants. Many dolabellanes display significant biological activity such as cytotoxicity and antimicrobial and antifungal activity. Additionally, examples show antiviral, antimalarial, molluscicidal, ichthyotoxic, and phytotoxic effects and may be associated with antipredation as constituents of defensive secretions of marine invertebrates. Representative natural products are shown in Figure 1 and include the dolabelladiene 1, first described in the isolation studies from the sea hare in 1976. ~Metabolites 2, 4 3, 5 and 4 6 illustrate common variations among this class. In 1998, Rodrfguez and coworkers authored a comprehensive review that reported a compilation of known dolabellanes. 7 In recent years, the number of dolabellane natural products has continued to grow steadily. The family also includes the neodolabellanes, which display a transposed bridgehead methyl substituent in a vicinal relationship to the isopropyl group of the bicyclo[9.3.0]tetradecane system. Examples of this variation are illustrated by 4,5-deoxyneodolabelline (5) 8 and [3-neodolabellenol (6) (Figure 2). 9 The isolation of dolabellanes is not infrequently accompanied by the identification of related tricyclic diterpenoids. In the marine environment, dolastanes, and clavularanes, which feature a common 5-7-6 ring skeleton, are often isolated. A recent review has summarized progress in the structure determination and synthesis of members of these classes of natural

Ac_O

H3COH3

H3C

k,/tO H

.

, ~ 7 ca3 ' 'CH3

~

H--~ OH 3

~ . ~ ca3 OH3

1 Dolabelladiene ~ 2 (+)-Acetoxyodontoschismenol 4

CH.

ca3 a " ~ "

;OH3 O

3 Clavulactone 5

FIGURE 1. Representative examples of dolabellanes.

H3C _H HO

.OH3

H3 5 (+)-4,5-Deoxyneodolabelline 8

H

OH 3

6 ~-Neodolabellenol 9

FIGURE 2. Examples of neodolabellanes.

O~.

dc~a~ Ca3 "'OH

4 Stolonidiol 6

o

7

SYNTHESIS STUDIES OF DOLABELLANES AND TRANSANNULAR PROCESSES

~4nAcOcH~

' ~ ~

~' TM

7 Dolatriol

8 Dolastadiene~5

HO . ~ ~

C

9 Cotylenol~6

10Anadensin~7

245

~

H3C ,H H3 ~'"a~ca3 0 , ~ . ",baCH3 H~'L~OCH3 CH3

FIGURE 3. Representative examples of dolastanes and fusicoccanes.

products. ~~Dolastanes demonstrate cytotoxic and antibacterial effects, and are isolated from brown algae and from the sea hare Dolabella, which feeds on this source as well as from the soft coral of Clavularia inflata. Fungi and liverworts produce 5-8-5 tricyclic diterpenes classified as fusicoccanes, which display potent activity associated with phytotoxicity, plant growth, and plant cell morphology. ~ Several metabolites within this group, such as fusicoccin A, 12 are characterized as diterpene glycosides and stimulate potent H+-ATPase activity involving a cell signal transduction mechanism. ~3Representative examples of dolastane and fusicoccane diterpenes are described in Figure 3, and suggest structural similarities with the dolabellanes that have been probed by biosynthesis studies. II.

Biosynthesis

A summary overview of the biogenesis of these natural products in Figure 4 describes the central importance of the dolabellane nucleus as a progenitor of an array of secondary metabolites. For the sake of simplicity, we have illustrated these relationships by diagraming a common enantiomeric series. The biosynthesis from geranylgeranyl pyrophosphate 11 provides two important insights. Initially, an eleven-membered carbocyclization occurs, and evidence of this event is found in the vibsane metabolites 13. ~8Formation of the dolabellane ring system is accomplished via the rt-cation cyclization from 12 to provide 14. The fate of cation 14 yields a variety of dolabellanes, resulting from the following events: (a) the direct capture by water, as shown in 15; (b) an elimination to yield a terminal alkene or a tetrasubstituted exocyclic alkene; and (c) the 1,2-hydrogen shift to produce a corresponding C 14 cation followed by hydration or elimination to give a C10-C14 olefin. Neodolabellanes 16 arise via the backbone migration through consecutive, stereospecific 1,2-shifts from 14. This rearrangement accounts for the vicinal disposition and syn-relationship of the bridgehead methyl and isopropyl substituents in these substances.

246

DAVID R. WILLIAMS

8 H a C ~

H3C ,k);--OH3

4-"~" OPP OH3 11 Geranylgeranyl

H3C 8 ,'N/;--CH3 H30"~/~1 ~ + ~k~ ~1 / l b ~ 3 13 CH3 12 Vibsyl cation

H3C ~-CH 3

HsC/~ ~

.... CH3 CH3 13 Vibsane

skeleton

pyrophosphate

H3C

8

H3C',,,+~CH3 H

-

H 3 C ~

CHa 17 Dolastanes

H

:30~

H3

5r3 /~/,1CH312 CH3 14 Dolabellane

Cation

CH3 16 Neodolabellane system

H3c ~~

H3c H H

'oOH

Hat H3oH3

~[]~2 ~'!~'~"'~4JL'rl3 ...~ \

~CH3 CH3

~ ~ H - CH3

18 Secodolastanes

15 Dolabellanes

19 Fusicoccanes

FIGURE 4. Biosynthesis of dolabellanes and related diterpenes illustrated in a common enantiomeric series.

Transannular reactions of the eleven-membered dolabellane carbocycle produce many novel terpene structures. Since these compounds are often independently characterized, carbon-numbering schemes from isolation and structure determination reflect issues of skeletal nomenclature rather than a common biogenetic origin as depicted in Figure 4. This aspect may lead to some confusion when one attempts to identify a common pathway for these natural products. For example, dolabellanes stemming from 14 are usually numbered beginning with the bridgehead carbon beating the angular methyl as C 1. Thus, the 3,7-dolabelladienes (15) represent the most significant pattern of unsaturation within this family, and acid-catalyzed cyclizations of this motif yield the transfused 5-7-6 tricyclic dolastane systems 17. The loss of a proton from the dolabellane cation 14 can give

7

247

SYNTHESIS STUDIES OF DOLABELLANES AND TRANSANNULAR PROCESSES

rise to the C10-C14 unsaturation as illustrated in 17 (Figure 4). However, any of the options previously outlined for dolabellanes are also available including neodolastanes, which feature migration of the methyl group to the C 10 position (as in the case of the neodolabellanes). Further diversification is in evidence in the secodolastanes 18, which arise via oxidative cleavage of the cyclopentenyl ring with introduction of the bridging acetal resulting from allylic oxidation at C2. Finally, the co-occurrence of fusicoccanes 19 and 3,7-dolabelladienes is indicative of an alternative transannular cyclization event leading to the tricyclic 5-8-5 skeleton. Many fusicoccanes commonly exhibit an endocyclic C 10-C 14 olefin. It is noted that dolastanes, secodolastanes, and neodolabellanes are found among organisms in the marine ecosystem, whereas fusicoccanes are most often associated with fungi and liverworts. On the other hand, neodolastanes as represented by the guanacastepenes ~9 and related trichaurantins 2~ are of fungal origin, whereas the secodolastanes are produced by marine brown algae (Dictyota). 2~,22 Some examples of natural products that are classified as secodolastanes and neodolastanes are illustrated in Figure 5. The assignment of absolute stereochemistry for a newly identified metabolite can often be a perplexing and challenging issue in this natural product arena. Cyclases are undoubtedly involved in species-related stereospecific conversions. However, marine invertebrates present a host of problems stemming from dietary intake via seawater filtration and cohabitation of bacteria and fungi. Terrestrial sources can also present challenging issues as exemplified by recent reports of dolabellanes from higher plants that must be distinguished from the production of dolabellanes and fusicoccanes by symbiotic fungi. Overall, relatively few diterpenes within the dolabellane biosynthesis manifold have proven to be amenable for X-ray diffraction studies. At this time, it is apparent that dolabellanes are produced in either enantiomeric series as d- or/-antipodes based upon a species-specific macrocyclization to 12 (Figure 4). This pathway establishes the absolute configuration of the quaternary carbon at

0

H ~ '0

AcO

0

~'X (~H'~" (3H3

3

0 '

"

" "CH3

O

O -CH3 ~.~,/

0 . O~ HO,, H3C

20 GuanacastepeneA~9 21 Trichoaurantianolide D2~ 22 Linearol2~

23 DichotenolC 22

FIGURE 5. Representative examples of secodolastanes and neodolastanes.

248

DAVID R. WILLIAMS

the ring fusion of the dolabellane framework in 14 as (R)- or (S)-compounds. As investigations uncover new members of the dolabellane family, these studies must evaluate complex interrelationships among species. Studies leading to the assignment of absolute configuration can provide significant stereochemical information embedded in the downstream cyclization products of Figure 4. Ill. Preliminary Studies A. S Y N T H E S I S O F N E O D O L A B E L L E N O L

Our initial synthetic investigations recognized the status of the [9.3.0] cyclotetradecane core as a common platform for the preparation of functionalized dolabellane progenitors for a broad range of natural products. These early studies focused on the development of reliable techniques for direct access to eleven-membered carbocycles. Two pathways for intramolecular alkylation reactions proved highly successful. Ring closure from the cyanohydrin derivative 24 was performed by deprotonation to give an acyl anion equivalent for internal displacement of an allylic chloride, efficiently providing the racemic dolabelladienone 25 after a hydrolytic workup. Formation of the eleven-membered system demonstrated substantial regioselectivity, considering the extended conjugation of the enolate derived from 24 and the possibilities of direct displacement v s . allylic substitution. Further corroboration was obtained by hydride reduction, which produced a separable mixture of 0~-neodolabellenol (26) and 13-neodolabellenol (6) (1.6:1 ratio) for independent comparisons with natural material (Scheme 1). 23

EtO

CH 3 _O. . . ~ " A

cl.. 24

/,,,,

NaN(TMS) 2, THF, 35 ~ then Hydrolysis

c"v-

70%

,3,L

H3C O

'H

H H

OH 3 H3C

25

OH3 OH3 L LiEt3BH, THF,-78 ~

R2 HIH

\~--'

94% (1.6:1)

26 o~-Neodolabellenol R 1 = H, R2 = OH 613-Neodolabellenol R1 = OH, R2 = H

SCHEME 1. Cyclization via intramolecular alkylation of an acyl anion equivalent.

7

249

SYNTHESISSTUDIESOF DOLABELLANESAND TRANSANNULARPROCESSES

nBuLi, CuCN, Et20, -78~ then 65% (8:1" E/Z)

,BuMe2SiO.

. 27

~~028 HaC

CH3

CH 3

H

tBuMe2Si

CH 3 BzO ~ ~ , . , \

o~ ~.o Tol ~ S ~ . [ ~ 30

I'"CH3 ~--"/ OH3

LiN(SiMe3)2, Phil, 0.005 M, 22 ~ 25-35%

1) BzCI, Et3N, CH2CI2, 98% 2) Et3N.HF, CH3CN 22 ~ 92% 3) MsCI, LiCI, DMF collidine, 0 ~ 4) NaSO2Tol, DMF 22 ~ 87%

CH3 i

J

~

22~

Hj

Major: 32

Minor: 33 (epkC6)

1 h, 65% O

H

~,,,,L......~

O/I.[! Tol H

Na(Hg), ~ MeOH J

CH 3

CH3 i

/O, 6 /L:?:::::::~H3 ~

2s

(6:1 ratio; 32/33)

TolO2

H3C

CH3

H3C.

9 1 """

31

,o'

29

2) Dess-Martin Periodinane 0H2CI2, 22 ~ 85%

7

CH3

,

1) LiOH, MeOH 22 ~ 75%

O

HO

CH 3 H9

6

H

6 (13-Neodolabellenol)

TolO2S..---~

H

,-,

34

._,/

SCHEME 2. Synthesisof 13-neodolabellenol(6).

A complementary strategy for ring closure was designed to examine the attributes of an intramolecular Julia olefination process. 23 The efficient transformation of iodide 27 to prepare the aldehydic sulfone 31 was achieved as shown in Scheme 2. A key reaction involved the conversion of 27 into a mixed cuprate species that provided facile SN2' opening of the unsaturated epoxide 28. Our protocol established the E-allylic alcohol of 29 (E/Z ratio 8:1) in a single operation that conveniently led to aldehyde 31. The prospects of utilizing the Julia olefination reaction as a cyclization process suggested a number of potential problems. As a basis for ring closure, the intramolecular addition leading to formation of a new C-C bond at the expense of the aldehyde C = O must also overcome the loss

250

DAVID R. WILLIAMS

of entropy and the introduction of elements of ring strain in the elevenmembered dolabelladiene system. Our calculations indicated that consideration of the stability of the resulting metal-coordinated alkoxide v s . the starting cx-sulfonyl allyl carbanion was a significant factor. Concerns also included issues of kinetic acidity and the competition between formation of an aldehyde enolate v s . deprotonation adjacent to the sulfone moiety. These discussions logically evolved to focus on thermodynamic acidities and the similarities of pK~ estimates for y-deprotonation of an cx,[3-unsaturated aldehyde v s . the allylic sulfone. The situation for 31 (Scheme 2) appeared far more problematic owing to the presence of the nonconjugated dienal. Thus, we were delighted to observe a remarkable reaction providing the diastereomeric 13-hydroxysulfones 32 and 33 (6:1 ratio) in yields ranging from 25 to 35%. Several isomeric hydroazulene byproducts accounted for an additional 25% of the product mixture. Tedious purifications of the latter component afforded 34 (C7 unassigned) as the product of a formal intramolecular Michael addition. However, the possibility of an oxy-Cope rearrangement of the eleven-membered alkoxide of 32 presented a more attractive alternative for 34, which bolstered our enthusiasm for pursuing the Julia methodology. Finally, the purification and characterization of the major diastereomer 32 led to verification of this strategy upon successful reductive desulfonylation to yield pure [3-neodolabellenol (6). B. DOLABELLANES VIA THE JULIA CONDENSATION

Our decision to examine the Julia condensation as a strategy for ring closure of the dolabellane system was inspired by an earlier report of Yoshii and coworkers in their studies of tetronolide (see example 1; Figure 6). 24 However, the conditions for low-temperature kinetic deprotonations pursued in our early efforts produced poor yields of cyclized products and destruction of the starting aldehydic sulfone. We quickly learned that solvents such as tetrahydrofuran (THF) and diethyl ether, in addition to polar additives, such as hexamethylphosphoric triamide (HMPA), led to the formation of many products. Alternatively, the inverse addition of substrate into benzene solutions of various soluble bases resulted in a facile ring closure. Quenching reactions immediately with glacial acetic acid also improved product yields, and offered a general pathway for the efficient construction of functionalized carbocycles as summarized by the examples of Figure 6. From our limited survey 25 (Figure 6), we gained important insights impacting future efforts. It was noteworthy that saturated sulfones (examples 3 and 4) performed well in these intramolecular cyclizations.

MOMO

L.... ~ .... ,,,~

M~MO~"~h/CH (OMe)2 PhO S MeO'~"~~"''O

#Jl

NaOtAm, Phil,

0

18 ~

H3

25 rain, 100%

Reference 24

H3C-

~ / ' H V ",CH3 OMOM T o l O 2 S ~ O

H

C

121 _ "'OH3 OMOM

I

H3 OH3 O H C ~

H3C H OR ""

LiN(TMS)2 benzene +10 ~

,,,H

T~ OH3 \

OH3

Major Product

,0

T~

""">'~~Major Producf ~

H3C H OR HO ~ . . .' . ~ ."" ..._.~

,. quench HOAc TolO2S'". . . . (R = SitBuPh2) (71%) Only Product H3C , R .-., I H ' H3CI1.-"~ H ~1-13

H

TolO2

~

>

H

50%

""

O

Na+-O < benzene; 35 ~

T ~

OH

H

,ere:.,73"

CH3

H3C RO H

H O " ' ~

Na+-O < - benzene; 35~

To,o s .

OMOM

T ~

THF at -78 ~ 5O% [dr 1:1]

~

H

CH3

LiN(TMS)2

OM

P H3 H-3C OHC ' " ' ~ "

O

CH(OMe)2

H

LiN(TMS)2 benzene +10~ quench HOAc (R = SitBuPh2) (92%)

HO

H3C RO H . ~

TolO2S'" Only Product

FIGURE 6. Macrocyclizations via Julia condensations.

252

DAVID R. WILLIAMS

Hence, deprotonation produced an ~-sulfonyl carbanion in spite of the opportunities for enolate formation and the presence of a base-sensitive oxirane, which was known to readily isomerize to an allylic alcohol (example 4). The reisolation of small amounts of starting aldehydic sulfones in examples 5 and 6 gave no evidence of 7-deprotonation or epimerization of the bis-allylic methine although these cases may have been favorably biased by the presence of the bulky silyl ether. The resubmission of pure cis-and trans-~-hydroxysulfone diastereomers of example 3 to our reaction conditions produced small amounts of all four possible [3-hydroxysulfone isomers as well as traces of starting aldehyde. Unfortunately, considerable decomposition was observed in these experiments. Our findings suggest that the optimization of Julia condensations requires careful study. Better yields and improved diastereoselection are often observed with soluble bases in nonpolar solvent (benzene). We speculate that coordination of bases with the sulfonyl functionality provides for a complex-induced proximity effect (CIPE) in these deprotonations as detailed by Snieckus and Beak. 26 This is also supported by the fact that potassium bases have proven to be generally ineffective. This aspect of coordination could lead to highly organized crown conformations in transition states as illustrated for examples 3 and 5, and may account for the modest preference of diastereoselectivity. Our observations and such considerations generally argue in favor of a kinetic process. IV. Synthesis of a Dolastane via a Dolabellane Precursor

We recognized that transannular cyclizations of suitably functionalized dolabellanes would provide an efficient strategy for the synthesis of the tricyclic 5-7-6 dolastanes. Our rationale suggested that these processes required a trigger to initiate the stereocontrolled event of C-C bond formation and a conformational bias of the eleven-membered ring system. Fenical and coworkers had reported a significant precedent in their characterization studies of natural dolabellanes. For example, the epoxide 35 led to a regioselective acid-catalyzed cyclization yielding the isomeric dolastane alcohols 36 (17%) and 37 (20%). 27 Additionally, oxirane 38 produced a stereoselective transannular cyclization via initial deprotonation to the E(O) enolate to give the 5-8-5 system of 39 (58%) (Figure 7). 28 In either situation, a pronounced conformational bias of the elevenmembered carbocycle and epoxide opening with inversion led to a predictable outcome. We explored these ideas in the context of the design of a nonracemic total synthesis of dolastanes. 29 Investigations focused

7

253

SYNTHESISSTUDIES OF DOLABELLANES AND TRANSANNULAR PROCESSES

H.~

,,

H3C~

H3C

,, ,-,

H3C

BF3,Et20/ether

),,

0~

+

15 min HO CH3

H3C 35

H3C

HO CH3

36

H3C

H3C 37

nOH3

O nl-.I " '3

LDA/THF ,~ CHa

HaC

~

@

~

-76 ~ 1h "H" (58%) H3CHH"~CH3 39

38

0

H3C

FIGURE 7. Transannular cyclizations of dolabellane epoxides.

3~ H31 s H

"

12 "

HO' 40 (biogenetic numbering scheme)

CH3 11

H3C

_h CH3

CSA CH2CI2 -78 ~ -- r.t. (77%) ratio 3:1

41 (dolastane numbering scheme)

42

SCHEME 3. Stereocontrolled formation of dolastanes.

on the preparation of the dolabellane epoxide progenitor 40, which led to the re-cation cyclization (Scheme 3) with anhydrous camphorsulfonic acid at - 7 8 ~ yielding 41 and 42 as a 3:1 mixture of exo- and endocyclic olefins (77%). Subsequent separation and purification of the corresponding monobenzoates and LiOH saponification gave (-)-3~, 4[3-dihydroxyclavulara- 1(15), 17-diene (41). The initial stages of our synthesis of the dolabellane epoxide 40 addressed the contiguous chirality of the cyclopentane nucleus (Scheme 4). Hydroxide-induced ring opening of (+)-9,10-dibromocamphor (43) gave carboxylic acid 4430 for convenient conversion to cyclopentanone 45. Saegusa oxidation 31 of 45 resulted in smooth conversion to the corresponding enone for stereocontrolled 1,4-conjugate addition of the isopropenyl unit yielding ketone 46. The incorporation of the asymmetric quaternary carbon at C 11 (structure 40) from a chiral pool precursor was viewed as a distinct advantage of this approach. Unfortunately, the terminal disubstituted alkene of 46 proved to be too reactive under conditions that explored the reductive removal of the C 12 carbonyl group. This dilemma forced us

254

DAVID R. WILLIAMS

gr

H3C',~/__Br

HO~

2) Ph2tBuSiCI (2.2 eq.) Et3N; AgNO3 DMF ~

I'"H HOOC 44

3) LiAIH4 4) MEMCI i-Pr2NEt CH2CI2 (87%)

1) excess NaOH

~

DMSO; H20 (94%)

43

H3C O

5) NaN(TMS)2

.o~

THF; TMS-CI

then Pd(OAc)2 CH3CN(89%) 6) H3CC(Br)=CH2 nBuLi; THF CuBroDMS -78~ (92%)

7) NBS; THF (88%)

. ~H3C J~ RO" , , , , , ~ ) MEMO"

45 (R = SitBuPh2)

Ph2tBuSiO

H3C O -~ "H

'H

.E~O~ ~ Y -

Hr

-78oc (88%)

H3C Ph2tBuSiO ~H/-

9) Imid2C=S

"o"OH3

CI~ C I DMAP 22 o ---+ 180 ~ (85%)

49

48

H3C_

10) TsNHNH2 NaOAc

several steps

Ph2tBuSiO

THF; H20 75 ~ (99%) 11) Zn; EtOH (92%)

/Br

OH3 47

46 (R = SitBuPh2)

8) DIBAL

'IH

OH3

c , ~CH3

.oJ~~"H

R/O

50

OH3

51 R = CH(CH3)OC2H5

a) NaN(TMS)2

b) DIBAL

THF; +35~

CH2CI2;-78 ~

then H3O+ then aqu. NaOH (85%) ~

k

c

a

a3

(94%)

O

52 i

/

c) DEAD; PPh3 THF PhCOOH

CH3 CH

H

r

H

53

then DIBAL (65%) d) VO(acac)2 tBuOOH (88%)

HO" r,,~; T ~

SCHEME 4. Synthesis of dolabellane precursor 40.

40

.

7

SYNTHESIS STUDIES OF DOLABELLANES AND TRANSANNULAR PROCESSES

255

to consider masking the C =C in a protected form. To accomplish the task, we carried out an internal haloetherification using N-bromosuccinimide, which efficiently provided tetrahydropyran 47 via direct nucleophilic participation of the [3-methoxymethylene (MEM) ether and in situ dealkylation of the resulting oxonium species. In this fashion, we maintained the status of the protection of the primary alcohol, and we effectively replaced the conflicting double bond with an unreactive neopentyl bromide. The conversion of 47 to the cyclopentane core 50 was achieved in four steps. Diisobutylaluminum hydride (DIBAL) reduction to the [3-alcohol 48 and subsequent acylation with thiocarbonyldiimidazole led to a facile syn elimination producing alkene 49. Reduction of 49 with diimide as generated in situ from p-toluenesulfonohydrazide 32 and deprotection with zinc in refluxing ethanol gave intermediate 50. Our strategy for ring closure of the eleven-membered dolabellane was devised to utilize an intramolecular alkylation of the enolate from the cyanohydrin derivative 51. In a series of steps, each side chain was extended to incorporate the E-olefins of 51, and deprotonation with NaN(TMS)2 followed by hydrolysis gave the unsaturated enone 52 (85%). DIBAL reduction selectively produced the axial alcohol 53 (dr 10:1), which was anticipated based upon considerations of a rigid crown conformation of the 3(E),7(E)-dolabelladienone. However, this stereochemistry was opposite to that which was required at C3 of the dolastane natural product 41. Fortunately, the Mitsunobu inversion procedure successfully gave the corresponding equatorial alcohol, which guided the vanadium-catalyzed Sharpless epoxidation to 40. While the crown conformation of the dolabellane system guaranteed facial selectivity in the oxidation, the internally-directed characteristics imposed by vanadyl ester formation in this reaction precluded competing epoxidations at other C=C sites. Upon the communication of our efforts, this report featured the first total synthesis of a chiral, nonracemic dolabellane as well as the biomimetic conversion to the tricyclic dolastanes 41 and 42. A. T R A N S A N N U L A R C Y C L I Z A T I O N S OF D O L A B E L L A D I E N O N E S

From the outset of these studies, we envisioned a stereoselective pathway for isomerization of the nonconjugated dolabelladienone 52 to the dolastane framework via a conformationally controlled transannular olefin-enone cyclization. 33 In the event, the treatment of 52 with BF3-etherate produced diastereomeric enones 54 and 55 in a ratio of 6:1 (Figure 8). These ringcontracted products provided evidence of a highly specific reorganization

256

DAVID R. WILLIAMS

H3C

H3C C •

?H/~3

CH3

H3C Cl-i CH3

BF3.OEt [3 --,- 8] closure

H3C

55

54

52

H

,, ,. H

HILI

F3B-O ~'~ v H H OH3

OH3

52a (Chair-Chair) Steric Energy = 31.53 Kcal/Mole

56

+ .. H

I

3

~

---~" CH 3 O 54 (Major)

H

H 3 < ~ Z ~ t

/~_F3B" -U

H~['-(~a3 a CH 3

_~

H~d" 6H 3 CH3

52b (Boat-Chair)

55

(Minor)

Steric Energy = 32.40 Kcal/Mole Conditions: BF3.OEt2, CH2CI2, 0~ 22~ 18 h, 83%, 6:1 ratio FIGURE 8. Transannular cyclizations of dolabelladienone 52.

along the carbon backbone. Our global molecular mechanics extended (GMMX) calculations of 52 had indicated two low-energy conformers 52a/b, in which the chair-chair (crown) conformer 52a was found to be slightly more stable (by 0.9 kcal/mol). Thus, the rt-cation cyclizations (C3 --, C8) yielding boron enolates 56 and 57 were followed by successive suprafacial 1,2-hydrogen and 1,2-methyl migrations leading to a tertiary bridgehead cation for collapse to the ~,[3-unsaturated system. As a secondary process, the isopro-penyl substituent of 52 was isomerized to the tetrasubstituted, endocyclic double bond in 54 and 55. In the course of these investigations, we also examined the reactivity of the isomeric enone system presented in the neodolabelladienone 58 (Figure 9). Treatment of 58 with BF 3 etherate led to the rapid formation of two crystalline products in equal amounts. X-ray diffraction analyses provided confirmation of a transannular event leading to the 5-8-5 fusicoccane 59 and the unusual spirocyclic diterpene 60. A highly stereocontrolled reaction was postulated based on the consideration of chair-chair (crown) conformation 58a of the starting dolabellane. Stereoselective formation of the 5-8-5 cyclization product 59 was rationalized by initial formation of the tricyclic boron enolate 61. Two successive 1,2-hydride

7

SYNTHESISSTUDIES OF DOLABELLANES AND TRANSANNULAR PROCESSES

O H3C

?H..3

CH3

0

BF3.OEt2'

O.~.. ~ ~J.~...~.C . ..----..Y/ H.~ "h

CH2CI2, 0 ~ 60%, (1"1)

257

H

H"

3

C

~

H3C

58

6O BF3*OEt2

H3C

CH3 CH3 /

oH

58a Chair-Chair

61

1,2 HydrideShift H3C~ H 3 C ~ ,

0 ~ \

H

H

1,2 Hydride Shift

*OH3

59 CH

CH3 / .A~ 1,2 Carbon Shift

62

J

0

H3C

60 FIGURE 9. Acid-catalyzed rearrangement of dolabelladienone 58.

shifts with rearrangement of 61 to 62 was followed by regeneration of the conjugated enone system providing 59 as a single stereoisomer. Alternatively, formation of the spirocyclic product 60 occurred via a 1,2carbon shift from the common carbocation 62. Results gathered from these efforts demonstrated that the dolabelladienones were viable synthetic intermediates for the construction of more highly organized polycyclic diterpenes with formation of a new C-C bond. Aspects of stereochemistry for the transannular processes were affirmed and, in fact, may be manipulated, with a clear understanding of the conformational bias of the eleven-membered dolabellane carbocycle. Thus, the themes of this exercise suggested a general synthesis strategy focused on the formation of medium and large tings as an expedient pathway toward stereoselective preparation of functionalized, ring-fused systems.

258 V.

DAVID R. WILLIAMS

Fusicoccane Synthesis v/a a Dolabellane Progenitor: Total Synthesis of ( + ) - F u s i c o a u r i t o n e

The design of a concise synthesis route leading to dolabellanes and fusicoccanes was a significant goal for these studies. It was, in principle, intriguing to consider the preparation of an eleven-membered carbocycle as a key intermediate for the construction of a target molecule displaying an eight-membered ring. While the fusicoccanes have been discussed in this chapter as representative natural products featuring the fused 5-8-5 carbocyclic skeleton, fusicoccanes, 34 cotylenins, 35 and fusicoplagins 36 are diterpenoid examples, whereas ophiobolins 37 and ceroplastols 38 are sesterterpenes. These substances stimulated a number of notable synthesis efforts. Kishi 39 and Boeckman 4~ independently described syntheses of ophiobolin C and (_+)-ceroplastol I, respectively, and Paquette 41 subsequently reported a route culminating in the synthesis of (+)-ceroplastol I. Distinctive pathways for the synthesis of epoxydictymene were also described. 42 In addition, Kato, Takeshita and coworkers published a significant body of work culminating in the total synthesis of (-)-cotylenol. 43 Their studies developed a convergent approach featuring the preparation of two cyclopentane building blocks that were united prior to the latestage closure of the central eight-membered ring. Our strategy would be distinguished from the prior art by the identification of a dolabellane that would trigger selective formation of the fusicoccane 5-8-5 system. Fusicoauritone ( 6 3 ) 44 w a s chosen as the target for these efforts, and two dolabelladienones 64 and 65 were envisioned as key retrosynthetic precursors (Figure 10). The formulation of a Nazarov reaction of the cross-conjugated enone 65 was considered especially attractive since it dictated a ring-contraction event leading to the ring-fused cyclopentenone of 63. The promise of incorporating the desired stereochemistry at C7, and maintaining this element of chirality throughout the acid-catalyzed Nazarov cyclization process was another advantage of this scenario. The alternative exercise suggested a Lewis acid promoted cyclization of 64, which would

64

Fusicoau ritone (63)

65

FIGURE 10. Dolabellane precursors toward the synthesis of fusicoauritone (63).

7

259

SYNTHESIS STUDIES OF DOLABELLANES AND TRANSANNULAR PROCESSES

rely on a related sequence of stereocontrolled backbone migrations as illustrated in the previous section. To investigate a Nazarov application for the synthesis of 63, a route for the preparation of a functionalized cyclopentane was devised to incorporate three contiguous stereogenic sites corresponding to C10, C11, and C 14 of fusicoauritone. From this nucleus, we planned to extend the carbon chains with appropriate functionality for a ring-closing alkylation. This task was undertaken, as summarized in Scheme 5, beginning with the

O0.~1 H

) (EtO)3CCH3 H3CCH2COOH(cat.)

Et..O O'~'~

at 145oC

2) LiAIH4, Et20 at 0 ~

CH3

then MEM-CI iPr2NEt, DMAP CH2CI2 ~ OH3 ROt

"~"~

66

67

68 R - H (79%)

= 69 R = MEM (92%)

3)

BH3,THF at 0~ quench H202; -OH

H3C HO

RO

4) (COCl)2 DMSO; CH2CI2 at -78 ~ then Et3N at-50 ~

ROJ CH3

5) Li

MEMOJ

7(} R = H (75%) 71 R = C(O)C6H5

CH3

Et20 at-50 ~ H3C

7) Na~ HMPA t-BuOH (97%) HO"

,"1

MEMOJ

CH3

73 (70%)

H3c ToIO2S-v.P

,,,'~-./

MEMO...J OH3 76 (80% overall from 73)

72 (65%)

~I

6) (EtO)3CCH3 H3CCH2COOH(cat.) at 145 ~ then LiBH4, MeOH

3

10) HBF4 H20, MeOH (85%) 11) (COCl)2, DMSO CH2CI2at -78 ~ then Et3N 12) COOEt

8) TsCI, Et3N, DMAP CH2CI2 (89%) 9) Nal, ethylmethylketone at reflux then NaO3SC6H4CH3 DMF at 70 ~ (93%)

.30 ToIO2S.v.-J

CH3

COOEt

Ph3P=~cH 3

at 22 ~ (96%)

...,~--../

HaG.~

77 (82% overall from 76)

SCHEME 5. Preparation of the dolabellane precursor 77.

260

DAVID R. WILLIAMS

known optically active cyclopentenyl alcohol 66. 45 This compound served as an excellent substrate for the Johnson ortho-ester Claisen rearrangement. 46 The sigmatropic process occurred with high stereoselectivity via 67, avoiding nonbonded interactions with the isopropyl group, and the subsequent hydride reduction of the resulting ethyl ester and hydroxyl protection gave 69 (MEM = CH2OCH2CH2OCH3). Mehta and Krishnamurthy 47 had reported comparable results of a related Claisen rearrangement supporting our assignment of stereochemistry. Based on this information, we proceeded with the hydroboration of the exocyclic olefin, and the oxidative quench of the resulting alkylborane produced a single diastereomer 70. The impressive diastereofacial selectivity of this reaction was somewhat surprising, and led us to consider the possibility of coordination and internal delivery of the reagent via the MEM ether. However, replacement of the MEM ether in 69 with the corresponding tert-butyldimethylsilyl (TBS) ether gave a similar outcome suggesting a significant steric bias for this hydroboration. The relative stereochemistry of 70 was confirmed by single crystal X-ray diffraction of the benzoate 71, thereby establishing an efficient three-step operation for production of the cyclopentane moiety. Efforts to elongate the alkyl chain with incorporation of the remote stereochemistry at C7 were rewarded by further studies using the Claisen rearrangement. Toward this goal, (Z)-propenyllithium was prepared according to the Whitesides procedure, as and nucleophilic addition at -50 ~ with freshly prepared aldehyde from 70 yielded a mixture of (Z)-allylic alcohols (ratio 6:1; 82% yield). Upon chromatographic separation, the major product, later identified as 72, was obtained in 65 % yield. Efforts to assign the stereochemistry of the allylic alcohol gave ambiguous results. Attempts to prepare derivatives of 72 led to allylic rearrangements and elimination to the terminal diene, whereas Mitsunobu reactions failed entirely. On the other hand, each diastereomeric alcohol was independently subjected to the conditions of the ortho-ester Claisen rearrangement providing a pure ester as anticipated via a concerted transposition of stereochemistry. Hydride reductions gave a pair of primary alcohols as illustrated for the conversion of 72 to 73. These products were subjected to ozonolysis followed by a reductive sodium borohydride quench. Samples of nonracemic 2-methyl-1,4-butanediol were isolated and compared with optical rotation data for the known nonracemic substances. In this manner, (R)-2methyl-l,4-butanediol was obtained from the oxidative cleavage of the (E)-alkene 73 establishing the stereochemistry of the starting alcohol 72. The reduction of the hindered olefin in 73 proved to be problematic. While a variety of hydrogenation catalysts led to multiple products,

7

SYNTHESIS STUDIES OF DOLABELLANES AND TRANSANNULAR PROCESSES

H2 (1 atm)

s,

EtOOC

,,,,~/

MEMO/J OH3 74

Rh-AI203 (90%)

261

E,oo o MEMOJ OH3 75

SCHEME 6. Rhodium-catalyzed alkene migration.

rhodium on alumina efficiently transformed 74 into the isomeric tetrasubstituted (C10-C14) alkene 75 (Scheme 6). After some experimentation, the reduction of 73 via a dissolving metal procedure using sodium in a concentrated solution of HMPA and tert-butyl alcohol 49 successfully provided the saturated alcohol in high yield (Scheme 5). This reaction was convenient and amenable to large-scale preparation, which greatly aided our efforts for the straightforward conversion to the sulfone 76. The final tasks for functionalization leading to a cyclization precursor were met by hydrolytic deprotection of the acetal, Swern oxidation and direct Wittig olefination with a stabilized ylide yielding the (E)-0t,[3-unsaturated ethyl ester 77. Carbocyclization of the eleven-membered [9.3.0]cyclotetradecane system utilized modifications of the Julia condensation. Thus, the ester 77 (Scheme 5) was readily transformed into the aldehydic sulfone 78 (Scheme 7) for rapid addition into a vigorously stirred solution of sodium tert-amylate (0.15 M in benzene at 35 ~ Immediate quenching with glacial acetic acid provided the 13-hydroxysulfones 79 (dr 6:1) in yields ranging from 73 to 82% with the recovery of small amounts of starting 78 (12%). The major product proved to be the expected trans-f3-hydroxysulfone, distinguished by the large vicinal coupling constant (JAB = 10 Hz) as observed in the proton NMR spectrum indicating a diaxial disposition of methine hydrogens. Each sulfone diastereomer underwent desulfonylation with 6% Na-Hg in methanol to yield identical samples of the allylic alcohol 80 (X = H). Transformation of 79 to an appropriate divinylketone was initiated by Swern oxidation to produce diastereomeric at-sulfonyl ketones (5:1 ratio). Further characterization of this system was carried out by oxidation of the resulting enolate with the Davis oxaziridine 5~ yielding the 0t-diketone 81 (85%). This reaction was found to be an especially efficient process for

262

DAVID R. WILLIAMS

H3C ToI02S-,,.~

.,,~--/

H3C,~

H3C

Na + -O - ~

(COCI)2; DMSO CH2CI2at -78 ~

benzene at 35~ (2 min) then quench

CH3

HOAc (73%)

HO"

CHO 78

then Et3N then: add K§-OtBu

'CH3

THF at -78 ~ O ,/--'N p-CIP6" ~ (85%)

79 (X = SO2Tol) 80 (X = H)

cI~CI reflux

~J

.~c~

BF, Et20

HcH~Hy 3 : -

.

H

P~

H

CH3 82

81

H 3 C ~ O

~

HO

'CH3

83 (Ha5 2.92)

,. (77%)

H

O

~

o

84

[JAB- 11.5 Hz; Jgc = 5.5 Hz] S C H E M E 7. Dolabelladienone formation and Nazarov cyclization to 84.

the preparation of unsymmetrically substituted 1,2-diketones. Enolic tautomerization of 81 was not evident in the ~H NMR spectrum, however, we reasoned that Lewis acids would promote formation of the putative crossconjugated enone intermediate 82 for the Nazarov reaction. Upon brief treatment with B F 3 etherate, a facile cyclization produced enone 84 in varying amounts along with the isomeric 83. Ketone 83 was subsequently recognized as the first-formed Nazarov product with slow acid-catalyzed isomerization to the more stable cyclopentenone 84. Characterization of 84 with extensive use of NMR data confirmed the all syn stereochemistry of H A, H B, and H c in addition to the relationship to the bridgehead (C11) methyl group and the newly created ~-methyl ketone using NOE difference experiments. Our studies noted that irradiation of H B in 84 (6 = 3.00) produced a 6% NOE enhancement of H A (~ = 2.61) without affecting the bridgehead methyl substituent. In addition, the irradiation of H A gave an 8% NOE for H B (JAB = 6.0 Hz) and induced a negative 3% NOE of the axial H c (6 = 2.85).

7

SYNTHESIS STUDIES OF DOLABELLANES AND TRANSANNULAR PROCESSES

.30

R2."

1) NaHMDS THF at-78 ~ PhSeCI

-

then H202

H

_H

B

-

T~

263

2) DIBAL -78~ (95%) 3) sodium naphthalide THF at -78 ~ then MnO2 (70%)

(98%) L.) 85 R1= SO2Tol; R2= H

CH3 87

86 R1/R2= SePh; SO2Tol H

_H

B

HA L)

OH3

4) cat. TsOH CI~cI

H3C. '_-'/---~:' J ~

i.

reflux (92%) 5) t-BuOCI H20; acetone (40%)

~ _ ~ 0H3 ,.., CH3

DIBAL -78 ~

89 (X = H) 93 (X = OOH)

88

63 (X = OH) (Fusicoauritone)

H,ca,IH3 / - ~ cH3 HO

"CH3 90

0H2012

(98%) L

CH3

CH3

91

92

SCHEME 8. Synthesis of fusicoauritone (63).

These results provided encouragement for further efforts toward Nazarov studies of related dolabelladienones. In particular, we sought an outcome that would produce the enone regiochemistry as exhibited in fusicoauritone (63). Thus, the oxygen substituent of 82 (Scheme 7), originating from enolization of the carbonyl in 81, accelerated the Nazarov reaction, but it also dictated the stability of the resulting enone 84. Based on this rationale, we explored opportunities toward the less substituted divinylketone 88 (Scheme 8). Phenylselenation of ~-sulfonylketone 85 gave 86 for immediate oxidative syn-elimination exclusively producing the (E)-~,l~-unsaturated sulfone 87. The double bond geometry of 87 was established by NMR studies involving irradiation of aromatic hydrogens of the sulfone inducing an NOE of the vinylic hydrogen H B. A three-step sequence effected reductive removal of the electron-withdrawing sulfonyl substituent beginning with hydride reduction of the ketone 87, which provided the [3-alcohol. 5~ Desulfonylation with sodium naphthalide at-78 ~ in THF selectively occurred with retention of olefin geometry and a mild

264

DAVID R. WILLIAMS

allylic oxidation gave the (Z)-enone 88 as characterized by the vicinal coupling (JAB = 10.7 Hz) in the proton NMR spectrum. Finally, an efficient Nazarov reaction of 88 under protic or Lewis acid conditions resulted in formation of 89 and its C6 epimer. Flash column chromatography on silica gel led to the isolation of small amounts of pure 89 for characterization. However, these fusicoccane diastereomers readily isomerized under mildly basic or acid conditions. A number of experiments were attempted to investigate the general reactivity of this 5-8-5 tricyclic system. For example, DIBAL reduction of 89 gave a single diastereomer, which was assumed to be the ~-alcohol 90. We planned for a mild elimination reaction to yield the cyclopentadienyl system of 91 as a prelude to cycloaddition of singlet oxygen and base-induced fragmentation of the resulting endoperoxide to yield 63. However, the allylic alcohol 90 rapidly produced the stable diene 92 in the presence of mild protic acids. Upon standing for several days, chloroform solutions of 89 yielded large colorless crystals of a new substance. Further studies unambiguously identified this material by single crystal X-ray diffraction as the C6 hydroperoxide 93, and subsequent reduction with sodium hydrogen sulfite gave 63. The ease of C6 oxidation suggested a number of more efficient procedures, which led us to explore reactions of 89 and its C6 epimer with tert-butylhypochlorite in aqueous acetone. In the event, this direct oxidation provided a 40% yield of synthetic fusicoauritone (63), which proved to be identical in all respects with the natural fusicoccane metabolite. VI.

Summary

Dolabellanes represent a large class of diterpenes that are characterized by a [9.3.0] cyclotetradecane skeleton. Biosynthetic studies have shown that dolabelladienes serve as precursors for enzymatic conversions to a variety of diterpenoid natural products including neodolabellanes, dolastanes, and fusicoccanes. Recognizing the central importance of the dolabellane system as a progenitor of important classes of naturally occurring metabolites, our studies have explored intramolecular alkylation reactions to provide general approaches for the formation of functionalized elevenmembered carbocycles. Internal alkylations using enolates of cyanohydrin derivatives as acyl anion equivalents have produced nonconjugated dolabelladienones, and intramolecular Julia condensations of aldehydic sulfones have shown surprising versatility for the production of several dolabellane systems. Stereocontrolled, transannular reactions of dolabelladienes have been examined via the ~-cation cyclization, which is induced via ring

7

SYNTHESISSTUDIES OF DOLABELLANESAND TRANSANNULARPROCESSES

265

opening of an intermediate epoxide. These efforts have demonstrated the stereoselective formation of a dolastane natural product from its putative dolabellane precursor. Additional studies of nonconjugated dolabelladienones have documented Lewis acid-catalyzed transannular processes leading to 5-7-6 tricyclic enones as well as 5-8-5 tricyclic fusicoccanes. Finally, our investigations have utilized cross-conjugated dolabelladienones as key substrates for Nazarov cyclizations to produce the 5-8-5 fusicoccane skeleton, which has led to a total synthesis of (+)-fusicoauritone. Beginning with our early efforts, these studies were devised to specifically examine the regio- and stereocontrolled preparation of complex tricyclic diterpenes via the initial formation of appropriate mediumsized carbocycles. In this manner, our studies have highlighted the vibrant and fascinating chemistry of dolabellane natural products, and perhaps by undertaking this task, we have identified a departure point for a grand journey of new insights and future opportunities. Acknowledgments The studies described in this chapter are due to the efforts of a talented group of dedicated coworkers. I gratefully acknowledge the devoted efforts of energy and intellect by Dr. Leslie A. Robinson, Dr. C. Richard Nevill, Dr. Paul J. Coleman, and Dr. Jayachandra P. Reddy in these investigations. We thank the National Institutes of Health (GM42897) for financial support. References and Footnotes 1. (a) Ireland, C., Faulkner, D. J., Finer, J., Clardy, J., J. Am. Chem. Soc. 1976, 98, 4664. (b) Ireland, C., Faulkner, D. J., J. Org. Chem. 1977, 42, 3157. 2. For the synthesis of germacrane sesquiterpenes, see Minnaard, A. J., Wijnberg, J. B. E A., de Groot, A., Tetrahedron 1999, 55, 2115. 3. For some recent isolation reports, see (a) Cai, X. H., Luo, X. D., Zhou, J., Hao, X. J., Helv. Chim. Acta 2005, 88, 2938. (b) Morikawa, T., Xu, E, Kashima, Y., Matsuda, H., Ninomiya, K., Yoshikawa, M., Org. Lett. 2004, 6, 869. (c) Morikawa, T., Xu, F., Ninomiya, K., Matsuda, H., Yoshikawa, M., Chem. Pharm. Bull. 2004, 52, 494. (d) Cirne-Santos, C. C., Teixeira, V. L., Castello-Branco, L. R., Frugulhetti, I. C. E E, Bou-Habib, D. C., Planta Med. 2006, 72, 295. 4. Matsuo, A., Uohama, K., Hayashi, S., Connolly, J. D., Chem. Lett. 1984, 599. 5. Su, J., Zhong, Y., Zeng, L., J. Nat. Prod. 1991, 54, 380. 6. Mori, K., Iguchi, K., Yamada, N., Yamada, Y., Inouye, Y., Tetrahedron Lett. 1987, 28, 5673. 7. Rodriguez, A. D., Gonz~lez, E., Ramirez, C., Tetrahedron 1998, 54, 11683. 8. (a) Bowden, B. E, Coll, J. C., Mitchell, S. J., Stokie, G. J., Blount, J. E, Aust. J. Chem. 1978, 31, 2039. (b) Nakanishi, K., Dillon, J., J. Am. Chem. Soc. 1971, 93, 4058.

266

DAVID R. WILLIAMS

9. (a) Bowden, B. E, Braekman, J. C., Coil, J. C., Mitchell, S. J., Aust. J. Chem. 1980, 33, 927. (b) Kobayashi, M., Son, B. W., Kyogoku, Y., Kitagawa, I., Chem. Pharm. Bull. 1986, 34, 2306. 10. Hiersemann, M., Helmboldt, H., Top. Curr. Chem. 2005, 243, 73. 11. (a) For a review, see Muromtsev, G. S., Voblikova, V. D., Kobrina, N. S., Koreneva, V. M., Krasnopolskaya, L. M., Sadovskaya, V. L., J. Plant Growth Regul. 1994, 13, 39. (b) DeBoer, B., Trends Plant Sci. 1997, 2, 60. 12. Barrow, K. D., Barton, D. H. R., Chain, E. B., Ohnsorge, U. E W., Thomas, R., Chem. Commun. 1968, 1198. 13. Asahi, K., Honma, Y., Hazeki, K., Sassa, T., Kubohara, Y., Sakurai, A., Takahashi, N., Biochem. Biophys. Res. Commun. 1997, 238, 758. 14. Pettit, G. R., Ode, R. H., Herald, C. L., Von Dreele, R. B., Michel, C., J. Am. Chem. Soc. 1976, 98, 4677. 15. Konig, G. M., Wright, A. D., Fronczek, E R., J. Nat. Prod. 1994, 57, 1529. 16. Sassa, T. A., Takahama, A., Shindo, T., Agric. Biol. Chem. 1975, 39, 1729. 17. Huneck, S., Baxter, G., Cameron, A. E, Connolly, J. D., Rycroft, D. S., Tetrahedron Lett. 1983, 24, 3787. 18. (a) Fukuyama, Y., Minami, H., Takaoka, S., Kodama, M., Kawazu, K., Nemoto, H., Tetrahedron Lett. 1997, 38, 1435. (b) Kubo, M., Chen, I. S., Fukuyama, Y., Chem. Pharm. Bull. 2001, 49, 242. (c) E1-Gamal, A. A. H., Wang, S. K., Duh, C. Y., J. Nat. Prod. 2004, 67, 333. 19. Brady, S. F., Singh, M. P., Janso, J. E., Clardy, J., J. Am. Chem. Soc. 2000, 122, 2116. 20. (a) Knops, L., Nieger, M., Steffan, B., Steglich, W., Liebigs Ann. 1995, 77. (b) Invernizzi, A. G., Vidari, G., Vita-Finzi, P., Tetrahedron Lett. 1995, 36, 1905. (c) Benevelli, E, Carugo, O., Invernizzi, A. G., Vidari, G., Tetrahedron Lett. 1995, 36, 3035. 21. (a) Ochi, M., Miura, I., Tokoroyama, T., Chem. Commun. 1981, 100. (b) Teixeira, V. L., Tomassini, T., Fleury, B. G., Kelecom, A., J. Nat. Prod. 1986, 46, 570. 22. Ali, M. S., Pervez, M. K., Ahmed, E, Saleem, M., Nat. Prod. Res., 2004, 18, 543. 23. Williams, D. R., Coleman, P. J., Tetrahedron Lett. 1995, 36, 35. 24. (a) Yoshii, E., Takeda, K., Urahata, M., Takayanagi, H., Ogura, H., J. Org. Chem. 1986, 51, 4735. (b) Yoshii, E., Nakamura, H., Kawanishi, E., Takeda, K., Tetrahedron Lett. 1991, 32, 4925. 25. Williams, D. R., Coleman, P. J., Nevill, C. R., Robinson, L. A., Tetrahedron Lett. 1993, 34, 7895. 26. Whisler, M. C., MacNeil, S., Snieckus, V., Beak, P., Angew. Chem. Int. Ed. 2004, 43, 2206. 27. Look, S. A., Fenical, W., J. Org. Chem., 1982, 47, 4129. 28. Shin, J., Fenical, W., J. Org. Chem. 1991, 56, 3392. 29. Williams, D. R., Coleman, P. J., Henry, S. S., J. Am. Chem. Soc. 1993, 115, 11654. 30. Hutchinson, J. H., Money, T., Piper, S. E., Can. J. Chem. 1986, 64, 854. 31. Ito, Y., Hirao, T., Saegusa, T., J. Org. Chem. 1978, 43, 1011. 32. For a description of this diimide modification, see Hart, D. J., Hung, N. P., Hsu, L. Y., J. Org. Chem. 1987, 52, 4664. 33. Williams, D. R., Coleman, P. J., Tetrahedron Lett. 1995, 36, 39. 34. (a) For an overview, see Petasis, N. A., Patane, M. A., Tetrahedron 1992, 48, 5757. (b) Ballio, A., Casinovi, C. G., D'Alessio, V., Grandolini, G., Randazzo, G., Rossi, C.,

7

35. 36. 37. 38. 39. 40. 41. 42.

43.

SYNTHESISSTUDIESOF DOLABELLANESAND TRANSANNULARPROCESSES

267

Experimentia 1974, 30, 844. (c) For recent examples, see Kim, S., Shin, D. S., Lee, T., Oh, K. B., J. Nat. Prod. 2004, 67, 448. For a leading reference, see Sassa, T., Ooi, T., Nukina, M., Kato, N., Biosci. Biotech. Biochem. 1998, 62, 1815. Hashimoto, T., Tori, M., Taira, Z., Asakawa, Y., Tetrahedron Lett. 1985, 26, 6473. Nozoe, S., Morisaki, M., Tsuda, K., Iitaka, Y., Takahashi, N., Tamura, S., Ishibashi, K., Shirasaka, M., J. Am. Chem. Soc. 1965, 87, 4968. Iitaka, Y., Watanabe, I., Harrison, I. T., Harrison, S., J. Am. Chem. Soc. 1968, 90, 1092. Rowley, M., Tsukamoto, M., Kishi, Y., J. Am. Chem. Soc. 1989, 111, 2735. Boeckman, R. K., Jr., Arvanitis, A., Voss, M. E., J. Am. Chem. Soc. 1989, 111, 2737. Paquette, L. A., Wang, T. Z., Vo, N. H., J. Am. Chem. Soc. 1993, 115, 1676. (a) Jamison, T. E, Shambayati, S., Crowe, W. E., Schreiber, S. L., J. Am. Chem. Soc. 1994, 116, 5505. (b) Paquette, L. A., Sun, L. Q., Friedrich, D., Savage, P. B., Tetrahedron Lett. 1997, 38, 195. For leading references, see (a) Kato, N., Okamoto, H., Takeshita, H., Tetrahedron 1996, 52, 3921. (b) Kato, N., Nakanishi, K., Takeshita, H., Bull. Chem. Soc. Jpn.

1986, 59, 1109. 44. (a) Liu, H. J., Wu, C. L., Becker, H., Zapp, J., Phytochem. 2000, 53, 845. (b) Zapp, J., Burkhardt, G., Becker, H., Phytochem. 1994, 37, 787. (c) Huneck, S., Baxter, G., Cameron, A. E, Connolly, J. D., Rycroft, D. S., Tetrahedron Lett. 1983, 24, 3787. 45. Using slight modifications of a known procedure, (S)-limonene oxide was converted into 66 in 50% overall yield. White, J. D., Ruppert, J. E, Avery, M. A., Torii, S., Nokami, J., J. Am. Chem. Soc. 1981, 103, 1813. 46. Castro, A. M. M., Chem. Rev. 2004, 104, 2939. 47. (a) Mehta, G., Krishnamurthy, N., Tetrahedron Lett. 1987, 28, 5945. (b) For a related [2,3]-sigmatropic rearrangement from 66, see Wright, J., Drtina, G. J., Roberts, R. A., Paquette, L. A., J. Am. Chem. Soc. 1988, 110, 5806. 48. Linstrumelle, G., Krieger, J. K., Whitesides, G. M., Preparation of alkenes by reaction of lithium dipropenylcuprates with alkyl halides: (E)-2-Undecene. In Organic Synthesis, Masamune, S. J. (Ed.), Wiley: New York, 1976, vol. 55, pp. 103-113. 49. Whitesides, G. M., Ehmann, W. J., J. Org. Chem. 1970, 35, 3565. 50. For a leading reference, see Williams, D. R., Robinson, L. A., Amato, G. S., Osterhout, M. H., J. Org. Chem. 1992, 57, 3740. 51. The stereochemistry of this reduction was proven by further treatment with 6% Na-Hg in ethanol yielding a diastereomeric allylic alcohol compared to 80 (Scheme 7).

STRATEGIES AND TACTICS IN ORGANIC SYNTHESIS, VOL. 7 9 2008 Elsevier Ltd. All rights reserved.

Chapter 8 DIPHENYLOXAZINONES" VERSATILE TEMPLATES FOR THE ASYMMETRIC SYNTHESIS OF a-AMINO ACIDS, PEPTIDE ISOSTERES, AND NATURAL PRODUCTS Robert M. Williams and Cameron M. Burnett Department of Chemistry Colorado State University Fort Collins, CO 80523, USA

Introduction Preparation of the Template Glycine Electrophile A. Direct Substitution B. Diphenyloxazinones as Glycine Phosphonates IV. Glycine Enolate A. Monoalkylation B. Dialkylation C. Synthesis of Phenylalanine Analogs D. Glycinate Aldol Condensations V. Other Manipulations A. Oxazinones as Glycine Radicals B. Oxazinones as Glycine-Based Azomethine Ylides C. Direct Nitrogen Substitution D. Carbonyl Manipulation and Peptide Isosteres VI. Conditions to Remove the Chiral Auxiliary VII. Conclusion Acknowledgments References and Footnotes I.

II. III.

268 270 272 272 278 280 280 287 290 295 304 305 306 312 314 322 324 324 324

I. Introduction

Amino acids are central to the chemistry of nature, serving as the constituents of proteins, participating in nitrogen metabolism, and providing building blocks for numerous primary and secondary metabolites. ~The readily available proteinogenic amino acids serve as chiral starting materials for

8

269

DIPHENYLOXAZINONES

Ph

Ph

Ph",l~ O

Ph@o

RN,v~ O

RNv-~ O

(+)-1, R = Cbz (+)-2, R = Boc

(-)-1, R = Cbz (-)-2, R = Boc

FIGURE 1. Enantiomers of the diphenyloxazinone chiral auxiliary.

a variety of synthetic applications, 2 while the number of nonproteinogenic, naturally occurring amino acids, often with important biological activities, continues to increase. The stereocontrolled preparation of both natural and unnatural amino acids is an important problem in organic chemistry. Widely used methods include the hydrogenation of chiral or prochiral dehydro amino acid derivatives; carbon-carbon bond formation on chiral glycine equivalents via either enolate alkylation or carbocation substitution; electrophilic or nucleophilic amination of chiral carbonyl derivatives; and enzymatic or cell-based synthesis. 3 Herein is presented the development and use of chiral diphenyloxazinone glycine templates 1 and 2 (Figure 1) that allow the synthesis of amino acids of diverse structure in either the L or D series. 4 Our interest in this area began with our preparation of the electrophilic glycine anhydride derivatives 3, which were used to access precursors to the branched and oxidized isoleucine moiety required for the bridging framework of bicyclomycin (Scheme 1).5 While numerous glycine cation equivalents were known prior to 1985, none produced amino acids asymmetrically; we felt that a six-membered, rigid glycinate similar to 3 would be able to do so. o

o ....

--

I

3

=

4

HO'"

HN'~'-~n

bicyclomycin

SCHEME 1. Electrophilic glycinate used in the bicyclomycin synthesis provides inspiration for the development of a chiral, nonracemic electrophilic glycine template.

After exploring several chiral amino alcohols from which to develop a chiral glycinate and inspired by the bicyclomycin methodology, the use of erythro-2-amino-l,2-diphenylethanol (5), as reported by Kagan in 1968, seemed to offer a number of advantages. Kagan reported the condensation of (+)-5 with dimethylacetylene dicarboxylate (6) to yield

270

ROBERT M. WILLIAMSAND CAMERON M. BURNETT

Ph Ph,,,I~.OH /+/,

Ph,,,l~O

+ MeO2C - -

Ph,.,T~O

etoH

~

O,~.., CO2Me

Raney Ni

O~

Pd(OH)2

OMe

OMe

7

8

6

O-

"

OMe

SCHEME 2. Kagan's 1968 landmark asymmetric synthesis of aspartic acid.

the Z-dehydrolactone 7. 6 Hydrogenation of the conjugated enamine on Raney nickel in dioxane, followed by catalytic hydrogenation over Pearlman's catalyst in ethanol, furnished S-methyl aspartate (9) in >98:2 er (Scheme 2). The straightforward hydrogenolysis of the oxazinone 8 suggested that a structurally analogous heterocycle would function well as a glycine template. II. Preparation of the Template

The lactones are easily prepared starting from commercially available (1S,2R)- and (1R,2S)-2-amino- 1,2-diphenylethanol ((+) and ( - ) - 5 ) . 7 Reaction with ethyl bromoacetate in an Et3N/THF mixture gives the N-alkylated product in high yield in 1 h (Scheme 3). 8 The amine is protected with di-tert-butyl dicarbonate or benzyl chloroformate, and cyclization with recrystallized p-toluenesulfonic acid in toluene gives ( + ) - or (-)-lactones 1 (R = Cbz) or 2 (R - Boc) with no chromatography required. The lactones are soluble in most organic solvents, are stable for years to shelf storage at ambient temperature, and are commercially available. 9

Ph Ph

1. H2 / P d - C 2. Ph OH C~O2H/

OH iec-~procal ~ NOH recryst. ~ , ~ (__.)-10

Ph

Ph Ph,,,y/'&--OH .

Ph',,l~ O

/

NH2 (+)-5 Ph Ph~//"OH NH2 (+5

~

1. BrCH2CO2Et 2. acylation 3. TsOH, Phil, A R = Boc: 75% R = Cbz: 86% no chromatography

RN-...~ O (+)-1, R = Cbz (+)-2, R = Boc Ph Ph~o

RN,v~ O (-)-1, R = Cbz (-)-2, R = Boo

SCHEME 3. Procedure for the preparation of each oxazinone 1 and 2.

8

DIPHENYLOXAZINONES KCN

PhCHO

271 lipase TL|

,.~ benzoin

NH2OH-HCI optically active benzoin

vinyl acetate

Ph

Ph-H:_~__OH H2,Pd-C

optically active benzoin oxime

I H H2N (-)-5 optically active amino alcohols >99:1 er Ph _

Ph.... ~ O H H2N (+)-5

SCHEME 4.

Lipase TL|

resolution of benzoin and amino alcohols

4.

A racemic mixture of erythro-2-amino-l,2-diphenylethanol ((+)-10) can also be produced from benzoin oxime via hydrogenation. ~~ Reciprocal crystallization of the corresponding S-glutamate salts or, more recently, the employment of the corresponding (-)-mandelic acid salts allows the resolution of each enantiomer in high optical purity (>99:1 er). A preparative procedure for the synthesis of these agents has been published in Organic Syntheses. ~~Alternatively, benzoin has been prepared in high optical purity through a lipase-based enzymatic resolution procedure on a multigram scale that has been deployed in an economical synthesis of each antipode of the amino alcohols 5 (and thus, 1 and 2) (Scheme 4). ~ In addition, Sharpless has reported a preparative method to make the diphenylamino alcohols in >99:1 er by asymmetric dihydroxylation (AD) of trans-stilbene (Scheme 5). 12 p h / ~ / ~ Ph

SharplessA-D reaction ~. ~ P h

Ph (MeO)2CO, NaOH = Ph"~. OPh/.,,,O~O HO OH 12 13

11 _Ph 1. LiN3 2. H2, Pd-C

Ph'"~"OH NH2 optically active amino alcohols

(+)-5

>99:1 er

Ph Ph~o

H

NH2

(-)-5

SCHEME 5. SharplessAD reaction

on trans-stilbene.

272

ROBERT M. WILLIAMS AND CAMERON M. BURNETT

O ph"'~H 14

HCN

OH

,.._

O ~

oxynitrilase~Ph~'.CN "-ph.,,~CN

OTHPH O ..~Nv.,LLoMe Ph Ph 18

15

NH 1. PhMgBr ... THPO,_~)._~,_,

2. MeOH "-16, (> 99 % er)

Pn

r'h

1.

H2NCH2COOMe

2. NaBH4

17

1. p-TsOH OH Cbz O Ph ,.__ p h . ~ ~ v . . U . . . O M e p-TsOH ~ Ph~oII21 MeOH "cyclohexane C b z N ~ H I O 2. CbzCI Ph 19 (-)-1

SCHEME 6. Alternative oxynitrilase-based synthesis of the oxazinone by Brussee et al.

An alternative and potentially preparative synthesis of the N-Cbzprotected Williams oxazinone has been reported by Brussee and co-workers (Scheme 6). 13 The key step involves the oxynitrilase-mediated asymmetric cyanohydrin-forming reaction on benzaldehyde. Phenyl Grignard addition to the protected nitrile (16) provided an imine (17) that underwent transamination with glycine methyl ester followed by a stereospecific sodium borohydride reduction, providing the protected syn-(erythro) amino alcohol (18) that was acylated and cyclized to the oxazinone ((-)-1). The procedure reported in this paper furnished a little more than 9 g of the final oxazinone; the overall yield from benzaldehyde was 48%. The oxynitrilase utilized in this study (E.C. 4.1.2.10) is present in almond meal. llI. Glycine Electrophile

The diphenyloxazinone template was originally conceived as an electrophilic glycine equivalent. The methodology developed for this application also allowed conversion of the lactones to glycine phosphonate equivalents. A. D I R E C T S U B S T I T U T I O N

Reaction of oxazinone ( - ) - 1 with N-bromosuccinimide in refluxing carbon tetrachloride gave the anti-bromide 20 in quantitative yield. The reaction of 20 with a variety of nucleophiles in the presence of mild Lewis acids gave ~-substituted oxazinones (21) in good yield and with generally excellent diastereoselectivities (Scheme 7, Table 1). 14'15 The key C-C bond-forming reaction is believed to proceed through the agency of a highly reactive N-acyl iminium ion species that is attacked from the leasthindered face of the oxazinone, anti to the two phenyl substituents.

8

DIPHENYLOXAZINONES

27 3

Ph Ph Ph P~,~O NBS,COl4D.P OCbzN h ~O ~L Nu ~ P h ~ o deprotection + OCbZN.v~0 quant. : conditionsCbZN@o_ ,. HaN-@OR (-)-1

Br

A

20

21

22

SCHEME 7. Bromoglycinate homologations.

TABLE 1 Electrophilic reaction of N-Cbz-oxazinone-bromide 20 Entry

Nucleophile

a

OTBS --~

b

OEt //-.../SiMe 3

c

//~./SiM%

d

MeZnC1

Conditions

% Yield (21)

Amino acid (22)

% Yield 22 (ee)

ZnC12/THF, 25 ~

74

Ethyl aspartate

85 a (>98:2)

ZnC1JTHF, 25 ~

66

Norvaline

93 a (>99:1)

ZnC1JTHF, 25 ~

66

Allylglycine

908 (>95.5:4.5)

THE 78 ~

46

Alanine

100a (>98:2)

THF/Et20, 78 ~

48

Norleucine

52 a (>99.5:0.5)

72

Homophenylalanine

91a (>98:2)

-

e

Bu2Cu(CN)Li

-

f

E~,

g

/OSiM% ZnC1JMeCN, 25 ~ ,SiMe,

ZnClJTHF, 25 ~

82

Cyclopentylglycine

91 a (>98:2)

.SiM%

ZnC1JTHE 25 ~

82

Cyclopentenylglycine

94 b (>98"2)

ZnC1JTHF, 25 ~

64

(2-Tetrahydrofuryl)glycine

89b (>98"2)

ZnC12/THF, 25 ~

66

Dihydrofuranomycin

89 a (n.d.)

(~ h (~ i

j

1~_/CH 3

apdCl2 (0.3 equiv.), H 2 (20 psi), ETCH. bLi/NH3.

While the C-C couplings generally produced the anti configuration, the TBS enol ether of ethyl acetate gave the syn adduct. Further experimentation revealed that electron-poor nucleophiles react with the putative iminium ion from the less hindered face, while electron-rich nucleophiles react via direct SN2 displacement of bromide. The more

274

ROBERT M. WILLIAMS AND CAMERON M. BURNETT

basic organometallic reagents showed reduced yield due to a competing one-electron reduction of the bromolactone. The auxiliary was cleaved in most cases via catalytic hydrogenation over palladium chloride in good to excellent yields to give amino acids 22 (Table 1). Dissolving-metal conditions were also successfully employed to produce the amino acid derivatives and, in the case of unsaturated side chains, obviated saturation of olefinic residues (cf Table 1, entries b/c and g/h). Dissolving-metal reduction of the corresponding homologated N-Bocoxazinones 24 provided the first direct asymmetric synthesis of the corresponding N-Boc-protected amino acids 25 (Scheme 8, Table 2). This methodology was applied to a straightforward synthesis of [3carboxyaspartic acid (Asa), a natural, post-translationally modified amino acid first reported by Koch and co-workers. ~6This amino acid is unstable, being sensitive to decarboxylation and elimination of ammonia. The silyl enol ether of dibenzyl malonate was condensed with the bromide 20 to give the syn-coupling product 26 in modest yield and diastereoselectivity (Scheme 9). Catalytic hydrogenation of the major diastereomer under mildly acidic conditions gave a mixture of Asa as well as the decarboxylation product D-aspartate; purification with ion-exchange chromatography gave Asa in 30% yield and high optical purity. Importantly, no racemization was seen during the catalytic hydrogenation or workup.

ph_o y .~L . Ph

NBS,CCI4 P h ~

BocN.v~o

quant.

Ph O

B~

THF 13r 23

(-)-2

RM, ZnCI2 P h ~

Ph O

Li/NH 3

BooHN...,~OH ,...../-. 0 B~ _ ~ - " O EtOH / THF I~ 15, 24 25

SCHEME 8. Direct asymmetric synthesis of Boc-protected amino acids.

TABLE 2 Electrophilic reaction of Boc-oxazinone-bromide 23 Entry

Nucleophile

24 % Yield

Amino acid

25 % Yield (er)

~'-....~SiMe3 'SiMe3

63

Boc-allylglycine

70 (>98:2)

59

Boc-cyclopentenyl-glycine

70 (>99)

8

275

DIPHENYLOXAZINONES

Ph

OTMS

Ph~o

BnOJ~CO2Bn~

CbzN_~L-" O ZnOI2,THF [3r 53% (5.6:1 dr)

P/~

H2 (40 psi)

PhcbzN.7. ~O

O-

PdCI2(0.3 equiv.) ~ H3~I~cICO3

T 'O EtOH/THFo(0.04M) HO2C BnO2C"''CO2Bn 30

20

2H

6-carboxyaspartic acid (Asa)

26

SCHEME 9. Asymmetric synthesis of ~-carboxyaspartic acid (Asa).

Ph Phi..? CbzN~ O

Ph

2 ZnCI2 P h i " - O + Bu3SnC-C-R C014 = CbzN_~_..O

H2 (30 psi) + O PdCl2 (30-50 mol%) ~- H3N_.y~ O_ THF/EtOH (0.02- 0.06 M)

13r

R 273, R = Ph: 55% 27a, R = 06H13: 53%

20

""7 R

283, R = Ph: 57%, 97:3 er 28b, R = 06H13: 68%, 99:1 er

SCHEME 10. Alkynylation of bromo-oxazinone 20.

Ph

Ph Ph

~-

? + Bu3SnC-C-Me Br

Ph~o 2 ZnCI2 B o c N . ~ O ~ :

cc,. 61%

M~

_~

OH BooHN. ~ - . . O

iiI Me

23 29

M = Li: 79%, 82:18 er M = Na: 18%, >99:1 er

_

Me 30

SCHEME 11. Asymmetric synthesis of E-vinylglycine derivative 29.

Bromide 20 could also be reacted with organotin acetylides in the presence of zinc chloride to give the alkynyllactones 27 in moderate yield (Scheme 10); catalytic hydrogenation yielded the fully saturated amino acids 28.17 The analogous N-Boc-alkynyl oxazinone 29 was subjected to Birch reduction to yield the E-vinylglycine derivative 30 (Scheme 11). The Birch reduction with lithium proceeded in good yield but caused partial epimerization at the ~-stereogenic center (82:18 er). Switching to sodium eliminated the epimerization (>99:1 er) but also drastically reduced the yield. 18

276

ROBERTM. WILLIAMSAND CAMERONM. BURNETT

Ph Ph,,,1,/~O

Ph ArM o r ArH Ph,,,1,/~,O

B~

B~

ZnCI 2

Br

Ar

31

32

Ph

1. TMSI

Ph'"~OH

NalO4

+

O -

H20 / THF H3N,~ CIHoHNyCO2H =O 2.10% HCI pH 3 Ar THF, A Ar 33

34

SCHEME 12. Oxidative cleavage of the chiral auxiliary.

The arylated oxazinones 32 could be prepared via reaction of arylmetal or electron-rich aryl species with 31 (Scheme 12), but liberation of the amino acids required a new method for removal of the auxiliary, as neither catalytic hydrogenation nor Birch reduction were expected to selectively cleave the biphenyl auxiliary over the benzylic C-N bond of the arylglycine unit. The oxidative protocol reported by Weinges worked well: removal of the Boc group with TMS iodide allowed acidic ringopening of the lactone to produce acid 33, and periodate cleavage liberated two equivalents of benzaldehyde to give the free amino acids 34 in moderate yield. ~9 Unfortunately, some epimerization of the diastereomerically pure adducts was observed during biphenyl removal (Table 3). The furyl adducts 36a,b could be d!rectly hydrogenated to the free amino acids without reduction of the furan ring (Scheme 13), though

TABLE 3 Synthesis of arylglycines Entry

ArM/ArH

a

~

34

Conditions

33 % Yield a (er)

34 % Yield (er)

Et20/THF, -78 ~

56 (1:0)

52 (>91:9)

Et20/THF'-78 ~

55(1:0)

29(>97"3)

ZnClz/THF,25 ~

838(1:0)

62 (>95.5:4.5)

ZnC12/THF, 25 ~

50 (1:0)

26 (>95:5)

ZnClJMeCN, 25 ~

39 (1:0)

73 (>96.5"3.5)

CuLi

CuLi

b

(

~

c

MeO\~..OMe

d

I~/

e

o

OMe

ayield over three steps, byield after removal of Boc group.

8

277

DIPHENYLOXAZINONES Ph Ph Ph,,,,/~ O / CbzN

yA Br

~O O

R

ZnCI2 THF

"

Ph',,1,~-ON~ O H2 (1 atm.) 5% Pd/C (4 mol%) Cbz O

+ H3

O-~i ~o O

"

THF (0.07 M)

R

R 36a, R = H: 64% 36b, R = Me: 66%

35

37a, R = H: 57% 37b, R = Me: 82%

SCHEME 13. Removalof the chiral auxiliary without furan decomposition.

Ph

Ph O

1. D2 (40 psi) H,,Ph O Peel2 (0.30 equiv.) Ph~ O'D O ~H 5:1 THF/D20 (0.046 M) ,-" --

CbzN"~O13r

2. Dowex 54%

20

|

o|

H3N~ O H D

" L~

39, 84-90% atom-D 88.5:11.5-91:9 er

38

SCHEME 14. Asymmetricsynthesis of s-deuterium-labeled glycine. Ph 1.T2 (1 atm.) PdCI2 (096 equiv') | O,~ ~ Ph ~ O CbzN j,,O 6:1 THF/T20 (0.032 M) ~ H3N O -

13r 20

2. ion-exchange column 31%

H

T

40, 0.78 Ci/mmol 88:12 er

SCHEME 15. Asymmetricsynthesis of s-tritium-labeled glycine.

direct hydrogenation of the Cbz-adduct resulted in decomposition in other cases. Selective removal of the Cbz group from 36 followed by the ringopening/periodate protocol also gave the free amino acid without furan decomposition. The bromolactone 20 also served as a convenient template for the synthesis of chiral, isotopically labeled glycine. Reduction of putative iminium 38 under D 2 gas (Scheme 14) gave ~-deuterioglycine 39 with high isotopic incorporation of deuterium and good enantioselectivity. 2~Vederas showed that the procedure could be adapted to produce chiral, monotritium-labeled glycine 40 (Scheme 15). 21 This amino acid is difficult to produce in high isotopic purity by pyridoxal-dependent enzymatic exchange due to the large preponderance of the doubly-labeled byproduct.

278

ROBERT M. WILLIAMS AND CAMERON M. BURNETT

B. D I P H E N Y L O X A Z I N O N E S

AS GLYCINE PHOSPHONATES

Williams and co-workers used the lactone as a glycine phosphonate equivalent to synthesize the cyclopropyl amino acids. 22 Bromide 23 was treated with trimethylphosphite in THF at reflux to afford crystalline phosphonate ester (41) in 86% yield with the anti stereochemistry (Scheme 16). Upon treatment with base and an aldehyde, 41 provided the E-~,13-dehydrolactone adducts (42) in generally high yields (Table 4). The unexpected olefin stereochemistry probably results from minimization of steric interaction between the aldehyde R-group and the N-Boc residue in the betaine transition states. Cyclopropanation with (diethylamino)phenyloxosulfonium methylide gave cyclopropyloxazinones (43) in excellent chemical yields and high diastereomeric excess with addition

Ph Ph O BooNO \/~

Ph Ph Phi.. Ph~..-[... 1. NBS, CCI4, reflux O DMSO BocN~ O conditions BocN~ -1"-.-O ,, | 2. (MeO)3P,THF -HO ~" O 86% (2 steps) (MeO)2P"~O R'"/~ R R''''LL"R Ph-S-CH 2

(-)-2

41

Et/(~. Et

42

Ph Ph~

O

Li, NH3

B~ ~-_-__~ RO R' 43

-~

BocHN@,,~R'

HO2(f

""R 44

S C H E M E 16.

(~)

HCI, MeOH;

O~..-Me EtOH

HaN~H

""R 020E) 45

Asymmetric synthesis of aminocyclopropane carboxylic acids.

TABLE 4

Preparation of ~,[3-dehydrolactones (42), cyclopropanations (43), and ACC derivatives (44/45) Entry

R

R'

Conditions

42 % Yield

a b

H 2H

H 2H

NaH/THF LDA/THF

Quantitative 97

c

Me

H

LDA/THF

93

82 (1:0)

63

100 (1:0)

d e f g h

Et Pr i-Pr Ph p-NO2Ph

H H H H H

LDA/THF LDA/THF LDA/THF Nail/Phil Nail/Phil

92 82 19 96 84

79 (1:0) 88 (1:0) 96 (1:0) -

64 61 -

100 (1:0) 99 (1:0) -

43 %

Yield (dr)

96 (11:1)

44 % Yield 65

45 %

Yield (dr)

100 (92:8)

8

DIPHENYLOXAZINONES

279

occurring unexpectedly on the same face of the oxazinone as the sterically demanding phenyl groups, presumably due to rt-stacking interactions with the ylide phenyl group. Birch reduction of the adducts gave N-Boc-cyclopropane amino acids (44). Acidic removal of the N-Boc group and treatment of the resulting hydrochlorides with excess propylene oxide in EtOH generated the free cyclopropane amino acids (45) in essentially quantitative yield. The phenylcyclopropyl oxazinone 46 decomposed under both attempted dissolving-metal and periodate conditions previously employed for removal of the biphenyl auxiliary. Interestingly, a lead tetraacetate oxidative cleavage protocol gave the free amine, which was not isolated but immediately protected with a Boc group to give 48 (Scheme 17). Ester hydrolysis and Boc removal yielded the desired (E)-cyclopropylphenyl-alanine 49. 23 Ph

1. TFA, CH2CI2 2. LiOH (aq) EtOHI reflux

1. Pb(OAc)4 MeOH,CH2CI2 Ph O OH 2.1 M HCI (aq.), THF BocN.~ph O ,, HN /___~~O2Me ,,3. CH2N2,Et20/MeOH 3. Boc20, Et3N, THF H 85% (3 steps) H h 41% (3 steps) ~.

46

Ph

Ph~

47

BocHNU,%,, ph ..-

1.2.LiOHHcl,(aq)'MeoHEtOHCIH~

-.

MeO2C

_

H

87% (2 steps)

%

HO2C

48

H 49

SCHEME 17. Oxidative cleavage of the chiral auxiliary for synthesis of a phenylcyclopropyl amino acid.

Williams and Fegley used a diastereoselective [1,3]-dipolar cycloaddition of an azomethine ylide on the N-Cbz-~,13-dehydrolactone 50 (obtained from bromide 20) in their synthesis of S-(-)-cucurbitine. 24 A single diastereomer of cycloadduct 51 (Scheme 18) was obtained. Subsequent hydrogenation and ion-exchange chromatography yielded the free amino acid S-(-)-cucurbitine.

Ph

Ph O

1 (MeO)3P,THF Ph "2. Nail, THF;" Ph CbzN~ O O

CbzN_ " ~ O (CH20)n 13r 67% (2 steps) 20

v

50

(~ [/Ph //N~Q

hP].. P h .,y. O O CbzN

CH2CI2 0 ~ -> r.t.94% Ph--/ 51

(~) N(~,,4 H3 O

H2(60psi 5% Pd/C)

HCI (3.0 equiv.) HN--' ;

90%

SCHEME 18. Asymmetric synthesis of S-(-)-cucurbitine.

S-(-)-cucurbitine

280

ROBERT M. WILLIAMSAND CAMERON M. BURNETT

IV. Glycine Enolate

While originally conceived as an electrophilic template, the oxazinone was quickly adapted for use as a glycine enolate equivalent. A variety of alkali metal enolate conditions have been exploited for alkylation reactions and the corresponding boron, silicon, titanium, and aluminum enolates have been deployed in diastereoselective aldol condensation reactions. A. MONOALKYLATION

Treatment of the oxazinones (2) with either sodium or lithium hexamethyldisilazane gave a stable enolate. 25 Reaction with halides yielded the anti-alkylation products 52 (Scheme 19, Table 5), and the alkylated oxazinones could be deprotected as before to yield the free amino acids. 26 The N-Boc-oxazinone (2) generally gave higher yields and cleaner reactions than the N-Cbz-oxazinone (1). 27 The synthesis of N-Boc-alanine (1) (Table 5, entry a) was greatly improved via the enolate alkylation methodology, obviating the bromolactone reduction seen with the electrophilic coupling. N-Boc-allylglycine 52 (entry b) 28 served as a template for the synthesis of protected 2,7-diaminosuberic acid (see Schemes 55-56 a l s o ) . 29

Ph Ph

O

MHMDS RX

BooN.,.~ O

THF

Ph ~-

Ph

Li/NH 3

O B~

~

(-)-2

OH ~_ B o c H N . ~ o O EtOH / THF I~

52

53

SCHEME 19. Asymmetric synthesis of monoalkylated Boc-amino acids. TABLE 5 Production of amino acids via enolate alkylation of Boc-oxazinone 2 Entry

RX

M

52 % Yield

53 % Yield (er)

MeI I/-...~ Me

Na Li Na

91 86 84

54 (98.5:1.5) 50-70 (99:1) 52 (>99:1)

Na

70

76 (99:1)

Br/~'-Me

Br/~

8

281

DIPHENYLOXAZINONES

Ph Ph,,,

0

~

THF Br~CO2Et

CbzN"-~O (+)-1

CbzN

61%

H2 (20-50 psi) ~"

0

0

EtOH 2Et 71%, 98:2 er

2Et

54

Ph

Ph

,

55

OSiMe2t-Bu OEt

CbzN'...~ O quant. (+)-1

O-

CbZN@o Br 35

Ph",~'%0

H2, Pd=

ZnCl2,THF " C b z N ~ - " O 56 ~CO2Et

O+ H3N_~--. O .

~CO2 Et 57

SCHEME 20. Complementary syntheses of 13-ethyl aspartic acid by enolate and electrophilic templates.

An illustration of the complementarity of this approach is illustrated with the synthesis of R- or S-ethyl aspartate 55 or 57, respectively. Thus, via enolate alkylation of (+)-1 with ethyl bromoacetate and catalytic hydrogenation (Scheme 20) one obtains R-j3-ethyl aspartate. The electrophilic methodology yields S-[3-ethyl aspartate via the SN2 displacement of the bromide from 35 (Table 1, entry a). Thus, the same antipode of the oxazinone allowed synthesis of either enantiomer simply by the use of either electrophilic or nucleophilic chemistry. Williams and Aoyagi used the 13C/~SN-labeled N-Cbz-oxazinone 58 for the synthesis of ~3C/15N-labeled alanine. 3~Alkylation with methyl iodide proceeded in excellent yield to give 59 (Scheme 21), which was catalytically hydrogenated to give the doubly-labeled amino acid 60. Dong extended the reported alkylation of (-)-1 with ~,o~-alkyl diiodides to the five- and six-methylene units to give adducts 61, whose iodide was displaced to azides 62 (Scheme 22). Catalytic hydrogenation provided the desired ~,~o-diaminoalkanoic acids 63 in good yield. 31

Ph

NaHMDS,Mel Ph THF ,, Ph~15N~.c~O

Ph~o 15N. _ .~... Cbz~ -t~,--'O 58

-78 ~ 91%

0 = 13C

H2 (45 psi) PdCl2 (25 mol%)

" H315N+O'x~ -O

Cbz" "
SCHEME 21. Asymmetric synthesis of ~SN-1,2-[~3C]-S-alanine.

IVle 60

282

ROBERT M. WILLIAMS AND CAMERON M. BURNETT

Ph P h ~ j . ."~O . CbzN~/\- O (-)-1

Ph LiHMDS phy.J... O I(CH2)nl CbzN -- ~O THF/HMPA 71-74% i~) 5_6

NaN3 DMF, 90 ~ 87-89%

Ph phy~ O CbzN -- ~O

61

N / ) 5 _6 62

H2 (50 psi) PdCI2

+ OH3N~o

MeOH/THF/AcOH/H20 H2N(~) 5-6 (4:4:1:1) 71-78% 63

SCHEME 22. Asymmetric synthesis of a,m-diaminoalkanoic acids.

Ph

,o Cb

"~O (-)-1

Ph OPh,....J[.. H2 (40 psi) + O PdCl 2 (35 tool%) C l H ' H 2 N ' ~ " O CO2t_Bu "~ ~ ,- C b z N . . . ~ O -~ ~,./CO2t-Bu LiHMDS, THF 5:3 EtOH/THF (0.05 M) ]" 80% :""r'CO2t-Bu 74% CO2t-Bu

3L

64 CO2t-Bu

65

SCHEME 23. Asymmetric synthesis of a protected 7-carboxyglutamic acid derivative.

Hiskey conducted a Michael addition of the enolate of (-)-1 to di-tertbutyl methylenemalonate to give adduct 64 in good yield (Scheme 23). 32 Hydrogenation over palladium chloride yielded the 7-carboxyglutamic acid derivative 65 while preserving the tert-butyl protection on the side chain. Looper and Williams alkylated (+)-2 with trans-crotyl chloride in excellent yield (Scheme 24). Auxiliary removal under dissolving-metal conditions produced as a single diastereomer N-Boc-crotylglycine (66), a key component for the synthesis of the natural marine hepatotoxin 7-epicylindrospermopsin. 33 The nitro-aldol reaction between 73 and 74 could also be run under conditions to give all possible stereoisomers, from which cyclindrospermopsin was similarly prepared. The same methodology was applied to the first asymmetric synthesis of 7-deoxycylindrospermopsin, a natural metabolite reported from cyclindrospermopsin-producing blue-green algae. This work also served to correct the original structure proposed for this natural and highly hepatotoxic metabolite (Scheme 25). 34 Nolen alkylated ( - ) - 2 with the homologated ~-galactosyl iodide 80 in excellent yield (Scheme 26). Auxiliary removal under dissolving-metal conditions produced, after reprotection, the C-galactosyl N-Boc-S-serine derivative 82. The methodology succeeded with both 0t- and 13-galactosyl and glucosyl iodides; the 0t-glucosyl iodide gave a 93:7 dr while the other

8

DIPHENYLOXAZINONES

283

1. KHMDS, THF E-crotyl iodide . .~,. -78~ (92%) BocHN, y .OH

P~ Ph,,,,r O B~

2. Li~ NH3 THF, EtOH (68-87%)

(+)-2

H O1 ~ - M e"~ O'/~'-~ N~OQ

~"

H'"I-...~--~Me 2. LiAIH4, THF "(65% 2 steps) 66

Me O\ PhMe H~..N~b 200oC"-~ (78%) O-'~'-Oj

68

H2N_ /~ 1. BrCH2CO2Ph 1.AcCI, MeOH .._ H ' T "OH iPr2NEt,MeCN

69

'k"'~Me 67

(63-80%) 2. m-CPBA, Na2HPO4 CH2CI2, -78~ (84%)

1. DIBAL-H (87%) HO ~,. ' ~ ~ ....\OH TEMPO, CDC,3 2. PMBNH2, Me'~,K_ N'~:O Phl(OAc)2 H2/Pd/C; -~-N 1 tool% MsOH (P-O2NPhO)2CO PMB (75%) (X-ray) MeCN(81%) 70 H

HO,~.~,,,CHO

1. MeNO2, n-BuLi OzN...] 1. O2N--.] THF(84%) A c O ~ - - ~ .,' TFA, A (80%:AcO,~..--~,,, ,I TBAF, T H F ,,.._ ,,' Me4~"~- N ' ~ O 2. Ac20, DMAP v 4~1~,/N _ 2. Et3OBF4 .. ~ .1~ __. CHO ~N 0H2012; Me _ '~.O Cs2CO3 Me" T /~L)I-tM~J"~ "Z~l 71 'PUB NaBH4,EtOH 72 ----N CH2CI2 _~ ----N '~' II "" (67%) 'PMB (78%) /~ MeO/k~.N..:X..OMe OMe Nd 1. Pal(OH)2,H2 _OH Q OH MeO--/'k\ //~ 5~ "O~T~~~~~-Oso3-pI O 3 S O ~ O NH,"~' OH MeOH ~ ' M e I L . ~ K N y NH N.~.NH HO~.....---......~. 2. conc. HCI, A :--------~/NQ OH DMF, 3.& T _-T- "1,402 (32% 3 steps ) , (59%) Me'~I~K-N/~ Oet 75 H 76 : "--N

Med,~KNyNH

N.~NH

:------/N ~Q OH 'H 7-epi-cylindrospermopsin

SCHEME 24. Asymmetric synthesis of 7-epi-cylindrospermopsin.

0

Ph eh ~ "'~ CbzN

"O /

77 ~

~ O

M"

Me

BnO\ ~ " ' ~ H ]] ~T AcO. . . . NO2 N....f.N v --,.,,'v / ,,-,-, N OBn Me ,',,,.---OEt CsF,Ac20 :- ~/ = Me" 73- N MeCN (67%)

HO,~.~_~..~O ,~.,K.N%.d.NH HN.>~NH Me 2 /| II ~NH O 79 F3CCO2|

H H o2 N_ ...J.....J>..OBn l. NagH4, EtOH; "]// " ~ ~"~ Pd(OH)2,H2 ,J N_ /.N r ~ 2. conc. HCI, reflux (63% + C-8-diast. 2 steps) Y "~---OEt OBn : II L..._N 78

SO3.pyr,DMF, 3,&, (33%, C-8-diast. 33%)

r_

|

7 , ~ N . . . , ~ N H HN...,~NH Me . ~| "]]I" --NH

O

7-deoxFcvlindrospermopsin

SCHEME 25. Asymmetric synthesis of 7-deoxycylindrospermopsin.

alkylations gave a single diastereomer. 35 The axial benzyloxy group of iodide 80 would presumably block enolate attack from the back face of the iodide, while the all-equatorial orientation of the benzyloxy groups in 83 could allow some attack from the front face.

284

ROBERT M. WILLIAMS AND CAMERON M. BURNETT

BnO~_102n 1. Li/NH 3 EtOH / THF

Ph BnO-3~~80. BnO /

Ph Ph

O

B~

Ph~

i ,.

~O (-)-2

LiHMDS THF/HMPA -78 ~ -> r.t. BnO 87%, >99:1 dr

A co [ ~ O 2 0

o

2. CH2N2 ,, A c O ~ , , , , ~ BocN\_~ O AcO OBn 3. Ac20, py. 77% (3 steps)

NHBoc OMe O

82

B

OBn

Ph

Ph Ph@o

!~

BocN...v~ O (-)-2

LiHMDS THF / HMPA -78 ~ -> r.t. 71%, 93:7 dr

OO Bn

~

i 84

SCHEME 26. Asymmetric synthesis of amino acid sugar derivatives.

Ph Ph

1) TiCI4, Et3N

ph ~

Ph

h"" O

H2 (95 psi) 20% Pd(OH)2/C

O " CbzNL CbzN\...~O~ 2) (RO)3CH, CH2CI2 __ O -78 -> 0 ~ RO/~OR (-)-1

85a, R = Me (94%) 85b, R = Et (85%)

99%

O-

RO/~OR 86a, R = Me (99%) 86b, R = Et (99%)

SCHEME 27. Asymmetric synthesis of protected serinal derivatives.

DeMong and Williams reported the first use of the titanium enolate of (-)-1, alkylating with trimethyl orthoformate to give the corresponding dimethyl acetal (Scheme 27). 36 Auxiliary removal completed the first asymmetric synthesis of (2R)-~-formylglycine dimethyl acetal 86a. The total synthesis of capreomycin IB (see Scheme 62) necessitated use of the diethyl acetal 86b, synthesized by following the procedure with (+)-1 and triethyl orthoformate. 37 Lee arylated (-)-2 with the phenylmanganese complex 87 (Scheme 28); oxidative removal of the metal with NBS yielded aryloxazinone 89 in good overall yield. 38 Alkylation with KHMDS and an alkyl halide gave the disubstituted oxazinone 90. Removal of the N-Boc group with TFA allowed selective catalytic hydrogenation of the auxiliary to give ~-alkyl~-phenylglycines 92 in good yields. 39

8

285

DIPHENYLOXAZINONES

Ph Ph

O

NaHMDS Ph HMPA/ THF BocN -78 ~ /~ PF6 ~

Ph O LO

BocN...~O

Ph O BocN.,~..O

NBS

-Mn( L)3

59% (2 steps) 90% d.e. Mn(CO)3

D,

KHMDS -78 ~ -> r.t.

Et20

(-)-2

RX, THF;

89

88

Ph Ph~o

Ph TEA ~ P h ~ o

B~ P..~'-R O CH2CI2

H2 (1 atm.) Pd(OAc)2(50 mol%). H3N+ Oh..~ R -O

HNh P .~-"-RO

90

EtOH

p

91

92

SCHEME 28. Asymmetric synthesis of ~-alkyl-a-phenyl amino acids.

Ph

Ph

Ph O

c0z... o

O H2 (60 psi) NaHMDS,THF CbZN~o_ 10% Pd/C(50 mol%) -78 ~ : =I ~ 3:1 MeOH/THF (0.01 M) 93

OMe ~ Me

58% (-)-1

0-

84% Me~ --NB~ Me

Me.~-NBoc Me

94

95

SCHEME 29. Asymmetric synthesis of protected 5-hydroxylysine.

Brussee alkylated ( - ) - 1 with the chiral iodide 93 to give acetonide 94, whose catalytic hydrogenation yielded the protected (2S,5S)-5-hydroxylysine 95 (Scheme 29). 40 As either enantiomer of the iodide 93 could be obtained, this methodology provides access to all four possible diastereomers of 5-hydroxylysine. Allevi and Anastasia pursued the same target starting with ( - ) - 2 , which was alkylated with homoallyl iodide (Scheme 30). Epoxidation with mCPBA and opening with sodium azide gave azido-oxazinone 97 as a 1:1 mixture of diastereomers (epimeric at the 2 ~ alcohol stereogenic center). Removal of the N-Boc group and hydrogenation yielded the free 5-hydroxylysine 98. 41 Singh alkylated ( - ) - 1 with cyclopentyl iodide to give adduct 99 as a single diastereomer after recrystallization (Scheme 31). Catalytic

286

ROBERT M. WILLIAMS AND CAMERON M. BURNETT

Ph Ph~ - . . 0

Ph Ph~o B~ - ' ~ O

LiHMDS I ~

oc.. o _T

Ph Ph~...O ~__ BooN_...~O

c

19:1TFA/H20;H2 + _ O10%Pd/C(23mol%) H3N,..~ O__

8:3:1 MeOH/ H20 ,,..'~".../ HCI (0.0026M) H N3 88%

60%

U

(-)-2

96 S C H E M E 30.

HO" ~

NH2

v

97

98

Asymmetric synthesis of ~-hydroxy-S-lysine.

NaHMDS

Ph ~ ~ I Ph O ,,, CbzN..O/`~' '.. THF / HMPA -

P|h Ph

O

~00O o rt

+ H3N

O,~:O

1:2 MeOH / THF

,,4O,o

(-)-1 S C H E M E 31.

H2 (60 psi) PdCi2

@

99

100

Asymmetric synthesis of cyclopentylglycine via enolate alkylation.

Ph

Ph. ~

j[,.h P .O

NaHMDS,THF -78 ~

CbzN.v~O (-)-1

(Br ~ 101 MeO"Av~ L "O"L.~-O 40%

S C H E M E 32.

Ph~ ~ O CbzN O.....i--.,.> j.. :

O-/ "O" v 102

H2 (50 psi) ...~O+ PtCI2 (50 mol%) H3N -,,rO _ _THF 80%

"

O

~ OMe

"OMe 103

Asymmetric synthesis of a coumarin-containing amino acid.

hydrogenation gave S-cyclopentylglycine 100, though in lower yield than achieved with the electrophilic lactone methodology (Table 1, entry g).42 Leblanc alkylated ( - ) - 1 with 4-bromomethyl-7-methoxycoumarin (101) to give the ~,]3-unsaturated ester 102 as a single diastereomer (Scheme 32). 43 Catalytic hydrogenation in a mixed MeOH/THF system not only removed the auxiliary but also saturated the olefin. Fortunately, reaction in pure THF yielded the desired unsaturated product 103, presumably due to precipitation of the amino acid before the alkene could react. Combination of the phosphonate and enolate alkylation methods allowed a stereocontrolled synthesis of 2,3-methano-2,6-diaminopimelic

S

DIPHENYLOXAZINONES

287

Ph

Ph

,

2.03; DMS 79% (2 steps)

i"~

(-)-2

: (MeO)2P"~O

Et/o" Et

H~" O 105 BOCNv~,,, Ph (~ Ph O

O

104 OO

27%

41

Ph Ph~ " " O -

DMSO 83%

OH Li, NH3 63%

106

_ l z

: "Ph Ph

:

O HCI, THF

OH O ~ NHBoc 97% 107

-

O NH3 108 (2S,3S,6S)

SCHEME 33. Asymmetric synthesis of DAP stereoisomers.

acids. 44 Alkylation of ( - ) - 2 with 4-iodo-l-butene and ozonolysis of the alkene gave aldehyde 104 (Scheme 33). Horner-Wadsworth-Emmons condensation of 104 with phosphonate ester 41 produced bis-lactone 105 in poor yield, apparently due to retro-Michael decomposition of aldeyhde 104 under the basic conditions employed. The alkene of 105 was cyclopropanated as before, and Birch reduction and N-Boc removal yielded the cyclopropyldiaminopimelic (DAP) acid 108. Two other diastereomers were prepared by starting from (+)-2 in the synthesis of either aldehyde 104 or phosphonate 40. B. DIALKYLATION

Monoalkylated lactones can be deprotonated with potassium hexamethyldisilazane to effect a second alkylation with activated alkyl halides (Scheme 34, Table 6). The second alkylation proceeded anti to the phenyl rings as well, shifting the first substituent to the same side as the phenyl rings. Baldwin sequentially dialkylated the N-Boc-oxazinone (+)-2 with benzyl bromide and allyl bromide, using NaHMDS for both enolate formations, to give the disubstituted lactone 116 (Scheme 35). Birch reduction yielded amino acid 117, which was carried on to the bicyclic y-lactam dipeptide analog 118. 45

288

ROBERT M. WILLIAMSAND CAMERON M. BURNETT

KHMDS Ph OH R2X ~. P h~~ , . O Li/NH3 BocHN~, THF BOCNy~. EtOH/THF R( "R20 alrl-,,R20

Ph Ph@o

BocN_ @ O 109

Ph

i

Ph~o CbzN _ @

110 111 KHMDS Ph R2X Ph~.~ O Li/NH3 +~OCbzN.~~ ~" H3N. . . ~ THF RI~,,R:O EtOH/THF R1r "R20

O

i

al 112

113

114

SCHEME 34. Asymmetric synthesis of a,~-disubstituted amino acids.

TABLE 6 Production of a,0~-disubstituted amino acids from monosubstituted lactones 109 and 112 Template

R~

R2X

Alkylation % yield

Amino acid % yield (er)

a b

Boc Boc

Me n-Pr

i,/'-,,,,~ i,,,,---,,,~

87 90

70 (>99.5:0.5) 60 (>99.5"0.5)

c

Boc

Me

Me

80

65 (>99.5:0.5)

d

Cbz

Me

84

93 (>99.5:0.5)

e

Cbz

Me

80

95 (>99.5:0.5)

Entry

Br/"..,,~ Me Br/"'.,.]~ Br~

q J

Ph Ph NaHMDS Ph,,,~O ph,,,l~ O BnBr,THE B ~ O B~ 93% Bn 115

(+)-2

Li/NH3

,. t-BuOH/ THF 65%

BocHN

O,,,~

..,_ O Bn L 117

Ph NaHMDS Ph,,,1,~O BE~ BocN,,~ O 96%THE Bn" L 116 Bn.~//~,,,S,

" "~.N-...~

~.~ BocHN

O 118

/~NHi-Bu O

SCHEME 35. Asymmetric synthesis of a 7-1actam dipeptide analog.

8

DIPHENYLOXAZINONES

289

Ph

Ph,,, B~

O v"~'O

, SEMCI, THF

BocN

Ph _

O

76%

L..../TM S

(+)-2

Boc

15-crown-5, THF 2. Nal,refluxaCetone

119

O

/O120~

70% (2 steps)

I

TMS"

Ph _

Ph,,,./i~. O

O

O

T

B~

o

J'~L"ONa= 15-crown-5 75%

Ph .

.

.

.

Ph-

.

.

.

,TF

Ph

O--~__ TMs"P h

o

2. H 2, Pd/C 83% (2 steps)

O

-

H3N

121

~oa

I

NH3

122

SCHEME 36. Asymmetric synthesis of 6-hydroxymethyl-2,6-diaminopimelic acid.

Baldwin also reported sequential alkylation of Boc-lactone (+)-2 with SEMC1 and 3-chloro-l-iodopropane to give the disubstituted lactone chloride (Scheme 36), which was converted to the iodide 120 via a Finkelstein reaction. The iodide was used to alkylate the sodium enolate of a second oxazinone molecule, giving the bis-oxazinone 121. Removal of the protecting groups and hydrogenation gave 6-hydroxymethyl-2,6-diaminopimelic acid (122), a constituent of a naturally occurring antibiotic isolated from

Micromonospora chalcea. Aoyagi and Williams explored the synthesis of (S)-2-methylasparagine from either antipode of 1 2 3 . 46 Alkylation of the methylated oxazinone 123 with tert-butyl bromoacetate gave the disubstituted lactone 124 in good yield (Scheme 37). Cleavage of the tert-butyl ester to the acid and amidation gave amide 125, whose hydrogenolysis yielded (S)-2-methylasparagine (126). Sequential alkylation of the antipode (+)-1 in reverse order gave the disubstituted lactone 128, with the same stereochemistry at the ~-carbon

P h / o ~Ph CbzN_.,.~_..O Me 123

NaHMDS,-78 ~ BrCH2CO2t-Bu

Ph~o

15-crown-5,THF 71%

CbzN.,.~--. 2. EEDQ Me" ""[ 0 CO2t-Bu NH4HCO3 95% (2 steps) 124

!

Ph

H2 (60 psi)

+ O _ .a.."O PdCI2 (30 molYo) o 1. TFA, CHzCI2 Ph"~1/ " CbzN_--~ = H3N" - . ~ o ~ . 'O 2:1 EtOH Me "] THF (0.30 M) CONH2 94% 125

SCHEME 37. Asymmetric synthesis of (S)-2-methylasparagine.

Me" "[ CONH2 126

290

ROBERT M. WILLIAMS AND CAMERON M. BURNETT

Ph Ph,,,lr/~O CbzN

P_h NaHMDS

- v ' ~ O BrCH2CO2t-Bu THF, -78~

(+)-1

Ph

Ph,,,~,,~O NaHMDS,-78 ~

Ph,,,l~O

CbzN~

CbzN ~ . . O Me" ""l CO2t-Bu 128

O

15-crown-5 Mel, THF

"..CO2t_Bu 63%

81%

127

SCHEME 38. Asymmetric synthesis of (S)-2-methylasparagine.

but diastereomeric at the phenyl carbons (Scheme 38). Conversion to the amide proceeded in slightly lower yield, presumably due to the steric hindrance posed by the phenyl rings on the same side of the oxazinone, and hydrogenolysis again yielded (S)-2-methylasparagine. C. SYNTHESIS OF PHENYLALANINE ANALOGS

Williams and Im reported the synthesis of phenylalanine in the original oxazinone enolate paper (Table 5, entry d). Alkylation of the oxazinones with substituted benzyl halides has become one of the most prevalent uses of the lactone enolate chemistry, due to the typically high yields, high diastereoselectivities, and ease of cleavage of the oxazinone to the corresponding phenylalanine derivative. Schow reported alkylation of oxazinone (-)-1 with the benzylic bromide 129 to give, after deprotection with TBAF, alcohol 130 (Scheme 39). Catalytic hydrogenation with Pearlman's catalyst in THF gave cleavage of the N-Cbz and O-benzyl residues, but only partially cleaved the N-benzyl bond. Acidic catalytic hydrogenation was used to complete the synthesis of azatyrosine 131. 47 Ph H2 (50 psi) OTBDPS Ph...].. Pd(OH)2/C (9 mol%) +~O1. NaHMDS " ~ " ~ O 1:1 i-PrOH/THF (0.13 M); CIH~ O_ + CbzN .~r/'-~. =CbZN.v ~ ~ a THF (60%) O H2 (50 psi) O 2. TBAF, THF ~ Pd/C (4 mol%) CI Ph phVJ,,,O

(-)-1

B

(79%)

130

~J

I

HCl/H20 (0.06 M)

N/'..~'~'OH '

64%

131

H~

SCHEME 39. Asymmetric synthesis of an azatyrosine.

Solas reacted (-)-1 with the benzylic bromide 132 to give the alkylation product 133 in good yield (Scheme 40). 48 Catalytic hydrogenation over palladium chloride proceeded quantitatively to give amino acid 134 as its hydrochloride, which was further converted to the protected phosphotyrosine isostere 135.

8

DIPHENYLOXAZlNONES

291

CI OH OH Ph H2 (50 psi) .. + - [ FmocHN_ LiHMDS Ph~ . , , PdCI2 (50 m~176176 n31'~~ 0 _ --'T" "0 --T.-~ O . 2:1. EtOH/THF . . -:" " ] ~ CbzN...~O ~ OO / F--+I:'O78%P" (0.18 M) ]/EL ~/~l~(OEt)2 '/IL-.v~-h~P(OH)2 F (OEt)2 ~ 0 quant" F["F Fi "F

Ph B r"|h Ph~ ' ~ 0 i/~ CbzN' . v ~ + ~ (-)-1

132

133

(OEt)2

134

135

F " SCHEME 40. Asymmetric synthesis of a protected phosphotyrosine isostere.

Zhang also synthesized 135 indirectly, alkylating ( - ) - 1 with 4iodobenzyl bromide to give aryl iodide 136 (Scheme 41) and converting to the same difluoromethylphosphonate 133. Auxiliary removal at lower pressure with palladium on carbon gave zwitterionic amino acid 137. 49 Ph ph~[...O

-, | Ph H2 (1 atm.) + O~ n y ' ~ - O 100 Pd/C (25 mol%) H3N~L~ O -____._~._.,, -

Ph NaHMDS,ZHFph J [ -78 ~ -"~r ~ \ o

c z=

=:,

o

80%~/"~?~!

(-)-1

136i~ I~ //J... '1

" ~

Oq u a n t "

'/K"v~"'~'P(OEt)21"F

133 ~ J ' ~ / P(OEt)2 ~ , , F F

F 137

SCHEME 41. Asymmetric synthesis of a protected phosphotyrosine isostere.

Bender and Williams conducted a rapid synthesis of m-tyrosine via alkylation of ( - ) - 1 with 3-benzyloxybenzyl bromide (Scheme 42). 5o Catalytic hydrogenation of the alkylated oxazinone 138 removed the auxiliary and protecting groups to give m-tyrosine in excellent yield. LiHMDS Ph B r / " ' Y ' S " y O B n Ph Ph....JL. ~ PhiL. O '~ "O CbZN.v~,. , u (-)-1

THF

,.

I T CbzNy'~" O

87%

~o/OBn

H2 (50 psi) PdCl2 (30 mol%) 1"1 EtOH/THF (0.09 M)

+ H3N~o ~

O

"...~fOH

99%

138

meta-tyrosine

SCHEME 42. Asymmetric synthesis of m-tyrosine.

Paquette and co-workers modified the procedure slightly to produce the protected m-tyrosine 140 (Scheme 43), which was used for the synthesis of (-)-sanglifehrin A. 51

292

ROBERT M. WILLIAMS AND CAMERON M. BURNETT

NaHMDS B r ~ OB~

Ph ~ Ph O CbzNv.,,~--.0

THF / HMPA 60%

Ph H2 (50 psi) OH Phi.,.. O PdCI2 (30 mol%) CbzHN....~ CbzN~ 1"1 EtOH/THF (0.08 M); _ O -_ 0 ,IOBoc Cbz-O-Su ~~OBoc 92% (2 steps)

(-)-1

139

L ~

140

SCHEME 43. Asymmetric synthesis of protected m-tyrosine.

Seto alkylated ( - ) - 1 with the benzyl bromide 141 to give ketoester 142 (Scheme 44). 52 Auxiliary removal also reduced the benzylic ketoester to yield the corresponding ~-hydroxyester 143. Ph

Br.....~

Ph

H2 (1 atm.)

+

_

2:1 MeOH/91%THF(0.018 M)

O-

s

CbzN'-/'~'O ~ (-)-1

-78 ~ -> r.t.

O~

O

45%

2 14

141 Ot-Bu

"'"[/~

O

O

~

Ot-Bu Ot-Bu

143

OH

O SCHEME 44. Asymmetric synthesis of an ~-hydroxyester amino acid.

Roller alkylated ( - ) - 1 with the benzyl bromide 144 to give adduct 145. Hydrogenation gave amino acid phosphonate 146 (Scheme 45). 53

Br ~ P/~ Ph" 7 " O +

NaHMDS

CbzN" P~ ~'O Ph

H2 (36 Pd black (80psi) mol%)~

+ OH3N,~ O _

CbZN'v~o

(-)-1

~7 " ~ THF.78oc/HMPA %..0 .- 78% 144 (Ot-Bu)2

_ 5 14

~

MeOH (0.023 M) O ,, quant. P(OtBu)2

~

O

'/~.~x'....../P(Ot-g u)2 146

SCHEME 45. Asymmetric synthesis of an amino acid phosphonate.

Burke alkylated ( - ) - 1 with the benzyl bromide malonate 147 to give the adduct 148 in good yield and with high diastereoselectivity (Scheme 46). Auxiliary removal via hydrogenolysis and Fmoc protection gave the protected 4-(2'-malonyl)phenylalanine derivative 149. 54 The analogous enolate alkylation with the diethyl malonate derivative proceeded in only 14% yield. However, conversion of 148 to the corresponding diethyl ester allowed access to the Fmoc-protected amino acid containing the diethylmalonate side chain. 55

8

293

DIPHENYLOXAZINONES

Ph Ph

Br

Ph~o

Ph~o

+

LiHMDS, THF

CbzN _ @

0

y

CbZN.v~ O (-)-1

-78 ~ -> r.t. 66%, >97:3 dr t-BuO2

_

~

O2t-Bu 147

1. H2 (20-45 psi) Pd black (33 m o l % ) 1:1 EtOH / THF (0.28 M)

~

148

CO2t-Bu CO2t-Bu

~OH FmocHN.,_./~O_.

2. Fmoc-OSu, NaHCO3" 95% (2 steps)

~ ] ] ~ ~ C O 2

t-Bu

CO2t-Bu

149

SCHEME 46. Asymmetric synthesis of a protected malonyl phenylalanine derivative.

Garbay alkylated methyllactone 123 with benzyl bromides 150a/b to give dialkylated lactones 151 (Scheme 47); catalytic hydrogenation gave the disubstituted amino acids 152a/b. 56 Ph

Ph

Ph Br~ O + L

KHMDS

Ph~o CbzN e. . . ~ .

o

H2 10% Pd/C (6 mol%)

THF CbzN _ O -78 oC M~ c ~ ) O 1_ a:n=0,44% b:n=l 31% BuO2 150

O

H3'e"',~", O

1:1(0.025 EtOH /M) THF a:n=0,99% b: n = 1,90%

-

123

+

t-BuO2

1 151

t_BuO2C.

1 152

SCHEME 47. Asymmetric synthesis of a disubstituted phenylalanine derivative.

Burke and co-workers adapted the above procedure for the synthesis of the phosphorus-containing amino acid 156 (Scheme 48). 57 Taylor alkylated (-)-1 with the sulfonamide 157 to give 158 (Scheme 49). Auxiliary removal under a hydrogen balloon and reprotection of the free amine yielded protected amino acid 159.58 Jin and Williams alkylated (-)-2 with the benzyl bromide 160 to give 161 in good to excellent yield (Scheme 50). The benzyl ether of 161 could be hydrogenated to the free phenol 162 without cleavage of the diphenyloxazinone auxiliary. Further elaboration gave the N-methyl species 163,

294

ROBERT M. WILLIAMS AND CAMERON M. BURNETT

Ph Ph Ph Br\ O ~.. /~.. KHMDS CbzN Ph O + ~ THF M CbzN O -78 ~ Me / 86%

Ph Ph 10%Pd/C(13 mol%) OH O CbzN I:IEtOH/THF BocHN ..~--e.~O ~ (0.025M) M 2. Boc20,Et3N 61% (2 steps)

I

123

153

154 I

155 O',P(OEt)2

156 O"P(OEt)2

SCHEME 48. Asymmetric synthesis of a disubstituted phosphorus-containing amino acid. Ph

~. H2 (1 atm.) Ph ~ 0 PdC,2 (50 mol%) Fm~ ?H~o Ph [/ LiHMDS CbzN~~o 11 EtOH/ diox. (0.064M) Phi,.? + ~ I:ITHF:HMPA z'-....r ~ Fmoc-O-Su,Na2CO3 ~ J CbzN v ' ~ ' O F\"~O 80% ~[~.~F99~176 F~...~ 159 (-)-1 F">"'~-N(DMB)2 158 O=S:O O:S:O O I N(DMB)2 157 >98.5:1.5er N(DMB)2 Br

_

I

SCHEME 49. Asymmetric synthesis of a sulfonamide-containing amino acid.

Ph

Ph

Ph

OMe Ph~ Ph Me~ ~ /OBn O O + "~,..~ NaHMDS,- B~ _~-"O

BooN"v~O (-)-2

"~"T H F79-92~ Br ~ 160

~OBn 161

~\OMe

Ph~o

H2 (1 atm.) Pd/C = BooN..,..~ - O EtOH95%

=.

2-~OH 16

OMe

Ph Me Me Ph~o~ 1) H2Pd012(8psi) 0 FmocMeN ~Ho_ MeNlo 1:1 EtOH/ THF : = ~OTBS :"-,rJ"~--,rZOTBS 2) Fmoc-OSu,NaHCO3 163 DMF 164 ~ "OMe "OMe 80% (2 steps) Me Me SCHEME 50. Asymmetric synthesis of a highly substituted tyrosine derivative.

which suffered loss of the auxiliary under higher pressure hydrogenation. The resultant free amino acid was soluble in ether and was thus directly protected to give the N-methyl-N-Fmoc-tyrosine derivative 164. 59 Extension of this methodology to the tetrasubstituted benzyl iodide 165 provided 166 (Scheme 51). Elaboration as before gave the N-methyl-N-

8

295

DIPHENYLOXAZINONES

OMe .O n ~"~O (+2

-78~ 88% 165

OH .

NaHMDS BocN "

+ B~

Ph O == ~ O B n l I .~1 MeO~ ~ \OMe 166 Me

0

OTBS .J.L. /)... MeO" " ~ OMe Me 167

SCHEME 51. Asymmetric synthesis of a highly substituted tyrosine derivative.

Fmoc-tyrosine derivative 167. 60 These highly substituted tyrosine derivatives have been deployed by the Williams laboratory in total syntheses of members of the saframycin/ecteinascidin family of antitumor agents; 166 was utilized for both halves of these agents. Partial auxiliary cleavage from 166 by catalytic hydrogenolysis gave the free acid, whose reduction gave diol 168 (Scheme 52). Further Pictet-Spengler elaboration gave the diol 171, whose N-bibenzyl group could be removed by catalytic hydrogenolysis in absolute ethanol. Inclusion of Boc20 in the reaction mixture led to the protected amine 172 in excellent yield. We speculate that the nitrogen of 168 is deactivated by the carbamate-protecting group and thus is not reduced, while the dialkyl nitrogen of 171 is more electron-rich and thus more readily hydrogenated. Protection of the diol of 172 and removal of the Boc group gave free amine 173, which underwent peptide coupling with the acid chloride 174 derived from 167. The Fmoc-protected coupling product 175 was converted to the corresponding Boc-protected amine, and the primary O-TBS group was selectively removed. Oxidation of the resultant primary alcohol to the aldehyde allowed formation of the hemiaminal species and removal of the phenolic O-TBS group gave diol 176. Treatment of 176 with TFA converted the hemiaminal to the iminium ion, which underwent Pictet-Spengler-type cyclization to give the pentacycle. Reductive amination of the secondary amine provided the phenol 177, which was used as a common intermediate for the synthesis of both (-)-renieramycin G (Scheme 52) and (-)-jorumycin (Scheme 53). D. GLYCINATE A L D O L CONDENSATIONS

Miller and co-workers were the first to explore the utility of boron enolates of the oxazinone for glycine-based aldol condensation reactions. 61 Thus, formation of the boron enolate of ( - ) - 1 and reaction with aldehydes

296

ROBERT M. WILLIAMSAND CAMERON M. BURNETT OMe M e a l

Ph

Ph

~)~ ..'~

OBn ~

1 Pd/C(10%), H2, MeOH

MeO e

OH

Ph?BocN,v...~O NaHMDS'THFVMeO. i O 2"IBCF' NMM' M e O ~ B~ 88o/0 ~ ~"1 then NaBH4 ' 88% 2 steps HO (-)-2

Me"~//[''OBn

MeO ~ Me

168

OMe 161 MeO H_ OTBS EtO2CCHO ,.._ Me ~

"N

ph C.3CN,,0oc Ph

HO

92%

1. DBU,THF OTBS 50 ~

THF

_

HO O%EtPh

169

MeO H M e ~ O T B

Ph

v 2"TBSCI' NEt3' 0H2CI2 88% 2 steps

MeO S Pd/C(10%), H2 M e ~ . / - . ~

89% 2steps

170 OTBS1. KHMDS MeO TBSO allyI-Br j,. M e ~

MeO'//~"r~-N _ ~ P h i_ i : HO ~ 15h OH 171 (dr 1:0.3)

EtOH,Boc20 "- ...J-L /it... : .NBoc 69O/o(+23%otherMeO"HO"~/" "~ .... 2. TMS-OTf, M e O ~ NH diastereomer) 2,6 lutidine AllylO \ 172OH 97% 2 steps 173 OAIlyl Me OMe MeO. ~ .OMe ..~. 2,6-1utidine MeO TBSO ~ \--~ 1. piperidine,CH2Cl2 HO Me CH2CI2 Me"r/,'/'/'/'/~,.., _'/'~"~'~OTBS 2. Boc20, EtOH ,.._ MeO H OH ~H "- Me\/~/~.--hl-/JL.~Boc OMe 3. HCO2H,THF, H20 TI "T H 7" Cl[O~ MeO N~NHFmoc 4. Dess-Martin[O] . . . . ~ / / J - ~ N . / . Fmoc..N'''h AllylO \ O 5. TBAF,0 ~ Meu T ]r ]7 "H 174H r ~ ~OMe OAIlyl 175 77% 5 steps AllylO t,,OAIlyl O 176 OMe TBSO" "7:" Me OMe O Me OMe HO.~ ~ ..Me 1. Bu3SnH,Pd~ O HH 93% "r"/ ~ 2.2,4,6-TCB-CI M e ~ - : 0 TFA MeO H I. It NEt3' PhCH3 1. anisoCH2CI2,M e ~ O M e O _t N MeO" O~ ~,, O 2. HCHO, M e O / ~ "~"'1~'14 O..~1 Me NaBH3CN AllylO I,,, C) " H O Me 71% 2 steps OAIlyl 177

3. DDQ 42% 3 steps

(-)-renieramycin G

Me

SCHEME 52. Asymmetric total synthesis of (-)-renieramycin G.

OMe HO~Me

M

MeO e ~

1. LiAIH4, then aq. KCN, AoOH 2. Pd~ H .H .,L\ IJ,.,.OMe 3. Bu3SnH, DDQ M ~ =

MeO" "1>" "~"- "1]" "14 AllylO ~, O OAllyl 177

4.(49%3stepS)Ac20 5. AgNO3 (78% 2 steps)

OMe O O H H \ / , J ~ ~ Me :_- : Meo~]N O

\

___.~...H OH OAc

(-)-jorumycin

SCHEME 53. Asymmetric total synthesis of (-)-jorumycin.

~e

8

297

DIPHENYLOXAZINONES TABLE 7 Selectivity and yield of lactone aldol reactions % Recrystallized yield of major diastereomer (179) 57 38 42

17: 3:1 5:1 5:1

Me n-Pr i-Pr

Ph Bu2BOTf, Et3N CH2CI2 Phi'- O -78 ~ CbzN. v . ~ o RCHO (-)-1

Ph

Ph

l

Ph~ " . ~ ~0-O. -BBu CbZN~o'" "l 178

|

Ph

O 0 H2PdCI2 + OH (40 psi)_ H3N' OO THF/EtO~ R R" "OH 179 180

= CbzN

SCHEME 54. Aldol reaction of oxazinone 1.

gave the aldol adducts 179 with moderate diastereoselectivities (Table 7, Scheme 54); recrystallization provided the pure major diastereomers. Catalytic hydrogenation gave the corresponding [3-hydroxy-~-amino acids 180. Confirmation of the anti-diastereoselectivity (erythro) was secured through the synthesis of threonine itself. The Zimmerman-Traxler chair transition state shown mandates that the aldehyde approach from the more open face of the constrained E-enolate. This aldol methodology was quickly adapted for the synthesis of more complex natural products. Williams and co-workers reported aldol reaction between (-)-1 and the oxazinone-derived aldehyde 181 to give the aldol adducts 183 and 184 in a-~1:25 ratio (Scheme 55). Barton deoxygenation, catalytic hydrogenation, and cleavage of the methyl ether gave (2S,6S)-2,6-diamino-6-(hydroxymethyl)pimelic acid (6-HMDAP), a naturally occurring DAP acid analog (see above for Baldwin's shorter entrde to this natural product). 62 This methodology was then extended by Williams and Yuan to the synthesis of three stereoisomers of DAP (Scheme 56). 63 Thus, (+)-1 was alkylated with 4-iodo-l-butene to give 188, which after ozonolysis afforded the aldehyde 189. Aldol reaction with 190 provided the adduct 191. Subsequent Barton deoxygenation gave the dimeric species 192, which was hydrogenated as usual to give (2R,6R)-DAP (193) quantitatively. Mixing and matching the oxazinone templates 1 and 2 gave access to (2S,6S)-DAP and meso-DAP, a natural constituent of Gram-negative bacterial cell walls.

298

ROBERT M. WILLIAMS AND CAMERON M. BURNETT

Ph Phi--.

Ph O

3steps

Ph

_ Ph~,~? "-- CbzN

+

ph~[~. O ~

n-Bu2BOTf, NEt3 p h ~ . . . ~ O

4o,

c o, ooc

MeO/

[-..~,.H ,~C

8~

,., //0 HO H 0~_c~ ~/ ~ .-. //0 HO H 0\\ ,-, u .-" "-' u . w Ph~ ~ .....~' ~'"Ph + P h ~ ~ ~'"Ph ~,--N.N., \ N<. ~--N ~ H N~. Ph" L,DZ OMe Cbz "Ph -1 "25 ph/ Cbz C)Me Cbz "'Ph 183

PhOCSCI NaN(SiMe3)2 "THF -78 ~ RT 38%

184

S 05 ~ " ~/ ~ ~ Ph ph/

o

!,,H ~--O / i r . , ~ j ' N-\- ~

N . Cbz OMe Cbz

n-Ph3SnH,AIBN ~ .... Ph toluene / reflux

... "Ph

o Ph

.... Ph ph/

(81%)

Cbz OMe

185

I

(81%)

"'Ph

186

H2, PdCI2..._H O 2 C ~ O 2 EtOH-THF H2N

Cbz

n

H NH2

HO2C

1.48% HBr, h 2. <--.~Me

OMe

H2N

NH2 OH

EtOH, A

91%

187

CO2H

6-HMDAP

SCHEME 55. Asymmetric synthesis of 6-HMDAE

.<

o

_Ph

Ph LiHMDS Ph,, O CDZN.v~o THF / HMPA -78 ~ -> r.t. 47% (+)-1

Ph

Ph.,,T/~ O CbzN

03, MeOH Ph,,,~,... ph,,,.~O CH2CI2 CbzN. ~ + / = CH2CI2; CbzN/...~ -78 ~ Me2S H'" O OBBu2 55-62%

188 ~

79-84%

18

O

190

Ph O

Ph Ph,, A 1) H2 (60 psi) ],/ "O PdCI2 (2 equiv) ,,.,, ~LO 1. NaOH, CS2, Mel CbzN O 2:5 EtOH / CH2CI2 (0.012 M) _

Ph,,,

CbzN..

H

2. n-Ph3SnH,AIBN toluene/ reflux

-

CbzN~.O Ph

( 2 ) \ c

62-83%

CbzN.~O

~

O ~ , re-flux ~quant. 2 steps 193 (R,R-DAP)

Ph SCHEME 56. Asymmetric synthesis of 2,6-diaminopimelic acid.

8

299

DIPHENYLOXAZINONES Ph

v~

o

(-)-2

I

LiHMDS ~

O -78 ~ -> r.t. 58%

Ph

Ph

~

ooc 0o

03, MeOH Ph = I Me2S

194 ~ Ph

Ph Ph CH2CI2 O Ph,,,1 ~ O = ~ + -78 ~ C b z N / , ~ O B B u 2 56%

195 H/~.O Ph

HO

~

2. n-PhaSnH,AIBN OH

CbzN

O

Ph

toluene/ reflux

(

....V

EtOH/THF BocHN c

CbzN~ O

O -O-

....(

NH3

72%

+

198

Ph

SCHEME57. Asymmetricsynthesisof a differentiallyprotectedmeso-DAP derivative.

Employing the N-Boc-oxazinone ( - ) - 2 as one coupling partner in this protocol gave access to a differentially protected chiral form of meso-DAP (198, Scheme 57). This substance should provide access to numerous stereochemically defined peptides of DAP of biological significance. Interestingly, Williams and Yuan found that the differentially protected dilactone 197 could be selectively ring-opened with HC1 in dioxane to give, after esterification, amine 199 (Scheme 5g). 64 The selectivity stems from the protonation of the N-Boc-deprotected amine, making the adjacent lactone linkage more prone to hydrolysis. Oxidative bibenzyl cleavage furnished amino ester 200. Coupling of the free amine with either enantiomer of a protected glutamate derivative (201) yielded dipeptide precursors 202. Catalytic hydrogenation removed the chiral auxiliary as well as the glutamate-protecting groups, and HC1 in dioxane cleaved the methyl ester to give the 7-Glu-DAP dipeptide 203. DeMong and Williams conducted a stereocontrolled synthesis of (2R,3R) and (2S,3S)-[3-hydroxyornithine (205), beginning with the aldol reaction between ( - ) - 1 and protected 3-aminopropionaldehyde (Scheme 59). The resultant alcohol 204 was obtained as an 8:1 mixture of diastereomers, from which the minor diastereomer could be removed by recrystallization. High-pressure hydrogenation over palladium chloride, followed by neutralization with ammonium hydroxide, yielded

300

ROBERT M. WILLIAMS AND CAMERON M. BURNETT

Ph

Ph

h

", r

O

Ph Ph*~o ~0 t-BOCN /"

Ph OH ,,~ Ph /CO2Me diox. HCI 1. conc.

BocNIv ~O+cbzN v ~ o (-)-2

C l ~ l

HN'~"(

2. CH2N2 Cbz N

(+)-1

P

Ph 197

CH2CI2,MeOH

199

Cbz.N / ~ O

77o (two steps)

~ p

(

1. Pb(OAc)4

5

O 65% 0

H2N"-,,/Co2Me

O Ph

Ph

200

o

CO2H 201

COzH Et3N O ~ \\ N31P( OPh OPh 58%

~ NyCO2Me O ( CH2CI2/MeOH \ 2. conc. HCI,dioxane Cbz .-: 3. Dowex50x8-200 'N " O 92% ~ ~ O 202 Ph /

RT +H3N NH3+ 203, y-D-GlutamyI-L-meso-DAP

Ph

SCHEME58. Asymmetric synthesis of 7-glutamyl-L-meso-DAE

Ph

Ph@o

Ph Bu2BOTf, Et3N Ph~N.~.. CH2CI2 O -78 ~

1. H2 (78 psi) OPdCI2 (2 equiv) ,. I~,~~. 1:2 EtOH / THF (0.028 M) H3 O

O NHCbz Cbz ~.~ "O C b Z N ' v ~ O H...L[v.J O (N~:~

(-)-1

69% 8:1 dr

204

z

2. NH4OH, pH6

OH

68%

(NHOI-H 9 CI 205 (2S, 3S-[3-hydroxyornithine)

SCHEME59. Asymmetric synthesis of ~-hydroxyornithine.

(2S,3S)-[3-hydroxyornithine; the enantiomer was obtained simply by beginning with (+)-1. 65 Scott and Williams synthesized two of the four possible stereoisomers of [3-hydroxypipecolic acid as part of their investigations into elucidating the relative and absolute stereochemistry of the natural antitumor antibiotic tetrazomine. 66 Using the boron enolates derived from 1, both the (2R,3R)- and (2S,3S)-stereoisomers were obtained (Scheme 60). Their total synthesis subsequently revealed that tetrazomine was constituted from the (2S,3R)-isomer. 67 DeMong and Williams also examined Mannich-type reactions. The boron enolate of ( - ) - 1 reacted with imine 211 in low yield to form

8

301

DIPHENYLOXAZINONES

Ph 1. Bu2BOTf, Et3N

CbZN.v~ O

2. O CH2CI2, -78 ~ .,.,~. NCbz H -"[~""--.~ " Ph" Ph " 206 69%

(+)-1

~ ?

O H _OH ~ "-

Ph,,,lr~O

.~

1 03, CH2CI2,-78 ~ " 2. Me2S, 25 ~ 69%

.,.~,,~,fNCbz 15h 207

Ph"

O H OH H2 (1 atm), 5%Pd/C .~

O ~ . 1/ I -

CH2CI2 66%

O H .OH H2 (50 psi), PdCl2 ~

Ph'"'~"Y" N'v'~F'h 208

O H OH ? I L l - I~ O

~

HO~-

.N..~

EtOH, THF, 25 ~

209, (2R, 3R)-3-hydroxypipecolicacid

92% (er >99.5 90.5)

Ph Ph

O

4 steps =_~

CbZNv~ O (-)-1 25% overall yield (er >99.5"0.5)

,."~OH

H H_

O OH HO/[I,,,~

~HOHH_ O

HN.v.p

N

210, (2S, 3S)-3-hydroxypipecolic a c i d

(-)-tetrazomine

SCHEME 60. Asymmetric synthesis of the two anti-stereoisomers of 13-hydroxypipecolic acid.

(-)-1

1. LHMDS 2. Me2AICI 3. ~lBn TBSO/., ~ 211 "THF, lh., -78 ~ 60%

1.7% HF in MeCN

Ph O Ph CbzNO ~ .L. -.

Ph Ph O CbzNv . ~ O

Ph J. SMe Ph BocN,.j,,. NHBoc y" O CbzN- ' ~ ' O TEA, HgCI2 ~ TBSO/"'v'~NBn DMF, rt, o/n 212 67% 213 BocHN" " NBoc 3.3:1 dr (1H-NMR) Ph p h~..Jl... O H2NJ"--.--" 1. H2, PdCI2 " O--'OH DIAD,PPh3, THF CbzN.....~ O~-115 psi, 4d. _-Ph Phi..,,~ 7" O CbzN~ - - . . O _TBSO' ' ' ' ' v ~ NHBn

-

HO.,.-..,,v...~,NBn BocHN"~NBoc 214

0 ~ 15min, rt lh. 87o/o

I~ N B n L...N/~ Boc 215

reflux, 1.5h.

~N.. ~

2 HCI N

NBoc

95%

H

216, (2S,3R)-capreomycidine

SCHEME 61. Asymmetric synthesis of (2S, 3R)-capreomycidine.

a boron chelate that was resistant to hydrolytic cleavage. It was found that ( - ) - 1 could be converted to the corresponding aluminum enolate by treatment with LHMDS and transmetallation with dimethylaluminum chloride (Scheme 61). Mannich-type reaction with the protected imine 211 gave the protected amino alcohol 212 as a 3.3:1 mixture of diastereomers. Interestingly, the diastereoselectivity of this process provides the threo or

302

ROBERT M. WILLIAMS AND CAMERON M. BURNETT

I-

HPh

-1 /

Ph ~ ,: 0 r..''~OTBS |

|CBzN-~z:E~.. O AI,,,Me I

/

L

"ue/

A

an

__1

FIGURE 2. Zimmerman-Traxler transition state for the Mannich-type reaction of 1.

syn relative configuration, opposite to the selectivity seen with aldehydes. After conversion to the protected cyclic guanidine 215, high-pressure hydrogenation over palladium chloride and cleavage of the two N-Boc groups with hydrochloric acid yielded (2S,3R)-capreomycidine (216). 68 The syn-diastereoselectivity in this Mannich-type reaction can be explained by the Zimmerman-Traxler transition state (A) between the E-enolate and the E-imine as shown in Figure 2. DeMong and Williams then utilized the protected serinal species described above in Scheme 27 in conjunction with the asymmetric synthesis of capreomycidine (216, Scheme 61) to achieve a convergent and concise total synthesis of capreomycin IB, a natural tuberculostatic agent (Scheme 62). A Hofmann rearrangement was exacted on a late-stage pentapeptide in order to transform an asparagine residue into a diaminopropanoic acid residue. The serinal residue was utilized to construct the unusual 13-urea-containing unsaturated amino acid. This synthesis, which is the only modem synthesis reported since Shiba's first synthesis of NHCBz

1.1.5 M MeOH-HCI, A (97%) 2. N-methylmorpholine, CH2CI2 (92%) CBzHN _:..,~CO2H

BocHN ___y~/ONSu

:"" N 217 H2

BocHN/-...v j

O 218

3.2N NaOH, THF, M eOH, rt (99%)

I....LCO2H BocHN _ : ~ . . ~ N H

(

j

0

.

NHBoc

BocHN ..,.,..CO2H

1. EDCI, HOBt O Me Et3N, CH2CI2 (82%)BooHN _ . , ~ H'"~CO2H 219 ~CONH 2 H H2, 10% Pd-C :-...CONH2 B BnO2Cy,,.NH2"HCI 2. MeOH (98%) 220 i~leH .

+

SCHEME 62. (a) Synthesis of fragments A and B for the total synthesis of capreomycin IB.

8

DIPHENYLOXAZINONES

303 CbzHN....,fCO2H

1. TiCI4, Et3N, CH2CI2 (EtO)3C H, -78~176 (85%) 2. H2 (60 psi), Pd(OH)2 THF, EtOH (99%) 3.3N EtOH.HCI, A (96%) :.._

Ph Ph,,,l~O CbzN v . ~ O

4. A, EDCI, HOBt, NMM CH2CI2, (91%) BocHN ~ 5. H2, Pd(OH)2, EtOH (99%) _:

(+)-1

6

o0.o.t,7

.,

7. PhI(O2CCF3)2, py. MeCN, H20, DMF (70%)

;

1. H2, 10% Pd-C, EtOH 2.1N LiOH, EtOH 3. EDCl, HOAt, DMF CH2CI2, 20% (3 steps)

"O ..~ HN O BocHN _..r...---...T~N ~ .... O

4. 99% formic acid 5.2N HCI, acetone, A; urea, (50-90%)

II

HN~CO2Et

O

~"NHBoc

223

II

o

O O

.. CO Et

EDCI, HOBt DMF, THF, 0~ 86-89%

EtO..L o Et 221

, H H CbzHN ~ N...I ~ ) H,,,~. H BocHN 'rl" ~,, e

-

HN N

HBoc

H NH,.~,' " ,NH '~'NH'HCI

-

~ - NH L , ,.~---. 222 ' ~ J NH.HCl H 1 step from 216 (56%)

H2N% H , . J [ H ,,1 / -...~ N -'H'e BocHN O T ~

EtO" "OEt

O

Me h

XNH

O

NH2

H "UHN/%'O,

HN"" H

capreomycin IB

SCHEME 62. (b) Asymmetric total synthesis of capreomycin ]B.

capreomycin IB in 1977, required 27 steps and proceeded in 2% overall yield from (-)-1. Johns, Mori, and Williams have recently reported an interesting application of the diastereoselective aldol condensation reaction in an approach to quinine. 69 Aldol condensation of ketenesilyl acetal 224 with 225 produced the thermodynamic aldol adduct 226 in 76% yield. This possessed the correct relative and absolute configuration at C8 and C9 of quinine (Scheme 63). Further elaborations to the key piperidinone 233 set up a Pd-mediated SN2'-type cyclization that gave the quinuclidine having the desired relative configuration. The initial cyclization product proved to be unstable and the Pd-mediated reaction was quenched with diisobutylaluminum hydride to give 7-hydroxyquinine. This approach to quinine, which still faces the difficult deoxygenation at C7, is unique in that the C8/C9 stereogenic centers were installed in the very beginning of the

1"i21N

ROBERTM.WILLIAMSANDCAMERONM.BURNETT

304

Ph Ph -

CHO Ph NaHMDS, Ph MeO~ ...r TBSCl ..._ ' " ~ 1 + O "~ ~ ~ CbZN.v,~ O l THE,-78 ~ v CbzN/.~/L..OTBS ~ "N" 97% (+)-1 224 225

ph,,,./)~O

1. Pd-C, H2, EtOAc 2. ZnCI2, MeOH

H CO2Me B~

3. Pb(OAc)4 M e O ~ MeOH-CH2CI2 4. Pd-C, H2 (Boc)20, EtOAc 227 79% 4 steps OBn

Ph...., ~ O ~ o CbzN~OTES

1. TBAF THF, 0 ~

MeO~

2. TESOTf 76% 2 steps

226 (X-ray) H CHO DIBAL-H BocHN-~, .,OTES 1. B n O / ~ / ~ M g B r (81%; M e O ~

THF-ether,-78~ 2. Dess-Martinox. 3. DIBAL-H 50% 3 steps

228 OMs

Ac o, O.AP

OAc

BooHN--OTES 0H2012

n,,~"OC D '~''--~"

2. H2, Pd-C, MeOH MeO 3. MsCI, Et3N,CH2CI2 77% 3 steps

M e O ~

230

229 1. LAH, Et20

BzO

( H ~,'"OTMS

~. 2. TMSOTf,GH2GI2 2,6-1utidine MeO~ 3. i~'-..~-.~OBz 87% 3 steps

ES

232

1.1 M HCI, 0 ~ THF-H20 2. Swern ox. 90% 2 steps

n-Bu3SnF, tol. P(2-furyl)3 2. DIBAL-H 44% 2 steps 234

BzO~ , N ~ " ~-../ 0 ~--.,.,,/ ~ O H T ES MeO",~'~'/i I - -~----' 233

1. Pd2dba3,A MeO

OTES H

231

~OTMS

BzO TMSCI 59%

o N ' "~ ,OTES Nail BocN 'H~ THF, ~"--,, ... 99% Meu

~.

M e ~ H HO'~~N~ N~~ . ~ , , , O H

235, 7-hydroxyquinine

SCHEME63. Asymmetricsynthesisof 7-hydroxyquinine. synthesis, and features a novel C2/C3 C-C bond-forming strategy to access the quinuclidine ring system. V. Other Manipulations

The use of the oxazinones as glycine radical equivalents has been briefly explored. Substitution at either the nitrogen atom or the lactone carbonyl has also proven synthetically useful.

8

DIPHENYLOXAZINONES

305

A. OXAZINONES AS GLYCINE RADICALS

Williams showed that the bromolactone 20 could be reduced with Bu3SnD to give the corresponding syn-deuteriolactone with modest diastereoselectivity (Scheme 64). Subsequent hydrogenation gave chiral, deuterated glycine in 80:20 er. This protocol gives the stereochemistry opposite to that observed for palladium reduction of 20 under D 2 gas. z~

Ph O CbzN- ~ O 13r

1. Bu3SnD H3o NS ~O 2. H2, Pd/C

DH 80:20er

19

236

SCHEME 64. Asymmetricsynthesis of s-deuterium-labeled glycine.

Vederas attempted to synthesize meso-DAP using ~,[3-dehydrolactone 237 as an acceptor for conjugate addition of the 3,-radical species derived from the decomposition of 238 (Scheme 65). 7~ The initial adduct 239 was too sterically hindered to be reduced as expected by 1,4-cyclohexadiene, instead losing a hydrogen atom at the [3-position to give the unsaturated product 240 in low yield. O

-

Ph

Ph~o Ph \ BocN" "~O

238

~, 237

Ph

Ph

Ph~o

" Hcbz

Ph ~ " - -

CO2Me/2 =-

BocN~ < ' 0

Phil

32%

0 -

H9 BocN~ 0

k..~,,,NHCbz

K..~,,,NHCbz

CO2Me

241) CO2Me

239

SCHEME 65. Radical conjugate addition to an ~,13-dehydrolactone.

Kabat used the phenylselenolactone 241, prepared by alkylation of (+)-2, as a radical donor for conjugate addition to the ~,13-unsaturated glycine derivative 242 (Scheme 66). While the addition proceeded exclusively anti to the phenyl rings, the quench at the y-carbon exhibited poor diastereoselectivity, giving a 2:3 mixture of diastereomers 243:244. Catalytic hydrogenation of the auxiliary and protecting group removal gave the 2,4-diaminoglutamic acid hydrochlorides (245 and 246). 71

306

ROBERT M. WILLIAMS AND CAMERON M. BURNETT Ph

Ph,,,~,,~

H2,10% Pd/C ClH~ THF/EtOH,reflux; CIH~ "J O 6N HCI, reflux " 82% (2 steps) CO2H

.~ BocN AcHN~. -'~ Ph Ph , , , ~ O

LiHMDS, Ph THF;= P h , , , ~ O

~CO2Me 243 CO2Me 2 245 /NHAc 2 4 2 Ph "" CIH.H2N__/CO2H B o c N v ~ O PhSeBr B o c N . . ~ nBu3SnH,AIBN- Ph,,' ~ " 0 3 H2 10% Pd/C 84% O PhMe,80 ~ . ~ . THF/EtOH, reflux; CIH~ SePh BocN O (+)-2 74% 241 AcHN,,,~ 6N HCl, reflux 90% (2 steps) CO2H r 244 CO2Me 246

SCHEME 66. Asymmetric synthesis of 2,4-diaminoglutamic acids.

Ph Ph

O

RCHO,p-TsOH

H N . . v ~ O benzene, A

e.4.eL. s.~.4"~ ~'-O

=Ph

O

H2 / PdCI2/ EtOH

R1

002 R2

RI.~'N~o

THF, 25 ~ 40-60 psi . ., ".CO2Me R2= H MeO2C "CO2Me or 250

k__z, :

247

MeO2C

CO2Me

MeO2~ _

,

/

248, E/Z-~-endo

2. Pb(OAc)4,MeOH, CH2C[2 R2 = Me

SCHEME 67. Asymmetric synthesis of highly substituted proline derivatives.

B. OXAZINONES AS GLYCINE-BASED A Z O M E T H I N E YLIDES

Williams used the oxazinone template as a glycine azomethine ylide equivalent for the synthesis of highly substituted proline derivatives. 72 The oxazinone 247 is prepared by TFA removal of the N-Boc group from oxazinones 2. Condensation of 247 with various aldehydes proceeds under mild conditions forming the corresponding azomethine ylides 248 (Scheme 67, Table 8). Dimethyl maleate was found to be a generally reactive dipolarophile that underwent [1,3]-dipolar cycloaddition via the endo transition state to give the bicyclo[4.3.0] oxazinones 249. Formation of both E- and Z-ylides led to generally poor stereoselectivity at C7, but complete stereochemical control was observed at the other three stereogenic centers. Significantly, isobutyraldehyde gave a single diastereomer via the less sterically encumbered E-ylide. The cycloadducts resulting from saturated aldehydes could be subjected to catalytic hydrogenation to directly yield free amino acids 250 (R 2 = H).

8

307

DIPHENYLOXAZINONES TABLE 8 Preparation of substituted prolines (250)

Entry a b c d e f g

Rl

249 % Yield (dr)

Cleavage method (249-250)

250 % Yield (er)

H n-Pr i-Bu Ph 4-H2N-Ph 4-OzN-Ph 2-Furyl

71 32 (1.33:1) 52 (1:0) 70 (1.7:1) 71 (1"1) 71 (1" 1) 61 (1"1)

H2, PdC12 (0.04 M reaction concentration) H 2, PdC12 (0.02 M reaction concentration) H 2, PdC12 (0.02 M reaction concentration) HC1, MeOH/Pb(OAc) 4 HC1, MeOH/Pb(OAc) 4 HC1, MeOH/Pb(OAc) 4 H 2, PdC12 (0.02 M reaction concentration)

98 (>99.5:0.5) 93 (>99.5:0.5) 99 (>99.5:0.5) 57 (>99.5:0.5) 66 (>99.5:0.5) 56 (>99.5:0.5) 99 (>99.5:0.5) (R = THF)

Under these conditions, adducts with aromatic R l groups decomposed, and basic ring-opening resulted in extensive epimerization. Fortunately, an alternative oxidative cleavage protocol was employed for these substrates involving acidic methanolysis/ring-opening to the corresponding methyl esters. Subsequent oxidative cleavage of the amino alcohol auxiliary gave amino acid methyl esters 250 ( R 2 = Me). The C7 diastereomers could be separated by either column chromatography or preparative TLC. The C7 syn-diastereomers readily reacted under these conditions, while the anti-diastereomers gave the amino acid methyl esters in low yield and with attendant epimerization at the ~-stereogenic center. The furylpyrrolidine could not be obtained from 249 (R ~ = furyl); the furyl ring was saturated under catalytic hydrogenation conditions and decomposed under oxidative cleavage conditions. Wudl, Prato, and co-workers reported that N-alkylation of amine 247 and treatment of 251 with acid gave the corresponding azomethine ylide, which reacted with fullerene C60 to give the protected fulleroproline 252 in 26% yield (Scheme 68). 73 Ph

Ph~o Ph C1~OC8H17 P h ~ o 060 HN.v~ O Et3N,A = I / N v ~ o pTsOH= OC8H17 26% 247 251 Ph

PhyLo

N~'- O

252

SCHEME 68. Asymmetric synthesis of a protected fulleroproline.

308

ROBERT M. WILLIAMS AND CAMERON M. BURNETT

Sebahar and Williams reported a concise asymmetric synthesis of spirotryprostatin B, a cell cycle inhibitor isolated from Aspergillus fumigatus by Osada and co-workers. TM Reaction of oxazinone 253 with the masked prenyl aldehyde 254 gave the E-ylide selectively, in accordance with the results obtained earlier with isobutyraldehyde. The unsaturated oxindole 255 reacted as the dipolarophile to give the desired cycloadduct 257 in 82% yield (Scheme 69), setting four contiguous stereocenters in a single step. X-ray crystallography of the adduct revealed that the dipolarophile reacted in the exo-orientation from the [3-face of the dipole (256). Catalytic hydrogenation of the chiral auxiliary proceeded quantitatively, and further manipulations yielded spirotryprostatin B. Ph Ph,,,l~ O + Me\ J ~ ~

Me ,,,.,~OMe Swernox. ~u Me 89%~'~ Me Ph

Ha

254

J. ~

o

3A,,o,

2s6

II

820/0

9

-

Ha

4S ,

257

(x-ray)

o

~,.[ /~=0

-'N H isatin

Me HN -.

Me

EtO2C\

253

[I

eO4"

O H C v ~ ONe

p h , , , ~O. ~

Ph MePh. -~ MeO4 "'~"" "O 02etazomethineylideMe/~. . . . . N....~.._o [,,3]-dipolar Oi....~' "~'H" cycloaddition._ " ~". ~ ',,CO2Et

(Ph)aPCHCO2Et

diglyme, A

~'abnet~i-oeX~

te

Me MeO'7( H 1. D-Pro-OBn, OH.H,,.~~ M~ \ ....,("N""*CO2H'f;,U BOP,Et3N, Me. I ,y H2, PdCl2 MeCN (74%) MeO---~--...N J. ---~ O-.~~., n -" Me O ~ O "'CO2Et 2. H2, Pd-C, etOH THE, etOH 60 psi, 36h H N ~ 3. BOP et3N, MeCN Hlki'..~...""C02 St 99% ~ 94% 2 steps 258

259 __

1. Lil, py., A

TsOH O- H"I ~ 2 equiv Me " ~ \N / = --- N tol., A Me~ 0_~"'~ " N' " ~ H 0

(70-74%) 2. DCC, DMAP

8 _80o,o

,-,o-,,,y 260 ~

34-43o,0

S BrCCI3,A 3. NaOMe, MeOH

spirotryprostatinB SCHEME 69. Asymmetric total synthesis of spirotryprostatin B.

8

DIPHENYLOXAZINONES

309

Similar methodology was used to synthesize spirotryprostatin A. 75 The unstable exo-methylene oxindole was prepared in situ from 261 and reacted as above to give the cycloadduct 262 as well as a stereochemically homologous isomer that had suffered elimination of methanol (Scheme 70). Catalytic hydrogenation and thermodynamic epimerization of the ~-methine gave acid 263. The absence of the carboethoxy group reduced steric bulk and allowed lower catalyst loading and hydrogen pressure, compared with the procedure required for the hydrogenation of 257 for spirotryprostatin B. Coupling to D-proline, hydrogenation, and pentacycle formation gave 264. Elimination of methanol under acid-catalyzed conditions yielded spirotryprostatin A in 8% overall yield along with 9-epispirotryprostatin A. Me Ph Ph OHC.v~Lo Me Me Ph. i Me Ph 254 Me aeO4 "'~'-"0 aeO~2/~ H '"1,/'~O mol. sieves, tol. Me/~. N~...~O 1. H2, PdCl2 Me~ \ N. ,,CO2H HN"v~O TMS....] "~ O~"'H THF, EtOH~ O~" /

253 MeO/''''~" ~ H ~ ~''

O

H262 N~'~

2" HOAc n-prcHO

263 HN. ~

44% OMe + -20% MeOH elimination

261

OMe

.e

1. L-Pro-OBn, MeO-_-~ "''jN~ ...."'~O BOP'Et3N'MeCN),. Me O,k/~~. "

p-TsOH, H20

2. H2, Pd-C, EtOH 3. EDCl, Et3N, MeCN

3A, tol., 100 ~

HN~ I

~--

O... Me. Y IN Me~ ' " N ;'""~O

HN~ - - - ~

/I

264 OMe

OMe spirotryprostatin A (8% overall) + 9-epi-spirotryprostatin A

SCHEME 70. Asymmetric total synthesis of spirotryprostatin A.

Sebahar and Williams also conducted an investigation into the effect of varying the aldehyde side chain on the dipolar cycloaddition with ethyl oxindolylidene acetate (255). 76 In addition to the E-~-exo products 265, other products isolated resulted from endo dipolarophile approach (266) or from the Z-ylide (267) (Scheme 71). Bulky aldehydes favored the E-ylide and gave the standard cycloadducts 265. However, isobutyraldehyde (Table 9, entry d) gave all three products with an 8.6:1 ratio of diastereomers with the desired regioselectivity. Less branched aldehydes gave

310

ROBERT M. WILLIAMS AND CAMERON M. BURNETT

Ph RCHO 3,A.,PhMe

Ph

Ph,,.1S~O

Ph.

Ph

Ph

Et02C.

HNv-~ 0

@o

253

255

H

265

266

267

SCHEME 71. Dipolar cycloadditions of 253.

TABLE 9

Spirooxindole pyrrolidine cycloadducts Entry a b c d e f g h i j

R H CH2OBn CH2OBn i-Pr i-Pr i-Bu i-Bu HRCCMeR(OMe) H2CCMez(OMe ) p-MeOPh

Temperature 265 % Yield Reflux Reflux 60 ~ Reflux 60 ~ Reflux 60 ~ Reflux 60 ~ Reflux

28 44 54 43 74 84 86 29 82 60

266 % Yield

267 % Yield

dr (265:266)

ll 14 8 11 6 1 0 0 1 0

0 0 0 5

> 20:1 >20:1 8.6:1 >20:1 >20:1 >20:1 >20:1 >20:1 >20:1

trace 0 0 0 0 0

high diastereoselectivity but moderate e x o selectivity (Table 9, entries a-c), and bulkier aldehydes gave high diastereo- and facial selectivity (Table 9, entries f-j). Most reactions showed improved regio- and diastereoselectivity upon lowering of the reaction temperature from reflux to 60 ~ p-Anisaldehyde, though, required refluxing conditions to react, presumably because the electron-donating methoxy group slows formation of the azomethine ylide. Conversion of the cycloadducts 265 into the corresponding amino acids was generally straightforward, requiring high-pressure hydrogenation, and was followed by derivatization to the methyl esters 268 to aid characterization (Scheme 72, Table 10). Unfortunately, the p-anisaldehyde cycloadduct 265 (R = p-MeO-Ph) gave only a small amount of product under these conditions, and elevated temperatures or pressures gave decomposition. Pearlman's catalyst aided somewhat, while acidic conditions were found to give the desired amino acid in tolerable yield.

8

311

DIPHENYLOXAZINONES

Ph H Ph,, O 1. H2(70psi) R,, N...,CO2Me o R ~ O Peel2(1.0equiv.) O~ ''C ,, ~. O2Et [ '-. "'C02Et 11 MeOH/THF(0.05M) HN HN,~ 2. TMSCHN2 "~ 265

268

SCHEME 72. Cycloadduct auxiliary removal and methyl ester installation. TABLE 10 Conversion of cycloadducts 265 into methyl esters 268 Entry a b c d e f

Substrate

Method

% Yield

265a 265f 265h 265j 265j 265j

H 2, PdC12 H 2, PdCI 2 H 2, PdC12 H 2, PdC12 H 2, Pd(OH) 2 H 2, Pd/C, 1 M HC1

93 89 85 5 25 59

Attempts to cleave the bibenzyl auxiliary from 265 via oxidative conditions resulted in decomposition, presumably due to competing oxidation of the electron-rich oxindole. Schreiber and co-workers have extensively adapted this methodology to a solid-phase combinatorial platform as part of their diversity-oriented synthesis program. 77 Oxazinone 253 (or its antipode) is first reacted with resin-bound aldehydes with magnesium perchlorate in the presence of trimethyl orthoformate (Scheme 73). This is followed by Pd-catalyzed alkynylation, amine coupling, and N-acylation to give the tagged library pool 270. Cleavage of the resin from 270 allows direct assays, and some compounds, including 271, were found to have promoter activity for the actin polymerization inhibitor latrunculin B. Ahrendt and Williams used the oxazinone azomethine ylide in an approach to the manzamine-type alkaloid nakadomarin A (Figure 3). 78 The E-ylide formed from oxazinone- 253 and mannitol-derived aldehyde 272 reacts with the unstable alkene 273 in a E-~-endo fashion to give the cycloadduct 275 in moderate yield (Scheme 74). Catalytic hydrogenation removed the biphenyl auxiliary in excellent yield, and further elaboration gave ester 280, containing the ADE-ring core of the hexacyclic natural product.

312

ROBERT M. WILLIAMS AND CAMERON M. BURNETT

Tag-1 ~

Ph 1. amine PyBOP N-acylating i-Pr,, ,i-Pr + Ph'"l'~O 1 Mg(CIO4)2 1. alkyne Si'o__~__CHO HN'v~ O HC(OMe)3'PY. Pd(ll), Cu(I) i-Pr2NEt reagent 2.Tag-3 2.Tag-4 2.Tag-2 269 253 Ph Tag-1 Ph Ph,,,W-o TAG-. ~ Ph 2 y ~ % i-Pr\ i-Pr "'T/-\O .... , N 0 T A G ~ ~ ~ / ~ / si[ O-~,./N ,'~, O HO/~/O ~ ....i,,fO TAG-4 O HO O'~ N ' ~ H N v ~ 270

44

~/;'~,~,,,. a1

MeO~O

HO--~--CHO CHO O~ CHO + ~~/~]/CHO ~ . OHN~O OH OH " OH "[" CHO OH L,,,./O L,.,,./O L,,,/o I

271

o../~ o

SCHEME 73. Library of 3,520 spirooxindole cycloadducts.

FIGURE 3. Structure of nakadomarin A.

C. D I R E C T N I T R O G E N S U B S T I T U T I O N

Recent reports have explored substitution of the oxazinone nitrogen prior to auxiliary removal to give heterocyclic products. Funk alkylated the sodium enolate of ( - ) - 2 with his bromomethylvinyl ketone equivalent 281 to give 282 (Scheme 75). Thermal unmasking of the vinyl ketone 283 was tbllowed by removal of the N-Boc group to yield the free amine, which suffered conjugate addition in situ to give the 6,6bicyclic system 284. Stereoselective reduction of the ketone, followed by catalytic hydrogenation, yielded (2S,4R)-4-hydroxypipecolic acid 2 8 5 . 79

8

DIPHENYLOXAZINONES

o

313

o

Ph' "O "O ~ M e +

r~h,,,

~_

Ph"I"~'-"N/HH +

I'

13oc

253

272

o--o

[.

Ph,,, .O.T ,/_O . - - , Boc

]~ -'~O~=(~-~N-B~

H20"--/,/iPh'"''" N/'~M'~l---/~//O_.~ eO 1_

Me

~

1 44-60% PO.... O ' -

Me'~Me 275 Boc 20 mol%Pd(OH)2/C / " ' - - - ~ O /N--,~ 1. TMSCHN2 _k./ MeOH H2 (1 atm) ~ C l H. /---~'"\\ = MeOH/EtOAc(1:1) O NEt3' CH2CI2 O , ~ ~ " ,~, O 2.SnCI2 92% Me.C) "H ~ "CO2H 0~ - r.t. . ~__/"H ? "CO2H CH3NO2/H20 Me 276 75% Me L) v v -.~ 80% Boc Boc 277 Boc 'N '~'"--~k ~l , l Mes 'NA ~....~ MeslCN"'r/N" H /--~'"\\ PPh3, imidazole,12-~ H '_~/.." ~,. "O 5mo1% .Ru=~ HO~...~,.. 0 toluene,80 ~ ///'- "1',!" "'r,~ ,,I~ Cl 2Me "CO~e 60O~o l . L .... '0~2 h O ~ /// v v "O CH2CI2,reflux 278 279 280 65% \

273

h

274

. Boc

.~,.. ],

/-. ~'~

SCHEME 74. Dipolar cycloaddition approach to nakadomarin A.

Ph

P~"~O

NaHMDS HMPA, THF, -78 ~

B~ (-)-200

~ B r 281 86%

-v-

Ph BH3, THF TMSOTf, P h ~ o -78 ~ lutidine; ~ L____.~ MeOH " 59%

_ "[~:

0

O 284

83%

Ph Ph~o BocN _,~L.. O 150 ~ [ ' f ' ~ / " 282 v

Ph Ph ~JL.. 0 BocN _ @

PhMe 95%

283

O

o.. ,.o

H2 (50 psi) 5% Pd(OH)2 (20% w/w) EtOAc (0.05 M) 97%

0

~ H

-

OH

OH

.....~o 285

SCHEME 75. Asymmetric synthesis of 4-hydroxypipecolic acid.

Lee and co-workers alkylated (-)-2 with 5-chloro-l-iodopentane to give chloride 286 (Scheme 76). The N-Boc group was removed and cyclized to give the 7,6-bicyclic system 287. Catalytic hydrogenation and reprotection of the amine gave the protected (2S)-2-azepanecarboxylic acid 288. 80

314

ROBERT M. WILLIAMS AND CAMERON M. BURNETT

Ph

NaHMDS

o BocN v

Ph

h@'o "O HMPA/THF"-

73%

(-)-2

BocN

O

,~

C~

87%

45%

286

Ph

Ph~

H2 (50 psi) PdCI2 (30%w/w)

O

O 2:1 EtOH:THF (0.015 M) 96%

Boc20

Boc..,/\'?H N A,,.O

60% 288

287

SCHEME 76. Asymmetric synthesis of protected 2-azepanecarboxylic acid.

O O Ph [ ~ ' . ~ O E t Ph~ " ' O CH2 N2289 ,. r/N-,..~oCu(acac)2,.

Ph

Ph~o

eto cu':

/

HN-..~ O 247

Et3N 79%

N~O 290

Ph Ph~./]L...O EtO2C,,,~ ~::::O O~N----/

293

quant. ""/'/~CO2Et O 291

OEt

O/ ~/ 292

Ph THF= 97%

~N'[

.... z~OH

294 H

SCHEME 77. Asymmetric synthesis of a hydroxylated pyrrolidine.

Saba used the oxazinone in a highly unusual way as a template for synthesis of hydroxylated pyrrolidines. Conjugate addition of amine 247 to vinyl ketone 289 gave diazo species 290 (Scheme 77). Decomposition of 290 with copper(II)acetylacetonate yielded, after a Stevens rearrangement (291-292), the 7,5-bicyclo product 293. Reduction with LAH gave the tetraol 294 in excellent yield for the sequence. 8~ D. C A R B O N Y L M A N I P U L A T I O N A N D P E P T I D E ISOSTERES

Homologation at the oxazinone lactone carbonyl group has significantly expanded the scope and utility of these glycine templates and has

8

315

DIPHENYLOXAZINONES

Ph Ph ~ : O

NaN(TMS)2

~O HH IPh Ph -

THF, -78 ~ Me OTf

CbzN~ O _ Me j

Me./~75O/o

"~Me

Ph OTBS 1. DIBAL-H, CH2CI2 Ph.d.... H? J]... -78 ~ OMe ~ CbzN~ / ~ O A c ZnBr2, CH2CI2 2. Ac20, DMAP ; 58% Et3N' CH2CI2 Me.,~ ,,,.._

CbzN - . ~''~'O

(-)-1 Ph Ph ~ - O

+

295

Me 296

82%

Ph Ph ~ : O

OH O

1 KOH, dioxane MeOH

CbzN..~, ..../CO2Me Me\.r,"/: 297

CbzN_.~,.~~CO2Me H Me,,,~ 298 - 1 : 4, SYNANTI Me Me

Me.

2. Li~ NH3, THF n-BuOH 69 %

H

Me statine

SCHEME 78. Asymmetric synthesis of statine.

provided access to a number of peptide isosteres and related compounds. The original example of Williams, Colson, and Zhai began with alkylation of ( - ) - 1 with 1-isobutyl triflate to give lactone 295 (Scheme 78). Reduction to the lactol and acetylation gave the hemiacetal 296 in good yield as an equimolar mixture of diastereomers. Treatment of this mixture with the ketenesilyl acetal of ethyl acetate gave the masked 13-hydroxy esters 297:298 in moderate yield as a 1:4 mixture, favoring the desired relative and absolute stereochemistry. Hydrolysis of the methyl ester allowed dissolving-metal reduction of the auxiliary, giving the [3-hydroxy,/-amino acid statine. 82 Further investigation of the lactone carbonyl homologation revealed that Sakurai coupling of allyltrimethylsilane with lactol hemiacetals 300 proceeded in good yield with a slight level of stereoselectivity (Scheme 79). 83 _Ph Ph,,,.~O f CbzN O ~ '.L. '~

P_h

NaHMDS RX

1) DIBAL-H C H2CI2 -780C

Ph,,,~,~ 0 D,

THF_78 / HMPAoc

CbzN DMA_ Et3N' CH2CI2

R

(+)-2

299

Ph Ph,,,~jk.. O

CbzN.]~~-~

Ph Ph, "',t'/~O : == C b z N - - ~

R 301, R = H: 98% R = Me: 61% >99:1 d.r.

Ph

Ph,,,1~

R

302

R

BF3 9 OEt2, MeCN ~ SiMe3 -15 ~ (R = H) OAc -40 -> 0 ~ (R = Me)

300, R = H: 78% 68:32 d.r. R = Me: 76% 3:2 d.r.

0

Li/NH 3

+ OH O H 3 N ~ OOH EtOH/THF" R 303, R = H: 81% R = Me: 80% >99:1 d.r.

SCHEME 79. Asymmetric synthesis of hydroxymethylene peptide isosteres.

316

ROBERT M. WILLIAMS AND CAMERON M. BURNETT

The alkene was ozonized and oxidized to the [3-hydroxyacid 302, allowing dissolving-metal removal of the bibenzyl auxiliary to give the hydroxymethylene peptide isosters 303. Alternatively, a hydroboration-oxidation protocol yielded protected 7-hydroxy-]3-aminobutyric acid 302, whose dissolving-metal reduction gave hydroxyethylene peptide isostere 303 (Scheme 80). Formation of the unsubstituted hemiacetal and displacement with allyltrimethylsilane as previously gave alkene 306, which was converted to the acid 307 (Scheme 81). Hydrogenation at high temperature and pressure gave (S)-7-hydroxy-]3-aminobutyric acid ((S)-GABOB, (S)-308a),

Ph

Ph ~

Me 301

Li/NH3

' CbzN...]/~.'~OH Me O 302

~ H3N 80% >99:1 e.r.

OMe 303

SCHEME 80. Asymmetric synthesis of hydroxyethylene peptide isosteres.

Ph CbzN~.,r, ,'Ph

.o

L~ O

Ph CbzN--~,, ,Ph L..../,

RLi/CeCl3 THF,-78~ (45-73%)

(+)-1

-

~/--SiMe3 TiCI 4, 0H2012

.._ " -

R'

-78~ or BF3-Et20, MeCN

Ph ~

_Ph Me3Si"" I (~ /~

'"r CbzN~

304a R = H, R'= Ac -15 to -20 ~

R

305

304b, R = Me, R'= H

Ph

304c, R = n-Bu, R'= H

Ph C b z N ' ~ l ,,''Ph 1.03, MeOH/CH2CI 2, ~ Ph 1. 120 psi H2, PdCI2, O -78 ~ then M e 2 S CbzN "'t"" THF/H20, 75-80 ~ -

2. PDC, DMF, rt, 77-91% 306a, R = H (98%) 306b, R = Me (65%) 306r R = n-Bu (67%) H2NHO~ R ~.. v ~"OH

308a, R = H 308b, R = Me 308c, R = n-Bu

HO2C

2. DOWEX 50WX2-100 93-99%

307a, R = H 307b, R = Me 307c, R = n-Bu

t. aq. CH20, 10% Pd/C 80 psi H2, rt (97-99 %) 2. Mel, DMF, rt 3. Amberlist IRA-400 (HO-) (68-85%)

+ u HIiOR ,O, Me3N ~ ~ \ c ~

309a, R = H ((S)-(+)-carnitine) 309b, R = Me 309c, R = n-Bu

SCHEME 81. Asymmetric synthesis of (S)-(+)-carnitine and congeners.

O

8

DIPHENYLOXAZINONES

317

whose N-permethylation gave (S)-(+)-carnitine (309a). 84 Application of this protocol to substituted lactols (304b,c) produced the tertiary alcohol derivatives 308b,c and 3 0 9 b , c . Alternatively, ketenesilyl acetal addition directly to the oxazinone (+)-1 produced O-silyl lacto1311 (Scheme 82). 85Treatment with boron trifluoride effected elimination to 312, which was diastereoselectively hydrogenated to produce 313 (dr = 94:6). Hydrogenation gave (R)-GABOB 314, whose N-permethylation gave (R)-(-)-camitine. Ph Cbz- N ~.~,.,Ph L~o

OSiMe2t-Bu F EtO O -7 /~oet ,H H ~,._.2 SiMe2t-BuI "~ /Ph.... I ~ - '' O ~) (..~I

/

O

-20 ~

(+)-1

CH2CI2 0 ~

O O

312

Os, e

85%

t-Bu

bz" N L

Ph BF3.Et20 CbZ.N.~r,,,Ph

95%

I

Ph Cbz-. N/~1,,,, Ph

EtO""~O 310

Ph H2, PdCl2, EtOH ,,.. HCI.HN~r,,Ph HCI, rt, 99%

dr=94:6

311

_J

L~o

O~

313

1.1M NaOH, THF, rt, 12 h 2.1M HCI (till neutral) 3. H2, PdCI2, 75-80 ~ 3 h 4. DOWEX 50WX2-100 91% 4 steps

OEt

H2N

OEt aq. CH20 OH O 1. CH31,DMF, rt OH O 10%Pd/C,, M e 2 N ~ = OH OH 2. Amberlist 80 psi H2, rt IRA-400 (OH) 314 99 % 315 83% 2 steps

Q OH O Me3N~oQ (R)-(-)-carnitine

SCHEME 82. Asymmetric synthesis of (R)-(-)-carnitine and congeners.

The reaction of hemiacetal 316 with crotylsilanes was examined by Aoyagi and Williams (Scheme 83). 86 Yields of the coupling products 317 and 318 ranged from moderate to good (Table 11). The stereochemistry at the former carbonyl carbon was assigned on the basis of NOE experiments. The methyl-substituted stereocenter was not so easily assigned, and further derivatization was required. Ph

1"

Ph

~

CbzN ~ - ~ r , ^c,.,,.,

_ _ H

conditions

-15 ~ 316

Ph

"~ C b z N ~ anti

317

lVle

~

~

+ CbzN

syn

318

SCHEME 83. Coupling reaction of 316 with crotylsilanes.

Me

318

ROBERT M. WILLIAMS AND CAMERON M. BURNETT TABLE 11

Lewis acid mediated coupling reactions of hemiacetal 316 with crotylsilanes Entry

Crotylsilane SiMe3

Lewis acid

% Yield

317:318

B F3.OEt 2

96

74:26

48

74:26

a

Me~

b

Me~SiMe

3

TIC14

c

Me~ S i P h

d

Me

3

BF3"OEt2

54

84:16

/---SIMe3

BF3.OEt 2

93

38:62

e

Me~

SiPh3

B F3.OEt 2

70

37:63

f

Me

BF3"OEt2

81

39:61

/~TBS

An ozonolysis-reduction-mesylation protocol converted alkene 319 to the mesylate 320 (Scheme 84). Hydrogenation and intramolecular ringclosure yielded a highly polar pyrrolidine, whose isolation was aided by protection of the amine to give pyrrolidine 321. Surprisingly, the catalytic hydrogenation had cleaved both the Cbz and bibenzyl groups from nitrogen, while the C-O benzylic bond was untouched. Dissolving-metal conditions yielded the protected 3-hydroxypyrrolidine 322. _Ph

Ph,, <

Ph

O

~

C b z N ~

ph,,I~ ~

1) H2 (60 psi)

O

10% PdlC(100%w/w) Ph " Ph--~-O~

~ CbZNvJ~oMs

319

I:2EtOH/THF (0.022 M) 2) Boc20, CH2CI2 63% (2 steps)

320

Boc

LilNH3

HO ~N

t-BuOH / THF 79%

321

Boc 322

SCHEME 84. Asymmetric synthesis of 3-hydroxypyrrolidine.

A four-step sequence from alkene 319 gave protected diol-mesylate 323 (Scheme 85). Hydrogenation of the major isomer, though under somewhat different conditions, proceeded with the same chemoselectivity. N-Boc Ph

Ph

lf~ Ph... ~ O

-

= = Ph"'1~k'OI

I H 2 (60psi)

TBSO 319

323

Ph

20OloPd(OH)2/C / (3. IOMs MeOH (0.05 M) ~ 66% (2 steps)

Ph Li/NH 3 HO~t~ ~ ..... /,. / ~ ,OTBS t-BuOH/THF Boc 88% 'Boo ~IOTBS 324

325

SCHEME 85. Asymmetric synthesis of 4-hydroxy-2-hydroxymethylpyrrolidine.

8

DIPHENYLOXAZINONES

319

protection gave disubstituted pyrrolidine 324 in moderate yield, and Birch reduction gave the 4-hydroxy-2-hydroxymethylpyrrolidine 325. Finally, extension of the procedure to crotylsilated oxazine 326 gave the pyrrolidine 328 as an 86:14 mixture of diastereomers, favoring the syn relationship between the hydroxyl and methyl groups (Scheme 86).

Ph P h " ' l ' / ~' @ v CbzN

1. H2 (60 psi) Ph Ph Ph 20% Pd(OH)2/C ~ / ' ~== Ph,,,?/%0 O OMs MeOH(0.073M )MO. ~ e CbzN

Me 326

Me 327

2. Boc20,CH2CI2 70% (2 steps) 86:14 dr

Boc 328

H L i / N H 3 0M~ = t-BuOH / THF 96%

e

Boc 329

SCHEME 86. Asymmetric synthesis of 3-hydroxy-4-methylpyrrolidine.

Dissolving-metal conditions again removed the auxiliary to give the syn3-hydroxy-4-methyl-pyrrolidine 329. Aoyagi, Jain, and Williams also explored the homologation of lactols derived from 299 with cyanide. 87 The methyl-substituted lactone 299a was converted to the corresponding lactol as before and reacted with TMSCN in the presence of BF3.Et20 to give the coupling product 330a as a mixture of diastereomers (Scheme 87). Removal of the N-Cbz group aided in the difficult hydrolysis of the nitrile moiety, and treatment with KOH in ethylene glycol at elevated temperature provided the acid 332a as a single diastereomer. Epimerization at the a-stereogenic center occurs under the reaction conditions, placing the acid cis to the phenyl groups and trans to the methyl group. Hydrogenolysis in a H20/THF mixture gave (2S,3R)-isothreonine (333a), while conducting the reaction in methanol yielded the corresponding methyl ester. Starting with the analogous cyclohexyl lactone gave (2S,3R)-nor-C-statine (333b). Interestingly, the trans-substituted lactone 299 could be epimerized with a KHMDS/CO 2 system at low temperature to give all-cis lactone 334 as a single diastereomer (Scheme 87). Lactol formation and substitution gave the coupling product 335 as a single diastereomer, with the nitrile trans to the other lactone substituents. Removal of the N-Cbz group and hydrolysis of the nitrile provided the acid 337 as a single diastereomer. Hydrogenolysis in HzO/THF gave (2R,3S)-isothreonine (338a), while conducting the reaction in methanol yielded the corresponding methyl ester. Starting with the analogous cyclohexyl lactone gave (2S,3R)-nor-Cstatine (338b). Thus, two enantiomers of either isothreonine or nor-Cstatine could be synthesized from a common intermediate (299).

320

ROBERT M. WILLIAMS AND CAMERON M. BURNETT

OH HCI-H2N-.~CO2 H R 333 H2, 120 psi PdCI2, 75 ~ THF:H20 / Ph Cbz. N.~r,,,Ph

1. DIBAL-H Ph 2. Ac20. Et3N Cbz.. / ~ , , P h '

R ~ O 299 O

DMAP ~ 3. Me3SiCN BF3Et20 KN(SiMe3)2, 002 18-Crown-6, -78 ~

a ~--..~O CN 330

Ph

H2, Pd/C EtOAc

Ph

150-170 ~

J RI"~_ O _

331 CN

332 CO2H

(dr >95:5)

(1:1 diast. @ C2)

a series, R = Me b series, R = cyclohexylmethyl

Ph 1. DIBAL-H Ph CbZ..N.~r,,,Ph 2. Ac20. Et3N CbZ.N.~r,,,p h H2' Pd/C R'"~'O O

KOH

~__:~,,PhHO(CH2)2OH HN~, ,'Ph

DMAP "- R ' " ~ " O 3. Me3SiCN CN BF3Et20 334, dr>95/5 335

Ph KOH Ph HN--~,, ,Ph HO(CH2)2OH HN~ r , , ,Ph

EtOAc ~'R,,'~[. .-' ~ CN 336

150_170oc "- R , , , ~ O (dr >95:5) CO2H 337 ! H2, 120 psi | PdCI2, 75 ~ THF:H20 OH HCI-H2N_ ~ C O 2 H

338

IR

SCHEME 87. Asymmetric synthesis of isothreonine and nor-C-statine.

5 mol% CbzN v-"/~" O

(+)-1

TBAF / alumina (1 mol~ 5 equiv PMHS / toluene 92%

2

-

. CbzN

OH

339

SCHEME 88. Titanium-based reduction of oxazinone 1.

Buchwald reported a catalytic titanium-based method for reduction of lactones to lactols and applied it to (+)-1 to give the lactol 339 in excellent yield as a mixture of diastereomers (Scheme 88). 88 The procedure offers the advantage of using cheap polymethylhydrosiloxane as a stoichiometric reductant, but has not found synthetic application in this area. Interestingly, the procedure did not epimerize lactones with ~-stereogenic centers, suggesting possible application to carbonyl manipulation of the wide variety of ~-substituted lactones, though steric hindrance at the position appears to reduce the procedure's effectiveness. 89

8

DIPHENYLOXAZINONES

321

Baldwin reported that ( - ) - 2 could be substituted at the carbonyl with fluoride-activated TMS-CF 3 to give the substituted lactol 340 (Scheme 89). 9o Unfortunately, cleavage to the desired ~-trifluoroketone 341 failed under a variety of conditions. Additionally, ~-substitution with even a methyl group reduced the yield to 35%, and ~-substitution with a phenyl group completely blocked the addition. Ph ~ . ~ Ph 0 BocN.v~o

TMS-CF3 Ph CsFor -~ P h ~ TBAF/THF B~ 'O"OH-v~ 82% CF3 340

(-)-2

Ph ~r X = HO "'Pho B~

a

341

SCHEME 89. Trifluoromethyl substitution of oxazinone 2.

Williams and co-workers explored the stabilized Wittig reaction for homologation of the lactone carbonyl. 9~ Reaction of (+)-1 with (triphenylphosphoranylidene) acetonitrile in xylenes at 210 ~ gave the trisubstituted alkene 342 quantitatively, via tautomerization of the initial exo olefin (Scheme 90). High-pressure hydrogenation removed the Cbz group and reduced both the alkene and nitrile to give oxazine 343 as a single diastereomer in quantitative yield. Addition of hydrogen from the less hindered face of the olefin results in a syn-relationship between the phenyl rings and the side chain. Selective protection of the primary amine Ph Ph . . . < O (+)-1

Ph3P%/CN

Ph Ph...1~ O

xylenes, 210 ~ >99%

342

+

H N v ~ ....

/

344 1. Na~ NH3(0, EtOH

Ph ~ [O Obz"N- ~ O

~" NHCbz O

"~ 2. H2(80 psi), PdCI2, 80 ~

CN

120 psi, rt >99%

343

L

NH2-HCI

Ph

_Ph Cbz20, 2N NaOHO ) P,h ._. . ~ THF, rt 70%

Ph Ph.. ' ~ O _

H2, PdC12,HCl

H

DIPEA~-

Ph.. ~

O,~... ,,.0

120~ 75-85%

Ph

N ' ~ Cbz

OH

H O ' ~ ....~ ~ / N ~ N H NH2 hypusine

PhN ~

346

Ph

"" NHCbz

2

83% (two steps) SCHEME 90. Asymmetric synthesis of hypusine.

322

ROBERT M. WILLIAMS AND CAMERON M. BURNETT

allowed alkylation of the secondary amine 344 with known iodide 345, yielding dilactone species 346. Hydrogenation at elevated temperature and pressure allowed removal of both bibenzyl groups to give (+)-hypusine. Wittig reaction of (+)-1 with methyl (triphenylphosphoranylidene)acetate gave the trisubstituted alkene 348 quantitatively (Scheme 91). Catalytic hydrogenation removed the N-Cbz group and saturated the enamine. N-protection gave ester 349, which was reduced to the aldehyde and converted to the benzylimine 350. A chelation-controlled allylation (via 351) of the imine yielded alkene 352; the alkene was converted to the acid 353 and coupled to give hydrazide dipeptide 354. Global deprotection via hydrogenation gave (+)-negamycin. 92

Ph Ph'"l'~O

CO2Me PhaP=/

I

CbZN.v~ O xylenes, 210 o~ (+)-1 ph,,, / ~ O ' r Ph OMe __ CbzN "v~"' "/'~O 349, dr = 94:6

2. BnNH2, AI203 CH2CI2, 0 ~

HN-Bn CbZNv,J,,,, ~ 9 352, dr= 4.4:1 Ph Ph,,, .Bn CbZN'v~ ...." ~ H

~/

_Ph Ph'"f~O

Ph

Ph,,~ ~Bn

CbzN'/"'~"~O

"'

~

~

I~~~R"

Cbza

THF, -40 ~ ZnBr 351

Ph Ph,, f . ~ Cbz .Bn WSC,Et3N, HOBt, CH2CI2 "'I'" "? N" Me O CbzN..,..~ ....~ C O 2 H - v "JJ"oB N- n H2N" 353 80%

MeOH,H20, HOAc

SCHEME 91.

H

CeCl 3

1. 115 psi H2, PdCi2 MeOH-HCI,rt 2. CbzCI, EtsN DMAP, CH2CI2 96% 2 steps

350

40 psi H2, 10% Pd-C

"OBn

OMe

348

Cbz

2. 03, MeOH CH2CI2 3. PDC, DMF 56% 3 steps

MeO . E L

O/"N"

J 100%

83% 2 steps 1. Cbz-CI, NaOH

r

~

347

~

Ph',,~2"-0

] OMe

3bZN~o

1. DIBAL-H, CH2CI2

Ph

354

Ph 3h' " ~ "-O

OH N_H2 0 (+)-negamycin

Me O H

75%

Asymmetricsynthesisof negamycin.

VI. Conditions to Remove the Chiral Auxiliary A primary consideration when using the diphenyloxazinone template is the choice of method and specific conditions for removal of the bibenzyl moiety. As shown above, dissolving-metal conditions are appropriate when the target amino acid contains unsaturation, as well as for the direct deprotection of the N-Boc-lactones to the corresponding N-Boc-protected amino acids. The ring-opening/oxidative cleavage protocols are usually

8

DIPHENYLOXAZINONES

323

required when an aromatic side chain is introduced at the 0t-carbon, producing a new and potentially reducible benzylic C-N bond. Hydrogenation, while the most-used method for auxiliary cleavage, suffers from a lack of such clear-cut guidelines, especially in the choice of hydrogen pressure or catalyst. The report by Kagan that inspired our work is an example of the ambiguity inherent in choosing hydrogenolysis conditions: the alkene is saturated using Raney nickel in dioxane, while the auxiliary is cleaved with Pearlman's catalyst in ethanol, and an atmospheric pressure of hydrogen is sufficient for both reactions. In this case, the general rule that protic solvents increase catalyst activity appears to account for the differential reactivity, though the choice of catalysts remains unclear. Our general rule has been to first attempt the catalytic hydrogenation with 30 mol% of the catalyst and atmospheric pressure of hydrogen in either a protic solvent (MeOH, EtOH) or, if necessary for solubility, a mixture of protic and nonprotic solvents (e.g., EtOH-THF). The monosubstituted lactones acquired from either the electrophilic or enolate manifolds are often susceptible to catalytic hydrogenation at 14-20 psi of hydrogen. However, the procedure appears highly substratespecific, with similarly substituted adducts requiting varying pressure and catalyst loading. In extreme cases, deprotection of the same substrate is reported under different conditions (cf Schemes 40 and 41). The disubstituted lactones accessible via the enolate manifold require higher hydrogen pressure, presumably due to increased steric hindrance of the proximal benzylic C-N bond, which is typically the last residue to be reduced. In all cases, though, ambient temperature appears sufficient for the hydrogenolysis. Products of carbonyl homologation generally require higher hydrogen pressures and often elevated temperatures for successful hydrogenolysis. The absence of the lactone carbonyl group makes the ring oxygen more electron-rich, suggesting that the reduction should be more difficult. This, coupled with the increased steric hindrance presented by the 13-substituent, hinders reduction of the benzylic C-O bond, which usually is the first to be cleaved in lactone-containing systems. Dipolar cycloaddition products in which the pyrrolidine ring was substituted at every position required elevated hydrogen pressures for successful bibenzyl removal, though an unsubstituted methylene on the proline allowed for the use of atmospheric pressure. The limited examples of six- and seven-membered rings on nitrogen show that 50 psi of hydrogen allows catalytic hydrogenation of the auxiliary. Probably the most interesting case comes from Williams' recent synthesis of reineramycin and jorumycin. Hydrogenation in a mixture of

324

ROBERT M. WILLIAMSAND CAMERON M. BURNETT

protic and aprotic solvents selectively cleaves the bibenzyl C-O bond, while hydrogenation in neat ethanol is required for cleavage of the C-N bond. Also, while dialkyl substitution of the nitrogen appears to promote C-N cleavage, the N-methyl substrate in the same synthesis requires elevated pressure for auxiliary removal. In this case the TBS-protected phenol may impart sufficient steric hindrance to require the higher pressure. VII. Conclusion

The diphenyloxazinone template has been shown to possess a wide variety of reactivity manifolds that provides access to the synthesis of structurally diverse amino acids, their derivatives, and nitrogen-containing natural products. Current methods include glycine electrophilic reactivity, conversion to the phosphonate, enolate alkylation, radical chemistry, azomethine ylide dipolar cycloadditions, formation of heterocycles at nitrogen, and homologation at the lactone carbonyl, which has provided access to peptide isosteres and other nitrogen-containing substances in optically pure form. While some guidelines are given here for selecting hydrogenation conditions, the auxiliary removal procedure remains highly substrate-dependent and necessarily empirical. Efforts to further extend the breadth of reaction manifolds in these simple heterocyclic templates and the practical utility of these agents are an ongoing and active area of research in our laboratory. Current natural products whose syntheses are being pursued in the author's laboratory include palau' amine, axinellamine, quinine, tuberostemoninol, MPC-1001, ecteinascidin 743, lemonomycin, renieramycin I, bioxalomycin ~2, microsclerodermin H, quinine, nakadomarin A, zetekitoxin, and numerous isotopically labeled amino acids and their derivatives. Acknowledgments The authors are grateful to the National Institutes of Health (GM068011) and the National Science Foundation for financial support. Prof. David A. Evans of Harvard University is acknowledged for providing useful discussions at the earliest phases of this work.

References and Footnotes 1. Herbert, R. A., The Biosynthesis of Secondary Metabolites; Chapman and Hall: London, 1981. 2. Coppola, G. M., Schuster, H. E, Asymmetric Synthesis: Construction of Chiral Molecules Using Amino Acids, Wiley-Interscience: New York, 1987.

8

DIPHENYLOXAZINONES

3 25

3. Williams, R. M., Synthesis of Optically Active a-Amino Acids, Pergamon Press: Oxford, 1989. Vol. 7. 4. An earlier review on this topic has appeared: Williams, R. M., Aldrichim. Acta 1992, 25, 11. 5. Williams, R. M., Armstrong, R. W., Dung, J.-S., J. Am. Chem. Soc. 1984, 106, 5748. 6. Vigneron, J. P., Kagan, H., Horeau, A., Tetrahedron Lett. 1968, 54, 568 I. 7. Both enantiomers of 5 are available from Aldrich: (-)-5, catalog #331899; (+)-5, catalog #331880. 8. Dastlik, K. A., Sundermeier, U., Johns, D. M., Chen, Y., Williams, R. M., Synlett 2005, 693. 9. All four lactones are available from Aldrich: (+)-Cbz, catalog #331856; (-)-Cbz, catalog #331872; (+)-Boc, catalog #331813; (-)-Boc, catalog #331848. The (+)Boc-lactone is also available under catalog #331821. 10. (a) Weijlard, J., Pfister, K., Swanezy, E. E, Robinson, C. A., Tishler, M., J. Am. Chem. Soc. 1951, 73, 1216. (b) Williams, R. M., Sinclair, P. J., DeMong, D. E., Chen, D., Zhai, D., Org. Synth. 2003, 80, 18. 11. Aoyagi, Y., Agata, N., Shibata, N., Horiguchi, M., Williams, R. M., Tetrahedron Lett. 2000, 41, 10159. 12. Chang, H.-T., Sharpless, K. B., Tetrahedron Lett. 1996, 37, 3219. 13. van den Nieuwendijk, A. M. C. H., Warmerdam, E. G. J. C., Brussee, J., van der Gen, A., Tetrahedron: Asymmetry 1995, 6, 801. 14. Sinclair, E J., Zhai, D., Reibenspies, J., Williams, R. M., J. Am. Chem. Soc. 1986, 108, 1103. 15. Williams, R. M., Sinclair, E J., Zhai, D., Chen, D., J. Am. Chem. Soc. 1988, 110, 1547. 16. Williams, R. M., Sinclair, P. J., Zhai, W., J. Am. Chem. Soc. 1988, 110, 482. 17. Zhai, D., Zhai, W., Williams, R. M., J. Am. Chem. Soc. 1988, 110, 2501. 18. Williams, R. M., Zhai, W., Tetrahedron 1988, 44, 5425. 19. Williams, R. M., Hendrix, J. A., J. Org. Chem. 1990, 55, 3723. 20. Williams, R. M., Zhai, D., Sinclair, P. J., J. Org. Chem. 1986, 51, 5021. 21. (a) Ramer, S. E., Cheng, H., Palcic, M. M., Vederas, J. C., J. Am. Chem. Soc. 1988, 110, 8526. (b) Ramer, S. E., Cheng, H., Vederas, J. C., Pure Appl. Chem. 1989, 61,489. 22. Williams, R. M., Fegley, G. J., J. Am. Chem. Soc. 1991, 113, 8796. 23. Williams, R. M., Fegley, G. J., J. Org. Chem. 1993, 58, 6933. 24. Williams, R. M., Fegley, G. J., Tetrahedron Lett. 1992, 33, 6755. 25. Williams, R. M., Im, M.-N., Tetrahedron Lett. 1988, 29, 6075. 26. Williams, R. M., Im, M.-N., J. Am. Chem. Soc. 1991, 113, 9276. 27. Baldwin, J. E., Lee, V., Schofield, C. J., Synlett 1992, 249. 28. Williams, R. M., Sinclair, E J., DeMong, D. E., Org. Synth. 2003, 80, 31. 29. Williams, R. M., Liu, J., J. Org. Chem. 1998, 63, 2130. 30. Aoyagi, Y., Iijima, A., Williams, R. M., J. Org. Chem. 2001, 66, 8010. 31. Dong, Z., Tetrahedron Lett. 1992, 33, 7725. 32. Schuerman, M. A., Keverline, K. I., Hiskey, R. G., Tetrahedron Lett. 1995, 36, 825. 33. (a) Looper, R. E., Williams, R. M., Tetrahedron Lett. 2001, 42, 769. (b) Looper, R. E., Williams, R. M., Angew. Chem. Int. Ed. 2004, 43, 2930. 34. (a) Looper, R. E., Runnegar, M. T. C., Williams, R. M., Angew. Chem. Int. Ed. Engl. 2005, 44, 3879. (b) Looper, R. E., Runnegar, M. T. C., Williams, R. M., Tetrahedron 2006, 62, 4549.

326 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 58. 59. 60. 61. 62. 63. 64. 65. 66. 67. 68. 69. 70. 71. 72.

ROBERT M. WILLIAMSAND CAMERONM. BURNETT Nolen, E. G., Watts, M. M., Fowler, D. J., Org. Lett. 2002, 4, 3963. DeMong, D. E., Williams, R. M., Tetrahedron Lett. 2002, 43, 2355. DeMong, D. E., Williams, R. M., J. Am. Chem. Soc. 2003, 125, 8561. Lee, S.-H., Nam, S.-W., Bull. Korean Chem. Soc. 1998, 19, 613. Lee, S.-H., Lee, E.-K., Jeun, S.-M., Bull. Korean Chem. Soc. 2002, 23, 931. van den Nieuwendijk, A. M. C. H., Kriek, N. M. A. J., Brussee, J., van Boom, J. H., van der Gen, A., Eur. J. Org. Chem. 2000, 3683. Allevi, P., Anastasia, M., Tetrahedron: Asymmetry 2004, 15, 2091. Singh, S., Pennington, M. W., Tetrahedron Lett. 2003, 44, 2683. Kele, P., Sui, G., Huo, Q., Leblanc, R. M., Tetrahedron: Asymmetry 2000, 11, 4959. Williams, R. M., Fegley, G. J., Pruess, D. L., Schaeffer, E, Tetrahedron 1996, 52, 1149. Baldwin, J. E., Lee, V., Schofield, C. J., Heterocycles 1992, 34, 903. Aoyagi, Y., Williams, R. M., Synlett 1998, 1099. Schow, S. R., DeJoy, S. Q., Wick, M. M., Kerwar, S. S., J. Org. Chem. 1994, 59, 6850. Solas, D., Hale, R. L., Patel, D. V., J. Org. Chem. 1996, 61, 1537. Shen, K., Keng, Y.-E, Wu, L., Guo, X.-L., Lawrence, D. S., Zhang, Z.-Y., J. Biol. Chem. 2001, 276, 47311. Bender, D. M., Williams, R. M., J. Org. Chem. 1997, 62, 6690. Paquette, L. A., Duan, M., Konetzki, I., Kempmann, C., J. Am. Chem. Soc. 2002, 124, 4257. Chen, Y. T., Xie, J., Seto, C. T., J. Org. Chem. 2003, 68, 4123. Li, P., Zhang, M., Peach, M. L., Liu, H., Yang, D., Roller, P. R., Org. Lett. 2003, 5, 3095. Gao, Y., Burke, T. R., Jr., Synlett 2000, 134. Kang, S.-U., Worthy, K. M., Bindu, L. K., Zhang, M., Yang, D., Fisher, R. J., Burke, T. R., Jr., J. Med. Chem. 2005, 48, 5369. Liu, W.-Q., Vidal, M., Gresh, N., Roques, B. P., Garbay, C., J. Med. Chem. 1999, 42, 3737. Oishi, S., Kang, S.-U., Liu, H., Zhang, M., Yang, D., Deschamps, J. R., Burke, T. R., Jr., Tetrahedron 2004, 60, 2971. Hill, B., Ahmed, V., Bates, D., Taylor, S. D., J. Org. Chem. 2006, 71, 8190. Jin, W., Williams, R. M., Tetrahedron Lett. 2003, 44, 4635. Lane, J. W., Chen, Y., Williams, R. M., J. Am. Chem. Soc. 2005, 127, 12684. Reno, D. S., Lotz, B. T., Miller, M. J., Tetrahedron Lett. 1990, 31,827. Williams, R. M., Im, M.-N., Cao, J., J. Am. Chem. Soc. 1991, 113, 6976. Williams, R. M., Yuan, C., J. Org. Chem. 1992, 57, 6519. Williams, R. M., Yuan, C., J. Org. Chem. 1994, 59, 6190. DeMong, D. E., Williams, R. M., Tetrahedron Lett. 2001, 42, 183. Scott, J. D., Williams, R. M., Tetrahedron Lett. 2000, 41, 8413. Scott, J. D., Williams, R. M., Angew. Chem. Int. Ed. Engl. 2001, 40, 1463. DeMong, D. E., Williams, R. M., Tetrahedron Lett. 2001, 42, 3529. Johns, D. M., Mori, M., Williams, R. M., Org. Lett. 2006, 8, 4051. Sutherland, A., Vederas, J. C., J. Chem. Soc., Chem. Commun. 2002, 224. Kabat, M. M., Tetrahedron Lett. 2001, 42, 7521. Williams, R. M., Zhai, W., Aldous, D. J., Aldous, S. C., J. Org. Chem. 1992, 57, 6527.

8

DIPHENYLOXAZINONES

327

73. Maggini,M., Scorrano, G., Bianco, A., Toniolo, C., Sijbesma, R. E, Wudl, E, Prato, M., J. Chem. Soc., Chem. Commun. 1994, 305. 74. (a) Sebahar, P. R., Williams, R. M., J. Am. Chem. Soc. 2000, 122, 5666. (b) Sebahar, E R., Osada, H., Usui, T., Williams, R. M., Tetrahedron 2002, 58, 6311. 75. (a) Onishi, T., Sebahar, E R., Williams, R. M., Org. Lett. 2003, 5, 3135. (b) Onishi, T., Sebahar, E R., Williams, R. M., Tetrahedron 2004, 60, 9503. 76. Sebahar, E R., Williams, R. M., Heterocycles 2002, 58, 563. 77. Lo, M. M-C., Neuman, C. S., Nagayama, S., Perlstein, E. O., Schreiber, S. L., J. Am. Chem. Soc. 2004, 126, 16077. 78. (a) Ahrendt, K. A., Williams, R. M., Org. Lett. 2004, 6, 4539. (b) Ahrendt, K. A., Williams, R. M., Org. Lett. 2005, 7, 957. 79. Greshock, T. J., Funk, R. L., J. Am. Chem. Soc. 2002, 124, 754. 80. Dutton, E E., Lee, B. H., Johnson, S. S., Coscarelli, E. M., Lee, E H., J. Med. Chem. 2003, 46, 2057. 81. Saba, A., Tetrahedron Lett. 2003, 44, 2895. 82. Williams, R. M., Colson, P.-J., Zhai, W., Tetrahedron Lett. 1994, 35, 9371. 83. Aoyagi, Y., Williams, R. M., Tetrahedron 1998, 54, 10419. 84. Jain, R. E, Williams, R. M., Tetrahedron 2001, 57, 6505. 85. Jain, R. E, Williams, R. M., Tetrahedron Lett. 2001, 42, 4437. 86. Aoyagi, Y., Williams, R. M., Tetrahedron 1998, 54, 13045. 87. Aoyagi, Y., Jain, R. E, Williams, R. M., J. Am. Chem. Soc. 2001, 123, 3472. 88. Verdaguer, X., Berk, S. C., Buchwald, S. L., J. Am. Chem. Soc. 1995, 117, 12641. 89. Verdaguer, X., Hansen, M. C., Berk, S. C., Buchwald, S. L., J. Org. Chem. 1997, 62, 8522. 90. Walter, M. W., Thaker, N., Baldwin, J. E., Muller, M., Schofield, C. J., J. Chem. Res. (S) 2000, 310. 91. Jain, R. P., Albrecht, B. K., DeMong, D. E., Williams, R. M., Org. Lett. 2001, 3, 4287. 92. Jain, R. P., Williams, R. M., J. Org. Chem. 2002, 67, 6361.

STRATEGIES AND TACTICS IN ORGANIC SYNTHESIS, VOL. 7 9 2008 Elsevier Ltd. All rights reserved.

Chapter 9 RHODIUM-CATALYZED CYCLOISOMERIZATION REACTIONS OF ALLENES IN DIVERSITYORIENTED SYNTHESIS Kay M. Brummond and Branko Mitasev Department of Chemistry University of Pittsburgh Pittsburgh, PA 15260, USA

I. II. III.

Introduction Designand Synthesis of the Pivotal Allenic-Amino Ester Intermediates AllenicAlder-ene Reaction Affording Amino-Ester Tethered Cross-Conjugated Trienes IV. Diversificationof Cross-Conjugated Trienes via Diels-Alder Reactions: First Generation Triene V. Design and Synthesis of a Second Generation Triene VI. Synthesis and Stereoselective Diels-Alder Reaction of Novel Bicyclic Trienes VII. Summary Acknowledgments References and Footnotes

328 336 342 346 349 353 361 362 362

I. Introduction

Modem drug discovery efforts are based upon a deep understanding of the cellular pathways responsible for a particular disease. The extreme complexity of the cellular circuitry has made the goal of understanding all of its aspects a formidable challenge. Significant progress in this area has been made by using modem biochemical tools such as mutagenesis, which is commonly used to irreversibly modify proteins in order to understand their cellular function. However, there are many aspects of the cell's biochemistry that remain a mystery. These include important signaling pathways based on dynamic protein-protein interactions that are particularly difficult to study. ~ Therefore, there is a tremendous need for innovative strategies to discover new cellular pathways and probe the function of

9

RHODIUM-CATALYZED CYCI~OISOMERIZATION REACTIONS OF ALLENES

329

various protein targets. Small organic molecules are useful research probes for studying cellular pathways. Small molecules are capable of interacting with macromolecules such as proteins generally in a reversible manner, thereby modulating their function. 2 Observing the effects that result from such interactions in designed assays can often lead to an understanding of the role a particular target plays in the cell. Additionally, if the protein is a known therapeutically relevant target, lead compounds for drug discovery can be identified. The great number of relevant protein targets and their immense structural diversity dictates that the small molecules screened against these targets must also be structurally and functionally diverse. 3,4 Nature is one of the greatest sources of diverse small molecules with a broad bioactivity profile. However, a major limitation of screening natural products is their limited availability. 5 Therefore, synthetic small molecules constitute a major portion of the modem screening palette. Advances in the field of synthetic organic chemistry have led to development of many methodologies for efficient assembly of small molecules. The field of combinatorial synthesis has evolved in the past two decades as a consequence of the demand for diverse small molecules for biomedical research. 6 The synthetic strategy that is most commonly utilized in the combinatorial approach involves appending different building blocks around a common structural core. This approach has been greatly facilitated by the development of practical technologies that allow the streamlined parallel synthesis of large numbers of compounds (thousands) in a short period of time. Although this approach will undoubtedly continue to lead to identification of additional biological agents, researchers have questioned whether the level of structural diversity that is achieved is sufficient to complement the wide variety of modem biomedical targets. 7 The appendage diversity that is achieved by varying substituents around a common core is thought to limit the compounds to a narrow chemical space. Very often, and particularly in the pharmaceutical company setting, the molecules accessed in this manner are designed to fall within defined physico-chemical parameters that increase their chances of becoming drug candidates. 8 For example, the well-known Lipinski rules for drug-like molecules consider properties aimed at increasing bio-availability (molecular weight, solubility, number of hydrogen-bond donors and acceptors, etc.). 9 The need for novel, chemically diverse small molecules has been made clear in the National Institutes of Health (NIH) roadmap for medical research (http://nihroadmap.nih.gov/). According to these guidelines, the goal of biomedical research in the future is to identify a comprehensive set of small molecules that are capable of selectively modifying the function

330

KAY M. BRUMMOND AND BRANKO MITASEV

of the majority of biological targets in the human cell. Since the number of relevant biological targets continues to grow as a result of intense research, this important goal can only be achieved by effectively integrating the development of new synthetic technologies to generate novel chemically diverse entities, with assays against a broad range of biological targets. New synthetic technologies encompass methods for more efficient synthesis of small molecules, their purification, isolation and characterization. In recent years, organic chemists have become increasingly aware of these issues and have begun addressing them by innovative combinatorial strategies and diversity-oriented synthesis (DOS). ~~DOS aims to develop new and adapt existing synthetic methodologies for generating structurally diverse molecules specifically for biological screening. Because synthetic chemists generate novel compounds continuously, the term DOS is carefully assigned only to the designed and deliberate synthesis of collections of small molecules populating novel chemical space. 5 Contrary to the classical combinatorial approach, modern DOS efforts put their main emphasis on the diversity of the molecular scaffolds that are accessed and not on the numbers of compounds. ~ Therefore, the synthesis of libraries of compounds is generally limited to between tens to hundreds of compounds, and not thousands (or more) as in the classical combinatorial approach. ~2 Moreover, the compounds are accessed in milligram quantities, which allows a thorough assessment of their biological activity in a wide variety of biological assays. ~3 Characterizing the diversity of molecular libraries is another important aspect of D O S . 14 This is commonly done by using computational methods to predict various physico-chemical properties (i.e., molecular descriptors), which are then compared to those of existing libraries. 9 These can include not only the more common parameters such as number of rotatable bonds, H-bond donor and acceptor groups, solvent-accessible surface area and clogP (distribution coefficient for octanol/water), but also pharmacologically related ones such as the predicted affinity for serum-protein binding, intestinal permeability, metabolizable groups in the molecule, etc. Computational prediction of parameters like these allows the practicing DOS chemist to design libraries that possess a broad diversity profile. In designing such discovery libraries of novel compounds, at least three forms of structural diversity have been considered and include appendage, stereochemical and skeletal diversity. An ideal DOS strategy incorporates all three forms. Skeletal diversity is arguably the most important, but relatively difficult to achieve in an efficient manner. There are at least two conceptually distinct ways it can be accomplished. The first one involves

9

RHODIUM-CATALYZED CYCLOISOMERIZATION REACTIONS OF ALLENES

331

designing structurally similar substrates that under common reaction conditions undergo diverging transformations affording skeletally distinct products. The chemical information that leads to different products is therefore encoded in the substrates (substrate-based control). This is the most commonly utilized synthetic approach to skeletal diversity. ~5 One example from Schreiber involves a 1,3-dipolar cycloaddition between an indole dipolarophile and an in situ-generated cyclic oxonium ylide to give skeletally different products by varying the placement of the reactive functional groups around a common pyridone core (Scheme 1).~6

OMe

Me

S ~ )----N' OMe

O ; CO ~~~V~'N Me

"~,

"~0 OSi

1

4

0"~ ~-N20 ,0, ~ O S i Rh(ll)

MeH / ~ 0 " N '"H

H N ~ O S

Me

~ 5

TsN--~---~

i

0

Me HN /.~e H',~ ~ O N H tBu

o

o" N.

OSi

"%. ";0

L v j "lOSi!4 6

SCHEME 1

Despite the fact that a high degree of skeletal diversity can be accessed via the substrate-based approach, the efficiency of the overall process is compromised, since all of the precursors need to be synthesized independently. Therefore, a more efficient approach to skeletal diversity involves subjecting a common synthetic precursor to different reaction conditions to give skeletally unique products (reagent-based control). Although there are some applications of this concept, it remains the most difficult to achieve, particularly in the context of cyclic skeleton synthesis. ~7

332

KAY M. BRUMMOND AND BRANKO MITASEV

Accessing all three forms of diversity requires DOS strategies with branching reaction pathways available to common synthetic precursors. In this manner, structurally distinct scaffolds can be obtained from a small pool of reactants, thereby increasing the overall efficiency of the process. Therefore, there is a strong incentive to develop new chemical transformations and design strategies toward this goal. The development of many useful transition metal-catalyzed reactions in the last decade has opened the door for their application to DOS. ~8 Transition metal-catalyzed reactions are generally environmentally benign and economic synthetic processes, proceeding with high levels of selectivity (chemo-, regio-, and/or stereoselectivity) and minimize the use of raw materials and generation of byproducts. Among these, Pd-catalyzed coupling reactions (Heck, 19 Stille, 2~ Sonogashira, 2~ Suzuki, 22 etc.) and allylic substitution reactions 23 are regarded as some of the most important with the highest impact on the field of modern organic synthesis. Not surprisingly, these reactions have seen extensive application in combinatorial and DOS. The majority of classical combinatorial strategies in the past have been largely limited to utilizing transition metal-catalyzed coupling reactions only introducing appendage diversity (e.g., see Pd-catalyzed Suzuki, 24 Heck 25 and Stille 26 coupling reactions). More recently, intramolecular versions of these and related reactions have been used to create skeletal diversity in the synthesis of small- and medium-sized rings. 27 Related transformations such as the coupling of arylboronic acids and amides recently developed by Buchwald 28 are also becoming increasingly popular in the generation of cyclic skeletons. 29Another very important transformation that is widely used in combinatorial synthesis and DOS is the Ru-catalyzed ring-closing metathesis developed by Grubbs. ~5'3~ Transition metal-catalyzed reactions that transform relatively simple acyclic starting materials to cyclic (or polycyclic) products via a carbocyclization process are another important class of reactions that has received attention in the past decade. Carbocyclization, in general, refers to a cyclization process that involves carbon-carbon bond formation via a carbometalation, wherein a C*-M (Carbon-Metal) species delivers the carbon and metal component across an unsaturated bond (C=C) thereby affording a C * - C - C - M species. 31 In particular, carbocyclization reactions of precursors containing unsaturated functional groups (e.g., alkenes, alkynes) have been very useful in the syntheses of carbocyclic and heterocyclic molecules. 32 Examples of such reactions include transition metal-catalyzed enetype cycloisomerizations, 33-35 [4 + 2] and [5 + 2] cycloadditions 36 (7 to 10 and 7 to 11, respectively; Scheme 2) and [2 + 2 + 1] cyclocarbonylation 37

9

333

RHODIUM-CATALYZED CYCLOISOMERIZATION REACTIONS OF ALLENES

reactions of enynes (7 to 12). Aside from increasing molecular complexity, an important aspect of these reactions is that metal catalysis often allows for bond formation that would be difficult or impossible using conventional methods, to readily occur under mild conditions. A typical example is the intramolecular [4 + 2] cycloaddition of electronically unactivated dieneynes proceeding under Rh(I) catalysis (e.g., 7-10). 38 The vast potential for increasing molecular complexity and achieving skeletal diversity via metal catalyzed carbocyclization reactions (e.g., cycloadditon, cycloisomerization, cyclocarbonylation, etc.) remains an untapped resource for DOS.

C( 8 R1 ~ R 1 enyne ~.~esis

13

TR1= Hor alkyl R1= alkyl ~

R2 9

co

[2+2+1]~

~ ~[4+2]

7

R1

[5+2]

10

12

11 SCHEME 2

Although olefins and acetylenes are most commonly utilized in carbocyclization reactions, use of allenes as ~-components is becoming increasingly prevalent. For a relatively long time since their first synthesis, 39 allenes were considered no more than a chemical curiosity and remained underutilized. 4~ Intense research in the past decades, however, has resulted in many useful synthetic methods involving allenes. 4~-43The two cumulated double bonds of the allene display high reactivity toward a range of transition metals, and have been exploited in a variety of ways.

334

KAY M. BRUMMOND AND BRANKO MITASEV

Using allenes as olefin components in transition metal-catalyzed reactions often has the advantage of increased reactivity. This is largely due to the strain associated with having two cumulated double bonds, which is estimated at 10 kcal/mol. 44 Despite this fact, transition metal-catalyzed carbocyclizations of allenes (e.g., cycloaddition and cycloisomerization reactions) remain largely unexplored and underutilized in synthesis, presumably because there are no known control elements for effecting double bond selectivity other than substrate modification. 45,46 Recent studies by Brummond and coworkers have resulted in some of the first examples of reagent-based control of olefin selectivity in the allenic cyclocarbonylation (Pauson-Khand reaction) and cycloisomerization reactions. 47 Reagent-based control of double bond-selectivity in transition metalcatalyzed carbocyclization reactions of allenes is ideally suited for application to DOS since skeletally different products can be obtained. For example, selective engagement of the proximal olefin of allenyne 14 in a cyclocarbonylation reaction under Mo(CO)6-mediated conditions leads to an ot-alkylidene cyclopentenone 15 (Scheme 3). 48 Alternatively, the same transformation of the distal double bond under Rh(I) catalysis leads to a 4-alkylidene cyclopentenone 16. 49 Furthermore, a Rh(I)-catalyzed cycloisomerization reaction involving the distal double bond of the allene can lead to a cross-conjugated triene 17. 5o R2

Mo(CO)6

2R1

15 R2

R2 Rh(I), CO

CH2 R1

CH2R 1

16

14

R2 Rh(I)

-R 1 17 SCHEME 3

9

335

RHODIUM-CATALYZEDCYCLOISOMERIZATIONREACTIONS OF ALLENES

Each reaction results in increase of molecular complexity since relatively simple acyclic precursors are transformed to mono- or bicyclic skeletons. Furthermore, a novel, reactive moiety is generated (enone, cross-conjugated triene) that can be further exploited in diversity generating transformations. Therefore, we became interested in implementing a DOS strategy based on transition metal-catalyzed cyclocarbonylation and cycloisomerization reactions of allenes. From the beginning of this study in 2001, we have been interested in developing novel chemical transformations that can be applied to the efficient assembly of functionalized small molecules. As illustrated in Scheme 4, these studies have led us to diverse arrays of functionality. 51 Herein, we focus our attention on the design and synthesis of the pivotal allenic aminoester intermediates 18 and the details of their transformation into cross-conjugated trienes 19 via a Rh(I)-catalyzed cycloisomerization reaction. Furthermore, the diversification of these interesting molecules by using subsequent complexity-generating transformations is described. The ultimate goal of the study was twofold: (1) development of new synthetic methodologies for efficient assembly of complex small molecules; and (2) synthesis of collections of these compounds specifically for use as biological probes. R3 P~N'~ MeO20 R1 II 19A

P~N,.'~.~~ R3 MeO2C R1

R2 23

R.N,,~, ~ ~MeO2C~ R3 R1R2 22

Rh(I)

law

"~

O R3 P \ N - ' ~ "j MeO2C---~~ Rh(I),. R1 I[

R3

~

P-N

.K o

19B

18

Mo(CO)6 R3 P - N / ~ O

Me02C-~ -"~\ R1 ~-,.R2 21 SCHEME 4

MeO2C~O R1 2O

~2

336

KAY M. BRUMMOND AND BRANKO MITASEV

II. Design and Synthesis of the Pivotal Allenic-Amino Ester Intermediates

In designing a pivotal allene intermediate for our DOS strategy, three aspects were considered: (1) incorporation of a higher number of N and O heteroatoms was important since compounds containing them are more likely to have a desirable pharmacological profile and exhibit interesting biological effects via specific interactions with proteins; (2) potential for diversification of the molecular scaffolds by employing both front-end and back-end appendage diversity strategies; and (3) ease of preparation and availability of the precursors. Incorporating nitrogen- and oxygen-containing functional groups in the pivotal allenes is also important because they would allow for rapid attachment of pre-functionalized alkynes and alkenes used in the carbocyclization reactions, thereby incorporatingfront-end diversity into the scaffolds. Additionally, the reactivity of nitrogen and oxygen containing functional groups can be exploited in back-end functionalization of the scaffolds subsequent to their formation. With this in mind, several known methods for preparation of functionalized allenes were considered. 42~The Claisen rearrangement of propargyl ethers and esters is a versatile method for preparation of allenes. 52 Work by Castelhano and Krantz demonstrated that mild dehydrative conditions (Et3N, CC14, PPh3) effect the rearrangement of benzoyl-protected amino ester 24 to 4-allenyl-5-oxazolone 25, which is transformed to methyl ester 26 when treated with MeOH (Scheme 5). 53 Furthermore, Kazmaier reported an ester-enolate Claisen rearrangement of a variety of propargyl amino-esters 27 using LDA/ZnC12 affording allenic amino acids 29 with diastereoselectivities greater than 93%. 54 H R I ~ I ~ R2 PPh3,MeC 0014, N Et3N ~

BzHN/-..~O ~ O I

H -".,R2

--~~ R I ~ ~O O

I

MeOH,HCI ~

/,,~

R1 MeO2C NHBz

Ph 24

25

a3

26

H - R2

a3 R2

PHN~'" 0

LDA, ZnCI2, THF

R I ~ I ' ~ R2 PN

..P-~O

~Zn.----O 27

28

[3,3] sigmatropic rearrancjement ,,HO2C

,,,~R1 *NHP

29

P = Boc, Cbz, Tos SCHEME 5

9

RHODIUM-CATALYZED CYCLOISOMERIZATION REACTIONS OF ALLENES

337

The allenic amino acid derivatives obtained via these two methods appeared ideal for adapting to our DOS strategy Notably, the propargyl esters are easily obtained in one step by coupling the corresponding Nprotected amino acid and a propargyl alcohol, allowing for multiple points of diversity to be introduced. Therefore, the allenic amino esters obtained in this manner were elected as pivotal intermediates for development of a DOS strategy based on transition metal-catalyzed cycloisomerization and cyclocarbonylation reactions of allenes. Using Castelhano's protocol, we first prepared the phenylalaninederived allene 34 as outlined in Scheme 6. Esterification of N-benzoyl phenylalanine 30 with 3-butyne-2-ol by using N,N'-dicyclohexylcarbodiimide (DCC) and 4-dimethylaminopyridine (DMAP) gave ester 32 in 78% yield. The reaction proceeds via the intermediacy of oxazolone 31, which is then subjected to nucleophilic attack by the alcohol. Treatment of ester 32 with CCl4, PPh 3 and Et3N in acetonitrile affords the 4-allenyl-2-oxazolin-5-one intermediate 33, which is treated with MeOH/HC1 to give allene 34 in 74% yield as a 1.7:1 mixture of diastereomers (as originally reported). 53

Bn BzHN/~.OH O

OH DCC, DMAP 0H2CI2,rt, 78%

Bn)~/O N~.O

BzHN'~ O O

Ph 31

30

32 PPh3, CCI4, Et3N

MeCN, rt

H

H

MeOH, HCI, rt

Bn,~~

74% (2 steps)

MeO20

Ph

34

33

SCHEME 6

A serine-derived allene (38) was also synthesized following the route outlined in Scheme 7. Esterification of the known acid 35 with 3-butyn-2-ol in 73% yield was followed by removal of the Boc protecting group

338

KAY M. BRUMMONDAND BRANKO MITASEV

in 36 with TFA. Coupling of the primary amine of 36 with benzoyl chloride to afford amidoester 37 in 60% yield for the two steps. Claisen rearrangement of 37 afforded allene 38 in 87% yield as a --~2:1 mixture of diastereomers (determined by integration of the allenic methyl group resonances in the ~H NMR spectrum).

TBSO~ B~

L H 35

OH

TBSO.. ~"~. /

DCC, DMAP B O O , N]..~, ]O/ COOH 'CH2Cl2, 73% H II O

TBSO.. 1. TFA, CH2CI2, rt 1"'['/ 2. BzCI, Et3N, CHCI3 B Z ~ N ~ O 60% (2 steps)

36

H

n O 37 1. PPh3, CCI4, Et3N MeCN 2. MeOH, HCI, rt. 87%, dr = 2 : 1 H

.,~NHBz MeO2C" ~---OTBS

38 SCHEME 7

The alternative protocol for preparing allenic amino acids with high diastereoselectivity is the ester-enolate Claisen rearrangement reported by Kazmaier. All examples reported by Kazmaier involved internal alkynes (27, R 3 = alkyl, in Scheme 5) and consequently, all of the allenes contained an alkyl group at the proximal double bond of the allene (i.e., trisubstituted allenes). Therefore, our initial efforts focused on reproducing Kazmaier's protocol by preparing trisubstituted allenes. Propargylic esters 41a and 41b were obtained by coupling the corresponding acids 39a and 39b with alcohols 40a and 40b using DCC and DMAP in 88 and 77% yield, respectively (Scheme 8). Claisen rearrangement of 41a using the reported conditions (LDA, ZnC12, THE - 7 8 ~ to room temperature) proceeded to give the intermediate allenic acid, which was converted to the methyl ester 42a in 22% overall yield by treatment with MeI and KHCO 3. This low yield was surprising since neither of these reactions revealed byproducts by TLC. Indeed, when we performed the Claisen rearrangement reactions without purifying the intermediate acids, the

9

339

RHODIUM-CATALYZED CYCLOISOMERIZATION REACTIONS OF ALLENES

yields improved. Applying this strategy to the preparation of 42b resulted in 73% yield after purification. H ='=R2

R3 R1

1.LDA,ZnCI2,THF,-78~ to rt

4 0 a or 4 0 b

PHN.~OH O

DOC, DMAP,.. P H N " ~ O CH2CI2, rt

2.KHOO3,MeI,DMF,rt.

R3 MeO2' C ~ N l l p

O

39a P = Boc, R 1 = Bn 39b P = Cbz, R 1 = Me

41a

4 2 a - 2 2 % : P = Boc, R 1 = Bn,

- 88%

41b - 77% 42b

'~~C6H

14

R2= n-Hex, R3= Me - 73%: P = Cbz, R 1 = Me, R2= i-Pr, R3= n-Bu

04H1~~ ' ~ ~ ~

OH

OH

40a

40b

SCHEME 8

We were also interested in preparing 1,3-disubstituted allenes using this protocol. To this end, propargylic ester 41e was prepared in 97% yield from N-Cbz-alanine 39b and 3-butyn-2-ol (Scheme 9). Applying the two-step reaction sequence to this substrate resulted in the corresponding allenyl-amino ester 42e in 70% yield as a mixture of diastereomers in a 1:1 ratio as determined by ~H NMR. This result was disappointing since the high diastereoselectivity is considered the major advantage of utilizing Kazmaier's protocol. Since all examples of the Claisen rearrangement proceeding with high diastereoselectivity reported by Kazmaier contain an internal alkyne, the lack of diastereoselectivity in the case of 42e is attributed to the terminal alkyne. Nevertheless, the origin of this effect is not clear.

H Q

DCC, DMAP

CbzHN

~.,.. /

..OH O

39b

1.LDA, ZnCI2,THF,-78~ to rt 2.KHCO 3, MeI,DMF,rt

CH2012, rt, 97% C b z H N ~ O I I O 41c

SCHEME 9

70%,dr=1 9 1

MeO2C

NHCbz 42c

340

KAY M. BRUMMONDAND BRANKO MITASEV

To circumvent this problem, disubstituted allenes were prepared diastereoselectively by utilizing a trimethylsilyl (TMS) group to temporarily functionalize the alkyne terminus (Scheme 10). Following the original protocol, a solution of 41d in THF (kept at room temperature) was added to a solution of LDA at - 7 8 ~ followed by addition of ZnC12 (0.5 M in THF). The resulting Zn-enolate was then warmed to room temperature affording the intermediate carboxylic acid after aqueous work-up, which was converted to the methyl ester by addition of MeI and KHCO 3. Removal of the allenyl TMS group was accomplished by treatment of 43d with tetra-n-butylammonium fluoride (TBAF) in presence of a phosphate buffer (pH = 7.0) to give the disubstituted allene 42d in 49% yield for the three steps. 55 This yield was reproducibly obtained when the three steps were performed without purification of the carboxylic acid and allenyl-TMS intermediate 43.

TMS\ ~I~R2

H .-- R2 1. LDA, ZnCI2,THF, -78~ to rt 2. KHCO 3, Mel, DMF,rt ,- MTeM:c~,e.~NR~ TBAF' pH = 7.0 b uffe r, ~ THF

PHN@ 0 0

H .-- R2 H ~.e. ,,~R1 MeO2C ~'NHP

43d 43e 43f

41d P = Cbz, R 1 = Me, R 2 = Me

41e P-- Cbz, R1 = Me, R2= i-Pr 41 f P = Boc, R1 = Bn, R2= Me

42d - 49% (three steps) 42e - 48% (three steps) 42f - 60% (three steps)

SCHEME l0

Allene 42d was obtained as nearly a single diastereomer (diastereomer ratio of --~95:5 was determined by 1H NMR). The relative stereochemistry of the major diastereomer was assigned as syn, in accordance with Kazmaier's results. 54 This route was then applied to the synthesis of allene 42e (48%), which contains an isopropyl group at the terminal position of the allene. Attempts to prepare the Boc-protected allenic aminoester 42f using this protocol led to the formation of byproducts. It was reasoned that formation of unidentified byproducts in the Claisen rearrangement step could be minimized by keeping the temperature of the reaction at - 7 8 ~ during the addition of the propargyl ester 411' to LDA. Indeed, it was found that by cooling the THF solution of 411' to - 7 8 ~ and adding it to a solution of LDA simultaneously with ZnC12 (0.5 M in THF), the yield increased from ---45 to 60% for the three-step sequence (the protocols for the formation of the methyl ester and TMS removal were kept identical to that for the preparation of 42d and 42e).

9

341

RHODIUM-CATALYZED CYCLOISOMERIZATION REACTIONS OF ALLENES

Presumably, simultaneous addition of ZnCI 2 and the substrate to LDA is advantageous due to immediate formation of the stabilized Zn-enolate, which minimizes side reactions resulting from exposure of the propargyl ester to excess LDA. The next goal was to develop a general procedure for N-alkylation of the allenic aminoesters that would introduce the alkyne component of the precursors for transition metal-catalyzed carbocyclization reactions. It was quickly found that treatment of the amides or carbamates with Nail in DMF at room temperature for 2-5 min, followed by addition of the corresponding propargylic bromide, resulted in clean N-propargylation. Using this protocol on the Bz-protected substrate 34 and 1-bromo-2-butyne gave allenyne 44a in 83% yield (Scheme 11). We were interested in

a5

R4 PHN / ~ /~ . "*~...,,R3 MeO2C R1 "~" R2

Br\ R5 Nail, DMF, rt

PN MeO2C R

~R2

44a-I

a5 BzN MeO2CB'/X Bnn~e~,,H

/

R5

BzN M eTBO SO2 C ' ~ ~ |

CbzN ,,H

MeO2C'"~,I,-, \ ' ~ '"~

44a : R5 = Me, 83% 44b :R 5 = H, 73% 44c :R5= TMS, 75%a 44d :R 5 = Ph, 89% CbzN MeO2C',"M ' ~ e ~ e ~ H 44j: 86%

44f : R5= Me, 84% 44g: R5= H, 77% 44h : R5= TMS, 86%a 44i: R 5 = Ph, 68%

44e :94%

C bzN,v..~n-Bu MeO2C"~Vle " ~ 44k: 78%

a Conditions: KH, THF. SCHEME 11

,H L

B~ "

~

MeO2C'" ~Bn " ' ~ , ,H 06H13 441: 73%

342

KAY M. BRUMMOND AND BRANKO MITASEV

varying the substitution pattern of the alkyne in order to study the scope of the carbocyclization reactions. Therefore, a terminal alkyne was incorporated via N-alkylation with propargyl bromide affording 44b in 73% yield. Alkylation with 3-phenyl-l-bromopropyne gave a phenyl-substituted allenyne 44d in 89% yield. Attempts to prepare precursor 44e with a TMS group on the terminus of the alkyne led to desilylation, which was attributed to the presence of NaOH in the bulk Nail. This problem was circumvented by utilizing KH (in mineral oil) as a base and THF as a solvent, which gave 44e in 75% yield. The same protocols were applied to synthesize the Cbz-protected, alanine-derived substrates 44f-i in yields ranging from 68 to 86%. These two sets of allenynes were envisioned to serve as main model systems for studying the transition metal-catalyzed reactions and subsequent diversification of the scaffolds. We were also interested in examining the effect of allene substitution on the carbocyclization reactions so allenynes 44e, 44j, 44k and 441 were synthesized in 73-94% yields. III. Allenic AIder-ene Reaction Affording Amino-Ester Tethered Cross-Conjugated Trienes

Our synthetic investigations started with the Rh(I)-catalyzed allenic Alder-ene reaction. When allenyne 44a was submitted to the optimized reaction conditions (5 mol% [Rh(CO)2C1]2, toluene), cycloisomerization proceeded at room temperature in less than 10 min to give the expected cross-conjugated triene 45a in 80% yield (Scheme 12).5~"The structure of triene 45a was assigned based on the characteristic olefin resonances in the ~H NMR spectrum. Triene 45a was obtained as a single isomer of the exocyclic olefin, which is assigned Z-geometry in accordance with previous examples and mechanistic studies we reported. 5~ The scope of this transformation was next investigated by subjecting allenynes 44b-441 to the same reaction conditions. First, only allenynes substituted with a methyl group on the terminal allenic position were tested to avoid formation of E/Z isomers of the appending olefin. In all cases, the reaction proceeded in 10 min to afford the cross-conjugated trienes 45b-45i (Table 1). The reaction conditions were compatible with either a Bz- (entries 1-4) or Cbz- (entries 5-8) protected amine. Allenynes 44b and 44g with a terminal alkyne reacted to give 45b and 45g in 74 and 84% yield, respectively (entries 1 and 6, Table 1). Substitution of the alkyne terminus with either a TMS (entries 2 and 7, Table 1) or phenyl group (entries 3 and 8) resulted in the corresponding trienes in yields

9

343

RHODIUM-CATALYZED CYCLOISOMERIZATION REACTIONS OF ALLENES

//,.,••"f

5 mol% [Rh(CO)2CI]2

BzN

toluene, rt, 10 min, 80%

Bz. N ~ [ . - . MeO2C~'~~

Ha Hb

MeO2CB~nn~ i ' ~ , ,H 44a

45a He

Hb

Lo assigned resonance

chemicalshift ( p p m )

Hb He Ha Hd Hc

splitting/coupling constant

6.36 5.75 5.63 5.50 5.27

dd, J= 17.2, 10.7Hz s q, J = 7.0 Hz dd, J = 17.2, 1.5 Hz dd, J= 10.7, 1.5 Hz

SCHEME 12 TABLE l Scope study of the Rh(I)-catalyzed allenyl Alder-ene reaction

.R2

R2 5 mol% [ah(CO)2CI]2

P-NR~ ~o ~ MeO2C ,,H

toluene,rt,10 min

P\N'~ ~

MeO2C. _ ~ ~ , R1

44b-i Entry l 2 3 4 5 6 7 8

II

45b-i

Allenyne

P

R~

R2

Triene

Yield (%)

44b 44c 44d 44e 44f 44g 44h 44i

Bz Bz Bz Bz Cbz Cbz Cbz Cbz

Bn Bn Bn -CH2OTBS Me Me Me Me

H TMS Ph Me Me H TMS Ph

45b 45c 45d 45e 45f 45g 45h 45i

74 92 81 89 81 84 87 95

344

KAY M. BRUMMOND AND BRANKO MITASEV

ranging between 81 and 95%. Allenynes with a methyl (entries 5-8), benzy! (entries 1-3) and silyloxymethylene (entry 4) group in the amino acid side chain R ~also reacted without event. In all of the examples studied, the triene is the only product observed in the reaction and can be easily separated from the Rh-containing impurities by filtering the reaction mixture over a short plug of silica gel. The Bz-protected allenynes (44a-44e, entries 1-4, Table 1 and Scheme 12) were subjected to the reaction conditions as mixtures of diastereomers (dr --~1.7:1) while the Cbz-protected allenynes (entries 5-8) were nearly single diastereomers (all compounds are racemic). Nevertheless, all reactions resulted in the corresponding crossconjugated triene as a single isomer indicating that the exocyclic olefin geometry is not related to the relative stereochemistry of the allenyne, but is a result of the last reductive-elimination step in the mechanism of the reaction. The rate at which the reaction of these aminoester-tethered substrates proceeded was noticeably higher when compared to previously reported examples, some of which required up to 6 h for completion. 5~ The increased reactivity of the aminoester substrates is likely a result of a Thorpe-Ingold effect imposed by the quaternary center adjacent to the allene. 56 In addition, the carbomethoxy group may play an activating role by reversibly coordinating to the metal center. 5v Next, allenynes possessing different substituents on the allene moiety were tested (Scheme 13). Allenyne 44j, substituted with an isopropyl group at the terminal allenic position, reacted to afford triene 45j possessing a trisubstituted appended alkene in 95% yield. In this case, there is only one hydrogen atom that can undergo [3-hydride elimination. This does not affect the rate or the yield of the reaction, since 45j was produced in 95% yield after 10 min. Trisubstituted allenes 44k and 441 underwent the Alder-ene reaction affording 45k and 451 in 78 and 80% yield, respectively. In the case of 451, only the E isomer of the appended alkene was observed, characterized by a coupling constant of 16.0 Hz for the vinyl hydrogens in the ~H NMR spectrum. This stereoselectivity is in contrast to earlier examples from the Brummond group, where Rh(I)-catalyzed reaction of alkyl allenes resulted in mixture of appended alkene isomers in E/Z ratios in the range of 3-6:1. The geometry of the appended alkene is determined during the 13-hydride elimination step of the reaction, which requires a coplanar arrangement of the Rh-C bond and the C-H bond that is being

9

RHODIUM-CATALYZEDCYCLOISOMERIZATION REACI'IONS OF ALLENES

5 mol % [Rh(CO)2CI]2

345

P\N -

toluene,rt, 10 min

MeO2C

R R11

R1

L 2 Na

Ra

44j (P = Cbz, R 1 = Me, R2 = R3 = Me, R4 = H) 44k (P = Cbz, R 1 = Me, R2 = R3 = Me, R4 = n-Bu) 441 (P = Boc, R 1 = Bn, R2 = H, R3 = C5H11, R4 = Me)

BOC~N.~

CbZ-.N.~ M e O ~ . ~ l

MeO2~

45j 95%

45k 78%

MeO2C-~.....~ 13nl 1.L...~~ 451 80%

(E isomer only) SCHEME 13

broken. Therefore, [3-hydride elimination in the formation of 451 can occur via rotamers A and B as shown in Scheme 14. Rotamer B, which would lead to the Z-olefin isomer, posseses an unfavorable steric interaction between the appending alkyl group and the methyl substituent on the ring and is strongly disfavored. Therefore, 13-hydride elimination via rotamer A leads to the E-isomer 451, exclusively. In summary, pivotal .allenynes possessing nitrogen and oxygen heteroatoms have been prepared using either the Castelhano or Kazmaier Claisen rearrangement protocols. The former provides benzamide-protected allenynes in high yields as mixture of diastcreomers. The latter gives N-Cbz and N-Boc protected allenynes in moderate yield as single diastereomers. All pivotal allenynes 44a-1 underwent the Rh(I)-catalyzed cycloisomerization reaction in minutes to give trienes 45a-1 in high yield. Although the triene products themselves may serve as useful biological probes, we were concerned about the reactivity of the triene moiety in biological systems. Instead, it was reasoned that this reactivity can be exploited toward the efficient assembly of more complex molecular scaffolds. Some of these efforts are described next.

KAYM. BRUMMONDANDBRANKOMITASEV

346

/

Boc-N

4 Bn "~ "H CHaHbR 441 I Rh(I) BocN/"~Rh L

BocN'~'T~Rh Ln R

A - favored

B - disfavored

-

BocN"~H b

BocNAx~"~Ha

MeO2C~H

MeO2C~H

BnlHbt/J~C5H11 (Z)-451

(E)-451

notobserved SCHEME 14

IV. Diversification of Cross-Conjugated Trienes via Diels-Alder Reactions: First Generation Triene

Sequential Diels-Alder reactions of acyclic cross-conjugated trienes 46 in order to give functionalized decalin systems 48 were initially studied by Tsuge, 58 utilizing bis-silylenolether 49 as the triene, and Fallis, using monosubstituted triene 50 (Scheme 15). 59 However, tandem Diels-Alder reactions with these acyclic trienes are difficult to control and typically afford complex mixtures of regioisomers. In addition, the synthesis of the acyclic triene starting material is not straightforward. This may, in part, be due to their instability and tendency to polymerize. 6~ Thus, despite their potential, synthetic applications of these compounds have been limited. 62 In a related example, Sherburn and coworkers recently reported that an acyclic cross-conjugated tetraene ([4]dendralene-51) can participate in a tandem Diels-Alder reaction. 63 For example, the reaction of 51 with excess N-methylmaleimide affords a mixture of mono-, di-, and tri-cycloaddition

9

347

RHODIUM-CATALYZED CYCLOISOMERIZATION REACTIONS OF ALLENES

E E

[4+2]

[4+2] "

"

E

~

E

E E

E 46

E

47

48

so# Ph~OTMS

R

49

50

51

SCHEME 15

products. Nevertheless, this example underscores the fact that rapid increase in molecular complexity can be obtained via tandem cycloaddition reactions of cross-conjugated polyenes. It was reasoned that the cyclic trienes 45a-i would not pose regioselectivity issues in these tandem cycloaddition reactions because one diene is locked in an unreactive s-trans conformation. Therefore, we became interested in exploring the feasibility of the cycloaddition pathway illustrated in Scheme 16. Furthermore, exploring these pathways offered an opportunity to study the reactivity of cyclic cross-conjugated trienes as novel chemical entities. Finally, it was reasoned that these rigid and conformationally-defined polycycles would serve as interesting biological probes. 64 R2

R3

~

P\N

R3

R1 MeO2C

[4+2]

II 45

./.~

P'N

R2

O~.,,. 2l~'~ ~ ~'~ R1 Me a3" " ~ R3 52

P = C bz or Bz SCHEME 16

]~ R4 R4 ,, [4+2]

R2

P.~ R1 1 MeO2 R3

R4 R4

R3 53

348

KAY M. BRUMMONDAND BRANKOMITASEV

Our investigations began with the reaction of triene 451' with Nphenylmaleimide (Scheme 17). Intermediate Diels-Alder cycloadduct 54 was not isolated and, instead, immediately underwent a second Diels-Alder reaction to afford a 83% yield of pentacycle 55 as a 5:2:1 mixture of diastereomers. The ratio of diastereomers was determined by HPLC, which allowed for their complete separation. 0

~ MiboZ~//2C

0

m

_

N-Ph

o

CbZ..N~J/

0 II toluene,reflux,2hr M e 0 2 ~ 83%

0

45f

. Me 0 2 ~

....~0

~ -

54

55 5 92 91 mixture of diastereomers

-

(major diastereomer shown) S .....

4

,,~.

,%

....~

.::

.....~

~

.,~. :,~.~.,-........ 9 .

.

X-raY;55ctu re

SCHEME 17. (See color insert.)

The relative stereochemistry of the major diastereomer, as determined by X-ray crystallography, results from endo approach of the dienophile from the same face of the triene as the methyl group in the first cycloaddition, while the second equivalent of dienophile approaches in endo mode from the less-hindered convex face of the newly formed diene (Scheme 17). All attempts to isolate cycloadduct 54 by reaction of 45t' with an equimolar amount of the dienophile, gave tandem Diels-Alder cycloadduct 55 and recovered triene. This result can be attributed to the higher reactivity of the diene of 54, compared to the starting triene,

9

RHODIUM-CATALYZED CYCLOISOMERIZATION REACTIONS OF ALLENES

349

since 54 is locked in an s-cis conformation. Other dienophiles (maleic anhydride and 4-phenyl-[1,2,4]-triazole-3,5-dione) also reacted with 45f to give mixtures of diastereomeric products similar to 55. Although the cycloaddition reaction of 45f affords a complex molecular scaffold in a rapid manner, obtaining the product as a mixture of diastereomers was discouraging. Furthermore, biological testing of compounds as diastereomeric mixtures is not ideal due to variability in the assay concentrations. Even though separation of the diastereomers by HPLC was feasible, it would be costly and time consuming when preparing a larger-scale library. Therefore, controlling the chemo- and diastereoselectivity of the Diels-Alder reaction of the triene was important, and a new strategy for tandem intermolecular cycloaddition was considered. V.

Design and Synthesis of a Second Generation Triene

Our efforts to control the selectivity of the Diels-Alder reactions focused on designing a new triene. It was reasoned that constraining the appended ester as part of a ring would reduce the steric bulk from the C2 position and, therefore, increase the reactivity of the diene involved in the first Diels-Alder reaction. Moreover, the cyclic constraint of the ester may block one face of the sterically biased triene with the R ~ group. Finally, the rate of the second Diels-Alder reaction could be slowed by placing an electron withdrawing carbonyl group at the C6 position. Structures such as the novel imidazo-pyridinone triene 58 address all of these issues (Scheme 18). Putting the synthesis of 58 into practice required examination of the Rh(I)-catalyzed cycloisomerization of amide-tethered allenyne 56 to form 8-1actam triene 57. Traditionally, lactams are synthesized via carbon-nitrogen bond formation. For example, lactams are formed via dehydration of amino acids, 65 by cyclization of an amide onto an alkene, 66 alkyne 67 or an allene, 68 and intramolecular vinylation of amides. 69 Alternatively, lactams can be synthesized from ketones by a Schmidt or Beckmann rearrangement, v~ There are very few examples of lactam syntheses via transition metal catalyzed carbon-carbon bond formation, and most involve a ring-closing metathesis. 7~ Synthesis of lactams via cycloisomerization reaction appears particularly attractive, since additional functionality is generated in the course of the reaction (a cross-conjugated triene in this case). There are a few examples of lactam formation via cycloisomerization reactions, and they are strictly limited to preparing y-lactams. For example, in 1999,

350

KAY M. BRUMMOND AND BRANKO MITASEV

carbonyl slows the rate of the second cycloaddition

control of facial

selectivity ~ J ~

O

l

~

improved reactivity of the diene due to reduced steric hinderance at C2

0

0

HN

....

56

-~

R2

0 - - - "~

57

R3 -

N

i

R2

l

58

SCHEME 18

Lu reported a Pd(0)-catalyzed tandem cyclization/amination of dienyne 59 leading to ~-alkylidene-~,-lactam 60 (Scheme 1 9 ) . 72 More recently, Zhang reported an enantioselective Rh(I)-catalyzed cycloisomerization of amide-tethered enyne 61, affording y-lactam 62 with >99% ee. 73 Notably, both reports used a benzyl protected amide, and Zhang reported that the reaction did not proceed with the unprotected amide. To test the feasibility of a Rh(I)-catalyzed formation of ~i-lactams, amides 56a-56d were synthesized by Boc-deprotection of amine 42f to give 63 in 85% yield, followed by coupling with alkynoic acids 64a-d (Scheme 20). Employing a DCC/DMAP coupling protocol (conditions A) proved useful in preparing amides 56b and 56d in sufficient amounts for testing the subsequent cycloisomerization reaction (~100mg). Nevertheless, this protocol resulted in the formation of byproducts including dicyclohexylurea, which made the purification difficult and the yields irreproducible. To circumvent this issue, an alternative protocol was applied for the preparation of 56a and 56e. Treatment of the alkynoic acid with isobutyl chloroformate and N-methylmorpholine afforded a mixed anhydride, which was treated in situ with amine 63 to give the

9

~

O.

H 10 mol% Pd(OAc)2

B

351

RHODIUM-CATALYZEDCYCLOISOMERIZATIONREACTIONSOF ALLENES

n

-

N

+

Phi

+

Bn-N.~

"

Ph

20 mol% PPh3 MeCN, 80~

59

60

(•

O [Rh(cod)CI]2,(R)-BINAP

Bn-N

,..

"~

Bn--N

AgSbFs,rt,91%,>99%ee

61

(-)-62

SCHEME 19

corresponding amides. This proved a robust protocol that reproducibly gave 55-85% yield of the amides. O H O ~ BocH N. /..~ Meo~c,,y -.~H Bn

--

CF3COOH, CH2CI2

/111 10 min, 85%

H2N~, " MeO2C"B~n'~,H

o .),~.... R

R

64a-d

HN =" Me02C B n ~ e ~ ,'H conditions A or B

63

42f

entry

R

56a-d conditions

product

yield

1

Me

B

56a

85%

2

H

A

56b

82%

3

TMS

B

56c

72%

4

Ph

A

56d

55%

conditions A: DCC,DMAP,CH2CI2, rt. conditions B: /-BuOCOCI,NMM,-10oC, then amine 63.

SCHEME 20

With allenynes 56a-d in hand, the cycloisomerization reaction was tested (Table 2). 74 When 56a was subjected to the optimized conditions for triene formation (5 mol% [Rh(CO)=C1] 2, toluene, 0.3 M), reaction did not occur at room temperature (entry 1). Instead, triene formation was effected by heating 56a to 90 ~ With this information in mind, the

352

KAY M. BRUMMOND AND BRANKO MITASEV TABLE 2 Optimization and scope study of the Rh(I)-catalyzed Alder-ene reaction of propiolamides affording 6-1actams

~ R [Rh(CO)2CI]2 MeO20~o HNI~n ~i "H

toluene(0.03M)' 90~ '-

56a-d Entry 1 2 3 4 5 6 7

0 R HN"J~ "j MeO2C _ @ ~ Bn II 57a-d

Allenyne

R

Catalyst (tool%)

Time (rain)

Triene

Yield (%)

56a 56a 56a 56a 56b 56c 56d

Me Me Me Me H TMS Ph

5 2.5 5 l0 10 l0 l0

90 45 30 120 15 30

57a 57a 57a 57a 57b 57c 57d

0~ 18b 47 b 92 45 77 66

~Condition" 5 tool% [Rh(CO)2C1], toluene, rt. bIncomplete reaction" starting material was recovered.

catalyst loading was varied between 2.5 and 10 mol% (entries 2-4) to establish its effect on the efficiency and yield of the reaction at 90 ~ All reactions were performed in toluene at 0.03 M concentration since it was found that increasing the concentration led to the formation of by-products and a lower yield of the triene. TM With 2.5 mol% of the catalyst, the reaction was relatively sluggish and consumption of the starting material was incomplete after 90 rain, affording triene 57a in only 18% yield (along with ---50% recovered starting material). Increasing the catalyst loading to 5 mol% resulted in a shorter reaction time of 50 min and increased the yield of the triene (47%). A further increase in catalyst loading to 10 mol% resulted in complete consumption of the starting material in less than 30 min, and 92% yield of the triene 57a (this reaction was performed on --~1 g of 56a, demonstrating the scalability of the reaction). Next, the cycloisomerization of propynamide 56b was tested with 10 mol% of catalyst. Consumption of the starting material occurred after 2 h and resulted in the isolation of 57b in only 45% yield (entry 5). Trimethylsilyl (56e) and phenyl (56d) substituted propynamides also underwent the cycloisomcrization reaction to afford the corresponding trienes 57e and 57d in 77 and 66% yield, respectively (entries 6 and 7). Notably, the reaction of the TMS-substituted alkyne required only 15 min.

9

353

RHODIUM-CATALYZED CYCLOISOMERIZATION REACTIONS OF ALLENES

The higher temperature required to effect the cycloisomerization of these unprotected amides can be attributed to a preferred trans-conformation of the secondary amide, placing the reactive termini away from each other (Scheme 21). 75 To confirm this, N-benzoyl protected allenyne 56e was synthesized in 69% yield by treatment of 56a with BzC1 at 70 ~ (Scheme 22). The cycloisomerization reaction of this precursor using 10 mol% of [Rh(CO)2C1]2 occurred in less than 1 h at room temperature, affording cross-conjugated triene 57e in 75% yield.

__~

~--NH ...~\

O,~

Rh(I),90~

HN

,~ ../S \;~., 'v'~"J2L' Bn

'H

MeO2C Bn "~,' 'H

MeO2CBn~

|

56a trans-amide

56a cis-amide

favored

disfavored

57a

SCHEME 21

~. / / / .

BzCI,Etmm,DMAP

0 Ph~ / ~ ~ ' / ~

Ph 0 I 0 mol o/o[RH(CO)2CI]2,. 0..~ m / ~ , ,

moecuarseves

Bn 56a

'

Bn

69% 56e

II

57e

SCHEME 22

VI. Synthesis and Stereoselective Diels-Alder Reaction of Novel Bicyclic Trienes

With a synthetic route to lactam 57a, we next moved toward the synthesis of a facially differentiated precursor for the Diels-Alder reaction. Saponification of the methyl ester in 57a occurred within 5 min upon treatment with LiOH. After aqueous work-up, the acid was immediately coupled with glycinemethyl ester using EDCI, HOBt and DMAP to give diamide 65a (Scheme 23). In order to complete the synthesis of hydantoin 58a, amide 65a was reacted with phosgene (COC12). Unexpectedly, imino-oxazolidinone 66a was the only product isolated (55% yield for three steps) but was expected to provide the same steric and electronic control elements as the hypothetical hydantoin 58a. The structural assignment of 66a was made later based on an X-ray crystal structure of the

354

KAY M. BRUMMOND AND BRANKO MITASEV

Diels-Alder product 67a, vide infra. An additional substrate functionalized as an isobutyl amide 66b was prepared using the same protocol in 41% yield over three steps. To our knowledge, oxazolidinones with this substitution pattern have not been reported. 76

1.LiOH,THF/H20,rt. 2.RCH2NH2,EDCI,DMAP HOBt,CH2CI2,rt.

O HN MeO2C Bn 57a

II

, ~ ~ H HN ( R

Q ~,~ ,,-~'~'N

phosgene,Et3N CH2CI2,_10oC. ~

"

/,

"~

O R

65a,R =-CO2Me 65b,R = i-Pr

66a, R =-CO2Me 55% yield (3 steps) 66b, R = i-Pr 41% yield (3 steps)

O

O

=eO2J 58a or 58b not observed SCHEME 23

We suspected that the imino-oxazolidinone 66a resulting from Oacylation of the appending amide was a kinetic product of the reaction, and a consequence of the preferred conformation of the amide side chain in the precursor 65a. To confirm this, computational modeling of 65a was performed using Cache. 7v The energy-minimized model placed the amide oxygen 02 and lactam nitrogen N1 of 65a in the same direction, confirming the observed reactivity (Figure 1). Finally, it should be noted that the newly obtained imino oxazolidinone ring in 66a and 66b proved relatively sensitive to silica gel; prolonged exposure during chromatography led to ring opening giving the diamides 65a and 65b, respectively. The newly obtained bicyclic triene 66a was also modeled in order to visualize its three-dimensional structure, vv As illustrated in Figure 1, the fused bicyclic structure is relatively planar due to the presence of multiple spZ-hybridized atoms and the bottom face is blocked by the benzyl substituent. Therefore, cycloaddition reaction with the diene was expected to occur selectively with the top face (all compounds are racemic). To test this hypothesis, triene 66a was reacted with N-phenylmaleimide (1.3 equiv.) in toluene. Reaction occurred in less then 1 h at 90 ~ to afford the cycloadduct 67a in 73% yield as a single diastereomer

9

355

RHODIUM-CATALYZED CYCLOISOMERIZATION REACTIONS OF ALLENES

o'

o

4

/--N ~"' 002 Me

II

M e02C

65a

66a

;

.

..~

FIGURE 1. Cache minimizations of 65a and 66a. (See color insert.)

(Scheme 24). An X-ray crystal structure of 67a confirmed that the [4 + 2] cycloaddition occurred with endo-selectivity and the diene approached from the face opposite the benzyl group. Attempts to effect a tandem cycloaddition reaction of 66a by using excess dienophile still gave only 67a. Similarly, cycloaddition of 66b with N-methylmaleimide resulted in formation of 67b in 95% yield as a single diastereomer. Additional experimentation demonstrated that this cycloaddition is limited mainly to using maleimides. 78 Other dienophiles (diethylfumarate, p-benzoquinone and dimethylacetylene dicarboxylate) generally gave low yields of the cycloadduct (~50%). The X-ray crystal structure of 67a revealed several interesting features. As a result of the endo-cycloaddition occurring from the concave face of 66a, the product adopts a folded shape with the N-phenylpyrrolidinone moiety projecting directly above the new diene (Figure 2). Moreover, the two double bonds of the 1,3-diene are twisted out of planarity as evidenced by a dihedral angle of--~40 ~ In addition, the ~,13-unsaturated amide is also twisted in the opposite direction by a dihedral angle of ~42 ~ These steric and electronic features of the new diene significantly lower its reactivity toward electron-poor dienophiles. For example, the second cycloaddition could not be effected using electron deficient

356

KAY M. BRUMMOND AND BRANKO MITASEV

a2 i

-

O.

N/R 2 -

o_ ~ _ ~ o R1 o ~ ~.;_;#~

1.3 equiv.

\N

toluene, 1h,90~

;--',-

67a, R 1 = CO2Me,R 2 = Ph, 73% 67b, R 1 = i-Pr, R2= Me, 95%

66a, R 1 =-CO2Me 66b,Rl= i-Pr

SCHEME 24

......,i) .."~ ......3

....

a.

....

-

ph/N

"iD

67a

"--~',-----",:

torsional angle of diene = 39.9 ~

,

torsional angle of o~,13-unsaturated amide = 42.4 ~

FIGURE 2. X-ray crystal structure of 67a. (See color insert.)

dienophiles, except with diethyl fumarate. 78 Therefore, ethyl vinyl ether was examined as a small, electron rich dienophile to better match the character of the diene. Heating 67a in a mixture of toluene/ethyl vinyl ether at 90 ~ afforded 70% yield of pyran 68a as a single diastereomer. Few inverse electron demand hetero-Diels-Alder reactions of ~,13unsaturated amides have been documented, and they generally result in the formation of an aromatic compound (e.g., indole, thiazole, pyrazole). 79 Since lanthanide Lewis acids (e.g., Eu(fod) 3) can been used to catalyze hetero-Diels-Alder reactions, we tested this reagent on the cycloaddition of 67a and ethyl vinyl ether, s~ With 10 mol% of Eu(fod) 3, the reaction proceeded at room temperature, giving 68a in 95% yield as a single diastereomer (Scheme 25). Hydrolysis and isomerization of pyran 68a afforded aldehyde 69a.

9

357

RHODIUM-CATALYZED CYCLOISOMERIZATION REACTIONS OF ALLENES

OEt

.o

t

0

,

HO,

Me02C ph/N---'~0 67a

10 mol%Eu(fod)3 M 1,2-dichloroethane rt,95%

Ph/N'~;

ph/N%

68a

69a

SCHEME 25

Next, hydrolysis of the oxazolidinone moiety in 67a to the parent diamide was explored as a means to introduce structural diversity in the products and increase the number of hydrogen-bond donors. It was anticipated that this transformation would increase the water solubility of these compounds, and improve their pharmacological profile. 8 When 67a was heated to 70 ~ in 1 M HC1/dioxane (1:1) for 1 h, only starting material was recovered in ---80% yield. The stability of this imino oxazolidinone to acidic conditions is in sharp contrast to the bicyclic triene 66a, which is readily hydrolyzed in presence of aqueous acid. Next, we attempted to cleave the imino oxazolidinone under Lewis acid conditions (BF3-OEt 2 and Me2S), 81 which also gave recovered starting material in 86% yield. On the contrary, treatment of 67a with LiOH in THF/H20 caused complete decomposition. 82 Because strongly basic conditions caused decomposition of the starting material, primary amines were examined as weaker bases/nucleophiles. To this end, a solution of 67a in CDC13 was treated with benzyl amine, and the reaction was followed by ~H NMR. Although cleavage of the oxazolidinone moiety was evident by the appearance of new amide and urea N-H resonances in the downfield region (8-9 ppm), gradual disappearance of both olefinic peaks of the diene was also observed. Based on these observations, it was speculated that the final product of the reaction was 71a resulting from cleavage of the imino oxazolidinone and 1,4-addition of benzylamine to the ~,[3unsaturated amide, and subsequent isomerization of the remaining olefin into conjugation within the ~-lactam ring (Scheme 26, via intermediacy of 70a). Unfortunately, 71a was obtained as a mixture of diastereomers. Since this 1,4-addition side process was not desired, reduction of the diene in 67a and 67b was attempted using Pd/C and H 2 (1 atm). Interestingly, reduction of either substrate after 4 h at room temperature,

358

KAY M. BRUMMOND AND BRANKO MITASEV m

O

m

Bn"NH O

O

-NL/

BnNH2 (10 equiv.) CDCI3, rt

/--N

olefin

MeO2

NHBn

isomeriza%

O n

"

/

N-'"-~ _H II

NHBn

~

MeO2C

67a

70a

71a

SCHEME 26

gave the 0~,[3-unsaturated amides 72a and 72b in 80% and 95% yield, respectively (Scheme 27). This result was attributed to the steric hindrance around the diene. Presumably, reduction of the more accessible exocyclic double bond led to an intermediate [3,7-unsaturated amide followed by isomerization of the remaining olefin into conjugation. Next, solutions of 72a and 72b in CDC13 were treated with a primary amine (isobutylamine and allylamine, respectively), which clearly effected opening of the imino-oxazolidinone to ureas 73a and 73b. In addition to isobutylamine and allylamine shown in Scheme 27, benzylamine and 2-methoxyethylamine were also used to afford ring opening products in > 80% yield. a3

0

0

03\

Oii

NH

0

,o 67a, R 1 = CO2Me, R 2 = Ph 6 7 b R 1 = i-Pr, R 2 = Me

72a, 80% 72b, 95%

73a, R 1 = CO2Me, R 2 = Ph, R3= i-Bu, 75% 73b, R 1 = i-Pr, R 2 = Me, R 3 = Allyl, 92%

SCHEME 27

The structure of 73a was assigned by ~H NMR (in CDC13) and was based on the presence of two downfield resonances at 9.02 ppm (t, J - 5.8 Hz, 1H) assigned to the urea N-H proton and 8.46 ppm (dd, J - 7.1, 4.6 Hz, 1H)

9

RHODIUM-CATALYZED CYCLOISOMERIZATION REACTIONS OF ALLENES

359

assigned to the amide proton. These N-H resonances were unusually sharp, suggesting an ordered secondary structure of the urea and amide side chains. The downfield chemical shifts of the amide and urea protons support the notion of intramolecular hydrogen bonding. 83 To examine this computationally, 73a was modeled using Cache. 77 The minimized structure of 73a resulted in arrangement of the side chains as shown in Figure 3, with two potential hydrogen bonds: (a) between the urea N-H and lactam carbonyl oxygen (distance 2.287 /k); and (b) between the appending amide N-H and the adjacent pyrrolidine-dione carbonyl oxygen (distance 2.140 ,~). As a result of this secondary bonding, the two side chains are presumably held rigidly, which accounts for the sharp N-H resonances in the NMR spectrum. It was reasoned that this feature may be useful in designing biological probes that project functional groups in specific three-dimensional space. For example, many potent protease inhibitors are small peptide-like molecules that possess a defined secondary structure, resulting in strong interaction with the enzyme. 84 We anticipate that the rigid amido-ureas may also prove as useful biological probes to study protein function.

2.287,~,

"~NH

0

Me02C O " ~ n

-

o

Ph'

2.140A

73a

energy-minimized model of 73a; the two phenyl groups were replaced with a methyl group for clarity purposes FIGURE 3. Energy-minimized model of 73a; the two phenyl groups were replaced with a methyl group for clarity purposes. (See color insert.)

360

KAY M. BRUMMOND AND BRANKO MITASEV

Our focus then shifted to designing a library of these polycyclic scaffolds using the synthetic protocols described above. As outlined in Scheme 28, this synthetic pathway offers at least five points of diversity to be introduced gradually as the complexity of the scaffold increases. The imino-oxazolidinone moiety was envisioned as a crucial part of the triene precursor 76 because it enables a highly stereo- and chemoselective Diels-Alder reaction with a number of maleimides. Furthermore, this moiety is used as a key diversity element, because the transformation of 77 to 78 results in the conversion of a molecule rich in hydrogen-bond

200 compounds of various scaffolds synthesized in solution phase 0

R\ NH

zR2

O

R2

R1 = Me, Bn R2 = Me R3 = various primary amines R4 = H, Me, Et, Ph, -CH2CO2Me R5 = various primary amines

HN

MeO2C---Xr~o. i~1 "~,,,H 74

78

Rh(I) l

t

a4

O R2 1. LiOH. ~..~ 2" R3NH2' EDCI' HN DMAP, HOBt. ,.

O O~

O ...~ N

R20.....~/1~1~..O

O

O

R2

3. phosgene, Eta N. Ra_ 76

75

77 R3 =

a5 _

H2N"~-.~

H2N

H2N'~'-~

H2N~CO2Me

H 2 N ~

H2N~CO2Me

H2N~

OMe

H2N~

SCHEME 28

OH

H2N N..J f"'O H2N~N'v

~

9

RHODIUM-CATALYZED CYCLOISOMERIZATION REACTIONS OF ALLENES

361

acceptors to one that contains two hydrogen-bond donor groups. This transformation is expected to bring about significant differences in the physico-chemical properties and potentially the biological activity of the compounds. Since both classes of compounds were envisioned as library members, a broad range of diversity was accessed via a relatively simple set of transformations. The library synthesis was put into practice by the staff at the University of Pittsburgh Center for Chemical Methodologies and Library Development (UPCMLD, http://ccc.chem.pitt.edu/). Using four points of diversity (allenic amino-ester, amine for oxazolidinone formation, N-alkyl maleimide and amine for the oxazolidinone opening reaction), the center synthesized 200 library members in quantities of 5-100 mg each. 85 These compounds are continuously being sent out for biological testing and are available to the academic and industrial researchers at no cost (see http://ccc.chem.pitt.edu/). VII.

Summary

In summary, we have successfully applied allenic Rh(I)-catalyzed cycloisomerization reactions to DOS. The overall goal of this study was to design robust and efficient methods for the synthesis of complex small molecules to be used as biological probes. As part of our broader DOS strategy, we have developed a synthesis of pivotal allenic amino-ester intermediates by utilizing a Claisen rearrangement of amino-acid propargyl esters. The scope of the allenic cycloisomerization reaction was thoroughly studied with respect to the substitution of the amino acid side chain, amine protecting group and the alkyne and allene reactive moieties. Consequently, we have gained efficient access to various polysubstituted cross-conjugated trienes that would be difficult to access via existing methods. Moreover, a cycloisomerization of amide-tethered allenynes was utilized to prepare novel 8-1actams. This achievement is important, since previously, only y-lactams have been available via transition metalcatalyzed cycloisomerization reactions. The resulting trienes that are accessible via these methods represent a novel class of compounds. Here it was demonstrated that these trienes can be utilized in sequential Diels-Alder reactions to gain rapid access to non-aromatic polyheterocyclic skeletons. The first generation of sterically and electronically undifferentiated trienes underwent the reaction without chemo- or stereoselectivity, thus affording products as mixtures of diastereomers. This obstacle was overcome by designing second generation bicyclic triene precursors that underwent a stereoselective Diels-Alder reaction to afford

362

KAY M. BRUMMONDAND BRANKOMITASEV

tetracyclic skeletons. The synthetic route was then applied to the synthesis of a library of compounds, thus fulfilling the goal of DOS. Acknowledgments

We are grateful to the National Institute of General Medical Sciences (NIGMS P50GM067982) for generous support of this project. B.M. would like to thank the University of Pittsburgh for an Andrew Mellon Fellowship. We also thank Dr. Donald A. Probst and Dr. Bingli Yan for their contribution to this project.

References and Footnotes

1. Arkin, M. R., Wells, J. A., Nat. Rev. Drug. Discov. 2004, 3, 301. 2. (a) Crews, C. M., Mohan, R., Curt. Opin. Chem. Biol. 2000, 4, 47. (b) Crews, C. M., Splittgerber, U., Trends. Biochem. Sci. 1999, 24, 317. 3. (a) Lokey, R. S., Curt. Opin. Chem. Biol. 2003, 7, 91. (b) Schreiber, S. L., Science 2000, 287, 1964. (c) Schreiber, S. L., Bioorg. Med. Chem. 1998, 6, 1127. 4. (a) Lee, M. L., Schneider, G., J. Comb. Chem. 2001, 3, 284. (b) Kim, Y. K., Arai, M. A., Arai, T., Lamenzo, J. O., Dean, E. E, Patterson, N., Clemons, E A., Schreiber, S. A., J. Am. Chem. Soc. 2004, 126, 14740. 5. Fergus, S., Bender, A., Spring, D. R., Curt. Opin. Chem. Biol. 2005, 9, 304. 6. (a) Bunin, B. A., Ellman, J. A., J. Am. Chem. Soc. 1992, 114, 10997. (b) DeWitt, S. H., Proc. Natl. Acad. Sci. U.S.A. 1993, 90, 6909. 7. Wess, G., Urmann, M., Sickenberger, B., Angew. Chem. Int. Ed. 2001, 40, 3341. 8. Lipinski, C. A., Lombardo, F., Dominy, B. W., Feeney, E J., Adv. Drug. Deliv. Rev. 2001, 46, 3. 9. For computational methods for predicting physico-chemical properties, see (a) van de Waterbeemd, H., Gifford, E., Nature Rev. Drug. Disc. 2003, 2, 192. (b) Jorgensen, W. L. Science 2004, 303, 1813. 10. (a) Burke, M. D., Schreiber, S. L., Angew. Chem. Int. Ed. 2004, 43, 46. (b) Arya, E, Joseph, R., Gan, Z., Rakic, B. Chem. Biol. 2005, 12, 163. (c) Burke, M. D., Bergen E. M., Schreiber, S. L., Science, 2003, 302, 613. (d) Beeler, A. B., Schaus, S. E., Porco, J. A. Jr., Curr. Opin. Chem. Biol. 2005, 9, 277. 11. (a) Wipf, P., Coleman, C. M., Janjic, J. M., Iyer, P. S., Fodor, M. D., Shafer, Y. A., Stephenson, C. R. J., Kendal, C., Day, B. W., J. Comb. Chem. 2005, 7, 322. (b) Hotha, S., Tripathi, A., J. Comb. Chem. 2005, 7, 968-976. (c) Hanessian, S., Kothakonda K. K., J. Comb. Chem. 2005, 7, 837-842. (d) Lei, X., Zaarur, N., Sherman, M. Y., Porco, J. A., Jr., J. Org. Chem. 2005, 70, 6474-6483. (e) Oikawa, M., Ikoma, M., Sasaki, M., Tetrahedron Lett. 2005, 46, 415-418. (f) Simon, R. A., Schuresko, L., Dendukuri, N., Goers, E., Murphy, B., Lokey, R. S., J. Comb. Chem. 2005, 7, 697-702. 12. For examples see the comprehensive survey of combinatorial library synthesis for 2005: (a) Dolle, R. E., Bourdonnec, B. L., Morales, G. A., Moriarty, K. J., Salvino, J. M., J. Comb. Chem. 2006, 8, 597. 13. Stockwell, B. R., Nature, 2004, 432, 846.

9

RHODIUM-CATALYZEDCYCLOISOMERIZATIONREACTIONSOF ALLENES

363

14. (a) Blaney, J. M., Martin, E. J., Curt. Opin. Chem. Biol. 1997, 1, 54. (b) Willett, E, Curt. Opin. Biotechnol. 2000, 11, 85. 15. For a recent example, see Spiegel, D. A., Schroeder, E C., Duvall, J. R., Schreiber, S. L., J. Am. Chem. Soc. 2006, 128, 14766. 16. Oguri, H., Schreiber, S. L., Org. Lett. 2005, 7, 47. 17. For recent examples, see (a) Cordeiro, A., Quesada, E., Bonache, M. C., Velazquez, S., Camarasa, M. J., San-Felix, A., J. Org. Chem. 2006, 71, 7224. (b) Wipf, E, Stephenson, C. R. J., Walczak, M. A. A., Org. Lett. 2004, 6, 3009. (c) Chiara, J. L., Garcia, A., Sesmilo, E., Vacas, T., Org. Lett. 2006, 8, 3935. (c) Aurrecoechea, J. M., Suero, R., de Torres, E., J. Org. Chem. 2006, 71, 8767. (d) Micalizio, G. C., Schreiber, S. L., Angew. Chem. Int. Ed. 2002, 41, 3272. 18. For comprehensive summaries of transition metal catalysis in the past decade, see (a) Hegedus, L. S., Coord. Chem. Rev. 1996, 147, 44. (b) Hegedus, L. S., Coord. Chem. Rev. 1997, 161, 129. (c) Hegedus, L. S., Coord. Chem. Rev. 1998, 168, 49. (d) Hegedus, L. S., Coord. Chem. Rev. 1998, 175, 159. (e) Hegedus, L. S., Coord. Chem. Rev. 2000, 204, 199. (f) S6derberg, B. C. G., Coord. Chem. Rev. 2006, 250, 300. (g) S6derberg, B. C. G., Coord. Chem. Rev. 2006, 250, 2411. 19. (a) Beletskaya, I. E, Cheprakov, A. V., Chem. Rev. 2000, 100, 3009. (b) Dounay, A. B., Overman, L. E., Chem. Rev. 2003, 103, 2945. 20. (a) Farina, V., Krishnamurphy, V., Scott, W., J. Org. React. 1997, 50, 1. (b) Duncton, M. A. J., Patterden, G., J. Chem. Soc., Perkin. Trans. 1 1999, 1235. 21. Negishi, E., Anastasia, L., Chem. Rev. 2003, 103, 1979. 22. Stanforth, S. E, Tetrahedron 1998, 54, 263. 23. Trost, B. M., Crawley, M. L., Chem. Rev. 2003, 103, 2921. 24. (a) Frenette, R., Friesen, R. W., Tetrahedron Lett. 1994, 35, 9177. (b) Xie, E, Cheng, G., Hu, Y., J. Comb. Chem. 2006, 8, 286. (c) Ma, Y., Margarida, L., Brookes, J., Makara, G. M., Berk, S. C., J. Comb. Chem. 2004, 6, 426. 25. (a) Hiroshige, M., Hauske, J. R., Zhou, E, J. Am. Chem. Soc. 1995, 117, 11590. (b) Kulkarni, B. A., Ganesan, A., J. Comb. Chem. 1999, 1, 373. (c) Coelho, A., Sotelo, E., J. Comb. Chem. 2006, 8, 388. 26. (a) Plunkett, M. J., Ellman, J. A., J. Am. Chem. Soc. 1995, 117, 3306. (b) Yun, W., Li, S., Wang, B., Chert, L., Tetrahedron Lett. 2001, 42, 175-177. 27. (a) Bolton, G. L., Hodges, J. C., J. Comb. Chem. 1999, 1, 130. (b) Yun, W., Mohan, R., Tetrahedron Lett. 1996, 37, 7189. (c) Zhang, H. C., Maryanoff, B. E., J. Org. Chem. 1997, 62, 1804. (d) Arumugam, V., Routledge, A., Abell, C., Balasubramanian, S., Tetrahedron Lett. 1997, 38, 6473. (e) Yu, Y., Ostresh, J. M., Houghten, R. A., Tetrahedron Lett. 2003, 44, 2569. (f) Krishnan, S., Schreiber, S. L., Org. Lett. 2004, 6, 4021. (g) Xiang, Z., Luo, T., Lu, K., Cui, J., Shi, X., Fathi, R., Chen, J., Yang, Z., Org. Lett. 2004, 6, 3155. (h) Cuny, G., Bois-Choussy, M., Zhu, J., J. Am. Chem. Soc. 2004, 126, 14475. 28. (a) Antilla, J. C., Buchwald, S. L., Org. Lett. 2001, 3, 2077. (b) Kwong, E Y., Klapars, A., Buchwald, S. L., 2002, 4, 581. (c) Yang, B. H., Buchwald, S. L., J. Organomet. Chem. 1999, 576, 125. 29. Masse, C. E., Ng, E Y., Fukase, Y., Sanchez-Rosselo, M., Shay, J. T., J. Comb. Chem. 2006, 8, 293. 30. Kim, Y., Arai, M. A., Arai, T., Lamenzo, J. O., Dean, E. E III, Petterson, N., Clemons, E A., Schreiber, S. L., J. Am. Chem. Soc. 2004, 126, 14740.

364

KAY M. BRUMMONDAND BRANKOMITASEV

31. (a) Negishi, E. In Comprehensive Organic Synthesis, Trost, B. M. (Ed.), Pergamon: Oxford, 1991, Vol. 5, pp. 1163-1184. (b) Schore, N. E., Chem. Rev. 1988, 88, 1081. (c) Trost, B. M., Angew. Chem., Int. Ed. Engl. 1986, 25, 1. (d) Tamao, K., Kobayashi, K., Ito, Y., Synlett 1992, 539. (e) Lautens, M., Klute, W., Tam, W., Chem. Rev. 1996, 96, 49. 32. For a review on synthesis of heterocycles, see Nakamura, I., Yamamoto, Y., Chem. Rev. 2004, 104, 2127. 33. (a) Trost, B. M., Angew. Chem. Int. Ed. 1995, 34, 259. (b) Trost, B. M., Science, 1991, 254, 1471. 34. Trost, B. M., Krische, M. J., Synlett, 1998, 1. 35. Aubert, C., Buisine, O., Malacria, M., Chem. Rev. 2002, 102, 813. 36. (a) Wender, E A., Jenkins, T. E., J. Am. Chem. Soc. 1989, 111, 6432. (b) Wang, B., Cao., P., Zhang, X., Tetrahedron Lett. 2000, 41,8041. (c) Murakami, M., Ubukata, M., Itami, K., Ito, Y., Angew. Chem. Int. Ed. 1998, 37, 2248. 37. (a) Brummond, K. M., Kent, J. L., Tetrahedron, 2000, 56, 3263. (b) Gibson, S. E., Stevanazzi, A., Angew. Chem. Int. Ed. 2003, 42, 1800. (c) Rivero, M. R., Adrio, J., Carretero, J. C, Eur. J. Org. Chem. 2002, 2881. 38. Jolly, R. S., Luedtke, G., Sheehan, D., Livinghouse, T., J. Am. Chem. Soc. 1990, 112, 4965. 39. Burton, B. S., Pechman, H. V., Chem. Ber. 1887, 20, 145. 40. Hendrickson, J. B., Cram. D. J., Hammond, G. B., Organic Chemistry, 3rd edn., McGraw-Hill: New York, 1970, pp. 104-105. 41. For the most recent review, see Ma, S., Chem. Rev. 2005, 105, 2829. 42. (a) Modern Allene Chemistry, Krause, N., Hashmi, A. S. K. (Eds.), Wiley-VCH: Weinheim, 2004. (b) The Chemistry of Ketenes, Allenes, and Related Compounds Part 1, Patai, S. (Ed.), Wiley: New York, 1980. (c) Allenes in Organic Synthesis, Schuster, H. F., Coppola, G. M. (Eds.), Wiley: New York, 1984. 43. For reviews on reactions of allenes, see (a) Hashmi, A. S. K., Angew. Chem., Int. Ed. 2000, 39, 3590. (b) Marshall, J., Chem. Rev. 2000, 100, 3163. (c) Zimmer, R., Dinesh, C., Nandanan, E., Khan, E, Chem. Rev. 2000, 100, 3067 (d) Bates, R., Satcharoen, V., Chem. Soc. Rev. 2002, 31, 12. (e) Ma, S., Topics in Organometallic Chemistry, Tsuji, J. (Ed.), Springer-Verlag: Heidelberg, 2005, pp. 183-210. (f) Sydnes, L. Chem. Rev. 2003, 103, 1133. (g) Brandsma, L., Nedolya, N. A., Synthesis 2004, 735. (i) Tius, M., Acc. Chem. Res. 2003, 36, 284. (h) Wei, L. L., Xiong, H., Hsung, R. P., Acc. Chem. Res. 2003, 36, 773. (i) Lu, X., Zhang, C., Xu, Z., Acc. Chem. Res. 2001, 34, 535. (j) Wang, K. K., Chem. Rev. 1996, 96, 207. 44. Padwa, A., Filipkowski, M. A., Meske, M., Murphree, S. S., Watterson, S. H., Ni, Z., J. Org. Chem. 1994, 59, 591. 45. For examples of 4+ 2. reaction, see (a) Wender, E A., Jenkins, T. E., Suzuki, S., J. Am. Chem. Soc. 1995, 117, 1843. For examples of 5 + 2. reaction, see (b) Wender, E A., B i, E C., Gamber, G. G., Gosselin, F., Hubbard, R. D., Scanio, M. J. C., Sun, R., Williams, T. J., Zhang, L., Pure Appl. Chem. 211112,74, 25. (c) Wender, E A., Glorious, E, Husfield, C. O., Langkopf, E., Love, J. A., J. Am. Chem. Soc. 1999, 121, 5348. For an example of a 2 + 2 + 2. reaction, see (d) Aubert, C., Llerena, D., Malacria, M., Tetrahedron Lett. 1994, 35, 2341. 46. (a) Wender, P. A., Fuji, M., Husfield, C. O., Love, J. A., Org. Lett. 1999, 1, 137. (b) Wender, E A., Zhang, L., Org. Lett. 2000, 2, 2323. 47. Please see reference 45a for an additional example of catalyst-based control of double bond selectivity in a 4 + 2. reaction.

9

RHODIUM-CATALYZEDCYCLOISOMERIZATIONREACTIONSOF ALLENES

365

48. (a)Kent, J. L., Wan, H., Brummond, K. M., Tetrahedron Lett. 1995, 36, 2407. (b) Brummond, K. M. In Advances in Cycloaddition Chemistry, Vol. 6, Harmata, M. (Ed.), JAI Press: Stamford, CT, 1999, pp. 211-237. (c) Brummond K. M., Wan, H., Tetrahedron Lett. 1998, 39, 931. (d) Brummond, K. M., Wan, H., Kent, J. L., J. Org. Chem. 1998, 63, 6535. (e) Brummond, K. M., Lu, J., Petersen, J., J. Am. Chem. Soc. 2000, 122, 4915. (f) Brummond, K. M., Lu, J., Petersen, J., J. Am. Chem. Soc. 2000, 122, 4915. (g) Brummond, K. M., Kerekes, A. D., Wan, H., J. Org. Chem. 2002, 67, 5156. 49. (a) Brummond, K. M., Chen, H., Fisher, K. D., Kerekes, A. D., Rickards, B., Sill, E C., Geib, S. J., Org. Lett. 2002, 4, 1931. (b) Brummond, K. M., Gao, D., Org. Lett. 2003, 5, 3491. 50. Brummond, K. M., Chen, H., Sill, E, You, L., J. Am. Chem. Soc. 2002, 124, 15186. 51. (a) Brummond, K. M., Mitasev, B. M., Org. Lett. 2004, 6, 2245. (b) Brummond, K. M., Curran, D. E, Mitasev, B., Fischer, S., J. Org. Chem. 2005, 70, 1745. (c) Brummond, K. M., Chen, H., Mitasev, B., Casarez, A., Org. Lett. 2004, 6, 2161. (d) Brummond, K. M., Chen, D., Org. Lett. 2005, 7, 3473. 52. (a) Spry, D. O., Bhala, A. R., Heterocycles, 1986, 24, 4641. (b) Ley, S. V., Gutteridge, C. E., Pape, A. R., Spilling, C. D., Zumbrunn, C., Synlett, 1999, 1295. (b) Aoki, Y., Kuwajima, I., Tetrahedron Lett. 1990, 51, 7457. (c) Henderson, M. A., Heathcock, C. H., J. Org. Chem. 1988, 53, 4736. 53. Castelhano, A., Home, S., Taylor, G., Billedeau, R., Krantz, A., Tetrahedron 1988, 44, 5451. 54. Kazmaier, U., G6rbitz, C. H., Synthesis 1996, 1489. 55. (a) Oda, H., Sato, M., Morizawa, Y., Oshima, K., Nozaki, H., Tetrahedron 1985, 41, 3257. (b) Oda, H., Sato, M., Morizawa, Y., Oshima, K., Nozaki, H., Tetrahedron Lett. 1983, 24, 2877. 56. For examples of rate acceleration of cyclization reactions as a result of Thorpe-Ingold effect, see (a) Grubbs, R. H., Chang, S., Tetrahedron 1998, 54, 4413. (b) Alexander, J. B., La, D. S., Cefalo, D. R., Hoveyda, A. H., Schrock, R. R., J. Am. Chem. Soc. 1998, 120, 4041. (c) Yamamoto, Y., Nakagai, Y., Ohkoshi, N., Itoh, K, J. Am. Chem. Soc. 2001, 123, 6372. (d) Okamoto, S., Livinghouse, T. Organometallics 2000, 19, 1449. (e) Buchwald, S. L., Hicks, E A., J. Am. Chem. Soc. 1999, 121, 7026. 57. Rh or Ir coordination to an alcohol or ester is proposed as a reason for stereoselectivity in hydrogenation reactions (a) Crabtree, R. H., Davis, M. W., J. Org. Chem. 1986, 51, 2655. (b) McCloskey, E J., Schultz, A. G., ,/. Org. Chem. 1988, 53, 1380. 58. (a) Tsuge, O., Wada, E., Kanemasa, S., Chem. Lett. 1983, 12, 239. (b) Tsuge, O., Wada, E., Kanemasa, S., Chem. Lett., 1983, 12, 1525. 59. (a) Woo, S., Squires, N., Fallis A. G., Org. Lett. 1999, 1,573. (b) Woo, S., Legoupy, S., Parra, S., Fallis, A. G., Org. Lett. 1999, 1, 1013. 60. For synthetic methods for preparation of acyclic trienes, see (a) Arisawa, M., Sugihara, T., Yamaguchi, M., Chem. Commun. 1998, (23), 2615. (b) Trahanovsky, W. S., Koeplinger, K. A., J. Org. Chem. 1992, 57, 4711. (c) Shi, M., Shao, L. X., Synlett 2004, 807. (d) Moriya, T., Furuuchi, T., Miyaura, N., Tetrahedron 1994, 50, 7961. 61. For a review on cross-conjugated polyenes, see Hopf, H., Angew. Chem. Int., Ed. Engl. 1984, 23, 948. 62. For example, Fallis' methodology was applied by Schreiber and coworkers to the diversity-oriented synthesis of a library of polycyclic small molecules: (a) Kwon, O., Park, S. B., Schreiber, S. L., J. Am. Chem. Soc. 2002, 124, 13402.

366

63. 64.

65. 66. 67. 68. 69. 70. 71.

72. 73. 74. 75. 76. 77.

78. 79.

80.

KAY M. BRUMMONDAND BRANKOMITASEV Forother applications of heteroatom-containing, cross-conjugated trienes, see (b) Spino, C., Liu, G., Tu, N., Girard, S., J. Org. Chem. 1994, 59, 5596. (c) Dion, A., Dub6, E, Spino, C., Org. Lett. 2005, 7, 5601. Payne A. D., Willis A. C., Sherburn M. S., J. Am. Chem. Soc. 2005, 127, 12188. Conformationally defined polycyclic small molecules often exhibit higher potency and specificity of binding to biological targets compared to their acyclic analogues, since they do not have to undergo conformational changes in order to adapt to a binding site. King, F. D., Strategy and Tactics in Drug Discovery. In Medicinal Chemistry: Principles and Practice, 2rid edn., King E D. (Ed.), The Royal Society of Chemistry, Cambridge, UK, 2002, pp. 342-346. See Richard, C. L. Comprehensive Organic Transformations. A Guide to Functional Group Preparations, 2nd edn., Elsevier: Oxford, 1870 pp. Yeung, Y.-Y., Hong, S., Corey, E. J., J. Am. Chem. Soc. 2006, 128, 6310. Serna, S., Tellitu, I., Dominguez, A., Moreno, I., SanMartin, R., Org. Lett. 2005, 7, 3073. Grimaldi, J., Cormons, A., Tetrahedron Lett. 1986, 27, 5089. Hu, T., Li, C., Org. Lett. 2005, 7, 2035. (a) Golden, J. E., Aub6, J., Angew. Chem. Int. Ed. 2002, 41, 4316. (b) Sharghi, H., Hosseini, M., Synthesis, 2002, 1057. For examples of ~,-lactam formation via ring-closing metathesis, see (a) Clayden, J., Tunbull, A., Pinto, I., Tetrahedron: Asymm. 2005, 16, 2235. (b) Badorrey, Cativiela, C., Diaz-de-Villegaz, M. D., Diez, R., Galvez, J. A., Tetrahedron Lett. 2004, 45, 719. For an example of 5-1actam formation via ring-closing metathesis, see (c) Niida, A., Tomita, K., Mizumoto, M., Tanigaki, H., Terada, T., Oishi, S., Otaka, A., Inui, K., Fujii, N., Org. Lett. 2006, 8, 613. (a) Xie, X., Lu, X., Tetrahedron Lett. 1999, 40, 8415. (b) Xie, X., Lu X., Liu, Y., Xu, W., J. Org. Chem. 2001, 66, 6545. Lei, A., Waldkirch, J. P., He, M., Zhang, X., Angew. Chem. Int., Ed. 2002, 41, 4526. Brummond, K. M., Painter, T. O., Probst, D. A., Mitasev, B., Org. Lett. 2007, 8, 347. Stewart, W. E. III, Siddall, T. H., Chem. Rev. 1970, 70, 517. For the unexpected formation of a similar oxazolidinone, see Granier, T., Vasella, A., Helv. Chim. Acta. 1998, 81,865. Geometry optimization was performed with Cache Worksystem Pro. version 6.1.10. First, global minimum search was performed using MM3 parameters, followed by optimization of the side chain conformation with AM I parameters. Then, energy minimization of the sample was performed using AM 1 parameters. Mitasev, B., Yan. B., Brummond, M, Heterocycles, 2006, 70, 367. (a) Tacconi, G., Iadarola, P., Marinone, E, Righetti, E P., Desimoni, G., Tetrahedron 1975, 31, 1179. (b) Rudnichenko, A. V., Timoshenko, V. M., Chernega, A. N., Nesterenko, A. M., Shermolovich, Y. G., J. Fluorine Chem. 2004, 125, 1351. (c) Burdisso, M., Desimoni, G., Faita, G., Righetti, E, Tacconi, G., J. Chem. Soc. Perkin Trans. 2, 1989, 7, 845. For use of Eu(fod) 3 to catalyze hetero Diels-Alder reactions, see (a) Bednarski, M., Danishefsky, S., J. Am. Chem. Soc. 1983, 105, 3716. (b) Dujardin, G., Leconte, S., Coutable, L., Brown, E., Tetrahedron Lett. 2001, 42, 8849. (c) Hughes, K. D., Nguyen, T. L. N., Dyckman, D., Dulay, D., Boyko, W. J., Giuliano, R. M., Tetrahedron: Asymm. 2005, 16, 273. (d) Dujardin, G., Rossignol, S., Brown, E., Synthesis 1998, 763.

9

RHODIUM-CATALYZEDCYCLOISOMERIZATIONREACTIONSOF ALLENES

367

81. Sanchez, I. H., L6pez, E J., Soria, J. J., Larazza, M. I., Flores, H. J., J. Am. Chem. Soc. 1983, 105, 7640. 82. For an example of hydrolysis of oxazolidinones using LiOH, see Davies, S. G., Hermann, G. J., Sweet, M. J., Smith, A. D., Chem. Commun. 2004, (9), 1128. 83. (a) Lagenhan, J. M., Fisk, J. D., Gellman, S. H., Org. Lett. 2001, 3, 2559. (b) Gardner, R., Liang, G. B., Gellman, S. H., J. Am. Chem. Soc. 1999, 121, 1806. (c) Dado, G., Gellman, S. H., J. Am. Chem. Soc. 1993, 115, 4228. (d) Gellman, S. H., Dado, G., Liang, G. B., Adams, B.,J. Am. Chem. Soc. 1991, 113, 1164. 84. (a) Han, Y., Giroux, A., Colucci, J., Bayly, C. I., McKay, D. J., Roy, S., Xanthoudakis, S., Vaillancourt, J., Rasper, D. M., Tam, J., Tawa, E, Nicholson, D. W., Zamboni, R. J., Bioorg. Med. Chem. Lett. 2005, 15, 1173. (b) Quibell, M., Benn, A., Flinn, N., Monk, T., Ramjee, M., Ray, E, Wang, Y., Watts, J., Bioorg. Med. Chem. 2005, 13, 609. (c) Sperka, T., Pitlik, J., Bagossi, P., Toezser, J., Bioorg. Med. Chem. Lett. 2005, 15, 3086. (d) Verhelst, S. H. L., Bogyo, M., ChemBioChem 2005, 6, 824. (e) Grimm, E. L., Roy, B., Aspiotis, R., Bayly, C. I., Nicholson, D. W., Rasper, D. M., Renaud, J., Roy, S., Tam, J., Tawa, P., Vaillancourt, J. P., Xanthoudakis, S., Zamboni, R. J., Bioorg. Med. Chem. 2004, 12, 845. (f) Helal, C. J., Sanner, M. A., Cooper, C. B., Gant, T., Adam, M., Lucas, J. C., Kang, Z., Kupchinsky, S., Ahlijanian, M. K., Tate, B., Mennitti, E S., Kelley, K., Peterson, M., Bioorg. Med. Chem. Lett. 2004, 14, 5521. (g) Johansson, E O., Chert, Y., Belfrage, A. K., Blackman, M. J., Kvarnstroem, I., Jansson, K., Vrang, L., Hamelink, E., Hallberg, A., Rosenquist, A., Samuelsson, B., J. Med. Chem. 2004, 47, 3353. (h) Reid, R. C., Pattenden, L. K., Tyndall, J. D. A., Martin, J. L., Walsh, T., Fairlie, D. E, J. Med. Chem. 2004, 47, 1641. 85. Chambers, P., Turner, D., Werner, S., Mitasev, B., Brummond, K. M. Synthesis of a 200-member Library of Polycyclic Oxazolines and Amides. Manuscript in preparation.

515

COLOR PLATE SECTION

0

..~

2 / N ~

/ C02Me

Me02C

65a

66a

X,

Please refer to Figure 1 of Chapter 9 in text for figure legend.

0

0

F_~O~N.,H .4 Me02C

0==~ P 67a

torsional angle of diene = 39.9 ~

-

torsional angle of o~,l]-unsaturated amide= 42.4 ~

Please refer to Figure 2 of Chapter 9 in text for figure legend.

516

COLOR PLATE SECTION

2.287A

"~NH

O

73a

energy-minimized model of 73a; the two phenyl groups were replaced with a methyl group for clarity purposes Please refer to Figure 3 of Chapter 9 in text for figure legend.

c~~

0

~ MiboZ~2c

N-Ph

0 II toluene,reflux,2hr MeO2(~ ,~ 83%

45f

0

~oN_ph c0z.N~ .... . MeO2~

J

....~-Ph

~

-

54

55 5:2:1 mixture of diastereomers (major diastereomer shown)

-

/

X-ray structure of 55

~

..~

Please refer to Scheme 17 of Chapter 9.

STRATEGIES AND TACTICS IN ORGANIC SYNTHESIS, VOL. 7 9 2008 Elsevier Ltd. All rights reserved.

Chapter 10

HYDROGEN-BOND-MEDIATED SYNTHESIS

IN THE SOLID

ORGANIC STATE

Leonard R. MacGillivray Department of Chemistry University of Iowa Iowa City, IA 52242, USA

I. II.

III.

Introduction and Background The Problem of Crystal Packing Template-Controlled Solid-State Reactivity Using Resorcinol A. Prior Work B. Resorcinol as a Template

IV.

V.

Target-Oriented Synthesis in the Solid State A. [2+2] Paracyclophane B. Ladderanes Other Templates A. 1,8-NAP B. 2,3-NAP C. Reb-Im

Conclusion Acknowledgments References and Footnotes VI.

368 369 371 371 373 374 375 376 378 378 379 380 381 381 381

I. Introduction and Background

Covalent bond-forming reactions (e.g., carbon-carbon bond) lie at the heart of organic synthetic chemistry. ~ Such reactions are used to construct molecules of remarkable complexities. Methods that form covalent bonds in the most efficient manner possible - in high yields, with limited byproducts, and with minimal waste- are highly valued. 2 Organic chemists continue to search and develop improved ways to control the formation of covalent bonds. In this context, the organic solid state represents an intriguing medium within which to control the formation of covalent bonds. 3 The solid state is sufficiently flexible to allow atoms to move and react, yet

10

HYDROGEN-BOND-MEDIATED ORGANIC SYNTHESIS IN THE SOLID STATE

369

sufficiently rigid to enable reactions to proceed with remarkable stereocontrol. The homogeneous nature (i.e., 3D regularity) of the solid state means that covalent-bond-forming reactions that occur can proceed in up to quantitative yield. The benefit of controlling the formation of covalent bonds in such a solvent-free environment is apparent. 4 The solid state also enables molecules to adopt geometries impossible to achieve in the liquid phase. This means that opportunities exist to form molecules in the solid state that may be inaccessible in solution. Although covalent bonds have been known to form in the organic solid state for more than a century, 3 it has only been within the last decade that chemists have developed tools and concepts that are enabling the solid state to be used as a medium to construct molecules by design. 5 Rapid advances in the field of X-ray crystallography (e.g., CCD X-ray diffraction technology) have permitted structural data to be collected and analyzed on timescales of minutes-to-hours as opposed to days-to-months using more traditional instruments (e.g., single-point detectors). Moreover, conceptual advances in the field of crystal engineering 6 and the related field of supramolecular chemistry 7 have resulted in marked improvements of our understanding of how intermolecular forces influence the organization of molecules such that the design and construction of solids with predetermined physical properties (e.g., reactivity, optical activity) is now becoming a reality. It is with these ideas in mind that work performed in our laboratory during the past 5 years that has been aimed to achieve control of a carboncarbon-bond-forming reaction in the solid state; namely, the [2+2] photodimerization, 8 will be discussed. It will be shown that control of this cycloaddition reaction can be achieved by using small organic molecules, termed linear templates, 9 which are used to assemble and preorganize alkenes in geometries suitable for the photoreaction. The templates operate by assembling the alkenes via hydrogen bonds. It will be demonstrated that this supramolecular approach to covalent bond formation can be used to code the construction of carbon-carbon bonds with virtually perfect control of regiochemistry. Whereas our initial focus was to expand the nature of the templates, we are beginning to learn how this method can be employed to construct molecules with architecturally rich frameworks. II. The Problem of Crystal Packing

The field of crystal engineering finds its origins in the early work of Gerhardt Schmidt who, through a large number of crystallographic investigations, determined specific geometry criteria for a [2 + 2] photodimerization

370

LEONARD R. MACGILLIVRAY

to occur in the solid state. 8 By systematically studying a series of cinnamic acids, Schmidt determined that two carbon-carbon double (C =C) bonds should be aligned parallel and separated by less than 4.2 A to react to form a cyclobutane product. Thus, UV-irradiation of ~-trans-cinnamic acid, a polymorph of cinnamic acid with olefins arranged head-to-tail, was shown to produce ~-truxillic acid while irradiation of ~-trans-cinnamic acid, with olefins arranged head-to-head, produced [3-truxinic acid (Scheme 1). Cinnamic acid itself undergoes cis-trans isomerization in the liquid phase; thus, the fact that the acid underwent a dimerization in the solid state represented an intriguing departure from solution-phase chemistry. Schmidt also showed that the rigid environment of the solid state enables the dimerization to occur stereospecifically such that the molecular structure of the product is consistent with the geometric relation of the olefinic reactants in the crystal lattice. From a purely topological standpoint, the photodimerization provides a simple way to crosslink two stacked molecules.

/---cO2H hv ~ CO2H solid 13-cinnamicacid HO2C

~

truxinicacid hv

CO2H

CO2H CO2H

solid

HO2C ~~'--

o~-cinnamicacid

CO---~

truxillicacid SCHEME 1

Whereas the work of Schmidt provided criteria for a [2+2] photodimerization to occur, Schmidt also showed how the extreme sensitivity of solid-state organization to subtle changes to molecular structure severely hampered the synthetic value of the covalent-bond-forming process. ~~In particular, Schmidt showed that members of a homologous series of cinnamic acids did not exhibit homologous reactivities in the solid state. Such differences in reactivity were shown to result from the steric and electronic influences of the organic functional groups on the crystal packing. Thus, whereas p-methylcinnamic acid was found to form a solidstate structure that was photoactive, o-methylcinnamic acid formed a structure that was photostable. These observations lie in stark contrast to the

10

HYDROGEN-BOND-MEDIATED ORGANIC SYNTHESIS IN THE SOLID STATE

371

liquid phase, where such subtle changes to the structures of reactants typically do not have such adverse effects on geometry and reactivity. III. Template-Controlled

Solid-State Reactivity Using Resorcinol

In 2000, our group embarked on studies to determine whether it is possible to separate the [2+2] photodimerization from the effects of crystal packing. We hypothesized that by effectively decoupling reactivity from packing, we could broaden the synthetic scope of the solid-state reaction. In particular, we expected that the photoreaction could be decoupled from packing using small, ditopic molecules that function as linear templates. 9 In this design, a template would assemble and preorganize, via relatively strong directional forces such as hydrogen bonds, two functionalized olefins in a geometry that is both suitable for the photoreaction and independent of crystal packing (Scheme 2). In the minimalist case, a single template could assemble two monofunctionalized olefins for the photoreaction. Additional copies of the template could also be integrated into the proposed reactivity scheme so as to exploit the process of selfassembly. Owing to the fact that the template would assemble along the exterior of two stacked olefins, we expected that the approach could be eventually used to modify the structure of the olefin (e.g., addition of a functional group) so as to apply the templates to a variety of reactants and, therefore, construct a variety of products.

2 E

+2R--R

crystallize

--R--R--

E r

-R--R-

]

~

--R--R--

E, ,3 - -"

dissolve

-

"~

2

E

+ I I "R'--R -"

= linear template, R - - R = reactant

SCHEME 2

A. P R I O R W O R K

Prior to our proposed work, there had been two studies that suggested that the [2+2] photodimerization could be separated from the effects of long-range packing. First, Campbell and Feldman had described an olefinic 'J'-shaped dicarboxylic acid that self-assembled in the solid state to form a dimer held together by four O - H - - - O hydrogen bonds (Scheme 3). ~ The C = C bonds

372

L E O N A R D R. M A C G I L L I V R A Y

of the monomer units conformed to Schmidt's geometry criteria for a photodimerization within the dimer. Moreover, that the carboxylic acid formed a 0D hydrogen-bonded assembly (i.e., where the strongest intermolecular forces occurred within the dimer) meant that the arrangement of the olefins, and reactivity, would be largely independent of the effects of packing. When irradiated with UV radiation, the solid produced the expected cyclobutane product in quantitative yield.

"H-O x

O'H"O~V hv

"

~.0""H'O.

lsolid state

SCHEME

3

Second, Ito and Scheffer described a series of crystalline diammonium cinnamate salts in which the ammonium groups participated in N + - H - . - O - hydrogen bonds with the cinnamates (Scheme 4). 12 A number of salts were photoactive, which was accounted for by the ammonium ions adopting gauche conformations (e.g., ethylenediammonium) that forced the cinnamates to stack in positions to react. The ammonium ions also adopted anti conformations that forced the cinnamates into unreactive geometries. In some cases, the ammonium ions participated in N + - H . . . O - hydrogen bonds with other components of the lattice so as to give infinite hydrogen-bonded polymers that were photostable. Nevertheless, the studies involving the salts demonstrated that olefins could be forced into reactive geometries using a ditopic system.

NH; "OOC ,~~Ar (CH2)n'rr"~"NH~"OOC ~.,,A r

"OOC'~"Ar state "OOC.,,~~Ar hv

solid

SCHEME 4

10

HYDROGEN-BOND-MEDIATED ORGANIC SYNTHESIS IN THE SOLID STATE

B. R E S O R C I N O L

373

AS A TEMPLATE

Our first studies involved developing 1,3-dihydroxybenzene, or resorcinol, as a linear template. ~2 Specifically, we anticipated that cocrystallization of resorcinol with trans-l,2-bis(4-pyridyl)ethylene (4,4'-bpe) would produce a four-component complex, of composition 2(resorcinol).2(4,4'-bpe), held together by four O - H . . - N hydrogen bonds wherein two molecules of 4,4'-bpe would be positioned for a [2+2] photodimerization (Scheme 5). In line with our strategy, cocrystallization of resorcinol with 4,4'-bpe produced the 0D complex 2(resorcinol)-2(4,4'-bpe). The two C =C bonds of the two olefins were oriented parallel and separated by 3.65 A, an ideal position for a photodimerization. UV-irradiation of the solid produced the anticipated photoproduct, rctt-tetrakis(4-pyridyl)cyclobutane (4,4'-tpcb). The product formed stereospecifically and in 100% yield. The structure of the product was determined using single-crystal X-ray analysis in which 4,4'-tpcb was shown to assemble with resorcinol in the three-component complex 2(resorcinol).(4,4'-tpcb). We also determined that the template-controlled solid-state reaction could be conducted in gram-scale amounts.

1

o...

.....

.......

template

~0

hv

-H..... N ~ " ~ X

-H..... N -,........J

solid state

template

J,~'~tN ..... H-O

~~-~N

..... H-O

SCHEME 5

Following our initial report, we demonstrated that resorcinol could be used to assemble an olefin with a single 4-pyridyl group. ~4 In particular, resorcinol was used to assemble two stilbazoles in (resorcinol).2(transl-(4-pyridyl)-2-(4-chlorophenyl)ethylene) in a head-to-head geometry for a regiocontrolled photodimerization that produced the corresponding

LEONARDR. MACGILLIVRAY

374

head-to-head product rctt-l,2-bis(4-pyridyl)-3,4-bis(4-chlorophenyl) cyclobutane in quantitative yield (Scheme 6). C1 ... C1 interactions also formed between the hydrogen-bonded complexes such that the nearestneighbor assemblies constituted six-component assemblies held together by both O - H . . . N and C1...C1 forces. As a consequence of these interactions, the olefins were organized in close proximity, the two C = C bonds being separated by 3.98 A. Unlike 2(resorcinol).2(4,4'-bpe), however, the olefins adopted an antiparallel orientation, which was expected to render the C =C bonds photostable.

O'H ..... N

o.,

~

CI

c,

..... ..,

hvlS~ state

_!-. .....

c'

O-H ..... N r

~

Cl

SCHEME6 We attributed the generation of the photoproduct to the olefins undergoing a pedal-like change in conformation in the solid state that enabled the C = C bonds to adopt a parallel orientation suitable to react and form the product. IV. Target-Oriented Synthesis in the Solid State

That resorcinol served as a template to direct the [2+2] photodimerization within crystalline assemblies prompted us to explore whether the method could be applied to more structurally diverse olefins. In particular, we decided to study di- and tri-olefins as reactants. We hypothesized that the self-assembly process involving the templates could adapt to a lengthening of the reactants with the placement of additional C - C bonds between the pyridyl groups.

10

HYDROGEN-BOND-MEDIATED ORGANIC SYNTHESIS IN THE SOLID STATE

375

A. [2 + 2] P A R A C Y C L O P H A N E

Specifically, we expected that we could use a resorcinol to construct a lengthened molecule in the form of a [2.2]paracyclophane. The paracyclophane framework, introduced by Cram, bears relevance to physical organic chemistry and materials science owing to its reactive, photochemical, and electrochemical properties. 15 Despite these features, however, the synthesis of molecules based on [2.2]paracyclophane remains a synthetic challenge. A retrosynthetic analysis of the targeted [2.2]paracyclophane suggested that the molecule could be constructed from the bifunctional diene trans,trans-l,4-bis[2-(4-pyridyl)ethenyl]benzene (1,4-bpeb). Cocrystallization of resorcinol, or a derivative, with 1,4-bpeb would yield the four-component complex, 2(resorcinol).2(1,4-bpeb), with the two dienes positioned for a double [2+2] photodimerization. UV-irradiation of the solid would generate the cyclophane target tetrakis(4-pyridyl)-l,2,9,10diethano[2.2]paracyclophane (4,4'-tppcp). In the ideal case, the target would form in the solid state in quantitative yield and gram quantities. We determined that co-crystallization of 5-methoxyresorcinol (5OMe-res) with 1,4-bpeb produced the four-component assembly 2(5OMe-res).2(1,4-bpeb) (Scheme 7). z3 Similar to 2(resorcinol).2(4,4'-bpe), the components were held together by four O - H . . . N hydrogen bonds with the C=C bonds being separated by 3.70 A. The assemblies packed

O ' H ..... MeO

N..r~-'~

~~-~t~N ...... H-O

.--~

--/?OMe b'H ..... N.,':'--~

~,~,

~,~ ....... ,,,.4 n'u

hvlS~ state O-H.....

~l..... H-O OMe

SCHEME 7

376

LEONARD R. MACGILLIVRAY Q

-0

O-

FIGURE 1. ORTEP perspective of the targeted [2.2]paracyclophane.

with C =C bonds of adjacent assemblies parallel and separated by 3.95 A, a distance also suitable for a photodimerization. As determined by ~H NMR spectroscopy, UV-irradiation of 2(5-OMe-res).2(1,4-bpeb) produced 4,4'-tppcp in 60% yield. In addition to the paracyclophane, a monocyclized dimer and indefinable products formed. We attributed the formation of the side products to cross-reactions between nearestneighbor hydrogen-bonded structures. Later, we showed that 4-benzylresorcinol (4-bn-res) yields 4,4'-tppcp stereospecifically in 100% yield. 16 Similar to 2(5-OMe-res).2(1,4-bpeb), co-crystals of 2(4-bn-res).2(1,4-bpeb) consisted of four-component assemblies with the two dienes positioned for a double [2 + 2] photodimerization. However, in contrast to 2(4-bn-res).2(1,4-bpeb), the closest separation between C - C bonds of adjacent assemblies was 5.4 *. This longer distance was outside the distance criterion of Schmidt for a photoreaction. We attributed the different packing of the hydrogen-bonded assemblies to steric and electronic influences of the benzyl group attached to the resorcinol template. UV-irradiation of 2(4-bn-res).2(1,4-bpeb) produced 4,4'tppcp in 100% yield. The structure of the [2.2]paracyclophane target was confirmed via single-crystal X-ray diffraction (Figure 1). B. L A D D E R A N E S

In addition to a [2.2]paracyclophane, we anticipated that a resorcinol could be used to construct [n]-ladderanes (where n = 3 or 5) in the solid state. Near the time of our proposed work, ladderanes had been discovered in intracellular membrane lipids of anaerobic ammonium-oxidizing, or anammox, bacteria. 17The ladderanes were shown to rigidify the membrane of an organelle that allows the bacteria to participate in the oceanic nitrogen cycle. A retrosynthetic analysis of a ladderane suggested that a template-controlled [2+2] photodimerization of the conjugated diene

10

HYDROGEN-BOND-MEDIATED ORGANIC SYNTHESIS IN THE SOLID STATE

377

trans, trans- 1,4-bis(4-pyridyl)- 1,3-butadiene (1,4-bpbd) would yield the corresponding [3]-ladderane, all-trans-tetrakis(4-pyridyl)-[3]-ladderane (4-tp-3-1ad). Likewise, a templated [2+2] photodimerization of the triene trans, trans, trans- 1,6-bis(4-pyridyl)- 1,3,5-hexatriene (1,6-bpht) would produce the corresponding [5]-ladderane all-trans-tetrakis(4pyridyl)-[5]-ladderane (4-tp-5-1ad) (Scheme 8). O-H ..... N ~" ' ~ ~ ' ' ~ ~ ~

~

9

Q.H

~

H0

. . . . . . .

..... N

....... H . ~

hvlS~ state

c -H ..... O-H ..... N ~,,,,,~

~ N - -

-H-O

SCHEME 8

We determined 5-OMe-res to be a template for the quantitative construction of both the [3]- and [5]-ladderanes. ~8In particular, co-crystallization of 5-OMe-res with either 1,4-bpbd or 1,6-bpht produced the four-component assemblies 2(5-OMe-res).2(1,4-bpbd) and 2(5-OMe-res).2(1,6-bpht), respectively. In each case, the C =C bonds of the polyenes were organized in appropriate positions for the photoreaction. As established via 1H NMR spectroscopy, UV-irradiation of each solid produced the targeted [3]- and [5]-ladderane in gram amounts and 100% yield. Each product was characterized via single-crystal X-ray analysis (Figure 2). The template-controlled

w n

wD..

..

..

' ,

0

w

.

n

n

n

..

o

" 2- L......... ! 1 1 I1-,,5 .~7 ~ :, ~-,r,%.,a FIGURE 2. ORTEP perspective of the [5]-ladderane.

.

378

LEONARD R. MACGILLIVRAY

solid-state [2+2] photodimerization of 1,4-bpbd in 2(5-OMe-res).2(1,4bpbd) provided the first example of a high-yielding solid-state synthesis of a [3]-ladderane while the reaction within 2(5-OMe-res)-2(1,6-bpht) represented the first example of a solid-state synthesis of a [5]-ladderane. V. Other Templates

That we could use templates based on resorcinol to construct a [2.2]paracyclophane and [n]-ladderanes in the solid state prompted us to further explore the scope of the method. We reasoned that having additional templates on hand could increase the synthetic flexibility of the method by providing a means, for example, to increase a product yield where a resorcinol may be less effective. Additional templates could also be used with recognition sites other than 4-pyridyl groups attached to the olefins. In addition to expanding the method in terms of the supramolecular synthesis, different recognition sites could be used to increase the structural diversity of the photoproducts. Since our initial studies involving resorcinol, we have determined that 1,8-naphthalenedicarboxylic acid (1,8-nap), ~9 2,3-bis(4-methylenethiopyridyl)naphthalene (2,3-nap), 2~and Rebek's imide (Reb-im) 21 function as linear templates in the solid state. A. 1,8-NAP

Our first study to extend linear templates beyond resorcinol involved 1,8-naphthalene dicarboxylic acid (1,8-nap, Scheme 9). 19 Similar to 2(resorcinol).2(4,4'-bpe), co-crystallization of 1,8-nap with 4,4'-bpe was expected to produce the four-component assembly 2(1,8-nap).2(4,4'-bpe). That the 1,8-naphthalene framework had been previously used to enforce face-to-face stacking of aromatics provided a measure of support for our hypothesis. 22 That 1,8-nap was known to form a hydrogen-bonded dimer in the solid state also supported our hypothesis. 23 We expected that each carboxylic acid group of each diacid would interact with the 4-pyridyl group of 4,4'-bpe via an O - H . . . N hydrogen bond. As expected, cocrystallization of the two components produced a four-component complex with two C = C bonds organized parallel and separated by 3.73 ,~. UV-irradiation of the solid produced 4,4'-tpcb stereospecifically in 100% yield.

10

379

HYDROGEN-BOND-MEDIATED ORGANIC SYNTHESIS IN THE SOLID STATE

O'H ...... N ~ ' ~ / ~ '

~

-~.

:.o O-H ...... N I

I

~

~ ~,~

~N ...... H ' O

o.~

~N .....

I

-H-O 9

I

hv 1 solid

1,8-nap

1,8-nap !

state

~--%..

........N , ; ~ ~ ,~Z~~ ...... o,

,~ ~/~O-H

--.-o\ ........N

N.........H-O SCHEME 9

B. 2,3-NAP Following our work involving resorcinol and 1,8-nap, it occurred to us that the hydrogen-bonding, or code, involving the templates and reactants could be reversed. Thus, whereas resorcinol and 1,8-nap function as hydrogen-bond-donors, we aimed to develop a template that functioned as a bifunctional hydrogen-bond acceptor. Following a study involving agentophilic forces, 24 we determined that code reversal could be achieved using 2,3-bis(4-methylenethiopyridyl)naphthalene (2,3-nap) (Scheme 10). 2~ In particular, co-crystallization of 2,3-nap with fumaric acid (fum) produced the four-component assembly 2(2,3-nap).2(fum) held together by four O - H . . - N hydrogen-bonds. Two acid molecules were juxtaposed by the

S

[ ~ [ ~

"

O

~

0

S ---~"~s~N ....... H'O -"~~2N

O'H ...... N~/'~-'~---/S

...... H"O~"

~""~r

...... N:~'~/~)'--" S

m

I

m

I

hv i

solid ~state

2,3-nap S ~"-~'~s~N ..... H

0<,,,,~ ' O

~

_/,2~ O'H ..... N~

. . . . . .

SCHEME 10

I

I

2,3-nap sj~--S

I

380

LEONARD R. MACGILLIVRAY

bipyridines, with the C - C bonds being separated by 3.84 ,~. UV-irradiation of the solid produced rctt-1,2,3,4-cyclobutanetetracarboxylic acid in up to 70% yield. The photoreaction was also shown to proceed via a rare singlecrystal-to-single-crystal (SCSC) transformation, 25 with the structure of the photoproduct being confirmed using single-crystal X-ray diffraction. C. R E B - I M

Whereas resorcinol, 1,8-nap, and 2,3-nap are symmetrical molecules, we expected that an unsymmetrical molecule could be used to direct the [2 + 2] photodimerization in the solid state. In the simplest case, an unsymmetrical molecule could assemble two identical and symmetrical olefins for the photoreaction. In such an assembly process, the two different 'hands' of the template would interact with two identical 'handles' of the olefins. Such an assembly would also represent a rare form of degeneracy in a molecular recognition process. Following early work of Rebek, 26 and the more recent work of Dietrich, 27 we hypothesized that Reb-im could be used to assemble two molecules of 4,4'-bpe for a [2+2] photodimerization. Prior work by our group had also shown that Reb-im self-assembles in the solid state, similar to 1,8-nap, to give a hydrogen-bonded dimer. 28As expected, co-crystallization of Reb-im with 4,4'-bpe produced the fourcomponent hydrogen-bonded assembly, 2(Reb-im)-2(4,4'-bpe), in which the two olefins were parallel and separated by 3.70 A (Scheme 11).

o 'H"

~'~s~N---H'O. / O-~-

~

,H . . . . .

~~/~8'H.... Nr

N

....

9. . . . .

H,

~',~N ......H g ? - - ~ SCHEME 11

10

HYDROGEN-BOND-MEDIATEDORGANICSYNTHESISIN THE SOLID STATE

381

A toluene molecule was also trapped within the crystal lattice. UV-irradiation of the solid produced 4,4'-tpcb stereospecifically in 100% yield. We also demonstrated that the reaction occurred via an SCSC reaction. VI. Conclusion

In this chapter, it has been shown that small organic molecules, in the form of linear templates, can be used to direct the [2 + 2] photodimerization in the solid state. The ability to achieve independent control of solidstate reactivity by exploiting the process of self-assembly has been used to construct molecular targets, such as a [2.2]paracyclophane and [n]ladderanes, in quantitative yield and gram amounts. With the fields of crystal engineering and supramolecular chemistry developing at rapid rates, it is expected that this method to control the formation of covalent bonds and direct the formation of molecules can be used to form targets of increasing structural complexity. Acknowledgments

The NSF (CAREER Award, L.R.M., DMR-133138) is greatly acknowledged for support of this work. I also thank my outstanding co-workers listed in the references below.

References and Footnotes

1. Nicolaou, K. C., Vourloumis, D., Winssinger, N., Baran, E S., Angew. Chem. Int. Ed. Engl. 2000, 3 9, 44-122. 2. Trost, B. M., Acc. Chem. Res. 2002, 35, 695-705. 3. Braga, D., Grepioni, E, Angew. Chem. Int. Ed. 2004, 43, 4002-4011. 4. Anastas, E T., Warner, J. C., Green Chemistry: Theory and Practice, Oxford University Press: New York, 1998. 5. MacGillivray, L. R., CrystEngComm 2002, 4, 37-41. 6. (a) Desiraju, G. R., Angew. Chem. Int. Ed. Engl. 1995, 34, 2311-2327. (b) Aaker6y, C. B., Salmon, D. J., CrystEngComm 2005, 72, 439-448. 7. (a) Fyfe, M. C. T., Stoddart, J. E, Acc. Chem. Res. 1997, 30, 393--401. (b) Lehn, J.-M. SupramolecuIar Chemistry, Wiley-VCH: Weinheim, 1995. 8. Schmidt, G. M. J., Pure Appl. Chem. 1971, 27, 647-678. 9. Anderson, S., Anderson, H. L. In: Templated Organic Synthesis, Diederich, F., Stang, P. J. (Eds.), Wiley-VCH: New York, 2000, pp. 1-38. 10. (a) Cohen, M. D., Schmidt, G. M. J., J. Chem. Soc. 1964, 1996-2000, (b) Cohen, M. D., Schmidt, G. M. J., Sonntag, E I., J. Chem. Soc. 1964, 2000-2013, (c) Schmidt, G. M. J., J. Chem. Soc. 1964, 2014-2021. 11. Feldman, K. S., Campbell, R. E, J. Org. Chem. 1995, 60, 1924-1925.

382

LEONARD R. MACGILLIVRAY

12. (a) Ito, Y., Borecka, B., Trotter, J., Scheffer, J. R., Tetrahedron Lett. 1995, 36, 6083-6086. (b) Ito, Y., Borecka, B., Olovsson, G., Trotter, J., Scheffer, J. R., Tetrahedron Lett. 1995, 36, 6087-6090. 13. MacGillivray, L. R., Reid, J. L., Ripmeester, J. A., J. Am. Chem. Soc. 2000, 122, 7817-7818. 14. MacGillivray, L. R., Reid, J. L., Ripmeester, J. A., Papaefstathiou, G. S., Ind. Eng. Chem. Res. 2002, 41, 4494-4497. 15. (a) Cram, D. J., Steinberg, H., J. Am. Chem. Soc. 1951, 73, 5691-5704. (b) Cram, D. J., Cram, J. M., Acc. Chem. Res. 1971, 4, 204-213. (c) Boekelheide, V., Acc. Chem. Res. 1980, 13, 65-70. 16. Friscic, T., MacGillivray, L. R., Chem. Commun. 2003, 1306-1307. 17. Sinninghe Damste, J. S., Strous, M., Rijpstra, W. I. C., Hopmans, E. C., Geenevasen, J. A. J., van Duin, A. C. T., van Niftrik, L. A., Jetten, M. S. M., Nature 2002, 419, 708-712. 18. Gao, X., Friscic, T., MacGillivray, L. R., Angew. Chem. Int. Ed. 2004, 43, 232-236. 19. Papaefstathiou, G. S., Kipp, A. J., MacGillivray, L. R., Chem. Commun. 2001, 2462-2463. 20. Friscic, T., MacGillivray, L. R., Chem. Commun. 2005, 5748-5750. 21. Varshney, D. B., Gao, X., Friscic, T., MacGillivray, L. R., Angew. Chem. Int. Ed. 2006, 45, 646-650. 22. Iyoda, M., Kondo, T., Nakao, K., Hara, K., Kuwatani, Y., Yoshida, M., Matsuyama, H. A., Org. Lett. 2000, 2, 2081-2083. 23. Fitzgerald, L. J., Gallucci, J. C., Gerkin, R. E., Acta Crystallogr. Sect. B: Struct. Sci. 1991, B47, 776-782. 24. Zhao, Y.-Z., Hong, M.-C., Liang, Y.-C., Su, W.-P., Cao, R., Zhou, Z.,-Y., Chan, A. S. C., Polyhedron 2001, 20, 2619-2625. 25. Friscic, T., MacGillivray, L. R. Z. Kristallogr. 2005, 220, 351-363. 26. Stack, J. G., Curran, D. P., Geib, S. V., Rebek, J. Jr., Ballester, P., J. Am. Chem. Sor 1992, 114, 7007-7018. 27. Faraoni, R., Castellano, R. K., Gramlich, V., Diederich, E, Chem. Commun. 2004, 370-371. 28. Gao, X., Friscic, T., Papaefstathiou, G. S., MacGillivray, L. R., J. Chem. Crystallogr. 2004, 34, 171-174.

STRATEGIES AND TACTICS IN ORGANIC SYNTHESIS, VOL. 7 9 2008 Elsevier Ltd. All rights reserved.

Chapter 11

TOTAL SYNTHESES OF NATURAL PRODUCTS USING THE COMBINED C-H ACTIVATION/COPE REARRANGEMENT AS THE KEY STEP

Huw M. L. Davies and Xing Dai Department of Chemistry University at Buffalo, The State University of New York Buffalo, NY 14260, USA

I. II.

Introduction Marine Natural Products from Pseudopterogorgia elisabethae A. General Overview B. Synthetic Challenges III. C-H Activation as a Strategic Reaction in Synthesis A. C-H Activation as a Surrogate of Classic Organic Reactions B. Combined C-H Activation/Cope Rearrangement C. Enantiodifferentiation: Model Study IV. Total Synthesis of (+)-Erogorgiaene V. 1btal Syntheses of (-)-Colombiasin A and (-)-Elisapterosin B A. Retro-synthesis of (-)-Colombiasin A B. Key Step: Combined C-H Activation/cope Rearrangement C. Total Syntheses of (-)-Colombiasin A and (-)-Elisapterosin B VI. Total Syntheses of (+)-Elisabethadione and Related Natural Products A. Total Synthesis of (+)-Elisabethadione B. Total Synthesis of A p-Benzoquinone Natural Product C. Total Synthesis of (+)-O-Methyl-Elisabethadione D. Total Synthesis of (+)-O-Methyl-nor-Elisabethadione E. Synthesis Toward (+)-Elisabethamine VII. Summary Acknowledgments References and Footnotes

383 384 384 385 387 387 388 392 393 394 394 395 397 399 399 400 401 402 403 404 404 405

I. Introduction

Effective methods for the functionalization of inactivated C-H bonds offer opportunities for developing new strategies for organic synthesis. ~ C-H activation has been a very active area of research for

384

HUW M.L. DAVIES AND XING DAI

the organometallic community. Since the seminal work of Bergman, who demonstrated a stoichiometric conversion of pentane to 1-pentene using a ruthenium complex, 2 considerable effort has been expended to develop catalytic and ultimately enantioselective methods for C-H activation. In these methods, the key step is usually the oxidative addition of a metal complex across a C-H bond and the challenge has been to develop systems capable of cleaving the strong C-H bond and of achieving catalyst turnover. Since 2000, a number of impressive advances have been made. Hartwig developed a very elegant iridium-catalyzed method for the borylation of alkanes with exquisite selectivity for primary C-H bond functionalization. 3 Murai has conducted extensive studies on aromatic C-H functionalization by using a neighboring group to direct the chemistry 4 and this general strategy has been expanded by many including Bergman~llman, 5 Sames, 6 Fagnou 7 and Sanford. 8 The application of this type of C-H activation chemistry to enantioselective reactions is still in its infancy. Sames described a very interesting example using a pre-formed chiral palladium complex, which was capable of a stereoselective C-H activation, leading to the synthesis of the natural product (-)-rhazinilam. 9 One of the most impressive catalytic enantioselective C-H activation methods has been reported by Bergman and Ellman in the enantioselective synthesis of a protein kinase C (PKC) inhibitor. 1~ For the last few years, our group has been exploring the possibility of developing practical methods for enantioselective intermolecular C-H activation by means of metal carbenoid-induced C-H insertion. ~ We have made considerable progress in this area such that the metal carbene C-H insertions can be considered as the most general methods to date for regio-, diastereo- and enantioselective C-H functionalization. 12 In this chapter we describe how a newly discovered transformation from our laboratories, the combined C-H activation~Cope rearrangement, can be broadly applied to the synthesis of a family of marine diterpenes isolated from the soft coral Pseudopterogorgia elisabethae. II. Marine Natural Products from Pseudopterogorgia elisabethae A. G E N E R A L O V E R V I E W

The species Pseudopterogorgia elisabethae (Bayer) is found in the tropical western Atlantic, off the coast Florida, the Bahamas, Cuba, Jamaica, Honduras, Belize and Mexico. In the 1990s, this octocoral became the target

11

385

TOTALSYNTHESES OF NATURAL PRODUCTS Me

H MeH

O Me

Me~ ~ ] H

Me

M

e

Me~e "Me

(+)-elisabethatriene (+)-erogorgiaene 1

2

O Me

O Me

OH Me

Me

(-)-colombiasinA

(-)-elisapterosinB

3

4

OH Me MeHM~~H Me,Me Me

R1O~ Me ],,~/~He

M~ Me Me Me ...~_..OH (+)-elisabethadione (+)-p-benzoquinone (+)-elisabethamine pseudopterosins 7 8 5 6 FIGURE 1. Representative marine natural products from Pseudopterogorgia elisabethae.

of extensive chemical investigations leading to the isolation and characterization of a remarkable number of diterpenoid secondary metabolites. 13Most of these newly discovered compounds are based on unprecedented carbon skeletons and often feature unusual structural characteristics (Figure 1). The diverse family of diterpenes, comprising from bicyclic to polycyclic systems are all derived biosynthetically from (+)-elisabethatriene (1). 14 Examples of these natural products are (+)-erogorgiaene (2), 15 (-)-colombiasin A (3), 16'17 (-)-elisapterosin B (4), 18 (+)-elisabethadione (5), 19 (+)-p-benzoquinone (6), 2~ (+)-elisabethamine (7) 21 and pseudopterosins (8) 22. Many members of this super-family display substantial biological activity as anti-inflammatory, anticancer, antitubercular and/or general antibacterial agents. B. S Y N T H E T I C C H A L L E N G E S

Due to the common biosynthetic ancestry of these natural products, 14 all have three distinctive stereocenters. From a synthetic perspective, these three stereocenters have represented considerable challenges. 13,~5-22

386

HUW M.L. DAVIES AND XING DAI

The stereochemical problems associated with these syntheses are exemplified in the synthesis of colombiasin A (3) (Scheme 1). All of the approaches use an elegant intramolecular Diels-Alder end-game strategy developed by Nicolaou and proceed through the common diene intermediate 9. The first approach to 9, devised by Nicolaou, used a palladiumcatalyzed allylation to introduce two of the key stereocenters but 11 was formed with poor regiocontrol. 16b,16c Furthermore, 11 was formed as the wrong diastereomer, and several additional steps were required to achieve the necessary epimerization. An alternative strategy has been an intermolecular Diels-Alder reaction of benzoquinone 12 with a diene 13. Due to the lack of stereocontrol, the exocyclic stereocenter in the diene needed to be stereospecifically introduced prior to the cycloaddition. In the initial process reported by Rychnovsky, the diastereoselectivity in the cycloaddition was low (1:1.7), 16dbut recently, Jacobsen has greatly improved this process by using chiral Lewis acids to influence the diastereoselectivity of

OMe Me

OMe Me

Me~.~-~.-~..._Ok__~ Pd(O) OMMe/'~/~U" 10 0

MeO~

Me~ ~ ' ~ f f O OM~[~",Me

Nicolaou's strategy R1

+

Me" -.[]/ O R3

11 O,

..~

I colombiasin A IMDA

R1

O Me

M e O ~ R2 [4+2] Me/~I1/~~~'R2 ~ O ,-P~Me / R3

13 Rychnovsky's strategy 14

12

Jacobsen's strategy

OBut Me

_Me -

e~ M

Me

O

-~ _~

OH 0 M

15

Harrowven's strategy

16

SCHEME 1

/

9

I1

387

TOTAL SYNTHESES OF NATURAL PRODUCTS

this cycloaddition. 16f Due to the stereochemical challenges of these natural products, an alternative approach has been to begin the synthesis with commercially available monoterpenes, which already contain at least some of the problematic stereocenters, 22,23 Harrowven has used this strategy, converting the monoterpene 15 to 16, followed by a cascade of reactions via 9 to form colombiasin A . 16e The synthetic challenges to these natural products are exacerbated because there are no convenient neighboring functional groups available to assist in controlling the stereochemistry of the three common stereocenters. A C-H activation methodology, avoiding the requirement of specific functional groups, would be an attractive solution to these stereochemical challenges. We conceived of such an approach, which has the potential to be a universal solution to address the stereochemical issues associated with these natural products (Scheme 2). The key step is the combined C-H activation/Cope rearrangement, which controls the configuration of the three challenging stereocenters in one step. This account describes how we developed this new strategy and illustrates how it is ideally suited for the synthesis of this class of diterpenes.

OR Me OR Me MeO~ MeO~ Rh2(R.DOSP)4D.M e ~ ' i / ~ ~ "H Me" ']/ ~ N2 OR Me~,~/JJ~-CO2Me _ OR OR I~1~Me 17 18 MeO2C 19

O Me

R40~ ~ ]

_~ ~ M e / " ~ ~ "~'R5 Oj,~Me

Davies' strategy SCHEME 2

IIl. C - H Activation as a Strategic Reaction in Synthesis A. C - H ACTIVATION AS A S U R R O G A T E O F C L A S S I C O R G A N I C R E A C T I O N S

The vinyldiazoacetate 18, which will be used in the combined C-H activation/Cope rearrangement, is a precursor to a donor/acceptorsubstituted carbenoid. We discovered that donor/acceptor carbenoids are very effective at intermolecular C-H activation by means of carbenoid induced C-H insertion. Due to the presence of a donor group (e.g., vinyl or aryl), these carbenoids are much more attenuated in their reactivity and show greatly improved chemoselectivity than conventional carbenoids

388

HUW M.L. DAVIES AND XING DAI

that contain only electron-acceptor groups. 1~,~2,24,25 Furthermore, their reactions are routinely highly enantioselective when catalyzed by dirhodium complexes such as Rhz(DOSP)4 (Scheme 3). R1 R2-kC-H

EWG LnRh(II) N2==:~

EWG .

EDG

R3

LnRh:~

I /----~ J--IO-7-Rh R1

EWG

- R2--kC-~--H,

EDG

R3

EDG

/'" N / ~ ; ~1_!h SO2Ar J4 Ar = p-C12H25C6H4

EWG = CO2R,COR, PO(OR)2,SO2R, CN, NO2 EDG = vinyl or aryl

Rh2(R-DOSP)4

SCHEME 3

As the C-H activation is strategically a very different approach to synthesis compared to conventional reactions involving functional group transformations, we have demonstrated how the C-H activation can be considered as a retrosynthetic equivalent of some of the classic reactions of organic synthesis. Examples illustrating the synthetic usefulness of donor/acceptor-substituted carbenoids in conjunction with Rh(II)-prolinate catalysts are selective C-H insertion reactions ~ to heteroatoms such as nitrogen and oxygen, and at allylic and benzylic sites. Using this reaction, equivalent transformations have been achieved to several of the classic reactions of organic synthesis, such as the aldol reaction, 26 the Mannich reaction, 27 the Michael addition, 28 the Claisen rearrangement, 29 the Claisen condensation 3~ and enolate alkylation 31 (Scheme 4). In each case, highly enantioselective examples have been developed. B. C O M B I N E D C - H ACTIVATION/COPE R E A R R A N G E M E N T

Exploration of a new area of chemistry can be very exciting because when the expected chemistry fails, new discoveries can still be made. This occurred when we explored the scope of allylic C-H activation. 32 When vinyldiazoacetates are reacted with allylic C-H bonds, an unexpected rearranged product 20 is formed in competition to the C-H activation product 21 (Scheme 5). This product is not formed by a C-H activation followed by a Cope rearrangement, because the C-H activation product 21 is the thermodynamically favored product. The term combined C - H

11 TOTALSYNTHESESOF NATURALPRODUCTS

389 Asymmetric Mannich Equivalent

Asymmetric Aldol Equivalent

OQ +

X

H

R

X

0 ...EL.R1 Aldol

O

OH

0Q +

~-

R

H

R2

-

-DI

O ii Mannich =- X ~

H

NHR i R1 R2

O

O "~

N2 +

OR2 -..LR1

X

Asymmetric Michael Addition Equivalent

+ [

R2

R

R1

AsymmetricClaisen Rearrangement Equivalent

O OO O Michael . ~ X. ~ + ~j~ R2 Addition X R

~.R1

O

"~R2 R1

70

R1 R2 R1 O'~~ Claisen O ~~ R2 Rearrangement -~ 70

R R1

OTIPS

x~N2R

+~

R2

R2 O

Asymmetric Claisen Condensation Equivalent X~'~"l +

O

Claisen

O

O

RO...J~R1 CondensationX - J ~ ' - . R 1 R2 R2 o

X'~

AsymmetricEnolate Alkylation Equivalent

R1CH2--X

Enolate H / Alkylation - R1CH2~k-'CO2Mex G + R~CO2Me R

/-q

N2 + O.,~O R2 R1

R1CH2_H+ N2~ CO2Me R

l

SCHEME4

al

Rh(ll) R~"

CO2Me

R:~" v :

R3-

(~O2Me

20

I

21

heat Cope rearrangement

SCHEME 5

l

390

HUW M.L. DAVIES AND XING DAI

activation~Cope rearrangement is used to describe this new reaction, because the reaction has the appearance of a C-H activation process while the stereochemical outcome is as expected for a Cope rearrangement proceeding through a chair transition state. An early example of the use of this chemistry is the asymmetric synthesis of the antidepressant (+)-sertraline (Zoloft) (24). 32 Doug Stafford discovered that the combined C-H activation/Cope rearrangement between the vinyldiazoacetate 22 and 1,3-cyclohexadiene is highly enantioselective forming 23 in 99% ee (Scheme 6). The conversion of 23 to (+)-sertraline (24) could be readily achieved using conventional steps. N2.~CO2Me S

c, 22 CI

CO2Me

NHMe

Rh2(S.DOSP)4

c,' 23 99% ee

24 CI. (+)-sertraline

SCHEME 6

Having made this interesting discovery, recent studies have been directed toward determining the scope of this chemistry. The reaction has been extended further so that two stereocenters can be controlled in the C-H activation/Cope rearrangement. It was found that the reaction is truly spectacular in terms of diastereoselectivity and enantioselectivity. Good examples of this are the reactions of vinylcarbenoid 25 with the cyclohexene derivative 26 and the unsaturated lactone 27 (Scheme 7). 33 In the case of the cyclohexene derivative 26, the combined C-H activation/Cope rearrangement products were produced with >98% de and 97% ee, although the direct C-H insertion was a competing reaction. High stereoselectivity was also exhibited in the reaction with the unsaturated lactone 27 (>98% de, 99% ee) but in this case, some direct C-H insertion and cyclopropanation also occurred as competing reactions. During the exploratory studies, Qihui Jin identified that dihydronaphthalenes were excellent substrates for the combined C-H activation/Cope rearrangement. This was seen in the Rhz(S-DOSP)4 catalyzed reaction of vinyldiazoacetate 29 with dihydronaphthalene 28 (Scheme 8). 34 The C-H activation/Cope rearrangement product 30 was obtained in >98% de and 98% ee. One of the signature features of the donor/acceptor-substituted

11

391

TOTAL SYNTHESES OF NATURAL PRODUCTS

Me [~

Ph Rh2(SDOSP)4 Me "~~_.~ PhCF3,0 ~ CO2Me CO2Me 68%

N2~ +

26

Ph

25

97% ee, >98% de

Me Ph Ph N2, ~ oO~ Rh2(S-DOSP)4 ~ e~-~\ ~ O + PhCF3,0 ~ ""CO2Me CO2Me 87% O 27 25 99%ee, >98% de SCHEME 7

carbenoids is that the trajectory of attack of reagents to the carbenoids appears to be very specific. Even though the actual trajectory is not known with certainty, an excellent predictive stereochemical model has been developed for these reactions. 11a,35 Applying this model to the C-H activation/Cope rearrangement reaction of vinyldiazoacetates with dihydronaphthalene 28 correctly predicts the stereochemistry observed in product 30 (Scheme 8). The substrate approaches the vinylcarbenoid (S-catalyst) from the front face and the constricted trajectory leads to excellent control of the two stereogenic centers.

CO2Me + N2=~/ Rh2(S'DOSP)4. "Me 23 ~ t Me Et 28 29 MeO2C 30

[~~

Me flip E "CO2Me

>98% de, 98% ee

r ueO.Gi i Etl lJ . ,Tj

9. 180~ [ ~ ~ MeO2C"~v "~MeH

(S-DOSP) Catalyst substrate approaching from front SCHEME 8

180~ ,. H ~ C O 2 M e Me

392

HUW M.L. DAVIES AND XING DAI

C. E N A N T I O D I F F E R E N T I A T I O N " M O D E L S T U D Y

Inspired by the results of the combined C-H activation/Cope rearrangement shown in Scheme 8, we recognized that this methodology could be a wonderful strategic reaction for the synthesis of the marine natural products shown in Figure 1. Presumably, the C-H activation/Cope rearrangement product 32, which contains the three common stereocenters, could be a universal intermediate to these natural products (Scheme 9). Initially we thought it may be necessary to begin the synthesis with enantiomerically pure dihydronaphthalenes. During the analysis of this synthetic problem, however, we recognized that the racemic dihydronaphthalene (+_)-31 could be used as starting material. Applying the predictive models to the Rhz(R-DOSP)4-catalyzed reaction of ( +)-31 with 18 suggested that the two enantiomers would lead to different products. Only (S)-31 would be capable of a matched, combined C-H activation/Cope rearrangement to form 32, whereas (R)-31 would be matched for a cyclopropanation to form 33 (Scheme 9). This would be a very exciting outcome because the dihydronaphthalene (+)-31 could potentially be used as the limiting agent, as both enantiomers would be consumed but would form different products.

Me

Me

R . , E ~ . f ~ ~,H S-enantiomer ~ " H

Me

,14' Racemic

> OMe

)

Starting +Material Rh2(R-DOSP) (, ,

CO2Me

\

18 Me

~

,-ha "od" e'ge t ransitionState ($ubs*-"tes approaching from front)

.,J... I~',, Me" v 9

Me 32

R-enantiomer

r._ f

Rh I (R-DOSP) Catalyst SCHEME 9

le

Me" v

O2Me ~ "~Me 33

393

l 1 TOTALSYNTHESES OF NATURALPRODUCTS

As this was a bold hypothesis, Abbas Walji conducted model studies to test the predictions. Rh2(S-DOSP)4-catalyzed reaction of dihydronaphthalene 34 with the phenylvinyldiazoacetate 35 gave a very promising result (Scheme 10). 36 A 1:1 mixture of the combined C-H activation/Cope rearrangement product 36 and the cyclopropane 37 were formed in a combined yield of 80%. Both products were produced in 98% ee and essentially as single diastereomers. The relative and absolute stereochemistry of 36 was determined by conversion of 36 into the crystalline p-bromobenzoate 38, whose configuration was confirmed by X-ray crystallography.

Me

Me -

N2 35

""H

---

2% Rh2(S-DOSP)4 34

Me

h

80%

O,~'

+

,~Ph

(1:1)

36 98% ee

CO2Me 37

98% ee

/

OMe i, H2, Pd/C ii, LiAIH4 iii, DCC, DMAP, p-BrC6H4CO2H Me _

9 ,1 a

Br\~...~

"Ph 0

SCHEME l0 IV. Total Synthesis of (+)-Erogorgiaene

Having successfully completed the model studies, attention was then directed toward the total synthesis of (+)-erogorgiaene (2). 36 The key step is the rhodium-catalyzed reaction between the vinyldiazoacetate 18 and the dihydronaphthalene (+)-39 (Scheme 11). The Rhz(R-DOSP)4 catalyzed reaction of vinyldiazoacetate 18 with (+)-39 resulted in enantiodifferentiation to form 41 with 90% ee. The other enantiomer of the dihydronaphthalene 39 preferentially formed the cyclopropane 40. Due to the high diastereoselectivity of the combined C-H activation/Cope rearrangement, a single diastereomer of 41 was formed with the correct configuration for (+)-erogorgiaene (2). Owing to the

394

HUW M.L. DAVIESAND XING DAI

tendency of 41 to undergo a retro-Cope rearrangement, the combined mixture of 40 and 41 was globally hydrogenated, and the ester was reduced to the alcohol 42, which was isolated in 31% overall yield from the dihydronapthalene (+_)-39 (62% yield from the matched enantiomer (S)-39). Completion of the total synthesis of (+)-erogorgiaene (2) was achieved by oxidation of 42 to the aldehyde with pyridinium chlorochromate (PCC) followed by a Wittig reaction. Most impressively, Abbas Walji, the graduate student who conducted this project, completed the whole synthesis of (+)-erogorgiaene, including the sixstep synthesis of the dihydronaphthalene substrate, in just 10 days!

Me

. ~ ~ Me

Me

MeSH

2%Rh2(R-DOSP)4

2,2-DMB, rt, 2 h ~ Me~

39 +

N2 Me02C1 8 . ~ M e

,,H H

s

Me+ Me (1:1)

]~ "Me

4 1 0 " ~ ~ 90% OMe ee

40

i,H 2,Pd/C ii, LiAIH4

Me

[ 31% (62%) ~ over 3 steps Me

i, PCC (89%)

S

e

ii, Ph3P=CMe2(82%) OH

(+)-erogorgiaene SCHEME l l V. Total Syntheses of (-)-Colombiasin A and (-)-Elisapterosin B A. RETRO-SYNTHESIS OF (-)-COLOMBIASIN A

(-)-Colombiasin A (3) is one of the most well-recognized members of these diterpene marine natural products. It was envisioned that the combined C-H activation/Cope rearrangement would again be effective as illustrated in the retrosynthetic analysis outlined in Scheme 12. There was no reason to diverge from the published [4 + 2] cycloaddition approach for the formation of the bridged ring. 16b'16c The precursor diene 43 was

11

395

TOTALSYNTHESESOF NATURALPRODUCTS

simplified to aldehyde 44. The aldehyde 44 could be accessed with relative ease from the ester 45. The pivotal step of the synthesis, of course, would involve a combined enantioselective C-H activation/Cope rearrangement between the 1,2-dihydronaphthalene 46 and vinyl diazoacetate 18. During this step, the C6-C7 bond would be formed and the stereogenic centers at C3, C6 and C7 would be installed with the desired absolute and relative stereochemistry.

O Me

O intramolecular Diels-Alder

M MeI

~"~ ~v

"Me

(-)-colombiasinA

OR Me Grignard addition elimination

"~'H

O/~M

Nic~176 strategy

3

Me

~

e

Me" ~ "I~H O~//%M

e

43

II

0

44

OR Me

OR Me M e O ~

OR Me , M e O ~ <

Me" ~ OR 47

v

\-0

asymmetric M e O ~ C-Hactivation Coperearrangement

< Me" ~

~

OR 46

, N2

Me/'~"~CO2Me 18 Rhz(R-DOSP)4

Me" "1~ 6"~H OR~M e O ~ ' 45 OMe

SCHEME 12 B. KEY STEP: COMBINED C-H ACTIVATION/COPE REARRANGEMENT

Initial studies on the synthesis of colombiasin A was conducted by a post-doctoral researcher, Dr. Mathew Long, while graduate student Xing Dai was responsible for bringing this project to fruition. Two dihydronaphthalenes with different protecting groups, methyl (46a), and t-butyldimethylsilyl (46b) were found to be the most suitable substrates for the crucial combined C-H activation/Cope rearrangement. 37 The synthesis of 46 started from p-quinone 48 following a [4 + 2] cycloaddition sequence (Scheme 13). 16b'16c Reaction of 48 with the diene 49 generated the cycloadduct, which upon isomerization to the corresponding quinol could be trapped under different conditions to afford the dimethyl derivative 50a or the disilyl derivative 50b. Acidic

396

HUWM.L.DAVIESAND XINGDAI O

Me ~

MeO~ +

Me" -]]i O 48

OR Me M e O ~

OR Me 1)EtOH,rt. M e O ~ 2) Condition OTBS aorborc

49

Me" "~ v OR

TFA "OTBS

Me" ~ ~ / "-O OR

a) Mel/K2CO3 R = Me 50a (78%) b) TBSCI/imidazole R = TBS 50b (86%)

CIEN..I.~N".Tf OR Me 51 ~f (1.3 eq)= M e O ~ NaHMDS(1.32 eq)

Me" "~

R = Me

R = TBS OR Me

Pd(PPh3)4' LiCIEt3SiH, THF

">" "omf

M e O ~ Me" "~

OR R = Me

47a (88%) 47b (84%)

~"

OR R=Me

52a (90%)

46a(94%)

R = TBS 46b (96%)

R = TBS 52b (75%) SCHEME 13

hydrolysis of the resulting tert-butyldimethylsilyl (TBS)-enol ethers in 50 gave rise to the ketones 47 in good yields. Synthesis of the dihydronaphthalenes 46 was then achieved by initially convening 47 to the vinyl triflate 52 followed by a palladium-catalyzed reductive coupling. 38 From the conception of the project, it was proposed that the functionality on the aromatic ring would not interfere with the combined C-H activation/Cope rearrangement step because the functionality was far removed from the site of the C-H bond involved in the initiation. This was the case with the dimethoxy derivative 46a, as the Rhz(R-DOSP)4 catalyzed reaction of 46a with 18 gave a 1:1 mixture of the C-H functionalization product 53 and the cyclopropane 54 as single diastereomers (Scheme 14). Furthermore, the C-H functionalization product 53 was formed with the correct relative stereochemistry for the natural products in 92% ee.

Me

OMe Me

OMe Me

"OMe+ ( ~'~ 46a

"[" ~ H + MeO2C. ~ ~ H O~MeMe 2% Rh2(R-DOSP)4, . . . ~Me OM~'r 2,2-DMB, rt, 1.5 h Me "~-k Me O . . ~ ' 53 (41%) 54 (43%) Me OMe 92% ee

N2 LI MeO2C~ ~ ~ I

A

O

SCHEME 14

Me OMe

11 TOTALSYNTHESESOFNATURALPRODUCTS

397

Similar results occurred with the disilyl derivative 46b. A 1:1 mixture of the C-H functionalization product 55 and the cyclopropane 56 was formed (Scheme 15). Since the two products could not be separated at this stage, the mixture was hydrogenated and then reduced to the alcohols 57 and 58. The desired C-H functionalization product 57 was isolated in 34% yield (68% in theory) as a single diastereomer in >95% ee over three steps. TBSO

TBSO

Me

MeO~ Me" ~ 46b TBSO +

2% Rh2(R-DOSP)4 2,2-DMB, rt, 1.5 h

MeO2C"<~ "1-10TBS

O~ j55'~

"%e56

OMe

MeO2C~Me TBSO

Me

~

MeO~ 2) LiAIH4

Me OTBS

~ O M e

TBSO f'%Me

N2

1) Pd-C, H2

Me

MeO~

Me" "]>" ~[~H TBSO .....M (>95% ee) - - ]/ "Me 34% (68%.) y ~-I over3steps OH 57

~

Me OTBS

OMe

f H. OTBSMe

HO

Me

58

SCHEME15 C. TOTALSYNTHESESOF (-)-COLOMBIASINA AND (-)-ELISAPTEROSINB The completion of the synthesis would require access to the quinone diene 62 and this was achieved using very standard steps (Scheme 16).37 PCC oxidation of 57 afforded aldehyde 59, followed by a Grignard addition, which generated the allylic alcohol 60. Conversion of 60 to the triflate followed by elimination generated the diene 61, which was readily desilylated and air oxidized to the quinone-diene 62. The final steps of the synthesis are well established as Kim and Rychnovsky have previously shown that the diene 62 can be converted to (-)-colombiasin A by an intramolecular Diels-Alder reaction, while treatment of 62 with boron trifluoride etherate generates (-)-elisapterosin B by means of a [5 + 2] cycloaddition (Scheme 17). 16d'37 When diene 62 was heated to 180 ~ in toluene, followed by demethylation of 63 with

HUWM.L.DAVIESANDXINGDAI

398

TBSO Me MgBr MeO~ MeO~ PCC,DCM, M e O ~ (1.2 eq) - Me" -,T,- -~, H 87% S ~ "~H "~Me Me"T B"~ Me"T B"(" S O ~~I<~H Me 85% TBSO [/~Me TBSO

Me

I

OH

TBSO

Me

II

57

TBSO

H~ 60

59

0 Me

O Me

MeO~ 1) Tf20/DTBP 2) NaHCO~ 75% DTBP= 2,6-ditert-butylpyridine

Me" T

MeO~ -~r~H

rBSO~M e

)/....

TBAF,THF 9. Me --r( ~F~H 89% O / ~ Me .)/,... 62

61

SCHEME16 O

Me

M e O ~ [4+2] MeO.~~~ HO 80 ~ toluene Me/~'/O ~ H 1 /[~""L /J AICI3'PhNMe2 / ' M e 88o/o Me~.~O~ ~.~ 70O/o Me/ ~

O

Me

H [~ H Me (-)-colombiasinA(3)

~v "Me 63

/~ "

62 OH Me

O

Me

MeO~ BF3eOEt2

MeO,~H

[5+ 2]

Me H Me (-)-elisapterosinB(4) 51%

63

H

21%

SCHEME17 A1C13, (-)-colombiasin A (3) was formed. 16d Treatment of diene 62 with boron trifluoride etherate a t - 7 8 ~ for 1 h resulted in a [5 + 2] cycloaddition to form (-)-elisapterosin B (4) in 51% yield and colombiasin A methyl ether 63 as a side product in 21% yield.

11 TOTALSYNTHESESOFNATURALPRODUCTS VI.

Total Syntheses of

(+)-Elisabethadione and

399 Related Natural Products

A. TOTALSYNTHESIS OF (+)-ELISABETHADIONE During the synthetic studies directed toward colombiasin A, we became aware that the combined C-H activation/Cope rearrangement products 53 and 55 not only had the desired stereochemistry, but also the ideal sidechain functionality for the synthesis of a variety of natural products. In order to demonstrate the general applicability of this chemistry, the synthesis of (+)-elisabethadione (5) was accomplished (Scheme 18). 39 The 1,5-diene in 53 was hydrogenated and then the ester group was reduced to alcohol 64 in 96% yield over two steps. Oxidation of 64 with PCC followed by a Wittig reaction on the resultant aldehyde furnished the alkene 65. Demethylation was achieved by heating the compound 65 with lithium ethanethiolate in dimethylformamide (DMF) at 180 ~ to form the bisphenol 66 in 85% yield (Scheme 18). 4o Oxidation of 66 with cerium ammonium nitrate followed by demethylation and bond reorganization of the resultant red o-quinone 67 under acidic conditions gave 5, the assigned structure of elisabethadione, in 96% yield as a yellow oil. Contrary to our expectations, the reported ~H NMR and ~3C NMR data for the natural product (+)-elisabethadione (5), while similar, were different from our synthetic compound 5. Either the assigned structure of the natural material or our synthetic material is incorrect. Another possibility could be errors in the reported data for the natural material.

~

OMe Me

1) H2,Pd-C Me V

"~H

oo , ;;

"

OMe Me

OMe Me

1) PCC(94~

Me" ~ . "~"H

96%over2 steps

e

M~ ~

2) Ph3P=C(CH3;2 e 80%

J

oTMe~,H Me

~ 64 OH

OMe OH Me LiSEt, DMF H MeO ~ H --85% O~Me

O CAN

Q ~ H

-~ Me 77~176

~ SCHEME18

Me

TsOH,benzene

O~Me

SL',

O

Me

9~~

O~

Me

400

HUW M.L. DAVIESAND XINGDAI

B. TOTAL SYNTHESIS OF A p-BENZOQUINONE NATURAL PRODUCT

In order to confirm the proposed configuration of the synthetic material as 5, the total synthesis of a second related natural product, the (+)-p-benzoquinone 6, was conducted using all of the steps that had been used in the synthesis of 5 during which epimerization could occur. The general outline of the synthesis is shown in Scheme 19. 4o The synthesis started from the primary alcohol 64, the same intermediate used in the synthesis of compound 5. The terminal alkene 68 was generated by application of Grieco's selenoxide introduction/elimination procedure. 41 Then, performing a similar sequence as was used in the synthesis of 5, 68 was converted to the quinone 70. Selective demethylation of 68 to form the bisphenol 69, followed by oxidation with ceric ammonium nitrate gave the o-quinone 70 in 84% yield. The subsequent isomerization of the o-quinone 70 gave the p-quinone, which was then protected by a TBS group to form 71 in 91% yield. Completion of the synthesis proceeded in a straightforward fashion. Installation of the allylic alcohol by a cross-metathesis reaction catalyzed by the Grubbs second-generation ruthenium catalyst, using Jacobsen's strategy, ~6rfollowed by deprotection of the siloxy group afforded the natural product 6 in 60% yield over two steps. The spectral data of synthetic and natural (+)-p-benzoquinone 6 were in complete accord.

OMeMe

OeO

Bu3P (95%) Me" ~I~oM~He 2) H202(85%) r"

OMe Me

,,olO ,C04SeC

-" M e - % e ~ H o-e M

64

~,

68

Me

O

Me

OH Me LiSEt, DMF 93% -" Me

CAN e 84% ~

69

OH O O O ~

1) TBSO __ZsOH,benzene( 9 5 % ) . ~

Me- "OOMe"~He2)TBSOTf'2'6-1utidine~'~M Me 70

91%

~..,,.J

3

1) Grubbs II I] ]'~H 2) TBAF 60% O ~71Me over 2 steps

Me

Me J

"Me

..~OH 6 (+)-p-benzoquinone

SCHEME 19

11

401

TOTALSYNTHESES OF NATURAL PRODUCTS

C. TOTAL SYNTHESIS OF ( + ) - O - M E T H Y L - E L I S A B E T H A D I O N E

Assuming that the natural product 6 is correctly assigned, these results imply that the assigned structure of (+)-elisabethadione is incorrect or the reported spectral data for elisabethadione contain errors. In order to resolve this discrepancy, we initiated a collaborative study with Kerr's group, who had originally isolated and proposed the structure of (+)-elisabethadione. Due to the lack of any remaining supplies of (+)-elisabethadione, an effort was made to re-isolate it from a new harvest of Pseudopterogorgia elisabethae by Kerr and Wan. 42 Unfortunately, no (+)-elisabethadione could be isolated, but two new closely related natural products, (+)-O-methyl-elisabethadione (72) and (+)-O-methyl-nor-elisabethadiene (73) were identified (Figure 2). (+)-O-methyl-elisabethadione (72) is a very useful probe to test which of the spectral data for (+)-elisabethadione (5) are most likely to be correct. As the only difference between the two compounds is a hydroxy versus a methoxy group, the spectral data for the two natural products would be expected to be very similar. A comparison of the 1H NMR reveals that the data for synthetic (+)-elisabethadione (5) is very close to the data for the newly isolated (+)-O-methyl-elisabethadione (72), while the published data for the natural elisabethadione (5) has many differences. This would suggest that the published data for the assigned structure of natural (+)-elisabethadione (5) is incorrect. Further confirmation of the assigned structures was obtained from synthetic studies. With synthetic (+)-elisabethadione (5) in hand, further confirmation of its structure could be obtained by its conversion to O-methyl-elisabethadione as treatment of synthetic compound 5 with

0 MeO~

Me

Me" ~1~ '"~He

0 MeO~ Me

Me He

Me~ !Me Me Me 72 73 (+)-O-methyl-elisabethadione(+)-O-methyl-nor-elisabethadione FIGURE 2. Structures of (+)-O-methyl-elisabethadione (72) and (+)-O-methyl-nor-elisabethadione (73).

HUWM.L.DAVIESANDXINGDAI

402

methyl iodide at room temperature, gave the desired (+)-O-methyl-elisabethadione (72) in 60% yield (not optimized) (Scheme 20). The spectral data of synthetic and natural (+)-O-methyl-elisabethadione were in full agreement. An alternative and more efficient synthetic route to 72 was possible starting from the aldehyde 59, as illustrated in Scheme 21. A Wittig reaction on the aldehyde 59 generated the alkene 74 in 77% yield, which upon treatment with two equivalents of tetra-n-butylammonium fluoride (TBAF) afforded (+)-O-methyl-elisabethadione (72) in 65% yield.

O

Me

O

Me

eO.J q Mel/K2C03 acetone, ~

M ~ M e

~ I.,L J M e / ' ~ ~]~H

60% (not optimized)

5

~

O f'~'Me

"

72

(+)-O-methyl-elisabethadione SCHEME20

TBSO

Me

TBSO

Me

O

Me

o

Me "(

~H

TBSI#Me 9

Ph3PCH(CH3)21/BuLi 0~

reflux,1h

Me "1>" "~H TBAF,THF 0~

I-B~

77%

Me

Me "IT" "r~H O r..~ M

e

65%

O 59

74

72

(+)-O-methyl- elisabethadione

SCHEME21 D. TOTALSYNTHESIS OF (+)-O-METHYL-NOR-ELISABETHADIONE During NMR studies of intermediates in the (-)-colombiasin A synthesis, Xing Dai observed that 61 in CDC13 underwent a double-bond isomerization (Scheme 16). This accidental reaction turned out to be very

11

403

TOTALSYNTHESES OF NATURAL PRODUCTS

TBSO

Me

TBSO

Me

HCi(1.0eq). Me "] Me ~c~ H rt, 30 min T B S O ~ Me TBS

61

O

H e

75

TBAF, THF = Me rt, 5 min 80% over two steps

"[[

Me [

O

H Me

73 ( + )-O-methyl-nor-elisabethadione

SCHEME 22

useful. Treatment of 61 with one equivalent of HC1 in ether at room temperature for 30 min gave complete isomerization of 61 to the more stable internal diene 75. Completion of the synthesis of the natural product 73 was achieved in 80% yield over two steps by desilylation followed by air oxidation of 75 (Scheme 22). E. SYNTHESIS TOWARD ( + ) - E L I S A B E T H A M I N E

During the course of these studies, we became concerned about the assigned structure of (+)-elisabethamine (7), another member of this class of diterpenes. N-alkylamino-hydroquinones are known to be extremely unstable in air43 and as no special anaerobic conditions were used to isolate (+)-elisabethamine (7), it was deemed unlikely that the structural assignment was correct. With the (+)-O-methyl-elisabethadione (72) in hand, it is straightforward to complete the synthesis of 7. Because of its ester-like reactivity, the methoxyl group in 72 could be replaced by amino groups upon direct interaction with the desired amines. Thus, upon treatment of 72 with an excess of methylamine in ethanol at room temperature, the (+)-O-methyl-elisabethadione (72) (yellow solution in ethanol) completely converted to the desired amino benzoquinone 76 (red solution in ethanol) in 20 h (Scheme 23). It is well known that the benzoquinone could be easily reduced to the hydroquinone in the presence of sodium hydrosulfite. Thus, the amino benzoquinone 76 was subjected to this reduction condition, and after 70 min at room temperature under argon, the red reaction solution turned colorless, which indicated that hydroquinone was formed. During workup, however, the red color returned and

404

HUW M.L. DAVIESAND XING DAI

O Me OH Me MeHNv.JJ..../~ MeHN~ II II I Na2S204 [I i[ [ MeNH2, EtOH Me.'~[]./"~H EtOAc/H20 M e S H Me" ~ ~[~H rt, 20 h O / ~ M e rt, 70 m,n O/~M e O~ Me 90% O

Me

MeO~.~

/L'

72

~ - ~ - ~ ~

(+)-elisabethamine

oxi"~" dized back duri work-u~~-~" ng SCHEME 23

the benzoquinone 76 was regenerated. This study indicates that 7, at least in its neutral form, is not stable under aerobic conditions. VII. Summary

In summary, we have demonstrated that the combined C-H activation/Cope rearrangement protocol is an exceptional method for the construction of the three stereogenic centers common to the numerous diterpenes isolated from Pseudopterogorgia elisabethae. The synthetic potential of this chemistry was demonstrated by means of efficient enantioselective syntheses of (+)-erogorgiaene (2), (-)-colombiasin A (3), (-)-elisapterosin B (4), the assigned structure of (+)-elisabethadione (5), the p-benzoquinone natural product (6), (+)-O-methyl-elisabethadione (72), (+)-O-methyl-nor-elisabethadione (73), and the oxidized form of elisabethamine (76). These studies also demonstrate the value of total synthesis to confirm the assigned structures of natural products. Acknowledgments The application of donor/acceptor-substituted rhodium carbenoid intermediates in organic synthesis has been a central focus of the Davies research group and the enthusiastic contributions of the many group members listed in the references, are gratefully acknowledged. Certain group members made very significant contributions to the material covered in this chapter and deserve special recognition. Brian Doan discovered the combined C-H activation/Cope rearrangement and Doug Stafford conducted the initial exploration of the scope of the chemistry with cyclohexadienes. Qihui Jin and Rohan Beckwith extended the chemistry to systems generating two stereocenters and Qihui Jin made the crucial discovery that dihydronaphthalenes were spectacular substrates for this chemistry.

11

TOTALSYNTHESESOF NATURALPRODUCTS

405

Abbas Walji achieved the 1O-day synthesis of (+)-erogorgiaene. Mathew Long conducted the initial synthetic studies related to Colombiasin A and Xing Dai completed its synthesis as well as all the other total syntheses described in the chapter. We are grateful to the National Science Foundation (CHE-030536) and National Institutes of Health (GM080337) for financial support of this work.

References and Footnotes

1. (a) Shilov, A. E., Shul'pin, G. B., Chem. Rev. 1997, 97, 2879. (b) Labinger, J. A., Bercaw, J. E., Nature 2002, 417, 507. (c) Ritleng, V., Sirlin, C., Pfeffer, M., Chem. Rev. 2002, 102, 1731. (d)Jia, C., Kitamura, T., Fujiwara, Y.,Acc. Chem. Res. 2001, 34, 633. (e) Fiedler, D., Leung, D. H., Bergman, R. G., Raymond, K. N., Acc. Chem. Res. 2005, 38, 349. (f) Godula, K., Sames, D., Science 2006, 312, 67. 2. Arndtsen, B. A., Bergman, R. G., Mobley, T. A., Peterson, T. H., Acc. Chem. Res. 1995, 28, 154. 3. Hartwig, J. F., Cook, K. S., Hapke, M., Incarvito, C. D., Fan, Y., Webster, C. E., Hall, M. B., J. Am. Chem. Soc. 2005, 127, 2538. 4. Kakiuchi, F., Murai, S., Acc. Chem. Res. 2002, 35, 826. 5. (a) Colby, D. A., Bergman, R. G., Ellman, J. A., J. Am. Chem. Soc. 2006, 128, 5604. (b) O'Malley, S. J., Tan, K. L., Watzke, A., Bergman, R. G., Ellman, J. A., J. Am. Chem. Soc. 2005, 127, 13496. 6. (a) Toure, B. B., Lane, B. S., Sames, D., Org. Lett. 2006, 8, 1979. (b) Wang, X., Lane, B. S., Sames, D., J. Am. Chem. Soc. 2005, 127, 4996. 7. (a) Lafrance, M., Rowley, C. N., Woo, T. K., Fagnou, K., J. Am. Chem. Soc. 2006, 128, 8754. (b) Campeau, L. C., Parisien, M., Jean, A., Fagnou, K., J. Am. Chem. Soc. 2006, 128, 581. (c) Leblanc, M., Fagnou, K., Org. Lett. 2005, 7, 2849. 8. (a) Kalyani, D., Dick, A. R., Anani, W. Q., Sanford, M. S., Org. Lett. 2006, 8, 2523. (b) Hull, K. L., Anani, W. Q., Sanford, M. S., J. Am. Chem. Soc. 2006, 128, 7134. (c) Deprez, N. R., Kalyani, D., Krause, A., Sanford, M. S., J. Am. Chem. Soc. 2006, 128, 4972. 9. Johnson, J. A., Li, N., Sames, D., J. Am. Chem. Soc. 2002, 124, 6900. 10. Wilson, R. M., Thalji, R. K., Bergman, R. G., Ellman, J. A., Org. Lett. 2006, 8, 1745. 11. For recent reviews, please see (a) Davies, H. M. L., Beckwith, R. E. J., Chem. Rev. 2003, 103, 2861. (b) Davies, H. M. L., Loe, O., Synthesis 2004, 2595 (c) Davies, H. M. L., Walji, A. M. In Modem Rhodium-Catalyzed Organic Reactions, Evans, P. A. (Eds.), Wiley-VCH: Weinheim, 2005, pp. 301-340. (d) Davies, H. M. L., Nikolai, J., Org. Biomol. Chem. 2005, 3, 4176. 12. Davies, H. M. L., Dai, X. In Comprehensive Organometallic Chemistry III, Crabtree, R. H., Mingos, D. M. E (Eds.), Elsevier: Oxford, 2006, Vol. 10, pp. 167-212. 13. For a general review on the isolation, synthesis and biosynthesis of these natural products, see Heckrodt, T. J., Mulzer, J., Top. Curr. Chem. 2005, 244, 1. 14. Coleman, A. C, Kerr, R. G., Tetrahedron 2000, 56, 9569-9574. 15. Erogorgiaene, (a) Isolation: Rodriguez, A. D., Ramirez, C., J. Nat. Prod. 2001, 64, 100. Total synthesis: (b) Cesati, R. R., Armas, J. D., Hoveyda, A. H., J. Am. Chem. Soc. 2004, 126, 96. Formal synthesis: (c) Harmata, M., Hong, X., Tetrahedron Lett. 2005, 46, 3847.

406

HUW M.L. DAVIESAND XING DAI

16. Colombiasin A, (a) Isolation: Rodriguez, A. D., Ramirez, C., Org. Lett. 2000, 2, 507. Total synthesis: (b) Nicolaou, K. C., Vassilikogiannakis, G., Magerlein, W., Kranich, R., Angew. Chem., Int. Ed. 2001, 40, 2482. (c) Nicolaou, K. C., Vassilikogiannakis, G., Magerlein, W., Kranich, R., Chem. Eur. J. 2001, 7, 5359. (d) Kim, A. I., Rychnovsky, S. D., Angew. Chem., Int. Ed. 2003, 42, 1267. (e) Harrowven, D. C., Pascoe, D. D., Demurtas, D., Bourne, H. O. Angew. Chem. Int. Ed. 2005, 44, 1221. (f) Boezio, A. A., Jarvo, E. R., Lawrence, B. M., Jacobsen, E. N., Angew. Chem., Int. Ed. 2005, 44, 6046. 17. For approaches toward colombiasin A: (a) Harrowven, D. C., Tyte, M. J., Tetrahedron Lett. 2001, 42, 8709. (b) Chaplin, J. H., Edwards, A. J., Flynn, B. L., Org. Biomol. Chem. 2003, 1, 1842. 18. Elisapterosin B, (a) Isolation: Rodriguez, A. D., Ramirez, C., Rodriguez, I. I., Barnes, C. L., J. Org. Chem. 2000, 65, 1390. Total synthesis: (b) Waizumi, N., Stankovic, A. R., Rawal, V. H., J. Am. Chem. Soc. 2003, 125, 13022. References 16d, 16e and 16f. 19. Ata, A., Kerr, R. G., Moya, C. E., Jacobs, R. S., Tetrahedron 2003, 59, 4215. 20. Rodriguez, A. D., Shi, Y. P., Tetrahedron 2000, 56, 9015. 21. Ata, A., Kerr, R. G., Tetrahedron Lett. 2000, 41, 5821. 22. Pseudopterosins: (a) Look, S. A., Fenical, W., Jacobs, R. S., Clardy, J., Proc. Natl. Acad. Sci. U. S. A. 1986, 83, 6238. (b) Lazerwith, S. E., Johnson, T. W., Corey, E. J., Org. Lett. 2000, 2, 2389. (c) Corey, E. J., Lazerwith, S. E., J. Am. Chem. Soc. 1998, 120, 12777. 23. (a) Johnson, T. W., Corey, E. J., J. Am. Chem. Soc. 2001, 123, 4475. (b) Davidson, J. P., Corey, E. J., J. Am. Chem. Soc. 2003, 125, 13486. 24. (a) Davies, H. M. L., Eur. J. Org. Chem. 1999, 2459. (b) Davies, H. M. L., Antoulinakis, E. G., Org. Reactions 2001, 57, 1-326. (c) Davies, H. M. L., J. Mol. Catal.: Chem. 2002, 189, 125. (d) Davies, H. M. L., Antoulinakis, E. G., J. Organomet. Chem. 2001, 617-618, 47. 25. Doyle, M. P., McKervey, M. A., Ye, T. Modern Catalytic Methods for Organic Synthesis with Diazo Compounds, Wiley, New York, 1998. 26. (a) Davies, H. M. L., Antoulinakis, E. G., Org. Lett. 2000, 2, 4153. (b) Davies, H. M. L., Beckwith, R. E. J., Antoulinakis, E. G., Jin, Q., J. Org. Chem. 2003, 68, 6126. 27. (a) Davies, H. M. L., Hansen, T., Hopper, D. W., Panaro, S. A., J. Am. Chem. Soc. 1999, 121, 6509. (b) Davies, H. M., Venkataramani, C., Org. Lett. 2001, 3, 1773. (c) Davies, H. M. L., Venkataramani, C., Angew. Chem., Int. Ed. Engl. 2002, 41, 2197. (d) Davies, H. M. L., Jin, Q., Org. Lett. 2004, 6, 1769. 28. Davies, H. M., Ren, P., J. Am. Chem. Soc. 2001, 123, 2070. 29. (a) Davies, H. M. L., Gregg, T. M., Tetrahedron Lett. 2002, 43, 4951. (b) Davies, H. M. L., Walji, A. M., Townsend, R. J., Tetrahedron Lett. 2002, 43, 4981. (c) Davies, H. M., Ren, P., Jin, Q., Org. Lett. 2001, 3, 3587. 30. Davies, H. M. L. Yang, J., Nikolai, J., J. Organomet. Chem. 2005, 690, 6111. 31. (a) Davies, H. M. L., Jin, Q., Ren, P., Kovalevsky, A. Y., J. Org. Chem. 2002, 67, 4165. (b) Davies, H. M. L., Jin, Q., Tetrahedron: Asymm. 2003, 14, 941. 32. Davies, H. M., Stafford, D. G., Hansen, T., Org. Lett. 1999, 1 (2), 233. 33. Davies, H. M. L., Jin, Q., Proc. Natl. Acad. Sci. U. S. A. 2004, 101, 5472. 34. Davies, H. M. L., Jin, Q., J. Am. Chem. Soc. 2004, 126, 10862. 35. Nowlan, D. T., III, Gregg, T. M., Davies, H. M. L., Singleton, D. A., J. Am. Chem. Soc. 2003, 125, 15902.

407

11

TOTALSYNTHESESOF NATURALPRODUCTS

36. 37. 38. 39. 40. 41. 42. 43.

Davies, H. M. L., Walji, A. M., Angew. Chem., Int. Ed. Engl. 2005, 44, 1733. Davies, H. M. L., Dai, X., Long, M. L., J. Am. Chem. Soc. 2006, 128, 2485. Scott, W. J., Stille, J. K., J. Am. Chem. Soc. 1986, 108, 3033. Davies, H. M. L., Dai, X., Tetrahedron 2006, 62, 10477. Dehmel, E, Schmalz, H. G., Org. Lett. 2001, 3, 3579. Grieco, E A., Gilman, S., Nishizawa, M., J. Org. Chem. 1976, 41, 1485. Dai, X., Wan, Z., Kerr, R. G., Davies, H. M. L., J. Org. Chem. 2007, 72, 1895. (a) Carpino, L. A., Triolo, S. A., Berglund, R. A., J. Org. Chem. 1989, 54, 3303. (b) Watanabe, T., Takeuchi, T., Otsuka, M., Umezawa, K., Chem.Comm. 1994, 4, 437. (c) Harger, R. N. J. Am. Chem. Soc. 1924, 46, 2540.

STRATEGIES AND TACTICS IN ORGANIC SYNTHESIS, VOL. 7 9 2008 Elsevier Ltd. All rights reserved.

Chapter 12 THE TOTAL SYNTHESIS OF AMPHIBIAN ALKALOIDS USING THE INTRAMOLECULAR SCHMIDT REACTION Kevin J. Frankowski, Aaron Wrobleski, and Jeffrey Aubd Department of Medicinal Chemistry University of Kansas 1251 Wescoe Hall Drive Lawrence, Kansas 66045- 7852

I~

Introduction and Background

II. III. IV.

A Tryout: (-)-Indolizidine 209B

Indolizidine 251F: Background and Synthetic Planning First Generation Work: Synthesis of the Key Enone and a Total Synthesis of Racemic Desmethyl 251F V. Intermezzo: A Few Words About Safely Working with Azides VI. Development of a New Route to a Key Enone and the Total Synthesis of Alkaloid 251F VII. Final Remarks Acknowledgments References and Footnotes

408 412 421 430 440 442 455 456 457

I. Introduction and Background

A continuing theme in our laboratory has been the development of new methodology, a predilection that the principal investigator acquired from his graduate school experience and retains to this day. Two events from the long-distant 1980s helped to cement this interest in synthetic methodology (throughout, the personal voice of this narrative will be that of JA). One was hearing several lectures by Sir Derek Barton, whose eloquence in describing the process of "inventing chemical reactions" was instrumental in convincing me that (1) there were still new reactions out there to be discovered (or at least useful variations of old reactions) and (2)thoughtful mechanistic reasoning might lead one to them. When asked to name the most important problem in contemporary

12

THE TOTAL SYNTHESIS OF AMPHIBIAN ALKALOIDS

409

organic chemistry in 1983, Sir Derek replied that it was "the 100% yield problem." Also influential was Paul Wender, who in personal conversation and public lectures spoke of powerful new reactions able to readily convert simple starting materials to advanced structures both complex and wonderful. In particular, Paul argued that despite much effort and some successes, the science is still a long way from the goal of developing truly efficient syntheses of sophisticated targets. For us, despite a few fun projects in the indole alkaloid area, total synthesis always seemed beyond the scope of our laboratory in its early days. This chapter describes how the discovery of a new reaction in 1991 led us to become more personally involved in total synthesis. Although we will not claim that our work has come close to achieving the Wender desiderata for strategy-level organic transformations, it is to his inspiration and example that this chapter is dedicated. The first few years of independent work in our group were dedicated to exploring the application of oxaziridine rearrangements in asymmetric synthesis. This transformation, shown in Scheme 1, provided a three-step means of converting a ketone to an N-substituted lactam. ~The keys to the utility of this process were (1) that it directly afforded fully substituted lactams as opposed to the N-protonated varieties obtained from Beckmann or Schmidt chemistry and (2) that the reaction allowed for a measure of stereochemical control unavailable to either of those older processes. The discovery by Greg Milligan, in 1991, that alkyl azides would undergo smooth intramolecular reactions with ketones to afford bicyclic lactams was of immediate interest because the variety of ring systems it afforded in one easy step were readily identifiable subunits in a nearly endless array of natural product targets (Scheme 1; the two structures shown in bold play a starring role in the present story). 2 Thus, a connection between our commitment to reaction development and our fascination with natural product synthesis was immediately evident. In this chapter, we relate two stories that show how this connection was made in the context of a particular, very large and very interesting, class of natural products. In so doing, we will try to give a sense of how our laboratory's divergent interests feed off of one another and demonstrate how the intramolecular Schmidt reaction has provided us with a continuing source of inspiration for new forays into the natural world. In one of the most celebrated and sustained efforts in recent natural product chemistry, John Daly, working with his colleagues at the National Institutes of Health, uncovered an enormous number of structurally novel compounds that display a wide spectrum of biological activity. In particular,

410

KEVINJ. FRANKOWSKI,AARONWROBLESKI,ANDJEFFREYAUBI~

Oxaziridine formation and rearrangement Me

0 1. imine formation 2. oxidation ,,.

Ph

N

Me Ph hv =

Ph

Ph

Ph

The intramolecular Schmidt reaction

,,.

~

N

HO N 2 ~

3

Ring systems prepared include the following"

O

0

o

O

o

0

o

SCHEME 1

the dendrobatid family of flogs is a highly prolific purveyor of alkaloids. Initial interest in identifying and characterizing these compounds was inspired by the legendary toxicity of the skin secretions of these amphibians, where they function as a formidable defense against predators. Certain individual compounds are such potent neurotoxins that the unrefined skin excretions of some frog species have been used by the South American Indians of Western Colombia to impart the lethal effect to their poison darts. 3 Particularly worthy of mention is the steroidal alkaloid batrachotoxin, which has shown selective and potent activity on the

12

411

THE TOTALSYNTHESISOF AMPHIBIANALKALOIDS

sodium channels of nerve and muscle. 4 Although initially it was assumed that these complex alkaloids were synthesized by their amphibian sources, it is now accepted that the frogs collect and accumulate the alkaloids from their diet of tiny arthropods. 3,5 To date, over 800 of these alkaloids have been isolated from skin extracts of the dendrobatid frogs. 6 Many of the individual dendrobatid alkaloids have been isolated from multiple frog species. Thus, a single alkaloid may be widely distributed amongst numerous species of dendrobatid frogs. Because of the large number of these alkaloids, a systematic naming system is used for their identification. Each alkaloid is identified according to its molecular weight, and if more than one alkaloid exists with the same molecular weight, a letter is also assigned to the alkaloid. Thus, alkaloid 251F has a molecular weight of 251 amu, and was the sixth alkaloid identified with this mass, as designated by the letter E Likewise, alkaloid 209B was the second alkaloid isolated with the molecular weight of 209 amu. This system sometimes makes it sound as though these compounds were named by Star Trek's Borg, but certain families have been given particular names. Of the nearly two dozen structural classes of alkaloids that have been identified, 6 the more well-represented classes include the batrachotoxins (the class of alkaloids used as dart poisons by South American Indians), the histrionicotoxins and pumiliotoxins (a large family of bicyclic alkaloids), and the pyridine alkaloids (most notably, epibatidine; Figure 1).7

M Me H .0.

Me

1 HO

batrachotoxin

283A histrionicotoxin

OH Me ~

M

~H

e

CI\ N

H N

I~le L NL__._~OH

323A

pumiliotoxin B

epibatidine

FIGURE 1. Representative dentrobatid alkaloids.

412

KEVIN J. FRANKOWSKI, AARON WROBLESKI, AND JEFFREY AUBI~

Me

Me Me~~N~

C5Hll (-)-Indolizidine209B(1)

Me ....2H

Me (-)-Indolizidine251F(2)

~

/

OH

....

Me (+)-3-desmethyl251F(3)

FIGURE 2. Targets for total synthesis.

Herein we describe in detail our efforts toward the synthesis of two such alkaloids, (-)-indolizidine 209B 1 and (-)-indolizidine 251F 2 as well as the (+_)-desmethyl analog of 251F 3 (Figure 2). This work spanned nearly a decade and highlights advances in the application of the azido-Schmidt reaction for the construction of natural products of increasing complexity. Thus, 209B is one of the simplest dendrobatid-derived alkaloids and was the very first total synthesis project that we undertook using the intramolecular Schmidt reaction, while our second total synthesis in this area, directed toward the significantly more complex 251F, was not begun until a number of years had passed. II. A Tryout: ( - ) - I n d o l i z i d i n e 209B

The discovery of the intramolecular Schmidt reaction was significant because it allowed an extremely straightforward route to bicyclic ring systems such as those shown in Scheme 1 above. Of course, nitrogen ring expansion reactions, such as the classical Schmidt or Beckmann rearrangements, have found utility in natural product synthesis for many years. A typical example of this strategy utilized by LaLonde and coworkers is shown in Scheme 2. 8 In this case, ring expansion of the ketone 4 affords the lactam 5 bearing the nitrogen that will eventually end up in one of the ring-fusion positions of the natural product. This necessitates a second ring closure via nitrogen alkylation. One main potential advantage of the intramolecular Schmidt is that the replacement of hydrazoic acid with an azide allows both rings to be constructed in a single step. Furthermore, the typical Schmidt reaction can typically lead to two regioisomeric products. Generally, the more highly substituted carbon migrates, but a double substituted ketone such as that shown in Figure 3 would almost certainly lead to a mixture of the two products shown. In contrast, the intramolecular version provides strict control over which carbon-carbon bond is subjected to nitrogen insertion. Thus, the example shown would afford

12

413

THE TOTAL SYNTHESIS OF AMPHIBIAN ALKALOIDS

Me

1. H2NOH.HCI 2. PCIs

Me +N,,'

Me

. m-CPBA 9NaH

~

r

H 0

0

OH

4

5

Me 1. Jones reagent 2. Pb(OAc)4, Cu(OAc)2

~~ Me

O

OAc

1. NaCNBH3

Me

, 3Fu~,,i 2. NaCNBH3

2. H2, Pd/C

Me 0

(_)-nuphar indolizidine

0 SCHEME 2

Classical Schmidt reaction

CI

CI

O

CI

O

Me HN Me Intramolecular Schmidt reaction N3 O

Me

=

Me'~N/

+

N Me not observed

FIGURE 3. Likely regiochemical outcomes of hypothetical Schmidt reactions.

414

KEVIN J. FRANKOWSKI, AARON WROBLESKI, AND JEFFREY AUB]~

only a single isomeric product (in the vast majority of cases, it is the carbon bearing the azidoalkyl tether that migrates). This is because the alternative bridged lactam is very rarely seen as a product in this chemistry. The first application of an intramolecular Schmidt reaction to a natural product synthesis was published by Professor Jean d'Angelo and coworkers, then at CNRS and later at the Universit6 de Paris. 9 This group was interested in the synthesis of the homoerythrina alkaloid skeleton 8 (Scheme 3). In the original route envisioned by these workers, ketone 6 was subjected to a variety of standard ring-expansion conditions but in no case was the desired lactam obtained. This was possibly due to the rather hindered nature of the ketone in this substrate. About this time, the first communication describing the intramolecular Schmidt reaction was published 2~ and Marie-Anne Le Dr6au, the graduate student working on the homoerythrina project, synthesized the azide 7. Treatment of this compound with acid provided the desired tetracyclic system 8 in good yield. Back in Kansas, we were gratified to see this work published because, after all, the aim of any methodology developer is to introduce chemistry that will be used by other workers. Furthermore, it provided a clear example of how the intramolecular Schmidt might not only shorten synthetic routes, but also provide an alternative for examples where the classical approaches might prove problematic.

MeO MeO" ~ ../-~'~,~'0 ....~ O M s

NH2OH,thenacid or HN3

no lactamproductobtained

6

M e O ~

TFA

MeO" ">" /,~<,~.'O ~ ..../~N3

85%

7

M e O ~O r , ~ -"

MeO" '">" ~~~,,~, 8

SCHEME 3

In 1992, Greg Milligan, the discoverer of the intramolecular Schmidt reaction, was still working in the group and that summer we were joined in the laboratory by Pat Rafferty, a pharmacy student at the University of

12

415

THE TOTAL SYNTHESIS OF AMPHIBIAN ALKALOIDS

Missouri-Kansas City. Although Pat had never done practical organic chemistry aside from the usual sophomore laboratory experience, she was obviously bright and motivated. Accordingly, it was decided that Pat, under Greg's mentorship, would take on a simple total synthesis problem for her summer project. There was another compelling reason to believe that Pat would prove particularly adept at this kind of problem. There is a common stereotype that organic chemists make good cooks, presumably based on similarities between chemical synthesis and food preparation techniques. So, although Pat had never worked in a laboratory like ours before, she had worked as a professional c h e f - affording a perfect opportunity to see if the converse of that old stereotype would be borne out. As it happened, Pat did a marvelous job in laboratory and was prodigiously productive in her 8 weeks of summer work in our group, during which time she completed all of the following project except for the final step, which was Dr. Milligan's work alone. The dentrobatid alkaloid 209B immediately stood out as a candidate for synthesis. The bicyclic alkaloid was structurally simple yet still possessed three stereogenic centers. Previous syntheses of indolizidine 209B l~ and related compounds ~ had confirmed the structure of the natural product as shown. Accordingly, a retrosynthetic analysis was proposed in which manipulation of the lactam carbonyl would provide a handle for introduction of the "bottom" C5 side chain, leading to the intramolecular Schmidt precursor 9 (Scheme 4). This led to a decision point: how would we synthesize the trans keto azide 9 in diastereomerically and preferably enantiomerically pure form? Me

Me

Me

N3

C5Hll 209B (1)

0 10

9

SCHEME 4

Our first thought was to use cyclopentenone as the starting material and install the methyl group and the side chain via a conjugate addition/alkylation sequence. This route had the virtue of shortness and would also reliably provide for the desired trans stereochemistry. Still, there were a couple of possible drawbacks, mainly that it might be hard to carry out the conjugate addition step in an enantioselective fashion. Although such

416

KEVIN J. FRANKOWSKI, AARON WROBLESKI, AND JEFFREY AUBI~

reactions were then known, we had no experience with that chemistry and decided to take another tack. As a side note, it would be necessary under the trapping scenario to introduce the three-carbon side chain as either allyl bromide or as a 1,3-dihalopropane derivative (Scheme 5). It is true that a more convergent synthesis of keto azides would arise if a 1-azido3-halopropane were used as a nucleophile. One concern with such a route is that the azide-containing alkylating agent would be somewhat low in molecular weight and accordingly a possible explosion hazard (see the discussion later in this chapter). In addition, we would later learn that enolates actually react with the unsaturated azide much more readily than the sp3-hybridized halide, preferentially forming triazoline-containing products instead of simple alkylated ketones. ~2

Br~l O

,. B r ~ , ,

0@ /

O ~],~ ""

Me ~'-

"Me@"

n

MeF-. 12t

X N3~I / // /

O ",, ~ / ~

~ N3~

MeF-; n Likely reaction course:

.•

N3~I

Me

N---N

O

Me~-~ SCHEME 5

An alternative route was based on the known ring contraction of pulegone 11 to chiral cyclopentanones bearing the right kind of substitution for our purposes (Scheme 6). ~3 The key step in this process is an in situ Favorskii rearrangement of a dibromoketone to form the desired fivemembered ring. Philosophically, by converting a chiral terpene into an alkaloid precursor, one establishes an appealing connection between two important classes of natural products. Finally, this route seemed readily scalable.

12

417

THF TOTAL SYNTHESIS OF AMPHIBIANALKALOIDS Me

Me Br2 ~

O

Me O NaOMe.~

HO B

pulegone(11) Me

1.0

NaOMe~

3

Me

2Me (2.372%stepsDMS~~)~ , C02Me 12 SCHEME 6

The conversion of pulegone to 12 worked just as advertised and provided a steady supply of starting material from which to examine the synthesis of the desired azido ketone. Initially, we expected that the ketoester 12 could be alkylated and then efficiently decarboxylated under either acidic or basic conditions to provide an appropriate azide precursor (Scheme 7). The alkylation was successful but we were only able to obtain traces of the desired decarboxylated product 14 or the fragmentation product 15 depending on the reaction conditions used. It seemed likely that the problem was the 13-methyl group, which shields the carboxylic ester in compound 13 from attack from external nucleophiles. The observation that basic conditions led to attack at the ketone carbonyl and ultimately retro-Dieckmann fragmentation is consistent with this view. Me HCI

NaH 12

Me

C02Me ,,~CI

/

~ ,

low yield

14 1. NaN3

66%

'~CI Me

0

13 41%

~

N3 CO2H 15

SCHEME 7

418

KEVIN J. FRANKOWSKI, AARON WROBLESKI, AND JEFFREY AUBI~

Although it seemed natural to attach the side chain via the doubly stabilized enolate, we were forced into another approach in which the ester group itself would be extended to provide the necessary azidoalkyl side chain. Thus, ketoester 12 was converted into the protected keto alcohol 16 by protection followed by reduction (Scheme 8). The resulting aldehyde was converted via a high-yielding synthetic sequence to the azido ketone 9, which was produced in a ca. 13:1 trans/cis ratio; the cis isomer was carried through several additional steps before removal by recrystallization. The alcohol 16 was oxidized and homologated to the ~,[3-unsaturated ester 17 without purification of the aldehyde intermediate. Dissolving metal reduction provided the unsaturated primary alcohol in good yield. 14 This was a nice reaction in which conjugate addition clearly occurred before comprehensive reduction of the ester to the alcohol. Finally, Mitsunobu conditions ~5 converted the alcohol to the azide 18.

Me ~ ' O '"CO2Me

1. ethylene glycol, H§ Me 2. LiAIH4 + O ' \OH 88%

o..)

12 Me +,O~I

16 1. Li, NH3 2. HN3, PPh3, Me CO2Et DEAD ,, ~ O . ~ ~ 84%

17

1. PCC 2. triethyl phosphonoacetate, DBU, LiBr 82%

1. LiBF4 2. H20/MeCN,. N3 93%

18

Me

Me N3

89-93% O

9

10 SCHEME 8

We were nearly ready for the key intramolecular Schmidt reaction. At this point, we were unsure whether it would be possible to deprotect the ketal in the presence of azide without also enacting the Schmidt reaction. Also, Craig Mossman of our group was then starting to examine the direct

12

419

THE TOTAL SYNTHESIS OF AMPHIBIAN ALKALOIDS

reactions of ketals with azides, ultimately showing that the conversion of an azido ketal to a lactam was possible using two steps (acid-promoted rearrangement followed by dealkylation) but that the reaction worked much better with dimethyl ketals than cyclic ones. ~6 As it happened, the ketal 18 was deprotected under mild Lewis acid conditions ~7without affecting the azide functionality to generate the azido Schmidt substrate 9 in very good yield. And, in the main event, the proposed intramolecular azido Schmidt reaction proceeded smoothly upon dissolution of the ketoazide 9 in trifluoroacetic acid to provide the bicyclic lactam 10 in yields ranging from 89 to 93% as a c a . 13"1 mixture of diastereomers. The mixture of isomers was separated by recrystallization from hexanes affording isomerically pure compound 10. At this point, the summer was nearly over and it was time for Ms. Rafferty to return to her classes. Obviously, she had done an outstanding job in bringing this synthesis, over the course of only a few weeks, up to its penultimate step. All that was now necessary was for Greg Milligan to add the side chain onto the bicyclic core (Scheme 9). Addition of pentyl Grignard followed by an acetic acid work-up afforded an iminium ion 19 that was not isolated. Instead, the direct addition of sodium borohydride to the reaction mixture and neutralization with potassium hydroxide afforded tertiary amine 1 directly. An obvious point concerns the stereochemical outcome of the reaction. The stereoelectronic principles of Stevens predicted that hydride attack from the pseudoaxial direction would be favored (Figure 4). ~8These predictions had previously been exploited by Polniaszek and Belmont in the synthesis of the related alkaloids 167B and 209D ~f as well as the previously mentioned nuphar indolizidine synthesis by LaLonde and coworkers (see Scheme 2). 8 Me

Me |

C5HllMgBr;

/ 0 10

Me

MgBrC5Hll 19

58% overall

: (~5Hll 209B (1)

SCHEME 9

The first evidence that the product possessed the desired relative stereochemistry was the observation of C-H stretches in the IR spectrum known as Bohlmann bands. ~9 These stretches are observed when a nitrogen lone pair is antiperiplanar to three C-H bonds as is the case in

420

KEVIN J. FRANKOWSKI, AARON WROBLESKI, AND JEFFREY AUBI~

HOe C5Hl1"

~

Me

19

L,.~

167B

209D

FIGURE 4. Pseudoaxial attack of hydride to the iminium ion 19.

Me

Me

Et

167A

Me

Me

Me

205A

237D

245D

FIGURE 5. Examples of other 5,8-disubstituted indolizidine alkaloids.

the alkaloids 209B and 251F (we will return to the subject of Bohlmann bands later in this chapter). In a more modern vein, the 1H and 13C NMR chemical shifts were in complete agreement with the values previously reported. 2~176Although a reliable value for the optical rotation of the natural product is unknown (owing to the limited quantity of material extracted from the frog skin secretions), our value was in agreement with the sign and magnitude of measurements on materials prepared in previous syntheses. ~~ This work was published in 1993, 2~ in a special issue of Heterocyles that honored the birthday of E.C. Taylor, another one of our heroes in the field of heterocyclic chemistry. Overall, the total synthesis was carried out in 22% yield over 13 steps. This compared favorably to previous syntheses, which at the time were relatively few in number. Since then, at least 14 more routes to 209B have appeared. 22 One can attribute this continuing interest to the same qualities that attracted us to this target: it is simple but interesting enough to constitute a good entrylevel target for those seeking to develop new methods for alkaloid synthesis. In addition, bicyclic lactam 10 is a potential precursor to over 30 naturally occurring indolizidines bearing a methyl substituent at C-8 (Figure 5). 6 It did briefly occur to us to run through a number of

12

THE TOTAL SYNTHESIS OF AMPHIBIAN ALKALOIDS

421

syntheses of related molecules. Although this activity would have certainly helped our collective publication records, we felt that so doing would not add much to our repertoire of new chemistry. Considering this, the fact that we were small in number, and that there was no shortage of exciting chemistry to do in other areas, we decided to not carry out this extension of the 209B work. It actually turned to be a number of years before another natural product in this overall class would attract our attention. I I l . Indolizidine 251F:

Background and Synthetic Planning

Compared to 209B 1, the dendrobatid alkaloid 251F 2 is a far more challenging target, possessing a fused tricyclic skeleton and seven stereogenic centers (six on contiguous carbon atoms; Figure 2). By the late 1990s, our group was interested in taking on more complex synthetic projects, and this particular alkaloid seemed appropriate for such an effort. As described below, the decision to synthesize 251F would not disappoint as a total synthesis workout, although this time most of the serious challenges turned out to be unrelated to our ostensibly key reaction. Although many dendrobatid alkaloids have been isolated from multiple frog species, the cyclopenta[b]quinolizidines, have thus far been detected in only one species. Originally identified as Dendrobates bombetes, 23 the taxonomic classification of this species has been changed to Minyobates bombetes: it is a small, poisonous Columbian f r o g . 24 Alkaloid 251F 2 was isolated as the major alkaloid of the skin extract of M. bombetes, and its structure was elucidated primarily via mass spectroscopy and NMR studies and reported by Daly and coworkers in 1992. 25 The skin extract from M. bombetes caused severe locomotor difficulties, muscle spasms, and convulsions upon injection into mice. 26 It is quite possible that the observed biological effects were due to other alkaloids present in the skin extract, so nothing is definitively known regarding the biological activity of 251F. In 1995, Taber and You reported the first synthesis of 251F utilizing a highly diastereoselective rhodium-catalyzed construction of a key cyclopentane intermediate. 27 This synthesis was a typically imaginative and efficient effort from the Taber group that relied upon their beautiful C-H insertion chemistry for the synthesis of a key intermediate. Thus, a seven-step sequence from enantiomerically pure [3-hydroxy ester 20 provided 0t-diazoester 21, which was subjected to the key diastereoselective rhodium-catalyzed carbene C-H insertion (Scheme 10).

422

KEVIN J. FRANKOWSKI, AARON WROBLESKI, AND JEFFREY AUBI~

The rhodium-catalyzed C-H insertion proceeded with retention of configuration, resulting in the formation of one diastereomer of 22 in 89% yield. Standard operations were then used in converting bicycle 22 to the substituted cyclopentane derivative 24. Overall, the 12-step sequence provided 24 in 14% yield. Scheme 11 shows how 24 was

rhodu

OH

Me

CO2Me

7 steps

Me

O-/~O

Me Me octanoate

O'/~'O

89% Me

Me

MeO2C

20

e

MeO2

21

22

Me Me 2. 1. BnBr LAH

~/~..,,,,,J M O'~'O e-

OH OTs 2. 1. TsCI, H3O+ Pyr

-~'.Me

12 steps, 14% overall yield

OBn

OBn 23

24 SCHEME l0

Me

1.03; PPh3 2. O

Me

Me.

6steoseeN3 EtO'P II

CO2Et

H

K2003, 24

3. PPh3;H20 Me

Me

M

25

CO2Et 26

CO2Et I - HO

27

1. BsCI 2. LHMDS

/

9

\"'

'"'

Me

Me

1. LAH 2. (PhS)2,Bu3P

CO2Et i|l,

Me \OBn

Me

28

29 SCHEME 11

3.NNH3

9

251F

12

THE TOTAL SYNTHESIS OF AMPHIBIAN ALKALOIDS

423

coupled to piperidine 27. In six steps, geraniol 25 was converted to azide 26 using a Sharpless asymmetric epoxidation to control the stereocenter of 26. Ozonolysis of 26 followed by Horner-Wadsworth-Emmons olefination, azide reduction, and conjugate addition generated 27. Under basic conditions 27 was coupled to 24 to provide the bicyclic alcohol 28. In the second key cyclization step, 28 was benzenesulfonated and subsequent enolate generation/alkylation provided the tricycle 29. Standard protocols were then followed for reduction of the ester to the corresponding methyl group and benzyl ether cleavage to the natural product. In 15 steps for the longest linear sequence, 251F was isolated in c a . 5% overall yield. Parenthetically, by using geraniol and species 21, which is derived from the terpene sulcatone, as starting materials, these workers also made a terpene ~ alkaloid connection similar to that described in the previous section. In fact, it was the appearance of the Taber and You publication that brought the existence of 251F to our attention. By this time, drawing retrosynthetic conversions of alkaloids to their potential intramolecular Schmidt reaction precursors had become something of a parlor game in the Aub6 group. In this case, as is usually so, two intramolecular Schmidt disconnections can be recognized by adding a lactam carbonyl group to either side of the target's tertiary amine (Scheme 12). Application of the intramolecular Schmidt retron to these two possible precursors leads to the two complex cyclopentanones shown. In addition to any issues arising from the Schmidt reactions themselves, it was clear that the main challenge in this synthesis would be preparing the necessary azido alcohol substrates with a high level of stereochemical control. Just as the natural product target contains six contiguous stereogenic centers, so do both of these precursors (there's also the lonely methyl group at C-3 (251F numbering)). However, application of classical retrosynthesis guidelines makes the choice between these two possibilities pretty obvious. In the branch shown to the left, all of the contiguous stereocenters are present on a [3.3.0]bicyclooctane skeleton, which immediately suggests the possibility of well-established means of controlling the stereochemistry using the bicyclooctane template. On the other hand, the alternative disconnection spreads these centers across two rings and also includes the awkward placement of the C-9 stereocenter on a one-carbon bridge. For these reasons, all further consideration was exclusively devoted to the bicyclooctane possibility. Taking the analysis further, it was clear that the left-hand side of the molecule 30 could well derive from a conjugate addition-alkylation trapping

424

KEVIN J. FRANKOWSKI,AARON WROBLESKI,AND JEFFREYAUBE

~

Me iiii /

o~

Me Me Indolizidine 251F 2

Me

Me

.... /

.... /

or

Me

OH

Me 0

Me

0

Me

Ji

Jl

Me

ee;?O

~' _~. . . . . Me

OH i1,1/

M

N3

N3

Me

SCHEME 12

strategy employed on an enone 31, provided that the methyl group and azide-bearing side chain could be added stereoselectively (Scheme 13). In general, stereocontrol in polycyclic cyclopentane-based ring systems has been intensively studied during years of research on the diverse and rich family of polyquinane natural products like hirsutene, capnelline, quadrone, and retigeranic acid. (The reader is advised to flip through the first couple "MEG"

Me /

X~)Rf~<~

~ .... ?TIPS

0

"

Me

Me N3

31

30

SCHEME 13

OTIPS

12

THE TOTAL SYNTHESIS OF AMPHIBIAN ALKALOIDS

425

of volumes of the Strategies and Tactics in Organic Synthesis series, published in 1984 and 1989, respectively, to get a sense of just how prevalent this area of research was in the 1980s - not to mention reading about some first rate science. Indeed, the honoree of the present volume, Paul Wender, burst onto the organic chemistry scene largely on the strength of his brilliant approach to such molecules using photoannulation chemistry; anyone who is not already well aware of this work could well start with the Wender/Temansky route to silphinene} 8) The first rule is that these ring systems almost always contain cis ring fusions because a trans-[3.3.0]octane ring system is highly strained. This leads to a cup-shaped bicyclic ring system, with a sterically disfavored endo position and a much more accessible exo face from which most reagents add. A second-order consideration is the minimization of steric interactions between substituents, that is, groups generally prefer a 1,2-trans disposition to the corresponding cis orientation, in which the two groups are essentially eclipsed. In the present case, manipulation of the proposed enone would require exo methylation of the "diquinane" enone 31 followed by placement of the azido side chain precursor in the more sterically demanding endo orientation. The first step seemed straightforward based on the analysis in the preceding paragraph, but the outcome of the second step would depend on the relative contributions of the endo/exo vs. the 1,2-cis/1,2trans issues. As the outcome of this was not at all obvious, MM2 calculations were performed on the in silico models shown (Scheme 14). These calculations indicated that the energy of the isomer with both groups exo and cis to one another was 2.4 kcal/mol greater than that of the trans isomer with an endo substituent. This suggests that minimizing 1,2 steric interactions is more important than having both groups in exo positions. Accordingly, this enone could serve as a reasonable intermediate en route to constructing 251F with the requisite stereochemical relationships: even if the desired stereochemistry were not obtained in a direct trapping experiment, we should be able to adjust the ~-keto stereocenter through a later equilibration. At the planning stage, we had another notion concerning a possible downstream consequence of this stereochemistry as it related to the intramolecular Schmidt reaction. In the event that a mixture of isomers had been obtained, or even if the undesired exo isomer had been the sole product, we had hopes that the situation might be corrected in the ring expansion step as shown in Scheme 15. The two possible stereoisomeric starting materials are shown, along with one possible Schmidt reaction intermediate and the resulting product for each. In the desired isomer

426

KEVIN J. FRANKOWSKI, AARON WROBLESKI, AND JEFFREY AUBI~

Stereochemical possibilities at C-10 endo,

Me

exo

Me

9 ,,i k

\

\

OTIPS

OTIPS Me~, 1 0

Me

Me

Me

N3

N3

Model system structures Me

Me OMe

Me/ " . . . . . . . . 0

OMe

/

Me,/~ ~ 0

Me

endo

.... /

Me exo

Relative energy (MM2)

endo: 0 kcal/mol

exo: 2.4 kcal/mol SCHEME 14

containing an endo-oriented azidoalkyl side chain, addition of the azide to the ketone can easily occur to provide a temporary seven-membered ring containing a methyl group in a pseudoequatorial orientation. Rearrangement then leads nicely to the ring system of the natural product, which also sports an equatorial methyl group; the azadecalin ring system in this structure has only a single axial substituent (the carbon with the methyl group attached). This methyl group suffers one 1,3-trans-diaxial interaction with a C-H bond due to the presence of the planar amide moiety. This situation stands in sharp contrast to that of the epimeric substrate, in which addition to the ketone will place an additional group into an axial position; in the intermediate shown, it's the methyl group that ends up in

12

427

THE TOTAL SYNTHESIS OF AMPHIBIAN ALKALOIDS

equatorial

enqb

i

",

TiC,n-I

Me

'\

. . . . .

Me'~ 0

o.2-1.

OTIPS

Me

Me

/

N3

TIPSO";

'--OTIPS

__1

axial

exq 9

~-I~ Me Me ~ N ~ ' - TiCln

Me '\

O .... / OTIPS OMeY

or

OTIPS M

_

N3

~ ~ N ~ oMe

TIPSO /

Me

OTIPS

SCHEME 15

this orientation (alternative structures resulting from cis vs. trans addition of the azide to the ketone or that involve boat-like conformations are possible but the ramifications are the same). Should this intermediate form and react, two of the possible product conformations are shown to the bottom right of the Scheme, indicating that at least one more large group will need to occupy a disfavored pseudoaxial position in the product. Since we knew that some conditions to carry out intramolecular Schmidt reactions result in epimerization of ot-keto groups due to transient formation of metal enolates (particularly when TiC14 is used to promote the reaction), 29 it was deemed plausible that even if the endo-oriented side chain was not the product of either kinetic or thermodynamic control in its formation, it might be persuaded to react preferentially from this orientation in the Schmidt step. Although we never had to test this idea, it was kept as a backup strategy as we began our work. Having reasonable confidence in our end game, it was time to deal with the non-trivial question of how to construct the desired enone. As a matter of principle, we strongly believe that even the most beautiful key transformation suffers if the surrounding steps are inefficient or awkward. The synthesis of the enone 31 was especially important to the overall viability of the project because it contained four of the seven stereocenters in the natural product. Indeed, these four particular stereocenters are contiguous and the two adjacent stereocenters in the natural product were planned to

428

KEVIN J. FRANKOWSKI, AARON WROBLESKI, AND JEFFREY AUBt~

be set using the conjugate addition/trapping sequence described above. Given that the additional stereocenter was planned to come in as part of a separate alkylating agent, it became especially important that whichever route we chose for the synthesis of this intermediate allow for its preparation in enantiomerically pure form so that the relative stereochemistry between the C-3 methyl group could be set relative to the other six centers (see Figure 6 below). At the time, we were already referring to enone 31 as the "diquinane" piece and, indeed, it was the PI's previous exposure to this field that ultimately suggested an intellectual framework for this part of the project. While a graduate student at Duke University, one of my favorite professors was Daniel Sternbach (to remind the reader, this is JA writing in the first person here). Dr. Sternbach is part of a distinguished lineage: his father was Leo Sternbach, inventor of Valium | and Librium | and he was a postdoc in the laboratories of both Albert Eschenmoser and R. B. Woodward. At the time, Dan was an assistant professor of organic chemistry interested in synthesis and, in particular, the quinane family of natural products. Furthermore, some of my good friends in graduate school, including Jeff Hughes and Carol Ensinger, worked on these projects. Germane to the present story is that we all talked a lot about chemistry, and so literally seconds after writing down the structure of compound 31 for the first time, it occurred to me that this molecule would be approachable by a signature Stembach approach to diquinane ring systems. Sternbach's approach to di- and triquinanes used the Diels-Alder reaction to generate relative stereochemistry in key cyclic intermediates. Once generated, the six-membered cyclohexene ring generated in the Diels-Alder reaction could then be oxidatively opened to afford two aldehydes, one of which would be perfectly situated to undergo aldol ring closure to form one of the five-membered ring in the natural product. Scheme 16 highlights this sequence en route to the total synthesis of silphinene. 3~ This method was also utilized in a synthesis of hirsutene (not shown) 3~ and in approaches to other targets in this field. Sternbach published an enjoyable account of this most elegant chemistry in the second volume of Strategies and Tactics. 32 In the present project, bicyclic enone 31 contains four consecutive stereogenic centers that demand stereocontrolled installation. A direct application of the Sternbach diquinane strategy to this system results in the retrosynthetic analysis shown in Scheme 17. Retroaldol of 32, along with a pro f o r m a oxidation of the primary alcohol in the side chain, leads to a tricarbonyl intermediate 33. Stitching the two aldehydes back together leads to the bicyclic ketone 34 that is immediately recognized as

!2

429

THE TOTAL SYNTHESIS OF AMPHIBIAN ALKALOIDS

Me 160~

Me Me M

1.03 2. DMS

Me~',0

MeOH, KOH,.

M_e M e~,,,CHO Me -

Me

'CHO = Me Me:o~CHO e ,, M Me M e ~ Me _

steps"

.

silphinene SCHEME 16

Me~ ,,,'~('~ 0

Me 32

~0 ~

H

>

Me 33

Me

0

34

SCHEME 17

a formal Diels-Alder adduct between cyclopentadiene and a simple enone dienophile. Furthermore, the stereochemistry of this transformation fell beautifully into place in two important ways. First of all, such a Diels-Alder adduct would certainly afford the desired relative stereochemistry of each stereocenter thanks to the Alder endo rule. Also, the 1990s were a time of much activity in the area of asymmetric Diels-Alder catalysis, and the go-to diene for nearly every paper in this area was cyclopentadiene. Thus, it seemed very likely that we would be able to obtain a fully stereocontrolled route to 32 using some variation of this overall plan. Despite the length of this section, most of these ideas arose and were solidified in the matter of only about an hour total, with probably another hour over the next few days taking a look at some of the relevant literature. Satisfied that we had a shot at a synthesis of alkaloid 251F that served the key intramolecular Schmidt step and would also permit an attractive route to the key enone 32, it was time to get to work.

430

KEVIN J. FRANKOWSKI, AARON WROBLESKI, AND JEFFREY AUBI~

IV. First Generation Work: Synthesis of the Key Enone and a Total Synthesis of Racemic Desmethyl 251F

I asked Dr. Kiran Sahasrabudhe, an excellent postdoctoral colleague who had joined my laboratory after completing a Ph.D. program at Penn State with Professor Ken Feldman, to begin work on the 251F project. Kiran had been working with us on a problem related to asymmetric ring expansions and was itching to get involved in total synthesis, so she viewed this as an attractive opportunity. We decided to initially carry out a model study of a racemic 251F derivative lacking the 3-methyl group as our first subgoal (Figure 6). The main advantages were that working out the chemistry pertaining to Me

Me

....?H

/OH ....

MeaN ~

Me

Me

(_+)-3-desmethyl251F(3)

(-)-251F(2)

Me

Me

. . . . . . .

/

. . . . . . .

H"~

/

Me

N3

N3 35

U

X H

30

..../O''

+

N3

S

<~~

..../

Me

O

Me

36

31

achiral material

racemic substrate

N3

0

37

\

Me 31

V

..)

enantiopure components FIGURE 6. Rationale for the model synthesis.

OTIPS

12

THE TOTAL SYNTHESIS OF AMPHIBIAN ALKALOIDS

431

the diquinane piece would initially be simpler and less expensive. Since the C-3 stereochemistry needed to be "matched" to the enone portion, the removal of the C-3 methyl group was necessary for working in the racemic series. As a side benefit, this would allow us to test the problem of C-10 side chain stereochemistry without the complicating presence of this additional substituents (see Scheme 15 and the surrounding discussion). Once the required chemistry was developed, individual asymmetric syntheses of both components needed for the 251F synthesis would commence. Although risky, we decided to initially prepare enone 32 using the specific precursor shown above in Scheme 17. It was most convenient to synthesize the methyl ketone 34 by carrying out the known Diels-Alder reaction shown 33 and process the adduct 38 through the Weinreb amide followed by methyl Grignard addition. Olefin 34 was ozonized to yield the corresponding dialdehyde 33; however, in no case could an intramolecular aldol reaction be induced to afford 40. We had thought a bit about the possible stereochemical outcomes of deprotonating this tricarbonyl compound (Scheme 18). For example, we figured that the two carbonylcontaining groups adjacent to C-2 would prefer to remain in their trans relative stereochemistries to minimize 1,2-steric interactions. And, while the C-2 formyl substituent might thermodynamically prefer the trans position relative to the acetyl side chain, an aldol reaction from the C-2 epimer would be extremely unlikely as it would result in the formation of a t r a n s - f u s e d [3.3.0] bicyclic ring system. As it happened, we were never able to recover any tractable materials by treatment of 33 with a large array of bases (we used Steven Burke's papers on quadrone synthesis as a source of various aldol conditions to try34). Most likely, we underestimated the lure of [3-elimination reactions for the various 1,5-dicarbonyl subunits within compound 33. One such process is shown, but similar decomposition modes are equally likely beginning with removal of any of the protons Ha-H c. As it became apparent that a direct ozonolysis/aldol strategy on substrates such as 34 was not viable, alternatives that would be less likely to undergo this kind of elimination were sought. This would be most straightforwardly accomplished by removing the C-4 aldehyde, that is, the one that was not needed for the desired aldol process. The two aldehydes were readily differentiated by ozonolysis of olefin 41 followed by borohydride reduction (Scheme 19). This resulted in the formation of the diol 42, which was not isolated but upon acidification spontaneously lactonized. Protection of the free alcohol yielded the lactone 43 in 70% overall yield for the three-step sequence. At this point,

432

KEVIN J. FRANKOWSKI, AARON WROBLESKI, AND JEFFREY AUBI~

Me

Me

9

steps

7

7

34

38 0

1.0 3

0

aldol

2. DMS

//

H

Me--,,~" ~M e H

O

33

Me 4O

(not observed)

One possible decomposition mode:

-'JlHa~He ,, O

H

O

, [ . ~ ....\0 /(, 13-eliminationof Hb=

Me"'~ 'Hc Me H

H~, , , , ~ H Me ~/O

O

unfortunate "products"

iMe '

SCHEME 18

we hoped that methyl lithium addition (1 equiv.) to lactone 43 would afford a hemiacetal 44 that could be oxidized to methyl ketone 45, the desired aldol precursor. In practice, the only product isolated upon addition of methyl lithium was the tertiary alcohol 46, a result of double addition to the lactone, along with recovered starting material. Not surprisingly, replacement of methyl lithium with the corresponding Grignard reagent gave the same results, as did carrying out the reaction at lower temperatures. Accordingly, an alternate tactic toward the methyl ketone 45 was developed. Instead of using a methyl lithium addition/oxidation sequence to introduce the extra carbon needed to generate the desired methyl ketone 45, an olefination of the lactone 43 was contemplated (Scheme 20). This would directly convert the lactone carbonyl into the enol ether corresponding to the desired methyl ketone. Initial experiments focused on using the Tebbe reagent 35 to convert 43 to the analogous enol ether 47.

12

433

THE TOTAL SYNTHESIS OF AMPHIBIAN ALKALOIDS

Me

OH / ii:[~

1.0 3 2" NaBH4-

7

....

--H

3 steps

2

OTIPS dk~T.~

70% HO"~'0

/1~

O

Me

41

..../ Me

43

42

OH /

%%

,, /

OTIPS

1 equivMeLi Me Me Me 46

>..../

.

.

.

.

H"J/""G

oxidation

....2 TIPS

. . . . . . . .

HO

.

O

OTIPS

Me Me not observed 44 .

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

"~'" ~Me 45

.

SCHEME 19

O~) O

'"'

/

OTIPS

CP2TiMe2~~ M O

"||

Me

e

43

47

/

OTIPS

/~_

silica gel

75% (3steps)

H J/'', G

,, ""~

/ OTIPS KOH,Bu4NOH

OTIPS

44

0

PCC

_ J--"x

O/ T ~ ..../ HO Me Me

<~~

..../

OTIPS

52% ~Me

o

Me

31

45

SCHEME 20

However, no positive results were obtained via this avenue. After some experimentation, it was discovered that using the Petasis reagent 36 (Cp2TiMe2) in place of Tebbe reagent provided the desired enol ether 47. Amy Howell at the University of Connecticut had substantial experience

434

KEVIN J. FRANKOWSKI, AARON WROBLESKI, AND JEFFREY AUBI~

with this kind of conversion, 37 and we benefited greatly from Professor Howell's guidance on preparing and using this reagent. The enol ether 47 was subsequently hydrolyzed to the hemiacetal 44 and immediately oxidized with PCC to generate the aldol precursor 45. After a number of conditions were examined to carry out the intramolecular aldol conversion of 45 to 31, reasonably efficient conditions were identified with the help of Carol Ensinger, who graciously shared the best aldol results that she had obtained while in the Sternbach laboratory. Ultimately, KOH and BuaNOH in refluxing THF/water provided the aldol adduct 31 in 52% yield. Compound 31 was accompanied by small amounts of an isomeric enone assigned as compound 49 below (Scheme 21). Apparently, the keto aldehyde precursor was not completely immune to epimerization and a small amount of the doubly epimerized keto aldehyde 48 built up in the reaction vessel, ultimately affording the isomeric enone in 5-10% isolated yields. Although this was not an undue hardship, thanks to our ability to remove this isomer by chromatography, the formation of 49 eroded the efficiency of the synthesis by removing some perfectly good keto aldehyde from the pipeline. O / OTIPS KOH, Bu4NOH 52%

("/ T

'~ .... /

0

Me

45

OTIPS major product

31 KOH, Bu4NOH

o

H

OTIPS ,e,,,/

H 48

,,.

O~M,

e ,,,

/

OTIPS isolated in 5-10% yield

49 SCHEME 21

At about this time, Dr. Sahasrabudhe had obtained more gainful employment and so the project changed hands. Aaron Wrobleski, at the

12

THE TOTAL SYNTHESIS OF AMPHIBIAN ALKALOIDS

435

time a second-year graduate student, was completing a "starter project" in which he had examined a series of intramolecular reactions of benzylic azides with ketones 3s and was ready for a more substantial challenge. Building on Kiran's early work, Aaron was able to scale up the route presented in Scheme 20 to afford quantities of 31 in excess of 2 g. Despite this heroic effort, numerous drawbacks using this strategy surfaced. The Petasis olefination product 47 was obtained cleanly and could be used in subsequent transformations without purification (though some titanocene byproduct could be seen in the crude NMR analysis). However, only after the aldol step could the product be purified chromatographically. The four-step sequence in going from 43 to 31 was routinely carried out with crude reaction mixtures, analyzing the intermediate products only with 1H NMR to ensure complete reactant consumption. Overall, the seven-step sequence from 41 could be carried out in 27% overall yield at best. At seven steps, this sequence was sufficient for carrying out preliminary experiments; however, more efficient means would ultimately be sought to streamline the synthesis. These nagging concerns were temporarily set aside and work was continued toward the construction of desmethyl 251F. Thus, the next goal of the model study became the stereocontrolled installation of appropriate substituents onto the enone moiety. As discussed above, with the cupshaped geometry of a cis-fused 5,5-bicyclic system, we expected that methyl cuprate addition would place the methyl group in the desired exo position. We furthermore dared to hope that quenching the resultant enolate with an appropriate electrophile would stereoselectively install a precursor to the endo azido side chain (trans to the methyl group). Toward that end, conjugate addition of MezCuLi followed by quenching with 1-chloro-4-iodobutane resulted only in installation of a methyl group at the [3-position of the enone (compound 50, Scheme 22). Under a variety of conditions examined using alkyl halides as electrophiles, no successful instances of incorporation of ~-substituents were observed. The use of even highly reactive electrophiles (MeI or allyl bromide) resulted only in the conjugate addition product. With an abundance of 50 collecting, we examined a two-pot sequence in which 50 was treated with a variety of bases (LiHMDS, NaHMDS, KHMDS, etc.) and quenched with alkyl halides. These experiments resulted only in the re-isolation of starting material (Scheme 22). Becoming convinced that using an alkyl halide to install the azido side chain was not a viable option, explorations into other possible electrophiles commenced. Because of the highly electrophilic nature of aldehydes,

436

KEVINJ. FRANKOWSKI,AARONWROBLESKI,ANDJEFFREYAUBI~

1. Me2CuLi OTIPS 2.1-chloro-4- Me iodobutane ~ ..../

<~~ 0

Me

@

31

Me { ~

/OTIPS

,,,,

OTIPS

,,

Me 50

,,b|/

~-~' O CI

Me

51 not observed

. . . . . . . . . . . . . . . . . . . . . . . . . . .

Me

~

|,,,

0

/

OTIPS

1. numerousbases 2. numerousalkyl halides 9

!

recoveryof 50

Me 50 SCHEME 22

we felt that a one-pot conjugate addition/aldol strategy could serve as an effective means for introducing both substituents. With that, conjugate

addition of Me2CuLi followed by quenching with an appropriate aldehyde was examined. To our delight, aldehydes proved to be efficient coupling

partners (Scheme 23). Accordingly, addition of MezCuLi from the most exposed face of 31 resulted in enolate 52, which when quenched with

.... /

OTIPS Me2CuLi

Me ~

/ OTIPS

O H-"JJ'~

,i,,

O

Me

LiO

31

Me 52

m

H

OBn

Me ~,~

OTIPS

,,,,

,,,,/

Me

0

H20 OTIPS

65 - 85%

OBn

iiii

Me

0

53

54 SCHEME 23

/

OBn

12 THETOTALSYNTHESISOFAMPHIBIANALKALOIDS

437

4-benzyloxybutanal resulted in aldol adduct 53 (not isolated). In situ dehydration of 53 led to a single olefinic isomer of 54. In other examples, we have had to force dehydration to occur in a separate step (e.g., see our synthesis of lasubine II29), but that turned out to be unnecessary here. Scheme 23 indicates the E stereochemistry for the aldol dehydration, but it was not until the asymmetric synthesis of 251F that the olefin geometry was rigorously proven. The in situ dehydration of the aldol adduct 53 to the exocyclic enone 54 precluded the direct formation of the sixth consecutive stereocenter. This point was rendered inconsequential as it was recognized that a dissolving metal reduction could serve two functions: (1) cleavage of the benzyl ether of 54 to the corresponding alcohol and (2) reduction of the enone moiety to the saturated derivative placing the azido side chain precursor in the desired endo position (Scheme 24). In this situation, the side chain stereochemistry could be set in two ways. On the one hand, our model study calculations suggested that thermodynamic protonation of the intermediate enolate 55 would favor placing the side chain trans to the existing adjacent methyl group. On the other hand, kinetic protonation of the intermediate enolate might selectively occur from the less hindered exo face. In practice, the dissolving metal reduction of 54 proceeded to yield one observable diastereomer of 56. However, the stereochemistry of this newly formed center could not be verified until later in the synthesis OBn

OBn

L ~

OTIPS

/

exo protonation

Na/NH 3

OT, S

iiii/

0

Me 54

55

Me /

-

OTIPS

OH 56

SCHEME24

438

KEVIN J. FRANKOWSKI, AARON WROBLESKI, AND JEFFREY AUBE

due to the lack of definitive NMR evidence. Rather than purify the alcohol 56, it was directly converted into the azide necessary for the final test of this model system. Progress toward the azido ketone Schmidt precursor was carried out via azidation of primary alcohol 56 under Mitsunobu conditions 39 to yield the azide 57 in ca. 50% overall yield from 54 (two steps, Scheme 25). There are a number of ways of installing azides into Schmidt reaction precursors. Typically, we like the Mitsunobu reaction because it requires only a single step from the precursor alcohol, although in some cases oxygen activation through mesylation or tosylation followed by sodium azide treatment has been used. Hydrazoic acid is a competent azide source in the Mitsunobu reaction, but has the obvious drawbacks of toxicity and explosion potential. Accordingly, the group has adopted the use of the zinc azide complex shown in the scheme for this purpose; 39 this reagent is touted as being safer than hydrazoic acid on grounds of both stability and toxicity. It is a solid reagent that can be readily made and stored without untoward effects, to the best of our knowledge.

Me

Me

' . .O. .~. M , e / OTIPS Zn(Na)2~ DEAD, PPh3 ~ ' ' .O.~ .M .I 50% (from 54, 2 steps) OH N3

/ OTIPS TfOH, CH2CI 2

y

56

52%

57

Me ,,,

OT,.s

Me

i

|||l/ O

58

OH

Me

59 SCHEME 25

With 57 in hand, the stage was set for carrying out the intramolecular Schmidt reaction. Thus, treatment of 57 with excess triflic acid (TfOH) smoothly converted the azido ketone 57 to the ring-expanded lactam 59 while simultaneously removing the silyl protecting group to reveal the primary alcohol in 52% yield. The successful use of triflic acid was actually

12

439

THE TOTAL SYNTHESIS OF AMPHIBIAN ALKALOIDS

quite surprising, given that the ideal distance between the ketone and the carbonyl in intramolecular Schmidt reactions is four carbons, which generates a nice six-membered ring azidohydrin intermediate. In this case, the presence of five intervening carbons means that one needs to form the seven-membered ring intermediate 58. Greg Milligan, in carrying out his initial studies of the intramolecular Schmidt reaction, had found that such reactions typically required more stringent conditions and in particular strong Lewis acids like TiC14 for their promotion. Thus, the utility of even the strong protic acid TfOH in the synthesis of 59 was unexpected. At this time, extensive 2D NMR analysis was used to corroborate the stereochemistry of all six stereogenic centers. Thus, the combined use of COSY, HMQC, and HMBC permitted the assignment of values to all protons and carbons of 59. A NOESY experiment was then performed to establish all NOEs in the molecule (Figure 7). The most revealing piece of evidence came from the observation of an NOE between the C-10 methine proton and the adjacent methyl group, indicating a cis orientation between the two or, alternatively, a trans relationship between the methyl group and side chain. Four other NOEs were observed and further corroborated that the Schmidt product had stereochemistry corresponding to that of the natural product. Had the methyl group and azido side chain been cis to one another (both groups exo), only four NOEs would be expected (60, Figure 7), and most importantly, the C-10 methine-methyl NOE would be absent. This afforded evidence that during the dissolving metal reduction of 54 to 56 (Scheme 24) exo protonation occurred and placed the side chain in the expected endo position, thus supporting the stereochemical outcome of this step. H~H~H / L ~

selectedNOEs consistent with desired stereochemistry

~N~o~

M

e

H

59

~

HO Me H~NH~~~ C-10 methine-Me would not be observed H ; ~ in this isomer ~ e 60

Me

O NOEs confirm trans orientation between Me and side chain

.Me NOEs disprove cis orientation between Me and side chain

FIGURE 7. Determination of the relative stereochemistry using selected NOE experiments.

440

KEVIN J. FRANKOWSKI, AARON WROBLESKI, AND JEFFREY AUBl~

Completion of the model synthesis required only the reduction of the lactam to the corresponding tertiary amine. Thus, treatment of 59 with LAH in ether afforded the desired amine 3, 3-desmethyl 251F, albeit recovered only in sufficient quantity to obtain mass spectral and IR characterization (Scheme 26). Further evidence supporting the stereochemical relationships assigned using the above-described NMR techniques came with the observation of Bohlmann bands ~9in the IR spectrum of the amine (see the discussion on the completion of the asymmetric 251F synthesis below for more details about the significance of Bohlmann bands). Me

Me /

OH

traces 0

Me

Me

59

3, 3-desmethyl 251F SCHEME 26

The racemic desmethyl analog of 251F was synthesized in 17 steps to yield detectable quantities of the compound, establishing the validity of the key intramolecular Schmidt reaction toward the synthesis of 251F. More importantly, many of the pitfalls encountered e n r o u t e to the key intermediate 31 that were overcome would be instrumental in the success of the natural product synthesis. The lessons learned about what did not work would free us to explore other, less traditional directions and ultimately arrive at a route to significant amounts of 251F. V. I n t e r m e z z o : A Few Words About Safely Working with Azides

This is as good a point as any to briefly discuss safety concerns with using alkyl azides. Although we have never had an explosion in our laboratory due to any azide, we are always cognizant of the potential of these compounds to explode. The tendency of a given azide to explode is correlated with how much "active nitrogen content" is present in a particular compound, often expressed as the number of C +O atoms over the number of N atoms in a compound; ratios over 3 are generally considered nonexplosive. A couple of representative azides are shown in Figure 8 below. Note that the Schmidt substrate shown to the left has a high (C+O)/N value; while we would never refer to any compound as completely safe, this one has a low probability of posing an explosion hazard. In contrast,

12

441

THE TOTAL SYNTHESIS OF AMPHIBIAN ALKALOIDS

Me ,|

O~MI

,,,/

OTIPS /N3 / N3

0 N3

NC+ NO NN

8.7

(probably safe)

1.7

(should be careful)

0.67

(extremely hazardous!)

FIGURE 8. Examples of organic azides and their relative stabilities.

4-azidobutyraldehyde has a ratio somewhat below 3 and could be considered a candidate for trouble. Accordingly, compounds of this type should be prepared and used in small scale using appropriate safety equipment ( g l a s s e s - a l w a y s ! - and shields for operations such as concentration). One should think very carefully before distilling a compound of this type or making it on multigram scale: if such operations are considered important, then it would benefit the user to contract for experimental assessment of the thermal stability of the azide through differential scanning calorimetry. No such assessment would be necessary for the Bad Actor shown at the far right: not only does it have an unconscionable (C +O)/N ratio below 1, but the presence of not one, but two azides to only four carbons as well as the highly energetic triple bond makes for a dangerous species. One of us has seen the aftermath of a violent explosion of this material and can attest first-hand to the inadvisability of making and working with this compound on any scale. These considerations are routinely taken into account during synthetic planning exercises. In principle, one could have envisioned a more convergent synthesis of our desired azide using the direct coupling of enone 31 and 4-azidobutryaldehyde as shown in Scheme 27. Concerns about chemistry aside (this route would require a chemoselective reduction of the enone in the presence of azide), the price for additional convergency here would have been the need to make the azidobutyraldehyde on scale. As a rule of thumb, we forestall incorporating the azide group until as late in a synthesis as possible to (1) avoid as many azide-containing intermediates as possible and (2) maximize the formula weight and therefore the safety of the azides that we do use. Even then, the need to use some low molecular weight azide reagent is nearly unavoidable, although using Zn(N3) 2 instead of HN 3 for the Mitsunobu reaction above was a choice driven by safety considerations.

442

KEVIN J. FRANKOWSKI, AARON WROBLESKI, AND JEFFREY AUBI~

1. Me2CuLi 2.

O ,.,JJ,~

OTIPS .... / O

N3

L.

N3

OTIPS

H

.... /

Me

O

31

Me 61

SCHEME 27

I am often asked about another matter of safety pertaining to the Schmidt reaction. In recent years, it has become widely understood that one should not carry out displacement reactions using sodium azide or other nucleophilic N 3 anion in chlorinated solvents, especially methylene chloride. Presumably, the nucleophilic azide is able to displace a halide atom in the solvent to generate a low molecular weight azide in situ. In the case of methylene chloride, the possibility of forming diazidomethane is especially concerning ((C +O)/N = 0.17!). However, this concern does not extend to alkyl azides, which are much, much less nucleophilic than azide anion. Thus, one can generally carry out alkyl azide reactions such as the intramolecular Schmidt reaction in chlorinated solvents without forming diazidomethane or similarly hazardous byproducts. VI. Development of a New Route to a Key Enone and the Total Synthesis of Alkaloid 2 5 1 F

Although the synthesis of desmethyl 251F helped develop a great deal of the chemistry needed to formulate a route toward the natural product, new challenges awaited in completing the total synthesis of 251F itself. First, the issue of carrying out an asymmetric synthesis had to be confronted. As noted earlier, this issue was not viewed as a serious problem as numerous routes toward both enantiomers of acid 41 had been previously developed. 4~ A more pressing task, however, was the desire to produce enone 31, or its equivalent, in a more direct route. Although we had synthesized ca. 2-gram quantities of this important intermediate, the number of steps and low overall yield were viewed as detracting from the overall attractiveness of the total synthesis. A satisfying solution to this issue ultimately would be found by employing metathesis chemistry. In addition, the 3-methyl group had to be incorporated with the proper stereochemistry. Finally, we had also set for ourselves a goal to produce meaningful quantities of the dendrobatid natural product for biological

12

443

THE TOTAL SYNTHESIS OF AMPHIBIAN ALKALOIDS

screening. Along these lines, we settled on c a . 100 mg of the natural product as a reasonable benchmark. Although a huge amount in the context of modern high-throughput screening techniques, we felt that making this amount would make a positive statement about this approach to dendrobatid alkaloids. We next turned our attention to the preparation of key intermediates in enantiomerically pure form. Scheme 28 highlights a few (of many) methods that have been developed for the synthesis of enantioenriched derivatives of acid 41. Both catalyst 4~ and substrate control (chiral auxiliaries) 42 have proven effective as means for delivering the desired Diels-Alder adduct with high levels of enantioselection. Two of these methods were attempted, with our initial attempts focused on catalytic processes. Thus, the achiral diene 62 and cyclopentadiene were treated with Cu(OTf)2and the indicated bisoxazoline catalyst system. 41 Possibly due to difficulties in keeping the reaction highly anhydrous, this reaction failed to deliver the desired Diels-Alder 63 adduct in our hands. Despite our confidence that we would have been ultimately able to carry out this well-known procedure, we were anxious to get moving in the

0

0

oA.A Me h_J

~;KMe

Cu(OTf), L*, C5H6

63

Me

C5H6

Me O

L* =

62

Me

oMe

0

7

,_ou

e

chiral

0

catalyst

pantolactone chiral auxiliary

o o -O, e

O

O

Lewis acid,

. ~ M e

C5H6

oxazolidinone chiral auxiliary

L__/ Ph"

"Me Ph

64

65 SCHEME 28

444

KEVIN J. FRANKOWSKI,AARON WROBLESKI,AND JEFFREY AUBI~

enantiomerically pure series and so turned in the interim to a chiral auxiliary based approach to the same intermediate. Success was quickly realized, when the (4R,5S)-(+)-4-methyl-5-phenyl-2-oxazolidinonecontaining dienophile 64 was treated with cyclopentadiene and Lewis acid. This readily afforded a useful supply of the necessary intermediate 65 and so we never bothered to return to the more modern catalytic means of making the Diels-Alder adduct. Turning to the asymmetric synthesis of the azide-containing side chain 37, we recognized that the same chiral auxiliary could be used in the synthesis of the methyl-containing precursor 70. Thus, oxazolidinone 66 was used to control the outcome of an asymmetric alkylation reaction to set the C-3 stereocenter (Scheme 29). We confess that the use of the same chiral auxiliary for the synthesis of both major chiral precursors was an unplanned bonus in the synthesis of 251F rather than a deliberate design element.

0

~U Me

0 Scheme 28

crotonyl chloride ,r

0

Me) h

H-N/~ 0

Ph

64 0

66

65 0

-,..,,~N,,J~O propionyl chloride

1. NaHMDS

Me~~Ph

68

"•-'••OBn Me

0

MeMe~-~ph

67

1. LAH 2. Nail, BnBr

0

2. allyl iodide ~ N , , J ' L O

68 cat. OS04, NMNO, Nal04 36% (from 66)

69

O~l~~-,~OBn H Me 70

SCHEME 29

The synthesis of the enantiopure aldehyde 70 commenced with conversion of auxiliary 66 to 67 via deprotonation (butyllithium, THF) and quenching with propionyl chloride. The enolate of 67 was generated with NaHMDS and subsequently reacted with allyl iodide to provide 68 as a

12 THETOTALSYNTHESISOFAMPHIBIANALKALOIDS

445

white solid that could be recrystallized to ->95% diastereomeric purity. Reductive removal of the chiral auxiliary with LAH afforded the enantiopure alcohol 43 that was immediately benzylated to the ether 69. Finally, oxidative cleavage of the olefin afforded the desired aldehyde 70 in 36% yield over the five-step sequence. Conversion of 64 to 65 was effected by treating a mixture of 64 and cyclopentadiene in methylene chloride at - 1 0 0 ~ with diethylaluminum chloride. The Diels-Alder adduct 65 was isolated in ca. 93-95% diastereomeric purity on a 2-g scale. With protocols in hand for the asymmetric synthesis of acid 41 and the provisions met for the stereocontrolled installation of the 3-methyl group, we began to think seriously about re-tooling the ozonolysis/aldol route toward the bicyclic enone intermediate 31, given the length of this route and our concerns that in particular the lactone olefination and the aldol steps were insufficiently robust for use with the more hard-won enantiomerically pure materials. Aaron was discussing this problem at a group meeting when I had a "V8 moment", referring to an ad campaign for the popular tomato juice concoction: "We could use a metathesis reaction for that!" The thought that ruthenium-catalyzed ring-opening/ring-closing metathesis (RO/RCM) could be utilized in the desired [2.2.1] --, [3.3.0] skeletal rearrangement had very strong literature precedent. Robert Grubbs, in 1990 (i.e., before ring-closing metathesis was the force of chemistry that it has become), published a route to the triquinane natural product capnellene using titanocene alkylidene complexes (Scheme 30). 44This synthesis, which

Me

Me

CP2Ti=CH2=

Me

Me

.

MeO

M 71

M e ~

Cp2

Me01~.O TiCP2

72

Me O ~ Cp2,,,- Me

_

M

~ OMe-

73

steps H capnellene

MeO 74

SCHEME30

446

KEVIN J. FRANKOWSKI, AARON WROBLESKI, AND JEFFREY AUBI~

is one of very few to come out of the Grubbs laboratory, uses the Stembach Diels-Alder/ring-shuffling strategy, but with a twist. Thus, the Tebbe reagent 35 served as a source of titanaethylene (CP2Ti=CH2), which cycloadded onto the endocyclic olefin 71 and converted to the titanocene 73. In standard Tebbe fashion, the titanocene 73 olefinated the ester carbonyl with loss of O=TiCP2 to provide the skeletally rearranged enol ether 74. It seemed very attractive to update this approach for the synthesis of enone 31 using the more recently developed ring-opening metathesis catalysts based on ruthenium. 45At the time, there were already a few applications of combining ringopening metatheses with the ring-closing variety but at about the time that our work came out, many other related examples of this chemistry were starting to emerge. 46 The student of total synthesis will especially want to study some very elegant examples of this strategy used by Andrew Phillips and coworkers working at the University of Colorado. 47 So inspired, it seemed plausible that with proper manipulation of the [2.2.1]bicycloheptane skeleton of acid 41, its transformation to the desired [3.3.0]bicyclooctane enone intermediate would be possible. To accomplish this, we would need to convert the methyl ketone used for the aldol transformation into an appropriate olefin. In building this substrate, the Diels-Alder adduct 65 was cleaved to the enantiopure acid 41 with LiOH/H202 .48 Reduction of the acid to the alcohol 75 (LAH), oxidation to the aldehyde, and subsequent vinyl lithium addition provided the first potential metathesis substrate 76 as a 1:1 mixture of diastereomers (Scheme 31). Under a variety of conditions

2. 1. LAH LiOH, H202,.

OM.M.M~e~~_( O

7

Me

. oxidation . vinyllithium

.~~ .

Me I

%-~'~OH

H Ph 65

75

76 Cl,,, P(CY)3

~ u--Nph

Grubbs (77)

HO

Me

CI P(Cy)3 77

78 not observed .

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.4

SCHEME 31

12

447

THE TOTAL SYNTHESIS OF AMPHIBIAN ALKALOIDS

examined, however, the desired RO/RCM did not provide 78 and only the starting allylic alcohol 76 could be detected. This was not especially discouraging as it seemed possible that the alcohol portion of 76 would have complexed the catalyst, thus inhibiting the desired reaction pathway. In addition, an additional oxidation step would have been necessary to convert 72 to the desired enone. A more economical strategy would be carrying out the tandem metathesis reaction on a substrate already at the enone oxidation state, thus avoiding the additional reduction and oxidation steps shown in Scheme 31. Thus, the conversion of 41 to the Weinreb amide 49 followed by addition of vinyl Grignard afforded 79 in 85% yield for the two-step sequence (Scheme 32). The enone 79 represented a substrate that was both free of a potentially troublesome alcohol and in the desired oxidation state of the targeted bicyclic enone 80.

1. HNMe(OMe).HCI, Me

2. vinyl Grignard

CH2Ct2

9.

85%

'"

ca. 30% ~"'~-'~/%O

41

=_

79

O

,\\ \\

+ oligomerized by-products

Me 80

SCHEME 32

At this stage it became possible to explore the use of Grubbs's ruthenium catalysts to effect the desired RO/RCM reactions. Standard reaction conditions 46a afforded the desired enone 80, albeit in poor yield. Under a variety of modified conditions examined, it appeared that a good deal of the substrate had been converted to oligomerized material. Though only a 30% yield of the metathesis product was initially isolated, we felt that this yield could be increased with the identification of more fitting conditions. Upon further examination of the literature 46b and a very productive breakfast conversation with Professor Grubbs, we soon learned that certain additives to metathesis reactions can have drastic effects on their overall efficiencies. In particular, it was recommended that we attempt the reaction in the presence of another olefin (3-hexene or ethylene) to allow for redirection of unproductive pathways back into useful chemistry (this point is discussed in detail below). As it happened, the reaction of 79 was

448

KEVIN J. FRANKOWSKI, AARON WROBLESKI, AND JEFFREY AUBI~

attempted using the Grubbs first generation catalyst 77 in methylene chloride with 5 equiv, of 3-hexene. The resulting product was a derivative of 80 (containing a 2-butenyl group in place of the vinyl substituent), however, the yields remained in the 30-40% range. Remarkable increases in reaction efficiencies were realized, however, when the tandem RO/RCM was carried out in ethylene-saturated methylene chloride. With a minimal amount of optimization, enone 80 was isolated in 93% yield using 5 mol% of ruthenium catalyst 77 on multigram scale reactions. Scheme 33 presents a likely mechanism for production of the desired enone 80.

79

5 mol% 77, C2H4, 0H2CI2

LnR

~

~ LnRu . ~ . . ~ . ~

~ O

~ O

81

82

_ LnRU~.,.~ "t~O 83

93% 80 SCHEME 33

Metallocycloaddition of 77 onto the endocyclic olefin of 79 generates metallocyclobutane 81, which upon retro-cycloaddition provides alkylidene 82. Typically, the norbornyl olefin is more reactive than less strained acyclic ones, and in this case the fact that the other double bond is conjugated with the ketone is expected to make it less reactive still. A second intramolecular metallocycloaddition onto the enone olefin provides metallocyclobutane 83, which again undergoes retro-cycloaddition to provide the targeted enone 80. The yield of 80 was better than tripled upon adding ethylene. As suggested earlier, an excess of ethylene present in the reaction allows for the recycling of unproductive intermediates (those leading to polymers and other byproducts) back into the desired pathway leading to enone 80. Scheme 34 represents one possible unproductive pathway (amongst n u m e r o u s others) that can lead to an alternative product. To retard polymer production, dimeric or oligomeric materials can also be reverted back to monomeric alkylidene intermediates capable of entering the pathway depicted in Scheme 33. This complex equilibrium of

12

449

THE TOTAL SYNTHESIS OF AMPHIBIAN ALKALOIDS

u/•Me

LnR

_ -

LnRU~...[~... Me

Me

- LnR~'~. undesired

1LC=H4 %~/Me

LnRu/~ [~.... Me

[Ru] %[~....f Me _ LnRu~ . ~ ] /

~o

~o 82

~~.O

Me LnRu ~./"-~O

80 desired

83

SCHEME 34

intermediates ultimately proceeds toward what is likely the most thermodynamically stable component, enone 80. The successful incorporation of a tandem RO/RCM in place of an ozonolysis/aldol route proved to increase the efficiency and ease of bicyclic enone synthesis tremendously. In only three steps, 41 was converted to 80 in c a . 80% overall yield (compared to seven steps and 27% overall yield using the ozonolysis/aldol strategy). Utilizing this sequence, multigram quantities of 80 were produced with relative ease. As developed in the model study, it was envisioned that a one-pot conjugate addition/aldol reaction would serve to install the e x o methyl group and ultimately the azido side chain. Advancing toward the key Schmidt precursor, the enone 80 was reacted with MezCuLi to afford the intermediate enolate 84, which was quenched with the aldehyde 70 to provide the dehydrated aldol adduct 85 (Scheme 35). The aldol adduct was isolated as a single olefinic isomer in 65% yield. The geometry of the newly formed olefin was investigated using 2D NMR techniques. The use of COSY, HMQC, and HMBC allowed for the assignment of values to all the protons and carbons of 85. With a NOESY experiment, it was found that an

450

KEVIN J. FRANKOWSKI, AARON WROBLESKI, AND JEFFREY AUBt~

NOE f-h

Me

.... ,,Me2cu.i O

Me

,o

'%

Li

Me

H Me

65% (from 80)

....

0

80

84

Me

85

SCHEME 35

NOE existed between the newly installed exo methyl group and the allylic enone protons (Scheme 35). For these groups of protons to be within NOE proximity requires that the exocyclic enone olefin have an E configuration. Apparently, the additional steric interactions encountered by placing the allylic methylene group close to the C-9 methyl group on the ring were overcome by the electronic benefit of placing the alkyl chain anti to the cyclopentyl ketone. Following the protocol developed previously, 85 was converted to the azido ketone Schmidt precursor 86 (Scheme 36). The multipurpose Na/NH 3 reduction converted 85 to the corresponding reduced alcohol (not shown) as an approximately 4:1 mixture of inseparable diastereomers epimeric at the ~-position. The major isomer was presumed to be that placing the side chain in an endo orientation. A subsequent Mitsunobu reaction 39 transformed the primary alcohol to the corresponding azide 86, also as an inseparable 4:1 mixture of diastereomers (50% yield, two Me

H Me

0

1. Na/NH3 2. Zn(N3)2~ pyr, Me Me DEAD'PPh3 = N3~,,, ~ ....% 50% (2 steps) Me 0

85

TfOH or

TiCl4 ....% Me

86

Me

0

87 SCHEME 36

Me

86 plus decomposed products

12

451

THE TOTAL SYNTHESIS OF AMPHIBIANALKALOIDS

steps). With the necessary functional groups in place, the key intramolecular Schmidt reaction was then examined. Treating the azidoketone 86 with either TfOH or TiC14 resulted only in re-isolation of starting material or apparent decomposition of the olefin functionality. None of the desired cyclized lactam product 87 could be detected. This was a momentary cause for concern, but this particular problem was overcome in fairly short order. We hypothesized that the olefin was providing another site for protonation and possibly even interaction with the azide in the acidic medium; such chemistry has been extensively explored by the group of William Pearson and coworkers at the University of Michigan. 5~ Because the olefin was serving as a masked hydroxymethylene equivalent, Aaron decided to convert this compound to its oxidized analog before re-attempting the Schmidt reaction. Initial experiments focused on a one-pot conversion of the olefin to the primary alcohol using an ozonolysis/borohyride reduction sequence (Scheme 37). This method yielded the desired alcohol 88; however,

M ~ ~ k Me

\

/ Me

1.0 3

Me ....

0

....

O

/

2"NaBH4" ~

Me

Me

Me 87

Me

86

Me

1.0 3 2. excess NaBH4

Me

Me

Me

....

0

....

HO

Me

86

90 SCHEME 37

Me

\ OH

OH

452

KEVIN J. FRANKOWSKI, AARON WROBLESKI, AND JEFFREY AUBt~

incomplete reduction afforded mixtures of the desired alcohol 88 and the ozonide 89. Even more alarming was that the alcohol 88 was only isolated in 20-30% yield. When adding a larger excess of reducing reagent (NaBH4), over-reduction took place to give the diol 90 (stereochemistry neither shown nor determined). An alternative route involved carrying out the above sequence in a two-pot protocol. Because of difficulties in trying to reduce the ozonide 89 in acceptable yields, direct conversion of 86 to the corresponding aldehyde 91 was investigated (Scheme 38). Toward that end, olefin 86 was ozonized and treated with DMS to afford aldehyde 91 in c a . 70% yield. At this juncture a few experiments were conducted on 91 to determine whether this substrate was a viable Schmidt substrate. Treating the azido ketoaldehyde 91 with TfOH resulted in the detection of a lactam product (as ascertained by ~H and ~3C NMR) among a variety of unidentified impurities. Accordingly, attention was then redirected toward carrying out the Schmidt reaction on the alcohol 88. Caution was exercised with this reduction, though, as once 91 was produced, three functional groups capable of undergoing reduction were present. Reduction of either the azide or ketone eliminates one of the functional groups necessary for the Schmidt reaction. Fortunately, addition of 1 equiv, of NaBH 4 chemoselectively reduced the aldehyde in the presence of both the ketone and the azide to cleanly afford the alcohol 88 in 50-55% yield over two steps. Conversion of the 4:1 diastereomeric mixture of 86 to the alcohol 88 allowed for facile chromatographic separation of the diastereomers at this stage.

Me

Me

....

1.0 3

=

M~~k, ' M e

~' O

N3

Me

86

1. NaBH4 2. separation Me of isomers

91

Me \

D,

50- 55% (2 steps)

O O

0

Me 88

SCHEME 38

OH

Me

12

453

THE TOTAL SYNTHESIS OF AMPHIBIAN ALKALOIDS

The key intramolecular Schmidt reaction was now tried on the alcohol 88. In this case, treatment of 88 with excess TfOH proceeded smoothly and yielded the ring-expanded Schmidt product 92 in 79% yield (Scheme 39). X-ray analysis of the crystalline lactam 92, which was to our knowledge the first carried out on any analog of 251F, nicely confirmed the tricyclic structure and orientation of all seven stereogenic centers in this penultimate intermediate. Me

Me ,||1

Me '||'k

N3

~

OH 0

79%

Me

"|'

0

88

\ OH

Me Me

92 Me

LAH iiii k

86-100~176 Me

OH Me

251F (2) SCHEME 39

As expected, LAH reduction generated alkaloid 251F 2 in 86-100% yield. Synthetic 251F was identical in all respects to published spectra (see Table 1 for comparison of 13C values). 24,26 X-ray crystallography and 2D NMR analysis are immensely powerful methods for determining molecular structure and stereochemistry as has been demonstrated throughout this chapter. Far less prevalent, though, is the use of IR spectroscopy to help determine the stereochemistry of natural products. Figure 9 provides the IR spectra of both the lactam 92 and alkaloid 251F. Excluding the lack of a carbonyl absorption for 251F (as expected after LAH reduction), there is a subtle difference between the two IR spectra apparent in the C-H stretching region. Noticeable in the IR spectrum of 251F are two small bands at 2756 and 2800 cm-~; clearly lacking in the lactam spectrum, these are known as Bohlmann bands. 19 Such stretches are observed when a nitrogen lone pair is antiperiplanar to three C-H bonds as is the case in 251F. The nitrogen lone pair places electron density into the antibonding orbitals of the C-H bonds. Consequently, the C-H bonds assume a different stretching frequency,

454

KEVIN J. FRANKOWSKI, AARON WROBLESKI, AND JEFFREY AUBI~ TABLE 1 Experimentally determined and reported ~3C values for alkaloid 251F Entry

Experimental J3C NMR for 251F (125 MHz) (ppm)

Daly's reported 13C NMR for 251F (ppm)

14.7 15.6 17.5 27.1 28.3 30.2 31.4 35.3 37.6 41.2 44.3 47.3 53.1 61.3 64.8 67.3

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

14.8 15.7 17.5 27.1 28.4 30.2 31.4 35.4 37.7 41.3 44.4 47.4 53.2 61.4 64.8 67.4

ME

I 60

.... \

55

Me 0

Me

OH

92 aO00

;r162 wt~e-,u,~b~t O::m.t)

ID 84

Me

10

i"

~o

MeJ~~~N~

\ .... OH Me

251F (2) w

~

{r

FIGURE 9. The IR spectra showing the characteristic Bohlmann bands. Note the two small absorptions at 2800 and 2756 cm-1.

12

THE TOTAL SYNTHESIS OF AMPHIBIAN ALKALOIDS

455

and it is only when the combined effects of three C-H bonds vibrating at a new frequency occur that these bands are observed. Bohlmann bands were absent in the lactam because a carbonyl occupied the position where a third antiperiplanar C-H bond was needed. In fact, the observation of Bohlmann bands in the racemic 3-desmethyl 251F synthesis was the first piece of evidence obtained, indicating that the azido side chain occupied the desired e n d o position. It was not until later that sophisticated 2D NMR techniques and X-ray crystallography confirmed what the presence of Bohlmann bands had already indicated. VII. Final Remarks

Overall, the asymmetric total synthesis of 251F proceeded in 13 steps with an overall yield of ca. 5-8%. Approximately 100 mg of the natural product was produced in this way. In Daly's report on the isolation and structural elucidation of 251F, it was noted that 100 frog skins were required to isolate only ca. 300 ~g of the alkaloid (or ca. 3 lag per frog). 25 Accordingly, 100 mg of 251F corresponds to the amount of alkaloid that would be harvested from in excess of 30,000 frog skins! The potential of natural sources to provide fascinating new compounds remains unchallenged. However, it is clear that the thoughtful use of synthesis to provide rare compounds in quantity is not only valuable for feeding downstream biological applications, but also can play a role in conserving precious natural resources. The synthetic materials have been submitted to various programs to determine their pharmacological profiles, but the results of these studies will have to wait for future publications. One interesting element of this project arises from the fact that we made synthetic 251F in enantiomerically pure form. However, the absolute stereochemistry of the naturally occurring material is unknown. This leads to the possibility that our enantiomerically pure (-)-251F would correspond to the less biologically potent version of the molecule (assuming that 251F is biologically active in the first place). This point can only be settled through experiment, but the availability of r a c e m i c desmethyl-251F could be valuable for estimating the biological potential of the series enantiomeric to our (-)-251F preparation. In conclusion, a study ending in a synthesis of racemic 3-desmethyl 251F allowed us to develop chemistry that ultimately culminated in an asymmetric total synthesis of alkaloid 251F. Key to both syntheses was the use of an intramolecular Schmidt reaction on advanced intermediates, which delivered the core tricyclic skeleton of the natural product. In addition,

456

KEVIN J. FRANKOWSKI,AARON WROBLESKI,AND JEFFREY AUBt~

a second-generation approach toward an important bicyclic enone was developed that greatly streamlined the synthetic sequence toward 251F. At seven steps and 27% overall yield, the ozonolysis/aldol strategy toward enone 31 proved satisfactory. However, replacing this route with a domino metathesis strategy better than halved the number of steps (total three) and tripled the yield (80%) e n r o u t e to the analogous enone 80. And of course, these syntheses demonstrated the use of the intramolecular Schmidt reaction in the synthesis of increasingly complex alkaloids. This was an extremely enjoyable project from beginning to end. In retrospect, the Schmidt chemistry was not really the highlight from the perspective of the principal investigator, but rather the recognition that we could explore other chemistry, not so familiar to us, in ways that would really make the Schmidt steps shine. For example, I had been impressed with the Sternbach approach to quinane natural products for years, and was finally able to utilize this elegant scheme to procure a product that was important to our program. In addition, we were able to plug into the RO/RCM whirlwind that has become such an everyday part of fine organic synthesis. The synthesis of 251F was an important milestone in the group in that it demonstrated that the intramolecular Schmidt reaction would not necessarily be relegated to the synthesis of simple bicyclic compounds. Acknowledgments The three authors would like to thank, first and foremost, the other members of the laboratory who played decisive roles in the chemistry described in this document. They are, in chronological order, Greg Milligan, Pat Rafferty, and Kiran Sahasrabudhe. None of this work could have been done without their contributions. We have all benefited from a most stimulating work environment at the University of Kansas provided by other members of the group and would like to acknowledge all of the group members who contributed to this positive atmosphere. Other important technical contributions made by Douglas Powell (Xray) and David Vander Velde (NMR) are also acknowledged, and we are grateful to all of the scientists, named throughout the chapter, who shared their insights with us along with the way and provided useful suggestions. We are also grateful to Dr. John Daly, who provided both insights and valuable comments on our publications at various times throughout this project. Dr. Daly is to be congratulated for also providing the worldwide community of organic chemists with a lifetime of challenges and inspiration through his singular research program. Of course none of this work could have been accomplished without the generous support of the United States taxpayers as administered by the National Institutes of Health. We are specifically grateful to the National Institute of General Medical Sciences for its continuous support of our alkaloid synthesis program through GM-49093.

12

THE TOTALSYNTHESISOF AMPHIBIANALKALOIDS

457

References and Footnotes

1. Aub6, J., Chem. Soc. Rev. 1997, 26, 269. 2. (a) Aub6, J., Milligan, G. L., J. Am. Chem. Soc. 1991, 113, 8965. (b) Aub6, J., Milligan, G. L., J. Am. Chem. Soc. 1995, 117, 10449. 3. Daly, J. W., J. Nat. Prod. 1998, 61, 162. 4. Albuquerque, E. X., Daly, J. W., Witkop, B., Science 1971, 172, 995. 5. Daly, J. W., Proc. Natl. Acad. Sci. USA. 1995, 92, 9. 6. Daly, J. W., Spande, T. E, Garraffo, H. M., J. Nat. Prod. 2005, 68, 1556. 7. Daly, J. W., J. Med. Chem. 2003, 46, 445. 8. LaLonde, R. T., Muhammad, N., Wong, C. F., Sturiale, E. R., J. Org. Chem. 1980, 45, 3664. 9. Le Dr6au, M.-A., Desma6le, D., Dumas, E, d'Angelo, J., J. Org. Chem. 1993, 58, 2933. 10. (a) Smith, A. L., Williams, S. E, Holmes, A. B., Hughes, L. R., Lidert, Z., Swithenbank, C., J. Am. Chem. Soc. 1988, 110, 8696. (b) Gnecco, D., Marazano, C., Das, B. C., J. Chem. Soc., Chem. Commun. 1991, 625. (c) Holmes, A. B., Smith, A. L., Williams, S. E, Hughes, L. R., Lidert, Z., Sithenbank, C., J. Org. Chem. 1991, 56, 1393. (d) Shishido, Y., Kibayashi, C., J. Org. Chem. 1992, 57, 2876. 11. (a) Yamazaki, N., Kibayashi, C., Tetrahedron Lett. 1988, 29, 5767. (b) Nagao, Y., Dai, W.-M., Ochiai, M., Tsukagoshi, S., Fujita, E., J. Am. Chem. Soc. 1988, 110, 289. (c) Yamazaki, N., Kibayashi, C., J. Am. Chem. Soc. 1989, 111, 1396. (d) Nagao, Y., Dai, W.-M., Ochiai, M., Tsukagoshi, S., Fujita, E., J. Org. Chem. 1990, 55, 1148. (e) Polniaszek, R. P., Belmont, S. E., J. Org. Chem. 1991, 56, 4868. (f) Polniaszek, R. P., Belmont, S. E., J. Org. Chem. 1990, 55, 4688. (g) Taber, D. E, Deker, P. B., Silverberg, L. J., J. Org. Chem. 1992, 57, 5990. 12. Yao, L., Smith, B. T., Aub6, J., J. Org. Chem. 2004, 69, 1720. 13. Marx, J. N., Norman, L. R., J. Org. Chem. 1975, 40, 1602. 14. Mirrington, R. N., Schmalzl, K., J., J. Org. Chem. 1972, 37, 2871. 15. (a) Loibner, H., Zbiral, E., Helv. Chim. Acta 1977, 60, 417. (b) Chen, C.-P., Prasad, K., Repic, O., Tetrahedron Lett. 1991, 32, 7175. 16. Mossman, C. J., Aub6, J., Tetrahedron 1995, 52, 3403. 17. Lipshutz, B. H., Harvey, D.F., Synth. Commun. 1982, 12, 267. 18. Stevens, R. V., On the stereochemistry of nucleophilic additions to tetrahydropyridinium salts: A powerful heuristic principle for the stereorationale design of alkaloid syntheses. In: Strategies and Tactics in Organic Synthesis, vol. 1, Lindberg, T. (Ed.), Academic Press: Orlando, FL, 1984, pp. 275-298. 19. (a) Bohlmann, E, Angew. Chem. 1957, 69, 541. (b) Bohlmann, E, Chem. Ber. 1958, 91, 2157. (c) Gribble, G. W., Nelson, R. B., J. Org. Chem. 1973, 38, 2831. (d) Carson, J. R., Carmosin, R. J., Vaught, J. L., Gradocki, J. J., Costanzo, M. J., Raffo, R. B., Almond, H. R. J., J. Med. Chem. 1992, 35, 2855. (e) Pearson, W. H., Gallagher, B. M., Tetrahedron 1996, 52, 12039. (f) For a review, see: Crabb, T. A., Newton, R. E, Jackson, D., Chem. Rev. 1971, 71, 109. 20. Daly, J. W., Spande, T. E Amphibian alkaloids: Chemistry, pharmacology and biology. In: Alkaloids: Chemical and Biological Perspectives, vol. 4, Pelletier, S. W. (Ed.), Wiley: New York, 1986, pp. 1-274. 21. Aub6, J., Rafferty, P. S., Milligan, G. L., Heterocycles 1993, 35, 1141.

458

KEVIN J. FRANKOWSKI,AARONWROBLESKI,AND JEFFREYAUBt~

22. (a) Momose, T., Toyooka, N., J. Org. Chem. 1994, 59, 943. (b) Aehman, J., Somfai, E, Tetrahedron 1995, 51, 9747. (c) Jefford, C. W., Sienkiewicz, K., Thornton, S. R., Helv. Chim. Acta 1995, 78, 1511. (d) Michael, J. E, Gravestock, D., Synlett 1996, 981. (e) Bardou, A., Celerier, J.-E, Lhommet, G., Tetrahedron Lett. 1998, 39, 5189. (f) Michael, J. E, Gravestock, D., J. Chem. Soc., Perkin 1 2000, 1919. (g) Back, T. G., Nakajima, K., J. Org. Chem. 2000, 65, 4543. (h) Michel, P., Rassat, A., Daly, J. W., Spande, T. E, J. Org. Chem. 2000, 65, 8908. (i) Shu, C., Alcudia, A., Yin, J., Liebeskind, L. S., J. Am. Chem. Soc. 2001, 123, 12477. (j) Song, Y., Okamoto, S., Sato, E, Tetrahedron Lett. 2002, 43, 8635. (k) Ma, D., Pu, X., Wang, J., Tetrahedron: Asymmetry 2002, 13, 2257. (1) Davis, E, Yang, B., Org. Lett. 2003, 5, 5011. (m) Toyooka, N., Dejun, Z., Nemoto, H., Garraffo, H. M., Spande, T. E, Daly, J. W., Tetrahedron Lett. 2006, 47, 577. (n) Davis, E A., Yang, B., Deng, J., Zhang, J., ARKIVOC 2006, 120. 23. Myers, C. W., Daly, J. W., Am. Mus. Novit. 1980, 1. 24. Myers, C. W., Papdis Avulsos Zool. S~to Paolo 1987, 36, 301. 25. Spande, T. F., Garraffo, H. M., Yeh, H. J. C., Pu, Q.-L., Pannell, L. K., Daly, J. W., J. Nat. Prod. 1992, 55, 707. 26. Daly, J. W., Garraffo, H. M., Spande, T. E, Amphibian alkaloids. In: Alkaloids, vol. 43, Cordell, G. A. (Ed.), Academic Press: New York, 1993, pp. 185-288. 27. Taber, D. F., You, K. K., J. Am. Chem. Soc. 1995, 117, 5757. 28. Wender, E A., Ternansky R. J., Tetrahedon Lett. 1985, 26, 2625. 29. Gracias, V., Zeng, Y., Desai, E, Aub6, J., Org. Lett. 2003, 5, 4999. 30. Sternbach, D. D., Hughes, J. W., Burdi, D. E, Banks, B. A., J. Am. Chem. Soc. 1985, 107, 2149. 31. Sternbach, D. D., Ensinger, C. L., J. Org. Chem. 1990, 55, 2725. 32. Sternbach, D. D., A new strategy for the synthesis of polyquinanes. In: Strategies and Tactics in Organic Synthesis, vol. 2, Lindberg, T. (Ed.), Academic Press: San Diego, CA, 1989, pp. 415-438. 33. (a) Hoffmann, H. M. R., Vathke-Ernst, H., Chem. Ber. 1981, 114, 2208. (b) Parlar, H., Baumann, R., Korte, E, Z. Naturforsch. B: Anorg. Chem., Org. Chem. 1981, 36B, 898. 34. (a) Burke, S. D., Murtiashaw, C. W., Dike, M. S., J. Org. Chem. 1982, 47, 1349. (b) Burke, S. D., Murtiashaw, C. W., Saunders, J. O., Oplinger, J. A., Dike, M. S., J. Am. Chem. Soc. 1984, 106, 4558. 35. Tebbe, E N., Parshall, G. W., Reddy, G. S., J. Am. Chem. Soc. 1978, 100, 3611. 36. Petasis, N. A., Bzowej, E. I., J. Am. Chem. Soc. 1990, 112, 6392. 37. Dollinger, L. M., Ndakala, A. J., Hashemzadeh, M., Wang, G., Wang, Y., Martinez, I., Arcari, J. T., Galluzzo, D. J., Howell, A. R., J. Org. Chem. 1999, 64, 7074. 38. Wrobleski, A., Aub6, J., J. Org. Chem. 2001, 66, 886. 39. Viaud, M. C., Rollin, E, Synthesis 1990, 130. 40. (a) Vandewalle, M., Van der Eycken, J., Oppolzer, W., Vullioud, C., Tetrahedron 1986, 42, 4035. (b) Fukuzawa, S., Matsuzawa, H., Metoki, K., Synlett 2001, 709. (c) Krotz, A., Helmchen, G., Tetrahedron: Asymmetry 1990, 1,537. (d) Evans, D. A., Barnes, D. M., Johnson, J. S., Lectka, T., yon Matt, E, Miller, S. J., Murry, J. A., Norcross, N. D., Shaughnessy, E. A., Campos, K. R., J. Am. Chem. Soc. 1999, 121, 7582. (e) Evans, D. A., Chapman, K. T., Bisaha, J., J. Am. Chem. Soc. 1988, 110, 1238. (f) Evans, D. A., Miller, S. J., Lectka, T., J. Am. Chem. Soc. 1993, 115, 6460. (g) Evans, D. A., Miller, S. J., Lectka, T., yon Matt, E, J. Am. Chem. Soc. 1999, 121, 7559.

12

THE TOTALSYNTHESISOF AMPHIBIANALKALOIDS

459

41. (a)Evans, D. A., Miller, S. J., Lectka, T., J. Am. Chem. Soc. 1993, 115, 6460. (b) Desimoni, G., Faita, G., JCrgensen, K. A., Chem. Rev. 2006, 106, 3561. 42. (a) Evans, D. A., Chapman, K. T., Bisaha, J., J. Am. Chem. Soc. 1988, 110, 1238. (b) Poll, T., Abdel Hady, A. F., Karge, R., Linz, G., Weetman, J., Helmchen, G., Tetrahedron 1989, 30, 5595. 43. Evans, D. A., Bender, S. L., Morris, J., J. Am. Chem. Soc. 1988, 110, 2506. 44. (a) Stille, J. R., Grubbs, R. H., J. Am. Chem. Soc. 1986, 108, 855. (b) Stille, J. R., Santarsiero, B. D., Grubbs, R. H., J. Org. Chem. 1990, 55, 843. 45. For reviews on ruthenium metathesis catalysts see: (a) Sanford, M. S., Love, J. A., Grubbs, R. H., J. Am. Chem. Soc. 2001, 123, 6543. (b) Trnka, T. M., Grubbs, R. H., Acc. Chem. Res. 2001, 34, 18. 46. Selected examples include: (a) Zuercher, W. J., Hashimoto, M., Grubbs, R. H., J. Am. Chem. Soc. 1996, 118, 6634. (b) Stragies, R., Blechert, S., Synlett 1998, 169. (c) Arjona, O., Csaky, A. G., Medel, R., Plumet, J., J. Org. Chem. 2002, 67, 1380. (d) Hagiwara, H., Katsumi, T., Endou, S., Hoshi, T., Suzuki, T., Tetrahedron 2002, 58, 6651. (e) Weatherhead, G. S., Ford, J. G., Alexanian, E. J., Schrock, R. R., Hoveyda, A. H., J. Am. Chem. Soc. 2000, 122, 1828. (f) Zuercher, W. J., Scholl, M., Grubbs, R. H., J. Org. Chem. 1998, 63, 4291. (g) Lee, D., Stello, J. K., Schreiber, S. L., Org. Lett. 2000, 2, 709. 47. (a) Minger, T. L., Phillips, A. J., Tetrahedron Lett. 2002, 43, 5357. (b) Pfeiffer, M. W. B., Phillips, A. J., J. Am. Chem. Soc. 2005, 127, 5334. (c) Chandler, C. L., Phillips, A. J., Org. Lett. 2005, 7, 3493. (d) Hart, A. C., Phillips, A. J., J. Am. Chem. Soc. 2006, 128, 1094. (e) Phillips, A. J., Hart, A. C., Henderson, J. A., Tetrahedron Lett. 211116,47, 3743. 48. Evans, D. A., Britton, T. C., Ellman, J. A., Tetrahedron Lett. 1987, 28, 6141. 49. Nahm, S., Weinreb, S. M., Tetrahedron Lett. 1981, 22, 3815. 50. For recent examples see: (a) Pearson, W. H., Hutta, D. A., Fang, W.-K., J. Org. Chem. 2000, 65, 8326. (b) Pearson, W. H., Hines, J. V., J. Org. Chem. 2000, 65, 5785. (c) Schkeryantz, J. M., Pearson, W. H., Tetrahedron 1996, 52, 3107.

STRATEGIES AND TACTICS IN ORGANIC SYNTHESIS, VOL. 7 9 2008 Elsevier Ltd. All rights reserved.

Chapter 13 SEDUCED BY A SIREN'S CALL" EXPANDING APPLICATIONS FOR AROMATIC COMPOUNDS AND THE SYNTHESIS OF (+)-RISHIRILIDE B Thomas Pettus and Todd Wenderski Department of Chemistry and Biochemistry University of California at Santa Barbara Santa Barbara, CA 93106-9510, USA

Burgeoning Interest in Aromatic Compounds Why Choose a Particular Phenol Skeleton for Oxidation? II. Developing Easy Access to Various Resorcinol Systems III. Total Synthesis of (_+)-Epoxysorbicillinol IV. Plans for Chiral Auxiliaries Disintegrate V. Stereochemistry Describes an Unexpected Transition State VI. VII. Beating the Clock, Tuning the Reaction and Improving Yields VIII. Controlling B-Dione Tautomerization and Protection IX. A Lactic Acid-Derived Directing-Protecting Directing Group X. New Methods for Cleavage of the Directing-Protecting Group XI. The Total Synthesis of (+)-Rishirilide B XII. New Horizons for This Method and Oxidative Dearomatizations Acknowledgments References and Footnotes I~

460 464 466 471 472 473 474 476 478 482 483 484 488 488

I. Burgeoning Interest in Aromatic Compounds

It is often tough to tell when a story truly begins. I will start at Virginia Tech, where I was working in the laboratories of Professor Tom,is Hudlicky, (T'Hud). My organic chemistry college professor, Dr. Maurice Maxwell (Max), was visiting VPI for a summer research program and had brought me along. I was assigned to the bench of Gustavo Seoane, who served as my immediate mentor and handler for the summer of 1988. Gustavo's mission, which no doubt doubled his workload, was to prevent me from hurting myself, others, and taking out a wall while carrying out an experiment. T'Hud had recently recognized the importance of a Gibson paper reporting

13

461

SEDUCED BY A SIREN'S CALL

the formation of (+)-cis-2,3-dihydroxy- 1-methyl-cyclohexadiene from toluene by a mutant strain (39D) of the soil microbe Pseudomonas putida, and he and his group were just beginning to finish enantioselective syntheses of terpenes and prostanoids from arene diols (Scheme 1). 1 T'Hud had not yet been mellowed by the birth of his son and in those days he ran about the lab with a cigarette in hand whipping his students into line.

I

HO

Pseudomonas putida

succ,n,c

HO"

o

Gibson 1970

anhydride

S C H E M E 1. Gibson's observation with PP-39D. 2

T'Hud publicized a list of broken glassware along with its past owner. The idea was that the list, posted above the sink at the first bench found when entering the lab, would enable glassware to be efficiently reordered and returned to its original owner. In reality, the list indicated those with the worst technique and I seemed to be its most frequent contributor. After adding some rather expensive separatory funnels, I was banished to the microbiology lab. The microbiology lab at that time was an old broom closet in the bowels of Davidson laboratory. Here, surrounded by plastic petri dishes, I was assigned the task of generating the group's supply of arene diol from styrene (Scheme 2). The process began by growing agar plates of the soil microbes for about three days, followed by incubation of the microbe in the presence of the styrene, and subsequent feeding of the styrene in correct proportions with oxygen to a sugary aqueous suspension of microbes. The feeding apparatus had been cobbled together from plywood, Erlenmeyer flasks, two fish tank pumps, and two airflow control devices. By mid-afternoon in the sweltering heat of Blacksburg, Virginia, the small non-air-conditioned microbiology laboratory reeked of unconsumed styrene from the past twenty-four hours of force-feeding. The microbes, now dead, stewed in their own arene diol excrement. After centrifuging

Pseudomonas. putida

o.

A, oBz zeylena

HO,,

steps = O : : ~ / I ~ L ~/

HO"

IMDA

"i --" Ph

1989

S C H E M E 2. Total synthesis of (-)-zeylena from styrene?

462

THOMAS PETTUS AND TODD WENDERSKI

2 L of PP-broth, and a labor-intensive emulsion prone extraction, Mother Nature rewarded me with 500 mg of the corresponding arene diol. In all likelihood, this material would polymerize, forming a beautiful pink solution if not immediately purified. By the end of the summer, I had learned many things, but three stood out. First, I hated the smell of styrene. Second, unless you wanted it to be shredded, never show T'Hud a spectrum containing ethyl acetate. Thirdly, an aromatic ring could be an extremely useful six-carbon building block, but the bugs could provide only one enantiomer of it. By the end of the summer, I inquired about my graduate school options. T'Hud provided me with two choices: "Work for me or for Dick Schlessinger at UR." From what little that I can recall of my graduate school experience beyond the truth of the adage "out of the frying pan and into the fire," my fascination with the economy and elegance of aromatic intermediates continued to grow. While working in the Schlessinger group on the 4th floor of Hutchinson Hall, I discovered my talent for the reverse process of my work in T'Hud's group and converted many delicate chiral building materials into useless aromatic compounds (Scheme 3). Needless to say, Schlessinger shared his trademark expression of disgust and utter reprehension by flicking a lit cigarette into my hood in the midst of an extraction with ether. O Br

a) KHMDS c) dimethylfumarate

TMSO CO2Me BC, " ~~~

OH Br~CO2Me H~

/r,. OMe

"SOMe

SCHEME 3. Destruction of chirality. 4

While collecting references in Carlson Library for an upcoming NIH proposal, I encountered some beautiful and inspirational syntheses starting from aromatic building blocks that employed oxidative dearomatization for the introduction of chirality. These syntheses included Delongchamp's ryanodol, 5 Danishefsky's calicheamicinone, 6 and Corey's ovalicin (Figure 1).7 All had chosen to begin with an aromatic compound and arrived at an ornate albeit racemic natural product. With Schlessinger's NIH grant on asymmetric Diels-Alder reactions of chiral amino furans renewed, I escaped the University of Rochester. I married my bench mate, Dr. Liping Hong, whose earlier defection from the

13

463

SEDUCEDBY A SIREN'S CALL

NO .~ ~ Me Me "~Me~~

OH I OH Me

Me

O

/~NHCO2Me /O Me Me H O(,/ " HO ~ H .~. I ?H!~O .~lMe

~o o. L-..,,NH (+)-ryanodol

MeSSS (-)-calicheamicinone

ovalicin

FIGURE 1. Structures built by oxidative dearomatization.

Schlessinger group had often served as a point of inspiration. Liping completed her Ph.D. with Rick Borch at University of Rochester and began work with Art Schultz at Rensselaer Polytechnic Institute. Professor Schultz coincidently was among Schlessinger's first Ph.D. students and his research focused on the reductive counterpart to oxidative dearomatization. In 1996, I moved to NYC and began work in Havemeyer Hall at Columbia University as an NSF fellow on a terpene called tricycloillicinone. Shocked by my survival of graduate school, and thankful that my new advisor did not smoke, so that I might now quit, I busied myself making many aromatic compounds with a Nicoderm CQ | patch attached to each of my arms. The initial [5+2] dearomatization strategy failed. Instead, we devised a radical cyclization (Scheme 4). However, neither Danishefsky, group members, nor I could imagine a straightforward nonracemic entry leading to the key o-quinol building block.

OTBDPS 50~ O O

OTBDPS ~

O

1) Snider cyclization

~

-

O~oo~O 2)carbonylerasure " ~

(_+)-tricycloilicinone1998

SCHEME 4. Total synthesis of tricycloillicinone.

After these diverse experiences with aromatic compounds, most would have wisely chosen to avoid them further. Sure, it may seem that aromatic materials offer attractive starting points for synthesis, but in reality the yields for most aromatic transformations are exceedingly low, certainly not as high as those found for most aldol and polyketide chemistry, pervasive throughout the chemical literature these days. Moreover, despite a hundred years of research, the control of regiochemistry during aromatic

464

THOMAS PETTUS AND TODD WENDERSKI

substitution remains difficult, and distinguishing between the phenols of a poly-hydroxylated aromatic compound is still problematic. Although metalmediated reactions have overcome some of the problems, the issue of regiocontrol is often pushed upstream in a strategy requiting selective access to a halogenated aromatic compound. Therefore, chemists still resort to chemically beating an aromatic compound to afford a mixture containing some of the desired substitution pattern and leave the rest up to a mechanical separation. Like a moth to the flame, I was seduced by the malleability embedded within the aromatic skeleton, despite intimate knowledge of these problems, and decided to find some new procedures for transforming them into useful building blocks. II. Why Choose a Particular Phenol Skeleton for Oxidation?

No general method for enantioselective oxidative dearomatization existed, but I was not alone in the pursuit of a procedure. From careful inspection of past literature, 8 four intractable obstacles seemed to have barred the development of enantioselective oxidative dearomatization processes from offering genuine utility for asymmetric synthesis. First, oxidative dearomatizations proceed in low yield and, as all synthetic chemists know, a low yield dooms a synthesis to inefficiency and anonymity. Therefore, the few notable examples of oxidative dearomatization in synthesis have applied the transformation at a very early stage when starting material was abundant or near the end when the target could be reached without much further exploration. Many have claimed improvements for oxidative dearomatization involving additives and/or solvent effects. However, upon critical examination and comparison, it appears that only kinetically favored intramolecular oxidative dearomatizations, such as the Wesley oxidation, consistently provide good yields (>65%); the rest is mostly rubbish. Secondly, oxidative dearomatization results in an oxygenated derivative, be it a p-quinol, o-quinol or an arene diol. Therefore, subsequent steps are required to protect the acidic hydroxy residue before most subsequent reactions. Since the protection step had to be followed at some later point by a deprotection step, the efficiency of the method is again lowered. Exceptions to this problem are application of the method at the conclusion of a synthesis or inclusion of some useful functionality such as the epoxide generated by the Wesley oxidation. Thirdly, oxidative dearomatization usually leads to derivatives of p-quinol, o-quinol, and arene diols that are very unstable and susceptible to re-aromatization. This problem is usually nucleus dependent. For example,

465

13 SEDUCEDBYA SIREN'SCALL

o-quinols are susceptible to dimerization, whereas p-quinols succumb to single electron transfer reduction and rearomatization; arene diols undergo elimination and/or reoxidation. None of these problems can be avoided altogether. However, electronic and structural features can be introduced to dissuade undesired reactivity. Fourthly, control of regiochemistry and stereochemistry in the subsequent reactions of the products afforded by oxidative dearomatization often proves challenging. The problem stems from the nature of the product, which as a rule contains at least four or five trigonal carbon atoms. Without many tetrahedral atoms or other steric control elements, the planar products undergo reactions from both faces of their rt-systems; the stereocenter usually exerts a sizable influence only on vicinal tfigonal atoms. Given these four problems, most organic chemists have wisely avoided the enchanting siren's call of the functionality embedded within aromatic starting materials. Those who have pressed-on have cobbled together a series of fixes to sidestep these impediments. For example, one of the first methods aimed at chemically producing non-racemic building blocks from aromatic compounds carried out the finicky dearomatization reaction early, thereby producing a less precious achiral starting material (Scheme 5). This material is then trans-ketalized with a chiral diol. In principle, several diastereomers could arise. However, positioning the larger substituent in the equatorial position of the newly created 1,3-dioxane is favored. Unfortunately, chirality is far removed from subsequent reaction sites. Thus, the chemoselectivity and stereoselectivity of subsequent reactions suffer. Nevertheless, the method has been applied in the synthesis of miroestrol, 9 LL-C10037R, ~~and (-)-jesterone. ~

R~~

OMe

oxidize,, yield R2 10-65%

OH O RI'"~.

O R2

R OH

MeO OMe R ~ O achiral

transketalization =

R2

partially diastereoselective reaction, some / chemoselectivity

-._1,2-addition of R3deketalization

RZ~.. R2 O

RI~~o R2 0

SCHEME5. Trans-ketalizationof a quinonemonoketalwith a chiral diol.

466

THOMAS PETFUS AND TODD WENDERSKI

In a related strategy, the phenol is subjected to dearomatization affording a symmetric p-quinol. The non-precious symmetric system is then subjected to a desymmetrization (Scheme 6). The newly added functionality is then incorporated into the final target. This method has been used in the synthesis of manumycin A. ~2

OH ~

0 0 oxidation10_40% ~ ~1~. Desymmetrization ~J~.b"pO=

R3

R3 "OP achiral

chiral

SCHEME 6. Desymmetrization after oxidation.

In a similar approach, the achiral p-quinol is temporarily desymmetrized. After some further selective reactions, the blocking group, which also serves to direct the stereochemical outcome of succeeding reactions, is then removed. This modification has been employed in the synthesis of diepoxin. J3 IH.

Developing Easy Access to Various Resorcinoi Systems

While some advances have been reported, it was evident to us that much work remained in order to develop a versatile method that would provide genuine applicability to a diverse range of molecules. Our initial plan aimed at addressing this problem seemed simple enough (Scheme 7)--construct a 4-alklyated resorcinol (C) with differentially protected phenols and then submit a mono-deprotected variant of C to chemical oxidation. In principle, the structure that emerged (A or B) O

~,~.. / oxidation P" . R A,P"= -H ~ OP" P"= -H

OP"

P'~ c

B, P'=-H SCHEME 7. o- and p-quinols via selective protection.

13

467

SEDUCED BY A SIREN'S CALL

should reflect which phenol oxygen atom was unprotected at the time of oxidation as well as the atom ortho or para to the unprotected phenol displaying the greatest cationic coefficient in the intermediate. After reviewing the chemical literature, we chose to target p-quinols (A), with the notion that this platform would be more easily adaptable to enantioselective processes because of the proximity of the - O P ' residue, as well as the perceived controllable and divergent reactivity of A. Our next step was to secure a route to differentially protected 4-alkyl resorcinols such as C. This notion proved to be harder in practice than we had ever imagined (Scheme 8). All attempts to cause the more accessible phenol in 1 to undergo a selective protection failed. The reactions produced a mixture of 2, 3, and 4 that could not be readily separated. In hindsight, this problem should have been anticipated. First, there is a solubility issue. The mono-protected resorcinols 2 and 3 are vastly more soluble than the starting resorcinol 1. Secondly, the A-values that are normally associated with steric size and often used to predict the outcome of a reaction apply to sp 3 systems. In an aryl ring, the steric effects are only felt in two planes. Careful inspection of Protecting Groups by Greene and Wuts reveals only one example of selective protection of 4-t-butyl resorcinol with pivaloyl chloride and a bulky amine. ~5 It is clearly a loaded example.

~ HO 1

OH

OP"

largeprotecting ~ group oragent amine HO transfer TBDPSCi-IPr2NEt , orPivCI,i-Pr2NEt 2

~

OH

P'"O

3

OP""

~ P"'O 4

SCHEME 8. Selective protection proves impossible. 14

Obviously, a better entry into the resorcinol starting material was needed for us to continue. 4-Bromo-resorcinol was commercially available, and for a time we considered metal-mediated coupling reactions. However, 4-bromo-resorcinol has rarely been utilized in a palladium mediated coupling because of the electron rich nature of the C-Br bond due to donation from the ortho and para C-O residues. Therefore, we concluded that development of new chemistry was necessary to address differentially protected 4-alkyl resorcinols in an efficient manner. Based on McLoughlin's observations regarding the sodium borohydride reductions of o-O-acetyl ketones, j6 we invented some simple procedures

468

THOMAS PETTUS AND TODD WENDERSKI

that could afford a vast range of differentially protected 4-alkylated resorcinols. McLoughlin speculated that the corresponding reduction involves an o-quinone methide intermediate. Our amended sequence proceeds from an o-O-acylated 2,4-dihydroxybenzaldehyde 5 to the elaborated material 7 in one or two-pot processes (Scheme 9). 17 The latter involves addition of a Grignard reagent to compound 6. The o-OBoc group was selected after considerable experimentation and best enables use of a wide range of alkyl magnesium and lithium reagents. OBoc

OBoc

BocO 5

1 pot

H R2

0/~"-- H

BHs.DMS~

R2MgB/ / in Et20 OBoc

BocO 6

R2= H Me etc R1= H Me Br

HO j

SCHEME 9. New access to 4-alkylated resorcinols via o-QMs.

With easy access to a range of 4-alkylated resorcinols, we next examined oxidative conditions intended to afford a p-quinol derivative in some protected form. In 1997, Pelter reported that exposure of a phenol to [bis(trifluoroacetoxy)iodo]benzene (PIFA) resulted in an ortho cyclization with the attached alcohol becoming embedded within the product (Scheme 10). At the time, it was the only example of an ortho cyclization. All of the other reports of intramolecular oxidative dearomatizations involved an ipso cyclization, in which the nucleophile attacked the aryl carbon atom to which it was attached. Although the order of events could not be known with certainty, ortho cyclization seemed assured, given the gem-dimethyl and cis olefin residues. OH

O

PIFA MeOH

~O

1997 Me

SCHEME 10. First example of ortho oxidative cyclization. ~s

13

469

SEDUCED BY A SIREN'S CALL

Pelter's loaded example inspired our first idea: oxidation of the carbamate 8. Nontoxic hypervalent iodide reagents have become the reagents of choice for oxidative dearomatization. In principle, exposure of 8 to PIFA or [bis(acetoxy)iodo]benzene (PIDA) would lead to a cation intermediate and a cyclization would ensue, whereupon addition of water to the iminium intermediate should provide the desired carbonate 9 (Scheme 11). PIFA is believed to involve cation intermediates, whereas PIDA tends to involve single electron transfer. Despite a Herculean effort on the part of then graduate student Ryan Van De Water, the reaction failed to produce any of the carbonate 9. Instead, paltry amounts of products were isolated, which indicated addition of a ligand from the oxidant to the ethylated carbon of the starting phenol.

OH

0

0 .

8

.

.

.

o, '

Et.N0_...0 ~ Et

.

0

l~t

o L.

Et-N,Et

0 9

S C H E M E 11. First idea fai|s. 19

Dreiding models suggested that this outcome reflects the failure of the carbonyl to adopt an alignment necessary for cyclization. Therefore, the chain was extended by one carbon atom and the corresponding amide was next examined. To produce this material, the differentially protected resorcinol 10 was coupled with the ~-chloroamide 11 and the remaining OBoc residue was cleaved to afford phenol 12 (Scheme 12).2oAn extraordinary number of conditions were examined for the oxidation of 12, which included an examination of additives, the use of solid supported oxidants, as well as phenol silylation. The yields of products, particularly the desired product, were meager. Thallium nitrate afforded 14 and 18, arising from ortho and para nitro addition to the intermediate cation. Koser's reagent [PhI+OH-OTs] gave the tosylate 15, corresponding to ortho addition. Lead tetraacetate gave a mixture of the bis-ortho acetate 16 and the para acetate 17. The latter product was also produced as the major product with PIDA. Based on the notion that introduction of water halfway through the reaction might lower the energy barrier by quenching the iminium intermediate, we explored the use of 2-iodoxybenzoic acid (IBX) as an oxidant, because it would generate an equivalent of water as Iv

470

THOMAS PE'I'TUS AND TODD WENDERSKI

N 11 1.

OH

O _--

oxidation;H20 O

OH

OH

~

~OTs

PO" "~

PO" ""~

14

15

El

Et

O

PO

O

O

.,,~OAc PO" "I>" 16

Ac

Et

20

21

O CF3 "~ 0

PO

,

18

O

~ "

O

~l~O PO

17

O

PO

~~O PO

Et

O

~/~ ~ PO

N.-0

19

O

OHn-lI ,,~'~

Fs PO

PO H

CF 3

22

Et O OF3

23

Et

24

SCHEME 12. The oxidation proves capricious.

was reduced to I III. However, instead of the desired product 13, Van De Water isolated the o-quinone 20. Derek Magdziak and Andy Rodriguez would subsequently develop this discovery into a general method for the synthesis of o-quinones and catechols from substituted phenols. 2~ Yaodong Huang and Dr. Jinsong Zhang later applied this discovery in the synthesis of ( +)-brazilin. 22 The best outcome was obtained using PIFA. However, the reaction gave highly variable amounts of the lactone 13, along with the biphenyl 24 and the trifluoroacetate 19. Addition of PIFA to 12 in tetrahydrofuran (THF) afforded adduct 21 in a 72% yield and showed participation by THE When PIFA was used in hexafluoroisopropanol or acetonitrile, similar solvent adducts 22 and 23 formed in respective yields of 11 and 20%. The latter required an aqueous work-up to saponify a trifluoroacetoxy intermediate. Of the conditions surveyed, a 42% yield of 13 proved to be the best and it occurred only once from addition of I. 1 equivalents of PIFA to 0.1 M solution of the 12 in CH3NO 2 at 0 ~ However, inconsistent amounts of compounds 19, 23, and 24 and 13 emerged from these conditions. In spite of the poor performance of the reaction, the adduct 13 had displayed great synthetic promise by partaking in a myriad of stereoselective and chemoselective reactions (Scheme 13).

13

471

SEDUCEDBY A SIREN'S CALL

0

Ph.. OH

II

,.-,~"./C02Me

__~

O~v~~ 25101 ~0~ I I 0~'---/' [J

t,....6~ u

27

dia?en~Ytll. ~L,~~ ~N~MgBr KHMDS/ ~ 50%~ O /70%

HR/~PPc~31CoVh ~ f~ / , jl Ij ~ 0/ ~ 72% l: ",w-~

C[O2Me

26

O

OMe

Danishefsk3fs diene JJ..... 87% BORSM= O" ~ v [-..0 "r"- ]~

/ ~ 13 X MeOH ~ O ~ ~ NBS Jl. .~ Rh(PPh3)3CI/ 63%~ 0 II~ "1 " H2/CH2CI2'2h .... N /JJ~ 0 _ ~' ~ 61~ Me~.~ 0

"0

28

o

0 SCHEME 13. Versatility of glycolic derived cyclohexa-2,5-dienone adducts.

IV. Total Synthesis of (+_)-Epoxysorbicillinol

Therefore, we focused this emerging methodology toward epoxysorbicillino136. 23 Sorbicillin 31 was prepared in three pots and smoothly transformed into 32 by a Mitsunobu coupling. One of the phenols was engaged in hydrogen bonding with the sorbate carbonyl and could not participate (Scheme 14). However, the remaining phenol partook in oxidation when exposed to PIFA. The desired cyclization proceeded in a surprisingly good 65% yield and was followed by an unexpected diastereoselective epoxidation when two equivalents of PIFA were used. The product was unequivocally established by X-ray analysis. The masked carbonyI was freed by two different procedures. As suggested by Kirby's treaty describing the effects of secondary orbital interaction upon rates of SN2 reaction, 24 the lactone 33 was first opened to afford the amide 37 by Weinreb's conditions. This material undergoes a rapid reaction with tin tetrachloride and affords epoxysorbicillinol 36, whereas 33 does not. In addition, 33 smoothly opens with cesium hydroxide and the resulting carboxylate 34 undergoes lactonization to afford 35 upon treatment with acid. Another saponification affords 36. A manuscript describing this six-pot synthesis was submitted in October 2000 to Angew. Chem. It was rejected and a twelve-step racemic synthesis submitted in November appeared in J. Am. Chem. Soc. in February. 25 Ours was published in Org. Lett. in January 2001. 26 C'est la vie.

472

THOMAS PETTUS AND TODD WENDERSKI

/

oJ coo "o0

I

II

_

?-!e~

> 90% yield -

"

]

"

OH sorbicillin (31) pots

inthree

I

9

g.,:o

H ]I r../

~o.,.~o

' "-" over three steps epoxysorbicillinol(36) one pot

33

o

CsO. / L H=O \ ' ~ /

L "~ -'-

0 ~ ' " T :-~ H4 . . . . / 0

"Phi=O"

32 ~

0.~ 0 0

-78 ~

/0 [" ...~ ~ 0 I~" ">

\

0%

~ ~ ' , , , o: ~176 .~o oJ ~ ;

0 .....

'''~

o..;

0 :

o~Oo ~ 35

"~OGs

34 / //~/"

~~oc.,

I

/

AIMe2

\~ ~-~

"T 11

o\ o-<,

~"-1

r_~

v__/

~

~.~

.O..)U"-O 'b 37

SCHEME 14. Total synthesis of (+)-epoxysorbicillinol in six pots.

The formation of the epoxide 34 was not intended. It forms as the diastereoselective product of a second oxidation. The first oxidation affords the desired 2,5-cyclohexadienone adduct. However, in the presence of a second equivalent PIFA, which exists in equilibrium with [PhI+-O-], an epoxidation proceeds with the electron deficient double bond. Kevin McQuaid would later establish the range of electrophiles participating in this transformation, 27 which resembles enone epoxidation with a basic solution of t-butyl hydroperoxide. The most significant observation from this synthetic foray was the surprisingly good yield for the oxidative dearomatization. The sorbicillin derivative 32 should have been more difficult to oxidize than 12. Yet, the reaction proved to be more efficient than its simplified resorcinol counterpart 12. The reason for the discrepancy would soon emerge. V. Plans for Chiral Auxiliaries Disintegrate

Because the amine residue was jettisoned during the course of the oxidation, we considered if an amine could be employed as a chiral auxiliary

13

473

SEDUCED BY A SIREN'S CALL

to induce enantioselectivity during the lactonization-dearomatization. We examined trans-2,5-dimethyl-pyrrolidine derivatives such as 41, Ender's proline derivatives 39, as well as others 37-38 (Scheme 15). For this study, an account penned by Kurth, which described iodolactonization of chiral amides, proved quite informative. 2s Of the assorted chiral amides (38-44) tested, all failed to provide good enantioselectivity in product 13 (<60% ee). Moreover, the yields for these ornate pyrrolidine derivatives were substantially lower (<30%) than the pyrrolidine derivative 12 (Scheme 12). 0 low ee's lower yields

Phl(OCOCF3)2; H20 PO

"

o

38-44

0

3

o. 38

H 40

39

42

43

\ .... 41

44

S C H E M E 15. All chiral auxiliaries tail.

VI. Stereochemistry Describes an Unexpected Transition State

While utterly useless from a synthetic perspective, the absolute stereochemistry of the preferential product suggested that our initial assumptions regarding the transition state were wrong. Based upon well-accepted steric arguments, the favored course for the transition of a non-C2 symmetric pyrrolidine amide into the 6-1actone 13 should proceed through a preferential rotamer placing the bulky R-substituent of the pyrrolidine proximal to the carbonyl, as opposed to the more encumbered methylene. Both of the transition states proposed in Figure 2 display this wellestablished hypothesis. Moreover, this understanding leads to the conclusion that amides derived from S-proline should favor reaction from the Re

474

THOMAS PETI'US AND TODD WENDERSKI

S

SR/2 ax-Boat, box-like favored dipoles cancel

O-

+

ax-Chair, step-like disfavored dipoles additive

0

FIGURE 2. The absolute stereochemistry illuminates favored transition state.

face of their respective carbonyl. Extrapolation of this reasoning led us to conclude that proline derivatives would undergo reaction from the a-face (bottom face) as depicted. However, we had failed to appreciate that the six-membered transition state was comprised of five sp 2 atoms. As such, the boat arrangement experiences little, if any, destabilizing steric interactions as compared with the corresponding chair. Moreover, the boat transition state benefits from a minimization of dipoles and secondary orbital overlap between the iminium and vinylogous ester residues. It turns out that the boat transition state is favored and best explains the absolute stereochemistry. V I I . Beating the Clock, Tuning the Reaction and Improving Yields

With poor enantiomeric excess and insufficient yields in the case of hydrido derivatives (R ~ = - H ) , our vision for a general chemical method for enantioselective oxidative dearomatization was fading. However, the cyclohexadienone adduct had proven versatile. Therefore, we targeted a new natural product. In 2001, Danishefsky reported the first racemic synthesis of rishirilide B (45) (Scheme 16).29 The strategy employed a precious dimethide intermediate with a racemic quinoid. We had found that Danishefsky's better known diene 51 undergoes a diastereoselective reaction with the less congested face of the electron-deficient enone displayed within our cyclohexa-2,5-dienone to afford the cis-decalin 52. Therefore, the union of Danishefsky's dimethide 47 with our racemic quinoid would have yielded the undesired regioisomer 48. To solve this problem, we began considering and examining other precursors for the anthracyclic ring in 45. The isobenzofuran 49 was expected to mirror the regiochemical outcome of 47 and produce 50. However, the directing effect of the remote -OP residue might be overridden by the use of the isobenzofuran enolate 53. Before undertaking the synthesis of 53,

13

475

SEDUCED BY A SIREN'S CALL

OTBS

0

O

OH H OTBSO ,OTBS

OH

OMe

O

R2 " ~ o

~""

..... BS

49

OP

O

pots

T

52

""

OP

~O

~ RI

O[

Li 53

OH 0

_9P........ O~ . . . . . . . . . . .

p2 ~

_ o

"0

"~'o.. J) 570M e

46

~)

~so~ o

~

/

o ;Y I

I

OU

54

~o55

o

\

,1 ?"

I rishidlideB (45)

SCHEME 16. A versatile racemic platform.

however, we investigated the union of the Hauser isobenzofuran enolate 55 with 46 and isolated the anthraquinone 56, instead of the desired product. While the dearomatization process could have been repeated, this notion was rather unappealing. Instead, Dr. Junhua Wang prepared the dimethide precursor 57, a compound that was inspired by the work of Diirst-Charlton. 3~When heated to 110 ~ compound 57 combined with cyclohexadienone 46 in a selective Diels-Alder reaction and yielded the chiral anthraquinoid 58 after treatment of the crude adduct with trace acid (>60% yield). In view of this success, we felt that a continued search for optimum oxidative conditions was warranted. Unfortunately, the reaction had proven too capricious and hampered data analysis in many cases. We therefore devised compound 59, which undergoes a competing reaction. In this manner, the ratio between the desired and competing processes could be used to more easily study the effect of incremental changes on the system (Scheme 17).

476

THOMAS PETTUS AND TODD WENDERSKI

1. coupling O OBoc RI-,,,~ H O ~ ' h

CI,,

N" R3

A R2

2. deprotection 3. oxidation

O

B O

R1 O

R40

+

n( 59

R1

Ph O

amide derivatives: O~ n" 1 > 2 >>>>>> 0 /4"..... oxidant: PIFA - PhlO+TMSOTf >>>> PIDA A/B ratio optimized solvent: CH3NO2 > CH2CI2 >> any other solvent additives showing little effect: propylene oxide, K2003, phenol silylation, Bu4N+BF4-, fluorinated solvent

SCHEME 17. Rapid optimization by measuring A/B ratio.

The identity of the R ~ substituent, the length of the tether (n), as well as the shape and electronics of the amine contributing to the nucleophilicity of the amide were all examined in combination with different oxidants, solvents, additives, and other tricks that would supposedly improve the yield. Many of the results were expected. For example, electron-rich amines with little steric encumbrance afforded a greater A/B ratio than amides derived from electron-poor amines or amides derived from congested amines. The most surprising result, however, was the effect of the R 1 substituent. The presence of this residue had a far greater influence on the success of cyclization than any other criterion. For example, oxidations of electron poor oxazolidinone amides afforded B exclusively for R 1 = -H, but gave an almost 3:1 A/B ratio when R ~ = - B r or -Me. This result most likely reflects fewer degrees of freedom in the transition state, which is caused by reorientation of the amide residue into an arrangement prone to the desired cyclization. VIII. Controlling B-Dione Tautomerization and Protection

Meanwhile, Dr. Wang's progression toward rishirilide B had stalled. Our plan entailed a very risky gambit of controlling the formation of a particular B-dione tautomer in order to manage the subsequent addition of an anion to one to the carbonyl residues. Our appreciation of this problem had been dimmed by our recent involvement with epoxysorbicillinol (36) - a tautomeric natural product that favors the tautomer displaying hydrogen bonding. Therefore, we naively expected 61, which was the hydrogenbonded and fully conjugated tautomer, to predominate. We further suspected

477

13 SEDUCEDBY A SIREN'S CALL

that its predominance would be reflected in the subsequent reactivity of this mixture (Scheme 18). _

~~

equilibrium solvent dependent OH O 1~

O

OMe J 60

1LiOH then HCI

i

then LiOH

OO~ I

~

_ R1 OH

HCI viathen ~'-Iactone

OMe.... ,r O-H

O

i

61 CH2N

i NCOCI

O

\O

62

O Et2N O/

OMe~I

R1 ...

H 63

0

'

SCHEME 18. Controllingtautomer formation and protection.31

Over time, the mixture of pseudo-acids 61-62 equilibrates to structure 61 (for R ~= - M e and-Br). However, the reaction of the predominating tautomer 61 with diazomethane affords a 1:1 mixture of the vinylogous esters 63 and 65. After considerable experimentation, Dr. Wang discovered that the carbamate vinylogous ester 64 could be formed selectively. We attribute this discovery to the notion that the less stable carbamate corresponding to acylation of 62 reverts to the more stable product 64 under the reaction conditions. Next, Dr. Wang found that addition of an excess of lithiated ethyl vinyl ether to 64 exclusively produced the desired diastereomer 66 upon work-up (Scheme 19). This stereochemical outcome had been expected, because we knew that the unprotected hydroxyl residue would direct the 1,2-addition to the adjacent carbonyl while the pseudoaxial isopentyl residue would control the stereochemistry of the methyl residue. Unfortunately, Dr. Wang was unable to cause the enol ether 66 to smoothly convert into the ethyl ester 67. The best yield was obtained using RuO 4 under biphasic conditions and it afforded a 30% yield of the desired ethyl ester 67 along with a considerable amount of product corresponding to oxidation of the electron rich naphthalene. A bis-~-hydroxy ketone, which was easily produced by epoxidation and ring opening of the enol ether 66, returned the tautomeric mixture of 61 and 62 on treatment with

478

THOMAS PETrUS AND TODD WENDERSKI

o

Et2N'JJ'O

OEt

0

~'~Li ..

O

69%yield 64

RuO4

.....

th e n NH4CI

" OH~o~E

\

.

"'-...-~

t

less than

30%yield

~ OMe

66

87

DMAP Et2NCOCl

o

....OH

/~OHo~E: --~

m-CPBA

tautomeric mixture of

pseduoacids61+62

HIO 4

OH0 Rq,,,%O H bis-*x-hydroxyketone R2

SCHEME 19. Stereoselective formation of the stereotriad.

periodic acid. This outcome was quite surprising, given the considerable number of reports involving hydroxycortisone, which had shown that cleavage occurred between the primary hydroxyl and hydrated carbonyl residues to give the corresponding ~-hydroxy aldehyde. 32 While the ethyl ester 67 may seem to have still been a viable intermediate, from the earlier Danishefsky account of failed saponification, we knew that we had reached an impasse. Therefore, we set the project aside and tackled the asymmetry. IX. A L a c t i c A c i d - D e r i v e d D i r e c t i n g - P r o t e c t i n g D i r e c t i n g

Group

The realization of the importance of the R ~ substituent (R~4: -H) had reinvigorated the project. Lupe Mejorado began to systematically examine chiral auxiliaries (Scheme 20). Though the yields had improved, the enantiomeric excess remained low. Only 60% ee was obtained with 68 and those for chiral oxazolidone derivative 69 hovered around 20%. Of some note was the surprising stability of the oxazolidinone-derived intermediate 70. This compound proved stable to work-up and chromatography. Upon stirring it in an ethereal solution over a slightly basic aqueous layer, the desired product 46 formed. Given the disappointment afforded by our auxiliary plan, we next investigated the effects of substituents among the intervening atoms of the protecting group. Since the glycolic protecting group would be jeuisoned, a stereocenter on the intervening carbon atom might prove useful, provided its source was inexpensive and the unit was easy to install and remove. We considered a variety of non-racemic a-hydroxy acid derivatives,

13

479

SEDUCED BY A SIREN'S CALL

0

poRI~R Phl(OeOCF3) H20"= 2;~ OH

R 1 = Me-, B r -

2 68-69

R1

2

- chiral amine e.r.

2:1-4:1

0

46 0

o

RI

.

,Pr ....

69

0

~ i OC(O:~)CF~ o.~R

70

SCHEME 20. Auxiliaries fail for glycolic derivatives.

which were attached to 71 by a stereochemical inversion using the Mitsunobu process. The stereointegrity of these couplings was carefully checked against the corresponding racemic standard. Mandelic derivatives were ruled out because of their propensity to undergo oxidation rather than inversion under Mitsunobu conditions. We setfled upon simple lactic derivatives and further suspected that the corresponding amides would prove more selective, because of the pseudo allylic strain incurred during the progression into products. Indeed, the ester derivative 72 afforded a diastereomeric ratio of 3:1 (Scheme 21). On the other hand, the pyrrolidine derivative 73 proceeded to the lactone 77 in good yield and diastereoselectivity. 33Only a trace amount of the undesired diastereomer was evident in the 400 MHZ ~H-NMR of crude product and it was easily separated by chromatography. The methoxyamine 74 afforded a slightly improved yield of 77. It is interesting to note that 75, which was prepared from the ester 72, could have afforded the quinamine corresponding to 77. However, the oxidation instead produced the oxime ester 76. The oxime ester 75 smoothly converted to the lactone under acidic conditions or upon treatment with Koser's hypervalent iodide reagent. Wenderski found that the enone within 77 smoothly reduced with potassium azodicarboxylate (PADA). The resulting product 81 underwent further hydrogenation and hydrogenolysis to afford the trans-fused decalin 79 (Scheme 22). The modification, which was borrowed from Corey's ovalicin synthesis, was a significant advancement for us. Earlier, we had employed a single step reduction with the glyco|ic derivative 13 using the Wilkinson catalyst. However, this and other metal-mediated reductions

480

THOMAS PETTUSAND TODDWENDERSKI OBoc

RI~

71 Me2AINHOMe

HO Mitsunobu i

I MeO.~ O

OH

R2

OH

~R2 ~ N ~ 72 o

~R2 I 73

OH

OH

oMe: oMe: O//

LR2

74

0

I

I

i~,,

good d.r.>l0:1 good yield > 65%

good d.r.>10:l better yields>70%

75

good d.r.>10:1 better yields>70%

l

0 R1= -Me, -Br

R ~

0 PhlOTsOH R2

o

.....~ 0 0

77

>70% yield

o .....i ~ 0 N

MeO"

a~ 76

SCHEME 21. A diastereoselective dearomatization with lactate derivatives.

proved to be more capricious for the corresponding lactic derivatives because of competing single electron transfer and rearomatization. Since the equatorial positioning of the [~-alkyl hydroxyl residue, compound 79 proved surprisingly robust toward most bases. However, upon saponification of the lactone with potassium trimethylsilanolate (KOTMS), the system underwent a ring flip and ~-elimination, producing the corresponding enone 80 on work-up. Another interesting stereoselective transformation, discovered by Hoarau for lactic derivatives, was the product afforded by the addition of primary amines. The lactone proceeded first to the corresponding secondary amide, and then underwent cyclization to afford the lactam 87, presumably via a cis addition of N-H across the olefin of the vinylogous ester. On the other hand, reduction of the cyclohexadienone with MAD/L-Selectride afforded the enone 86 as a single diastereomer. Further indications of significant differences between the chemical reactivity of lactic and glycolic backbones came upon prolonged exposure of 77 to tertiary amine bases or 1,8-diazabicyclo[5.4.0]undec-7-ene (DBU).

13 SEDUCEDBYASIREN'SCALL O

R~

~ ~ )

O

].,..~O,,,,/ R2

O KOTMS R2

O

78

O 77

Bu4NOH

ifR 1: -Br / //// or-C(O)R

J R3 OH ,., R11--I...~u

H 2 Pd/C if R1 =-Br

O R

cso ,-,

O

R2 OM O 81 \ pyrrolidine

O . O

R1

R1

O ~2 O

86

,i, ,

/

"""

MAD,"....~ ,1. O ~

O

H+

H20 O ~ O " Hd '"~. "..~R 2

I

L-selectride

80

O

~

HO ()/ 88 R2 R3NH2--J/

O

O

: DBU R ~

O

.

481

82

O

83

,~KOH SnCl 4

OH

O "~ N

85

84 SCHEME22. Lactic derivatives behave differently than glycolic

derivatives.

The equatorial methyl residue adjacent to the lactone carbonyl proved surprisingly easy to epimerize into an axial position, as shown by exclusive equilibration into adduct 78. At first, we suspected that perhaps the amine caused the lactone to open by an SN2' displacement and the system then re-closed to produce 78. However, after several experiments we attributed the complete epimerization to ~-deprotonation due to the presence of an attached electron-poor oxygen substituent and the release of torsional strain. When the R ~ substituent was electron-deficient and no proton source was available, a stereospecific electrocyclization occured during enolization to afford the x-lactone 88. Lactone 81 also underwent saponification and furnished the corresponding metal carboxylate 82. As opposed to an efficient lactonization of its glycolic counterpart, treatment of 82 with acid proceeded to the diastereomeric mixture of the x-lactones 83 in poor yield. Moreover, saponification of this mixture failed to produce a significant amount of the pseudoacid 84 and returned instead the metal carboxylate 82, which can afford the starting 8-1actone 81 during chromatography. While the lactone 81 could be opened with pyrrolidine to afford the amide 85, this lactic acid derivative failed to cleave with tin tetrachloride as opposed to its glycolic counterpart. Therefore, the chemistry prevented retention of

482

THOMAS PETTUS AND TODD WENDERSKI

the ketone oxidation state, making the non-racemic method unsuitable for both rishirilide B (45) and epoxysorbicillinol (36). X. New Methods for Cleavage of the Directing-Protecting Group

A solution to our cleavage problems was revealed while progressing toward 89, a structure reported by Cordell and to which anti-cancer activity had been attributed. 34At the time of isolation, the enone 89 was proposed as the biosynthetic precursor for 90 (Scheme 23), which suggested an unknown biological dienone rearrangement. Malik had reported another natural product, 91, which appeared to display a similar core ring system. 35 Our nonracemic method seemed applicable to either chiral enone. Hoarau built the core lactic derivative 92 and converted it into the lactol 93. All attempts to cause a [3-elimination resulted in an epimeric mixture of 89, protected as its corresponding silyl ether. We imagined that opening of the lactol and formation of the corresponding hydrazone would enable a ring-flip under milder conditions, and thus lead to the desired enone via [3-elimination. Surprisingly, the alkoxy residue 0~to the aldehyde functionality cleaved when forming the hydrazone and afforded the secondary alcohol 94. Hoarau converted this alcohol into its corresponding mesylate, and after elimination and deprotection isolated the target structure 89. However, its spectra did not match that reported for the natural product. On closer inspection, we determined that both 89 and the related structure 91 had both been misassigned. Our revision for 91 had been reported elsewhere as the natural product acremine A. 37 We requested that the authors, who had isolated acremine A, transfer their sample to the deuterated

O

0

OH ,,OH

-b-i~

Hd 89 reportedstructure 0 B r ~ O

R2

Ov ~

R2

OH

93

o

92

HO/I

90

0 steps ~ , , , O T B S

0 H2NNMe2" TMSC..~,,,OTBS HO" ~

91 reportedstructure steps

=

R2

He

94

S C H E M E 23. Accidental cleavage of the directing-protecting group. 36

89

13

483

SEDUCEDBY A SIREN'S CALL

solvent employed with 91. The ~H-NMR spectra agreed. The actual structure for the natural products misassigned as 89 and 91 is the corresponding regioisomer of the enone functionality. Thus, we had completed a rather sophisticated non-racemic synthesis of 89. XI. The Total Synthesis of (+)-Rishirilide B

Despite this setback, these cleavage conditions, in combination with a solution conjured for the endgame of rishirilide B (45), indicated that the synthesis was nearly complete. The remaining question was the choice of a protecting group for the naphthol. After methoxymethyl (MOM) and silyl ethers failed, we resorted to a benzyl residue and prepared a modified dimethide precursor. In this regard, the Comins method for regioselective alkylation of 3-alkoxybenzaldehydes again proved most useful in providing compound 96 (Scheme 24). 38 Subsequent application of the DfirstCharlton chemistry affords the benzylated compound 97. Upon heating, the dimethide was liberated and underwent a regioselective cycloaddition. The reaction occured at slightly elevated temperatures (150 v s . 110 ~ when compared to the earlier union between the methoxy and glycolic counterparts. At this elevated temperature, the initial adduct succumbed to immediate [3-elimination. Subsequent oxidation with DDQ afforded the chiral naphthalene derivative 99. The remainder of the synthesis passed without incident. Addition of the aluminum amide of dimethylhydrazine to 99 produced the 13-dione 100

.CHO

95 OBn

I Li/N~N

I/ ,

3 equiv PhLi then Mel

1. SO2, 450W Hg-hv ~ 2. M e O H , ~

~ _ . . . C HO

OBn 96 0

SO2

1. (-)-98,ZnO, 155 ~ 9h 2. DDQ (680/otwo steps)

OBn 97

OBn reo,,~ (-)-99 " ~ ~) Me

SCHEME 24. Rapid access to a viable intermediate.

484

THOMAS PETTUS AND TODD WENDERSKI

99 Me2AINH_NMe2

~o

",,

~

, - ~ Me 0

RI

n~

oA~"'~ =

Me

-

(o

HN ' N \

~

o ... M e

-..

0

100 Et2NC(O)CI HiJnig'sbase /

0

" OH OH

102

-78 ~ Li 1. EtO~-,,, 9m-CPBA

NEt2

0"'~"'0 ~ M e ,, "~

1) Na(OAc)3BH

0

i

9,'

0

101 i

0

1. NaOCI , ~ , , , . ' H 2. H2,Pd/C 90%~ L v~"O4Hi[o]/ -

2) Nal04on silica 75%overall

,,,

103 _~

SCHEME 25. Completion of the first synthesis of (+)-rishirilide B.

(Scheme 25). It should be noted that a direct cleavage from lactone derivatives proceeded only with the electron-deficient vinylogous ester functionality. However, the alkoxy cleavage process has been effective with all functional variations of the corresponding lactol derivatives, such as compound 93. The desired vinylogous ester 101 was obtained as before, by carbamylation of dione 100 and the remaining carbonyl in 101 underwent 1,2-addition with the lithium species shown to produce a vinyl ether on work-up. Treatment of the enol ether intermediate with peracid gave the bis-~-hydroxyketone 102. To circumvent the earlier cleavage problem, 39 the bis-~-hydroxy carbonyl was reduced with sodium tri-acetoxyborohydride, whereupon the least hindered diol succumbed to cleavage with periodic acid to produce the ~-hydroxy-aldehyde 103. Lindgren oxidation and debenzylation of the naphthol completed the first total synthesis of the (+)-rishirilide B (45); 40 the enantiomer that is found in nature. 41 XII. New Horizons for This Method and Oxidative Dearomatizations

When we began, our original goal was to devise a general enantioselective dearomatization technique that would prove useful for a variety of

13

485

SEDUCED BY A SIREN'S CAI.L

synthetic applications. Given our synthesis of (_+)-epoxysorbicillinol and (+)-rishirilide B, and our manipulation of these core ring systems, we would like to think that we have succeeded to some degree. We have fashioned a protected chiral cyclohexa-2,5-dienone adduct capable of serving in a myriad of subsequent chemoselective and stereoselective transformations. Schemes 26 and 27 should give readers a clear indication as to the breadth and scope of the chemistry available with these chiral adducts. A surprising interplay of reactivity has been revealed while examining reductions of the cyclohexadienone 92. As described earlier, reduction with diimide generated from PADA in CH3NO 2 proceeded as expected with the electron deficient enone to produce 117. On the other hand, 1,4-reduction of the vinylogous ester proceeded smoothly with MAD and L-Selectride | to give the ~-bromoketone 104, in which the bromine atom is positioned in an equatorial arrangement with respect to the six-membered cyclohexenone. These results were anticipated. However, the results of other reduction conditions have been quite surprising. O

1~

O

R3

? ~

Me2AIN-NMe2 71%

PADA CH3NO2 85% \ OH

:

~28~% ~''~

NaBH3CN ZnCl2, 80% O_H 3....,,,,~,,, Br

R3 a

0

124 OH

O.~

107~

R30"v")'" 106 OH .,,,,"~/"OH Me2N--NH2 _

--n5

O

~

",

O

i::o -. . . . .

N"

i~3 OH 123

NH2

R3

O

"" O.v ,,,j..... 122 (3H

.....

125 SCHEME 26.

O ri ] .'.,~.p,,,o,

..~

CHCI3, >95%

92 O

/NaBH4 CeCl~ Br MeOR 75%

O ' v " J .....

r

~

NaBH4 THF, H20 75% /

O \~O

105 O B~AL~,,,~ I)DI 2) MnO2 60% 2-steps

MAD

R\Xo O

Br

R30..~ .....

L-Selectride 75%

O /~~

118 ~ . . ~ B r

O

TMSCI Nal 90%

R30.~) ..... O

1 1 7 0 - ~ ) .....

O

Br

Reductions and subsequent removal of the chiral tether.

486

THOMAS PETTUS AND TODD WENDERSKI

9

O /~J'~,Br

OSiR3 Zn~ / ~ R3SiOTf

o

PADA CH3NO2 90%

'"o

111 O

R30..~ ..... 114 O DBU DMDO1 R3SiOTf / oOSi~R3

112~

O

O Pd/C,H2 >95%

o

R30,,~ ..... 104 O

I

"'9 = 2-J .....

o 1130 ~ O

'

o

_

Et2Zn,

/+

O

105

OSiR3

"?

.,,,.# R OH 115

116 '"O R30,~ ..... O

SCHEME 27. Enolate trapping, bicyclic and ring expanded products.

For example, exposure of 92 to diisobutylaluminum hydride (DIBAL) alone resulted in reduction of both the alkene of the bromo vinylogous ester and the carbonyl of the lactone to afford 122. While the double reduction of 92 with NaBH 4 in THF/H20, which afforded the allylic alcohol 108, may seem reasonable given the DIBAL result, the formation of the lactol 124 by Luche conditions is difficult to rationalize. Moreover, the reduction of 92 with NaBHsCN and ZnC12 to give the alcohol 125 seems stranger still. An explanation for this divergent reactivity begins to emerge upon examining the carbonyl frequencies in the infrared spectrum from compound 92. The ketone carbonyl appears at 1668 cm-~, whereas the lactone carbonyl resides at 1759 cm -~. The bromine atom in 104 cleaved on exposure to trimethylsilyl iodide (TMSI) to afford 105, whereupon both carbonyls were reduced upon treatment with excess DIBAL. Regioselective allylic oxidation with MnO 2 yielded the lactol 106. The directing-protecting group could be cleaved at this stage by treatment with dimethylhydrazine to produce the diol 107, or in the case of lactol 122, the anti-epoxy alcohol 123. Alternatively, as shown in our synthesis of (+)-rishirilide B, the lactic acid tether can be cleaved directly from lactone 117 upon treatment with

13

487

SEDUCEDBY A SIREN'S CALL

the corresponding aluminum amide to give the vinylogous acid 118. Compound 104 is an important branching point for many other transformations as well (Scheme 27). For example, it led to the silyl enol ether 112 as shown, as well as to its regiomeric enol counterpart 114. The latter resulted from the regioselective enolization of ketone 110, which occurred parallel to the trans-decalin-like ring junction. The Baeyer-Villiger oxidation of 110 also proved entirely regioselective and afforded the seven-membered lactone 116. In addition, these systems, while protected with the lactic acid group that had been used to direct stereochemistry, underwent a variety of other subsequent diastereoselective transformations such as the epoxidation of 114 that produced 113 and the cyclopropanation of 104 that afforded 109. However, in order to remove the directing-protecting group by a 13-elimination to produce 115, the lactone 110 must first be opened so that the underlying cyclohexanone can undergo a ring flip. While many future applications can be imagined, the simplest are syntheses of natural products displaying a chiral six-membered ring containing a tertiary alcohol- compounds that are not easily procured by other strategies, such as cleorindicin D, illudin J1, and oxysporidinone shown in Figure 3. However, one need not limit applications to ornate six-membered rings. Given the capacity to rupture the core ring system by Baeyer-Villiger and ozonolysis reactions, and the power to expand or contract the core cyclohexyl ring system, applications to acyclic arrays such as phoslactomycin and other fused cyclic natural products such as the trans-fused 7,5-ring system found in (-)-kessane are equally plausible. 0

0 oH

(+)-cleroindicin D

OH (-)-illudin J1

H3C (-)-kessane

HO 0 HO" \ 0

H2N

OH

i~

...~...)~..5~~..~J OH O~)_..,

oxysporidinone

FIGURE 3. Future targets.

phoslactomycin B

488

THOMAS PETTUS AND TODD WENDERSKI

Some could argue that a catalytic enantioselective chemical oxidative dearomatization process would be superior to that which we have developed. From serious contemplation of strategies leading to catalytic processes, we beg to differ. We speculate that as far as oxidative dearomatization processes are concerned, a catalytic process, if it were ever developed, would be fairly limited in scope and applicable to a small assortment of phenolic nuclei. The method that we have developed permits oxidative dearomatization of an assortment of resorcinols and affords a chiral p-quinol nucleus that is both protected and fortified as well as discriminant in its subsequent reactivity. When considering that new fermentation and separation processes have reduced the cost of R- and S-lactic acid directing group to less than $0.50 per pound as provided from renewable biosources such as corn, the issue of cost seems mute. However, others will likely succumb to the siren's call and with some luck may develop more efficient methods for unlocking the synthetic potential embedded within six carbons of simple aromatic compounds. Acknowledgments The development of this synthetic method and the completion of (+_)-epoxysorbicillinol and (+)-rishirilide B are all testaments to the well-known saying "any hard experience that one survives leads to greater strength and wisdom." This work is a testimony to the resiliency and ingenuity of many graduate and undergraduate students, as well as postdocs who nobly fought to make a round peg fit into a square hole. Postdocs Dr. Junhua Wang and Dr. Jinsong Zhang; and graduate students Ryan Van De Water, Lupe Mejorado, Christophe Hoarau, Todd Wenderski, Yaodong Huang, and Maurice Marsini; and undergraduate students Andy Rodriguez, Kevin McQuaid, Dave Freeman, and Simon Meek all contributed to aspects of this work. This narrative is as much their story as mine. However, I would like to bestow a special thanks to my many mentors including Joe Shelor, Maurice Maxwell, Tom,is Hudlicky, Dick Schlessinger, and Samuel Danishefsky and my dear wife Dr. Liping H. Pettus. These people played a large role in shaping my life. To all of these unusual individuals, and to other inspirational folks such as Peter Wipf, I pass on my heartfelt thanks and gratitude.

References and Footnotes 1. 2. 3. 4. 5.

Hudlicky, T., Luna, H., Barbieri, G., Kwar, L. D.,J. Am. Chem. Soc. 1998, 110, 4735. Gibson, D. T., Hensley, M., Yoshioka, H., Mabry, T. J. Biochemistry 1970, 9, 1626. Hudlicky, T., Seoane, G., Pettus, T. R. R., J. Org. Chem. 1989, 54, 4239. Schlessinger, R. H., Pettus, T. R. R., J. Org. Chem. 1994, 59, 3246. (a) Deslongchamps, P., Belanger, A., Berney, D. J. F., Borschberg, H. J., Brousseau, R., Doutheau, A., Durand, R., Katayama, H., Lapalme, R., Leture, D. M., Liao, C. C.,

13

6. 7. 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.

SEDUCEDBY A SIREN'S CALL

489

MacLachlan, E N., Maffrand, J. R, Marazza, R., Martino, R., Moreau, D., Ruest, L., Saint-Laurent, L., Santonge, R., Soucy, E, Can. J. Chem. 1990, 68, 127. (b) Deslongchamps, E, Belanger, A., Berney, D. J. E, Borschberg, H. J., Brousseau, R., Doutheau, A., Durand, R., Katayama, H., Lapalme, R., Leture, D. M., Liao, C. C., MacLachlan, E N., Maffrand, J. E, Marazza, R., Martino, R., Moreau, D., Ruest, L., Saint-Laurent, L., Santonge, R., Soucy, E, Can. J. Chem. 1990, 68, 115. Cabal, M. P., Coleman, R. S., Danishefsky, S. J., J. Am. Chem. Soc. 1990, 112, 3253. Corey, E. J., Dittami, J. P., J. Am. Chem. Soc. 1985, 107, 256. Magdziak, D., Meek, S. J., Pettus, T. R. R., Chem. Rev. 2004, 104, 1383. Corey, E. J., Wu, L. I., J. Am. Chem. Soc. 1993, 115, 9327. Wipf, P., Kim, Y., Jahn, H., Synthesis 1995, 1549. Hu, Y., Li, C., Kularni, B. A., Stobel, G., Lobkovsky, E., Torczynski, R. M., Porco, J. A., Jr., Org. Lett. 2001, 3, 1649. MacDonald, G., Alcaraz, L., Lewis, N. J., Taylor, R. J. K., Tetrahedron Lett. 1998, 39, 5433. Wipf, E, Jung, J. K., Org. Lett. 2000, 65, 6319. Unpublished results from these laboratories. Greene, T. W., Wuts, E G. M. Protecting Groups in Organic Synthesis, 3rd edition, John Wiley and Sons, Inc.: New York, NY, 1999, p. 265. McLoughlin, B. J., J. Chem. Soc. Chem. Commun. 1969, 540. Van De Water, R. W., Magdziak, D. J., Chau, J. N., Pettus, T. R. R. J. Am. Chem. Soc. 2000, 122, 6502. Pelter, A., Hussain, A., Smith, G., Ward, R. S., Tetrahedron 1997, 53, 3879. Unpublished results from these laboratories. Van De Water, R. W., Hoarau, C., Pettus, T. R. R., Tetrahedron Lett. 2003, 44, 5109. Magdziak, D., Rodriguez, A. A., Van De Water, R. W., Pettus, T. R. R., Org. Lett. 2002, 4, 285. Huang, Y., Zhang, J., Pettus, T. R. R., Org. Lett. 2005, 5, 5841. Sperry, S., Samuels, G. J., Crews, E, J. Org. Chem. 1998, 63, 10011. Kirby, A. J. Stereoelectronic Effects, 1st edition, Oxford University Press, Inc: New York, NY, 1996, p. 40. Wood, J. L., Thompson, B. D., Yusuff, N., Pflum, D. A., Matth~ius, M. S. E, J. Am. Chem. Soc. 2001, 123, 2097. Pettus, L. H., Van De Water, R. W., Pettus, T. R. R., Org. Lett. 2001, 3, 905. McQuaid, K., Pettus, T. R. R., Synlett 2004, 2403. Najdi, S., Reichlin, D., Kurth, M. J., J. Org. Chem. 1990, 55, 6241. (a) Allen, J. G., Danishefsky, S. J., J. Am. Chem. Soc. 2001, 123, 351. (b) Yamomoto, K., Hentemann, M. E, Allen, J. G., Danishefsky, S. J., Chem. Eur. J. 2003, 9, 3242. Durst, T., Kozma, E. C., Charlton, J. L., J. Org. Chem. 1985, 50, 4829. Wang, J., Pettus, T. R. R., Tetrahedron Lett. 2004, 45, 5859. (a) Kertesz, D. J., Marx, M. J., J. Org. Chem. 1986, 51, 2315. (b) Ashton, M. J., Lawerence, C., Karlsson, J. A., Stuttle, K. A. J., J. Med. Chem. 1996, 39, 4888. Mejorado, L., Hoarau, C., Pettus, T. R. R., Org. Lett. 2004, 6, 1535. David, J. M., Chfivez, J. E, Chai, H. B., Cordell, G. A., J. Natural Prod. 1998, 61,287. Aziz-ur-Rehman, Malik, A., Riaz, N., Nawaz, H. R., Ahmad, H., Nawaz, S. A., Choudhary, M. I., J. Nat. Prod. 2004, 67, 1450. Hoarau, C., Pettus, T. R. R., Org. Lett. 2006, 8, 2843.

490

THOMAS PETTUSAND TODDWENDERSKI

37. Assante, G., Dallavalle, S., Malpezzi, L., Nasini, G., Burruano, S., Torta, L., Tetrahedron 2005, 61, 7686. 38. Comins, D. L., Brown, J. D., J. Org. Chem. 1985, 50, 4829. 39. Mejorado, L., Pettus, T. R. R., Synthesis 2006, 3209. 40. Mejorado, L., Pettus, T. R. R., J. Am. Chem. Soc. 2006, 128, 15625. 41. Iwaki, H., Nakayama, Y., Takahashi, M., Uetsuki, S., Kido, M., Fukuyama, Y., J. Antibiot. 1984, 37, 1091.

INDEX

ab initio methods, 127 acetic anhydride(Ac20 ), 60, 106 acetonitrile, 470 acetylation, 60, 76, 81,315 acid-catalyzed cyclizations, 246 acremine A, 482 activation, 384 activation barrier, 131, 134 acyl anion equivalent, 248, 264 acylation, 52, 311 acyloin ring closure, 222 1,4-addition, 357 AD-mix-[3, 85, 244 Ag20, 94 aglaforbesin, 222, 228 Aglaia forbesii, 222 aglain, 220, 226 aglaroxin C, 219 AIBN, 192 Alder end() rule, 429 aldol, 8, 434, 436-437, 445, 449, 456 aldol addition, 16, 20 aldol condensation, 295, 233 aldol methodology, 297 aldol reaction, 388 alkenylsilane, 22, 25, 30 alkyl azides, 440 alkyl magnesium and lithium reagents, 468 4-alkylresorcinols, 467 allenes, 334 allenic amino acids, 338 allenic aminoester, 340

4-allenyl-2-oxazolin-5-one, 337 4-allenyl-5-oxazolone, 336 allenynes, 342, 351 allyl cuprate, 166, 195 allyl ether, 176 allyl magnesium bromide, 166, 196 allyl magnesium chloride, 158, 165 allylic chloride, 248 ~-allylic palladium complex, 78 allylic rearrangements, 260 allylic substitution reactions, 332 allyltrimethylsilane, 316 aluminum ate complex, 182 aluminum enolate, 301 aluminum reagents, 168 Amberlyst-15, 28 0~-amino acids, 268 3-aminopropionaldehyde, 299 (1R,2S)-2-amino-l,2-diphenylethanol, 270 ammonium hydroxide, 299 amphibian alkaloids, 408 amphilectane, 35 amphotericin B, 3 amphotericins, 1 anhydrous HC1, 47 anti conformation, 372 1,5-anti aldol addition, 30 anticancer, 220, 385 antifertility activity, 59 antifungal activity, 244 antifungal agents, 2 antimalarial, 244 antimicrobial, 244 antlproliferative and antileukemic activities, 220

491

492 1,5-anti stereoinduction, 8, 21, 23, 31 antitubercular, 385 an tiviral, 244 arene diols, 464 arene-arene interactions, 99 arginine, 146 aromatic C-H functionalization, 384 arsine ligands, 27 arylboronic acids, 332 3-aryl-3-hydroxy-l,2-indandione, 230 aspergillusfumigatus, 308 asymmetric Diels-Alder catalysis, 429 asymmetric dihydroxylation, 271 asymmetric photocycloaddition, 233 asymmetric synthesis, 268, 442, 444 asymmetric synthesis of spirotryprostatin B, 308 auxiliaries, 443 A-values, 467 axinellamine, 324 azatyrosine, 290 (2S)-2-azepanecarboxylicacid, 313 azides, 28 l 4-azidobutyraldehyde, 441 azidohydrin, 439 azido-Schmidt reaction, 412 azomethine ylide, 306--307, 310

Baeyer-Villiger, 73, 75 Baeyer-Villiger oxidation, 487 bands, 453 Barton deoxygenation, 297 Barton-McCombie protocol, 54, 192 base-catalyzed isomerizations, 43 base-induced fragmentation, 264 bathochromic shift, 116 batrachotoxin, 410 9-BBN-H, 73, 75, 87, 91 Beckmann or Schmidt chemistry, 409 Beckmann rearrangements, 412 benzo [b]cyclobutapyran-8-one, 221-222, 228, 230 benzo[b]oxepine, 220 benzo[k]fluoranthene, 116 benzocyclobutene, 107 benzoin oxime, 271 ( + )-p-benzoquinone, 385

INDEX benzoyl chloride, 338 benzyl bromide, 87 benzyl chloroformate, 270 benzyl ether, 76 benzylation, 68 3-benzyldimethylsilylpropynol, 31 benzyldimethylsilyl group, 31 3-benzyloxybenzyl bromide, 291 4-benzyloxybutanal, 437 4-benzylresorcinol, 376 benzylsilane, 27 betaine, 87 BF3 etherate, 262 BF3-Et~O, 44, 319, 357 BH.~, 90 BiBr.~, 200 BiBr3, Et~SiH, 201 [3.3.0]bicyclooctane, 423, 446 bicyclo[4.3.0] oxazinones, 306 bicyclomycin, 269 (S)-BINOL, 169, 233 biogenetic pathway, 220 biological membranes, 232 biomimetic conversion, 255 bioxalomycin <,2, 324 bipyridinium dicfition, 136 bipyridinium salt, 130, 134 Birch conditions, 64, 69, 76 Birch reduction, 41, 46-47, 76, 95,275, 287, 319 bis(acyl) azide, 38 bis (cyclope n tadienyl) zirconium dichloride, 68 bis(dibromomethyl)fluoranthene, 111 2,3-bis(dibromomethyl)naphthalene, 109 5,6-bis-methylene-2,3-benzonorbornene, 106 2,3-bis (4-m e thyle n e th iopyridyl) naphthalene, 378-379 bisoxazoline catalyst, 443 bisphosphoramide, 11, 20 boat conformation, 189 Boc20, 295 Bohlmann bands, 419, 440, 453, 455 borohydride reduction, 431,451,467 boron chelate, 22, 301 boron enolate, 20, 256, 295, 300 boron tribromide, 110 boron trifluoride, 317 boron trifluoride etherate, 398

INDEX 1,4-bpbd, 377 1,6-bpht, 377 ( + )-brazilin, 470 brevenal, 154 brevetoxins A and B, 154 bridgehead cation, 256 bromine, 65 4-bromomethyl-7-methoxycoumarin, 286 1-bromo-2-butyne, 341 1-bromo-3-methyl-2-butene, 60 4-bromo-resorcinol, 467 4-bromo-2-methyl-l-butene, 76 Bu~SnD, 305 Bu~SnH, 192 n-Bu4NI, 94 Bu4NOH, 434 1,4-butanediol, 86 n-BuLi, 90 t-BuOK, 42 t-BuOOH, 79 t-butyl bromoacetate, 289 t-butyl hydroperoxide, 472 t-butyldimethylsilyl ethers, 71 tert-butyl hydroperoxide, 62 i-tert-butyl methylenemalonate, 282 tert-butyl propionate, 69 tert-butyldimethylsilyl ethe, 87 tert-butyldimethylsilyl group(TBS), 86 tert-butyldiphenylsilyl group(TBDPS), 86 tert-butylhypochlorite, 264 (Z)-2-butenyltrichlorosilane, 13 butyllithium, 110, 444 3-butyn-2-ol, 339 [2.2.1]bicycloheptane, 446

C C-H activation/Cope rearrangement, 383 C-H activation, 387 C-H insertion, 390 camphorsulfonic acid, 253 capreomycin IB, 284, 302 (2S,3R)-capreomycidine, 302 carbenoids, 387-388 carbocyclization, 332 carbocyclization reactions, 342 carbomagnesiation, 71 carbonyl homologation, 315, 323

493 q-carboxyglutamic acid derivative, 282 carcerands, 100 (R)-( - )-carnitine, 317 (S)-(+)-carnitine, 317 cascade, 387 catalyst and substrate control, 443 catalytic hydrogenation, 270, 274, 281,284, 286, 290, 293, 297, 299, 305, 307, 309, 311-313, 323 catalytic titanium-based method for reduction, 320 catechols, 470 re-cation cyclization, 245, 253, 264 CBr 4, 61 cellular pathways, 329 cerium(III) chloride, 106 ceroplastols, 258 ( _+)-ceroplastol, 258 CF~CO2H, 69 CF~COzZnCH2I, 56 C-glycoside, 156 CHzBr, 208 CHiLi, 73 chair transition state, 189, 390 chair-like transition state, 177 chelation, 18 chelation-controlled allylation, 322 chemoselective epoxidation, 62 chemoselective hydrogenation, 94 chemoselectivity, 81,387 chiral acyl oxazolidinone, 8, 16 chiral auxiliaries, 232, 444-445, 475, 478 chiral benzamides, 39 chiral Bronsted acids, 234, 239 chiral glycinate, 269 chiral recognition, 122 5-chloro-l-iodopentane, 313 3-chloro-l-iodopropane, 289 CH-~ interactions, 102 ciguatera poison, 163 ciguatoxin, 154 cinchona alkaloids, 85 cinnamic amide, 220 CIPE, 252 cis-2,3-dihydroxy-l-methylcyclohexadiene, 461 C-ketoside, 173 Clz(PCy3) 2Ru=CHPh, 90 Claisen condensation, 388

494 Claisen rearrangement, 175, 260, 336, 338-340, 345, 388 Claisen rearrangement of amino-acid propargyl esters, 361 clavularanes, 244 cleorindicin D, 487 clips, 100, 102 CO 2, 319 cocrystallization, 373 Collins' reagent, 61, 73, 76 (-)-colombiasin A, 385, 387, 394 complex-induced proximity effect, 252 concave-convex topology, 102 conformation of the boron enolate, 21 1,4-conjugate addition, 20, 253, 435 conjugate addition of hydride, 82 conjugated enamine, 270 convergent synthesis, 441 Cope rearrangement, 387 copper (I) catalysis, 222 copper (II)acetylacetonate, 314 Corey-Fuchs addition reaction, 214 COSY, 439, 449 cotylenins, 258 (CpzTi[P(OEt)~]2), 206 Cram-chelate, 61 CrO 3 in pyridine, 76 CrO:~, 73, 90 cross-conjugated polyenes, 347 cross-conjugated tetraene, 346 cross-conjugated trienes, 346 cross-metathesis reaction, 400 crotylsilanes, 317 [ 18] crown-6, 107 crown conformations, 252 cryptophanes, 100 crystal engineering, 369, 381 Cu, 443 CuCI, 81 CuI, 71 cuprates, 158 Curtius rearrangement, 38 cyanohydrin acetonides, 3 4-cyanopyridine, 137-138 cyclases, 247 cyclic enol ether, 208, 212 cyclindrospermopsin, 282 [3 + 2] cycloaddition, 227 [4+ 2] cycloaddition, 332, 355, 394-395 [5+2] cycloaddition, 332, 397

INDEX [2 + 2]-cycloadditions, 83, 333 [3 + 2] cycloadduct, 230 cycloamphilectane, 35 cyclocarbonylation, 333 cyclodextrins, 100, 122 2,5-cyclohexadienone adduct, 472 cycloisomerization, 333-334, 342, 350, 352 cyclopenta [b] quinolizidines, 421 cyclopenta[b] tetrahydrobenzofuran, 220 cyclopenta [b] tetrahydrobenzofuran ring system, 219 cyclopenta[bc]benzopyran, 219-220 cyclopentadiene, 103, 443 cyclopentyl iodide, 285 cyclophane, 100, 138 cyclopropanation, 51,278, 390, 487 cyclopropane, 396 cyclopropyl amino acids, 278 cyclopropyl- oxazinones, 278 cyclorocaglamide, 219 [9.3.0] cyclotetradecane, 243, 248 [9.3.0] cyclotetradecane, 261 cystic fibrosis, 156

2D NMR, 439, 449, 453, 455 Danishefsky diene, 169-170 Darzens type reaction, 87 Davis oxaziridine, 261 DBU, 42, 480 DCC, 337-338, 350 DCNB, 126 DDQ, 25, 106-107, 116, 140, 483 dearomatization, 463 debenzylation, 95 decarboxylation, 274 dehydration of amino acids, 349 [4] dendralene, 346 dendrimer, 135 dendrobates, 421 Dendrobatid alkaloids, 443 dendrobatid frogs, 411 dendron, 134, 136 7-deoxycylindrospermopsin, 282 deoxygenation, 53, 192 deoxyneodolabelline, 244 deprotection, 86 rl-deprotonation, 252

INDEX dermostatin, 3 desilylation, 403 Dess-Martin periodane, 44 Dess-Martin periodinane, 79 desulfonylation, 261,263 deuterium incorporation, 207 a-deuterioglycine, 277 D-glucal, 183 (2S,6S)-2,6-diamino-6(hydroxymethyl) pimelicacid, 297 6-diaminopimelicacid, 297 diammonium cinnamate salts, 372 diastereoselective aldol condensation reaction, 303 Diastereoselective [1,3J-dipolar cycloaddition of an azomethine ylide, 279 diastereoselectivity, 172, 252, 310, 479 diastereoselectivity of the aldol addition, 18 1,3-trans-diaxial interaction, 426 diazidomethane, 40, 442, 477 diazomethane-Pd (OAc) 2, 45 DIBAL-H, 16, 64, 68, 90, 180, 182, 194, 197, 200, 212, 255, 264, 486 dibenzyl malonate, 274 1,1-dibromoethane, 211 (+)-9,10-dibromocamphor, 253 dibromonaphthocyclobutene, 109 2,3-dichloro-4,5-dicyano-1, 4-benzoquinone, 106 2,3-dichloro-5,6-dicyan o- 1, 4-benzoquinone, 107 Diels-Alder reaction, 386, 397 Diels-Alder, 354, 431,443, 444, 446 Diels-Alder adduct, 107 Diels-Alder reaction, 36, 102-103, 106-107, 109, 116, 140, 233, 346, 348-349, 353, 360, 428, 475 diepoxin, 466 diethyl fumarate, 355-356 diethylaluminum chloride, 445 (diethylamino)-phenyloxosulfonium methylide, 278 diethylmethoxyborane, 22 dihydronaphthalene, 391-392, 394 2,4-dihydroxybenzaldehyde, 468 1,3-dihydroxybenzene, 373 (-)-3a,4 [3-dihydroxyclavulara-1 (15), 17-diene, 253 diimide, 255, 485 diisocyanoadociane, 35-36

495 diisopropylethylamine, 17 1,2-diketones, 262 dimerization, 370 dimethyl dioxirane, 90, 157, 171,175 dimethylacetylene dicarboxylate, 269, 355 dimethylaluminum chloride, 301 dimethylhydrazine, 483, 486 dimethylsulfonium methylide, 71, 79 2,2-dimethylvinyllithium, 71 dinaphthonorbornadiene, 102, 126 di n-butylammonium tetrafluoroborate, 119 1,3,2-dioxathiolane, 230 dioxirane, 212 dioxolane, 61 diphenyloxazinone template, 272 diphenyloxazinones, 268 dipolar cycloaddition, 309, 323 [ 1,3]-dipolar-cycloaddition, 306, 331 [3 + 2] dipolar cycloaddition, 239 dipolarophile, 308-309, 331 diquinane, 428 discovery libraries, 330 dissolving metal reduction, 274, 315-316, 437 dissolving-metal conditions, 282, 318-319, 322 distribution coefficient, 330 diterpene glycosides, 245 diterpenes, 35 diterpenoid oxepanes, 59 di-tert-butyl dicarbonate, 270 dithiane, 60, 201 diversity-oriented synthesis (DOS), 311,328, 330, 361 divinylketone, 263 DMAP, 28, 60, 95, 337-338, 353 DMDO, 157, 165, 172, 182, 184, 186, 194, 196-197, 200, 212, 215 DMS, 196, 452 DMSO, 73 DNA, 100 DNN, 126 DolabeUa californica, 243 dolabelladiene, 244 dolabelladienones, 264 dolabellane, 264, 243 3(E),7 (E)-dolabelladienone, 255 dolastane, 244, 246-247, 252, 264-265 domino metathesis, 456

496 DOS, 332, 337 [2+2] photodimerization, 376 double-bond isomerization, 402 Dowex 50W, 28 D-proline, 309 D/irst- Charlton chemistry, 483 dynamic equilibration, 129 dynamic protein-protein interactions, 328

(E)-cyclopropylphenylalanine, 279 ecteinascidin 295, 324, 743 EDCI, 95, 353 effect, 100, 234 [?~-elimination, 181,480, 482-483, 487 elisapterosin B, 385 (+)-elisabethadione, 385, 399, 401 (+)-elisabethamine, 385, 403 ( + )-elisabethatriene, 385 (-)-elisapterosin, 397 enantioselective chemical oxidative dearomatization, 488 enantioselective intermolecular C-H, 384 enantioselective oxidative dearomatization, 464 enantioselectivity, 15, 235, 237, 277, 473 endoperoxide, 264 ene-type cycloisomerizations, 332 enol ether, 4-5, 76, 158, 432, 446 enol ether formation, 208 enol phosphate, 76 enol silyl ether, 52 enolate alkylation, 388 enolization, 487 enzymatic resolution, 271 7-ep/-cylindrospermopsin, 282 9-epi-spirotryprostatinA,. 309 epimerization, 276, 307, 309 epoxidation, 44, 69, 176, 285, 472, 487 epoxide, 71, 79 (_+)-epoxysorbicillinol, 471,476, 482, 485 (+)-erogorgiaene, 385, 393-394 erythro-2-amino-l,2-diphenylethanol, 271 Eschenmoser fragmentation, 73 ESIPT, 222, 230, 233, 239

INDEX esterification, 106, 110, 200 Et3N, 86, 138 Et~SiH, 200 ethanediol, 52 3 ~ ether, 199 ethyl bromoacetate, 270, 281 ethyl diazoacetate, 92 ethyl oxindolylidene acetate, 309 ethyl vinyl ether, 90, 356, 477 ethyltriphenylphosphonium iodide, 87 EtMgBr, 41 EtOzCCH2P(O ) (OEt)2, LiHMDS, 94 EtSH, 200, 214 Eu (fod)3, 356 Evans aldol, 10 E-vinylglycine derivative, 275 exchange processes, 129 Excited State Intramolecular Proton Transfer, 222 6-exo-trig-process, 67 7-exo-tet-process, 60 ex0-5,6-bismethylenenorbornene derivatives, 102

face to face r~-rcinteractions, 126 face-to-face stacking, 378 facial selectivity, 310 Favorskii rearrangement, 416 filipin III, 3 Finkelstein reaction, 289 flavanoids, 230 fluorescence, 122, 223 Fmoc, 292 forbagline, 219,222,226 force-field calculations, 125 formamides, 17 (2R)-~-formylglycinedimethylacetal, 284 free-radical, 192 Friedel-Crafts acylation, 51 fullerene C60, 307 furan, 110 (+)-fusicoauritone, 258-259, 264 fusicoccane, 245, 247, 258, 264-265 fusicoccin A, 245 fusicoplagins, 258

497

INDEX

galactosyl and glucosyl iodides, 282 Gambierdiscus toxicus, 154 gambierol, 154 gauche conformations, 372 geraniol, 423 geranylgeranyl pyrophosphate, 245 germacrenes, 243 Gibbs activation enthalpies, 119, 129, 131-132 glycal anhydrides, 157 Grieco's selenoxide introduction/ elimination procedure, 400 Grignard, 158 Grignard addition, 272, 431 Grignard reagent, 61, 65, 71, 76, 168, 183, 189 Grubbs catalyst, 90, 179, 192, 196, 211,400 Grubbs 2 catalyst, 190, 197, 400 Grubbs's ruthenium catalysts, 447

H H202, 90, 446 haloetherification, 255 HBr, 110 hemiacetal, 432 (_+)-hemibrevetoxin B, 154, 163 hetero-Diels Alder chemistry, 168 hexafluoroisopropanol, 470 hexamethyldisilazane, 280 HE 69 HF-pyridine, 194, 215 Hg(OCOCF3) 2, 90 HI, 90 hirsutene, 428 histrionicotoxins, 411 HMBC, 439, 449 HMDAP, 297 HMPA, 261 HMQC, 439, 449 HN(OMe)M, 95 HNs, 441 HOBt, 353 Hofmann rearrangement, 302 Horner-Wadsworth-Emmons (HWE) reaction, 5, 28, 90, 287, 423

hostguest complexes, 233 HS(CH2).~SH, 60 hydrazoic acid, 438 hydrazone, 482 rt-hydride elimination, 344-345 1,2-hydride migration, 179 1,2-hydride shifts, 245, 257 hydroazulene, 250 hydroboration, 71,179, 182, 200, 316 hydrogen bonding, 99, 102, 237, 239, 369, 372, 379, 471. hydrogen peroxide, 22 hydrogenation, 270-271,274, 279, 281-282, 285, 292-293, 299, 302, 305, 309-310, 316-318, 321,323, 479 hydrogen-bond acceptor, 379 hydrogen-bond donor, 357, 361,379 hydrogen-bonded complexes, 374 hydrogen-bonded dimer, 378 hydrogen-bond-mediated organic synthesis, 368 hydrogenolysis, 270, 289, 292, 295, 319, 323, 479 hydrolysis, 319, 356 hydrolytic kinetic resolution, 31 hydrophenanthrene, 40 hydrophobic, 100, 149 [3-hydroxy-rl-amino acid statine, 315 3-hydroxyflavone, 219, 222-223 [3-hydroxyketone, 22 hydroxylated pyrrolidines, 314 hydroxyl-directed reduction, 234 6-hydroxymethyl-2, 6-diaminopimelicacid, 289 4-hydroxy-2-hydroxymethylpyrrolidine, 319 (2S,4R)-4-hydroxypipecolic acid, 312 (2S,5S)-5-hydroxylysine, 285 5-hydroxy-2-pentanone, 60 5-hydroxylysine, 285 hydroxyl-directed reduction, 227, 230 (2S,3S)-[3-hydroxyornithine, 299 hydroxyquinine, 303 13-hydroxysulfones, 252, 261

IBX, 469 ichthyotoxic, 244

498 illudin J 1,487 iminium, 469 iminium ion, 419 indolizidine 209B, 415 (-)-indolizidine 209B, 412 (-)-indolizidine 251F, 412 3-iodo-2-propenol, 26 a-iodophosphonium ylide, 87 a-iodo ylide, 87 infinite hydrogen-bonded polymers, 372 inflammatory, 385 inhibitor, 384 insecticidal, 219 rr interactions, 100, 471 intermolecular forces, 369 intermolecular interactions, 100 intramolecular [4+2] cycloaddition, 333 intramolecular alkylation, 255,264 intramolecular Horner-WadsworthEmmons reaction, 92 intramolecular hydrogen bonding, 359 intramolecular Michael addition, 38, 250 intramolecular ring closure, 318 intramolecular Schmidt reaction, 408, 425, 438, 453, 455 intramolecular solvation, 136 intramolecular vinylation of amides, 349 inverse electron demand hetero-Diels-Alder reactions of a,[3 unsaturated amides, 356 iodolactonization, 473 4-iodo-l-butene, 287, 297 3-iodo-2-propenol, 26 5-iodopent-l-yne, 68 ion pairing, 99 ion-exchange chromatography, 279 ionic radius, 125 zPr~NEt, 95 iridium catalyzed, 384 lsobenzofuran, 474 isobutyl chloroformate, 350 isobutyraldehyde, 306, 309 lsocycloamphilectane, 35 lsonitrile groups, 35-36 isoprene, 40 lsopropylmagnesium chloride, 13 isothermal titration calorimetric (ITC) measurements, 119, 144 (2R,3S)-isothreonine, 319

INDEX

Jacobsen's chiral Cr(III) adamantane catalyst, 169 jesterone, 465 Johnson ortho-ester Claisen rearrangement, 260 Johnson-Bartlett procedure, 54 Jones reagent, 67 jorumycin, 295, 323 Julia condensation, 249-250, 261,264 juncunol, 37

K252a, 226 I~2CO3, 42, 50, 65, 69, 76, 87 KzCO>K~Fe(CN)6, 83 I~2OsO~(OH) 4, 83 K3PO4, 91 kaempferol, 230, 232 kalihinane, 35 karenia, 154 KCtt~SOCH> 62 (-)-kessane, 487 ketal cyclization, 202 ketalization, 76 ketene acetal, 11 ketene actal, 187 ketenesilyl acetal, 303, 315 keto-enol isomerization, 106 keto-enol tautomers, 228-229 13-ketoesters, 56 <,-ketol (acyloin) rearrangement, 226 c~-ketol rearrangement, 222, 228 ~-ketol shift, 226-228, 234 ketone equivalent, 312 KF, 81 KH, 342 KHCO> 338, 340 KHMDS, 284, 319, 435 kinetic resolution, 6 KOH, 319, 434 Koser's reagent, 469 Kosower salt, 125 KOTMS, 26, 480

INDEX

L-(+)-diethyl tartrate, 79 lactam, 349, 451,453 lactams via cycloisomerization reaction, 349 lactol, 315 [3]-ladderane, 376-378 [5]-ladderane, 377-378 [n]-ladderanes, 378, 381 LAH, 314, 440, 445-446, 453 lanthanide Lewis acids, 356 latrunculin B, 311 LDA, 50, 64, 336, 340 lead tetraacetate, 279, 469 lemonomycin, 324 Lewis acid, 272, 357 Lewis acid mediated rearrangement, 44 Lewis base catalysis, 13 Lewis-base-catalyzed aldol addition, 31 LHMDS, 301 LizCuCl4, 65 LiA1H4, 41, 44, 68, 76, 87, 92, 94 Librium, 428 LiHMDS, 28, 435 Lindgren oxidation, 484 Li-NH3 reduction, 40 LiOH, 28, 54, 138, 253, 353, 357, 446 Lipinski rules, 329 5-1ithio-2,3-dihydrofuran, 69 lithium, 40, 44, 47, 275, 280 lithium ammonia reduction, 41 lithium enolate, 21 lithium ethanethiolate, 399 LL-CI0037R, 465 longitblene, 49 L-Selectride, 46, 480, 485 Luche conditions, 486 2,6-1utidine, 52, 81 lysine, 146

M

macrolactonization, 6 MAD, 480, 485 magnesium perchlorate, 311 magnetic anisotropy, 117 (2'-malonyl)phenylalanine derivative, 292

499 manganese(IV) oxide, 16 Mannich reaction, 300, 302, 388 manumycin A, 466 marine sponges, 35 mass spectroscopy, 135 m-CPBA, 65, 69, 71, 75, 90, 176, 285 Me2A1C1, 68 MezCuLi, 435-436, 449 M%S, 357 M%M, 168, 169 MeI, 338, 340 MEM ether, 255, 260 membrane lipids, 376 MeMgBr, 69 MeMgC1, 61 menthol, 232 (-)-menthyl 3,4-bis(dibromomethyl) benzoate ortho-quinodimethane, 112 mercury(II)-acetate, 17 mesylation, 68, 71, 318 metal-ammonia reduction, 39, 56 metal-assisted [2 + 2] cycloaddition, 222 (R)-2-methyl-1,4-butanediol, 260 metal-catalyzed carbocyclizations of allenes, 334 metallocycloaddition, 448 metallocyclobutane, 448 metalloenamine, 207 metathesis, 157, 196 metathesis reaction, 169, 445 methanephosphonic acid dichloride, 138 methoxymethyl chloride, 78 methoxymethyl ether, 92 methoxymethylene triphenylphosphorane, 45 1-methoxy-l-trifluoromethylphenylacetyl (MTPA) chloride, 21 2-methyl-1,4-butanediol, 260 3-methyl-2-butenyllithium, 67 2-methyl-6-methylene-(E)-2, 7-octadien-l-ol, 64 2-methylpent-4-enoic acid, 87 2-methylpropenylmagnesiumchloride, 184 methyl cinnamate, 227, 231 methyl cyanoformate, 41, 50 (+)-O-methyl-elisabethadione, 401-402 methyl iodide, 106 methyl lithium, 432 methyl magnesium bromide, 199

500 (4R,5 &- ( + )-4-me thyl-5-phenyl2-oxazolidinone, 444 methyl triphenylphosphanylideneacetate, 10 methyl (triphenylphosphoranylidene)acetate, 322 methylalumination, 67-68 methylation, 137 methyldimethylphosphonate, 92 methylmagnesium bromide, 13 1,2-methyl migrations, 256 methyltriphenylphosphonium iodide, 73 methyltriphenylphosphonium ylide, 90 MgClz, 196 micelles, 232 Michael reaction, 39, 41, 49, 52, 56, 220, 228, 282, 388 Micromonospora chalcea, 289 microsclerodermin H, 324 Minyobates bombetes, 421 miroestrol, 465 Mitsunobu reaction, 177, 255,260, 418, 438, 441,450, 471 mixed cuprate, 249 MnOz, 486 Mo(CO)6-mediated, 334 molecular, 149 molecular complexity, 333 molecular LEGO, 107 molecular recognition, 99, 122, 138 molecular scaffold, 349 molecular tweezers, 100, 102 molecular tweezers and clips, 112 molluscicidal, 244 MOM group, 44 Montanoa tomentosa, 59 montanol, 59 Morita-Baylis-Hillman reaction, 233 Mosher analysis, 170 Mosher ester, 21, 23 MPC-1001, 324 MTPA ester, 22-23 Mukaiyama aldol coupling, 207 Mukaiyama-type addition reaction, 187 mycoticin A, 3 mycoticins, 1 myrcene, 65

INDEX N

n PrSLi, 54 N,N-dimethylaniline, 138 N~CHCOzEt, 92 Na, 450 NaBH~CN, 486 NaBH 4, 452, 486 NaCIO z oxidation, 40 N-acyl iminium, 272 N-acyl sulfinamides, 232 N-acyloxazolidinones, 232 NAD +, 142, 145 Nail, 73, 92, 341-342 NaHMDS, 87, 287, 435, 444 NaI, 109 NaIO4, 62 nakadomarin A, 311,324 N-alkyl maleimide, 361 N-alkylation, 341 NaN(TMS)~, 255 1,8-nap, 378 2,3-nap, 379 1-naphthaleneboronic acid, 111 1,8-naphthalenedicarboxylicacid, 378 natural products, 268 natural receptors, 112 Nazarov cyclization, 258, 262, 265 N-Boc-allylglycine, 280 N-Boc-crotylglycine, 282 N-Boc-oxazinone, 287 N-Boc-protected amino acids, 274 N-bromosuccinimide, 255, 272 NBS, 284 n-Bu4NE 62 n-BuLi, 68, 78, 87 (+)-negamycin, 322 neodolabellanes, 247, 264 a-neodolabellenol, 244, 248, 250 NF-~.B activity, 220 Ni (0)-catalyzed coupling, 69 N-methyl-4-cyanopyridinium iodide, 137 N-methylmaleimide, 346, 355 N-methylmorpholine, 350 N-methylnicotinamide iodide, 141 NMNA, 142 NMO, 87 NMR, 145 NMR titration experiments, 119

501

INDEX NOE, 317, 439, 450 NOE correlation experiments, 45 NOE difference spectra, 48 NOESY, 439, 449 noncovalent conformers, 129 non-covalent interactions, 100 norbornadiene, 107 (2S,3R)-nor-C-statine, 319 N-phenyl maleic imide, 106 N-phenylmaleimide, 348, 354 N-propargylation, 341 a-sulfonyl allyl carbanion, 250 n-tetrabutylammonium iodide, 87, 92 nystatin, 1

0-di (dibromomethyl) benzene, 109 odorine, 222, 232 olefin metathesis, carbonyl olefination, 161 o-naphthoquinodimethane, 109 ophiobolins, 258 0-quinone methide, 468 organoborane, 85, 87, 91 organotin acetylides, 275 ortho ester Claisen rearrangement, 260 OsO4, 87 ovalicin, 462, 479 oxaphosphetane, 87 oxazinone, 270, 272, 311 oxazolidinone, 354, 357 oxepane, 60, 62, 64, 67, 69, 71, 76, 79, 83, 87, 90, 158, 196, 198, 211-212 oxidation, 316 oxidative addition, 384 oxidative cleavage, 14, 62, 220, 226, 247, 279, 307, 445 oxidative cyclization, 222 oxidative dearomatization, 463, 464, 472 oxidative hydroboration, 90 oxidative hydrolysis, 202 oxidopyrylium, 219, 223-224, 226-227 232-233, 237, 239 oxindole, 311 oxocarbenium ion, 165, 174, 186, 188 oxonium species, 255 oxy-Cope rearrangement, 7, 250

oxynitrilase-mediated asymmetric cyanohydrin-forming reaction, 272 oxysporidinone, 487 ozonolysis, 260, 287, 297, 318, 423, 431, 445, 451,456, 487

paclitaxel, 220 PADA, 485 palau'amine, 324 palladium chloride, 274, 282, 290, 299, 302 palladium mediated coupling, 467 palladium on carbon, 291 palladium-catalyzed coupling, 26 palladium-catalyzed cross-coupling, 9, 31 palladium-catalyzed reductive coupling, 396 palladium-catalyzed Suzuki-Heck type couplings, 111 p-anisaldehyde, 310 [2.2] paracyclophane, 375-376, 378, 381 parallel synthesis, 329 Pauson-Khand reaction, 334 Pb (OAc) 4, 226, 228 PbC12, 160, 161 p-benzoquinone, 103, 355 PCC, 394, 399 PCC on alumina, 69 Pd (0)-catalyzed tandem cyclization/ amination, 350 Pd(dba) 2, 28 Pd (PPh~) 4, 87, 91 PDC, 92 Pd-catalyzed alkynylation, 311 Pd-catalyzed coupling reactions, 332 PdClz, 81 p-DCNB, 126, 129 Pd-mediated reaction, 303 Pd-mediated SN2'-type cyclization, 303 Pearlman's catalyst, 270, 290, 310, 323 peptide coupling, 295 peptide isosteres, 268, 314 peptide-like molecules, 359 periodate cleavage, 276 periodic acid, 478, 484 Petasis reagent, 158, 204, 433 Ph2MeSiH, 200 phenanthrene ring, 236

502 3-phenyl-l-bromopropyne, 342 4-phenyl-[ 1,2,4]-triazole-3,5-dione, 349 phenylalanine analogues, 290 phenylmagnesium chloride, 187 phenylselenation, 263 PhI+OH-OTs, 469 phorbol, 224 phosgene (COC12), 353 phoslactomycin, 487 phosphines, 17 phosphine oxides, 17 phosphoramide catalysis, 15 phosphorus tribromide, 28 photoannulation chemistry, 425 photochemical apparatus, 235 photochemical cycloaddition, 230, 232, 233 photocycloaddition, 221,222, 228, 231 [3 + 2 ] photocycloaddition, 237 [2 + 2] photodimerization, 369-371,374, 375, 380, 381 photoreaction, 377 phytoalexin, 37 phytotoxic effects, 244 Pictet-Spengler, 295 PIDA, 469 PIFA, 468-471 pinacol rearrangement, 165 pinacolic coupling, 222 pKa values, 223 PKC, 384 Plasmodium falciparum, 36 PMB ether, 171, 177 :PMB group, 25 ]~methylcinnamic acid, 370 POCI~, 138 poison darts, 410 polyepoxide cascades, 156 polyketide, 463 polymethylhydrosiloxane, 320 1,3-polyols, 8 potassium acetate, 65 potassium azodicarboxylate (PADA), 479 potassium hexamethyldisilazane, 287 potassium hydroxide, 107, 419 p-quinol, 466 preamphilectane, 35 prenyl chloride, 82

INDEX prenyllithium, 69, 95 primary amine, 321 propanal, 15 propargyl ethers, 336 propargylic bromide, 341 (Z)-propenyllithium, 260 propenyl magnesium chloride, 171,173, 175 propiophenone, 15 propynylmagnesium chloride, 186 protease inhibitors, 359 protected 2,7-diaminosuberic acid, 280 protecting group, 305 pseudo allylic strain, 479 Pseudomonas putida, 461 Pseudopterogorgia elisabethae, 384 pseudopterosins, 385 p-toluenesulfonohydrazide, 255 pulegone, 417 pumiliotoxins, 411 pyridine, 195 pyridine alkaloids, 411 pyridinium chlorochromate, 394 pyrrolidine, 481

Q quadrone, 431 quinane natural products, 456 quinine, 303, 324, 470 quinuclidine, 303

R

radical anion, 47 Ramberg-B/icklund reaction, 206 Raney nickel, 270, 323 RCM (ring-closing metathesis), 83, 177, 182, 184, 190, 195-196, 200, 202, 204, 211, 215, 349, 447 rctt-l ,2,3,4cyclobutanetetracarboxylic acid, 380 rctt-l ,2-bis ( 4-pyridyl )-3,4-bis ( 4-chlorophe nyl ) cyclobutane, 374 reaction, 387, 394, 431,450 reactive conformer, 189 reagent-based control, 331

503

INDEX Rebek's imide, 378 red tide, 154 Red-N, 10 1,4-reduction, 314-315, 485 reductive amination, 295 reductive cleavage, 76 reductive elimination, 85 regioselective allylic oxidation, 486 regioselective oxidation, 78 regioselectivity, 309 renieramycin I, 323-324 (-)-renieramycin, 295 (+)-resiniferatoxin, 224 resorcinol, 371,373, 375, 378, 380 retro-aldol, 41 retro-aldol reaction, 43-44 retro-cycloaddition, 448 retro-Dieckmann fragmentation, 417 retro-Michael decomposition, 287 retrosynthetic analysis, 83, 90, 375, 415 retrosynthetic conversions, 423 reverse-micelles, 232 R-GABOB, 317 Rh (I) catalysis, 333 Rh ( I)-catalyzed cycloisomerization reaction, 334-335, 345, 349, 350, 361 Rh (I)-catalyzed formation of 5-1actams, 350 Rh (I)-catalyzed reaction, 344 Rh 2 (R-DOSP)4, 396 Rh2 (DOSP)4, 388

Rh2(S-DOSP)4-catalyzed reaction, 393 (-)-rhazinilam, 384 rhodium on alumina, 261 rhodium-catalyzed cycloisomerization, 328 rhodium-catalyzed insertion, 92 ring strain, 250 ring-opening, 447 (+)-rishirilide B, 460, 474, 476, 482-483 RK-397, 1 rocaglamide, 219 roflamycoin, 3 rotation, 131 roxaticins, 1, 3 Ru-catalyzed ring-closing metathesis, 332 RuO4, 477 ruthenium, 446 ruthenium catalyst, 400, 448 (R)-~ethyl aspartate, 281

S

S-(-)-cucurbitine, 279 (S)-2-methylasparagine, 289 (S)-GABOB, 316 (S)-q-hydroxy-~-aminobutyric acid, 316 sadenosylmethionine, 146 Saegusa oxidation, 48, 194, 197, 199, 253 safety concerns, 440 saframycin, 295 Sakurai coupling, 315 SAM, 146 (-)-sanglifehrin A, 291 saponification, 69, 353, 478 scaffolds, 336, 342 Schmidt or Beckmann rearrangement, 349 Schmidt reaction, 409, 423, 440, 451-452 Schrock Mo catalyst, 173, 190, 197, 211 secodolastanes, 247 secondary metabolites, 245 secondary orbital, 471 secondary orbital overlap, 474 secondary structure, 359 selective, 284 selective cleavage, 87 selective cleavage of the benzyl ether, 69 selective protection, 321 self-assembly, 99, 138, 146, 381 SEMCI, 289 SeO2, 65, 79 (+)-sertraline (Zoloft), 390 Sharpless asymmetric dihydroxylation, 83, 90, 92, 95 Sharpless asymmetric epoxidation, 3, 78, 423 SiCI4, 17, 20 SiF4, 215 sigmatropic 1,5-hydrogen shift, 106 silanol, 22, 27 silicon-based aldol, 32 silver(I) oxide, 69 silvestrol, 219-220 silyl dienol ether, 31 silyl ether, 252, 482 silylene, 183 Simmons-Smith cyclopropanation, 45 Simmons-Smith reagent, 51 single electron transfer reduction, 465 singlet oxygen, 264 (+)-rishirilide B, 485

504 six-component assemblies, 374 six-membered transition, 474 skeletal rearrangement, 220 SN2 displacement, 273 SN2 reaction, 471 SN2' displacement, 249, 481 SnC14, 69, 71 sodium, 280, 285, 467 sodium azide, 442 sodium borohydride, 22, 106, 419 sodium channels, 411 sodium chlorite oxidation, 200 sodium cyanoborohydride, 53 sodium hydrogen sulfite, 264 sodium hydrosulfite, 403 sodium iodide, 106-107 sodium naphthalide, 263 sodium silanolate, 27 sodium tert-amylate, 261 solid state, 369 solid-state reactivity, 381 solvent, 146 solvent-free environment, 369 Sonogashira, 332 sorbicillin, 471 spectrofluorimetric titrations, 119, 122 spirotryprostatin B, 309 S-proline, 473 stabilized ylide, 261 re-stacking interactions, 231 stereocontrolled, 253 stereocontrolled synthesis of (2R,3R), 299 stereoselective and chemoselective reactions, 470 stereoselective reduction, 214, 312 stereoselective vinylogous aldol addition, 30 stereoselectivity, 71,199, 231 stereospecific 1,2-shift, 245 stereospecific electrocyclization, 481 stereospecific sodium borohydride reduction, 272 steric hindrance, 358 steric strain, 125 stevens rearrangement, 314 stilbazoles, 373 Stiles carboxylation, 222 Stille, 332 Stille coupling, 6, 31 stoichiometric conversion of pentane to 1-pentene, 384

INDEX structural diversity, 329 Strycker reagent [Ph~PCuH] 6, 82 sulcatone, 423 sulfides, 17 sulfones, 250 a-sulfonyl ketones, 261 supermolecules, 99 supramolecular capsules, 100 supramolecular chemistry, 369, 381 Suzuki, 332 Suzuki-Heck type coupling, 111 Suzuki-Miyaura cross-coupling, 85, 87, 90-91, 95 Swern oxidation, 71, 86, 195, 261 syn elimination, 255

syn-3-hydroxy-4-methyl-pyrrolidine, 319 syn-diastereoselectivity, 302 1,5-syn stereoinduction, 17, 32 synthetic receptors, 138 (S)-methyl aspartate, 270

T TADDOL, 233, 236-237, 239 Takai cyclization, 206 Takai protocol, 158, 179, 196, 208 Takai reagent, 160 Takai-Utimoto reagent, 173, 191-192, 208, 211 Takeda reagent, 206 tandem intermolecular cycloaddition, 349 tautomer, 476 tautomerization, 262 taxusin, 226 TBAF, 27, 87, 182, 290, 340 TBDMS ether, 185, 189 TBDPSC1, 87 TBS ethers, 195, 203, 400 TBS-enol ethers, 396 TBSOTf, 2,6-1utidine, 71 TCNB, tropylium tetrafluoroborate, 119, 126, 130 TCNQ, 122, 126 Tebbe, 204 Tebbe reagent, 158, 161,433, 446 temperature, 234 temperature of coalescence, 134 temperature-dependent spectra, 131

505

INDEX templates, 379 tetrabromonaphthalene, 110 tetrabromo-0-quinodimethane, 110 tetrabromo-0xylene, 106 tetrabutylammonium hydroxide, 76 1,2,4,5-tetracyanobenzene, 119 tetrakis (4-pyridyl) 1,2,9,10diethano [2.2] paracyclophane, 375 tetramethylammonium, 24 tetramethylethylenediamineTMEDA, 78 tetrazomine, 300 TFA, 284, 295 TfOH, 438-439, 451-453 thallium nitrate, 469 thermodynamic control, 39 thermodynamic protonation, 437 thiocarbonyldiimidazole, 255 Thorpe-Ingold effect, 344 Ti (OiPr)4, 78 TiC14, 160, 427, 439, 451 TIPS ether, 182, 188 tltanacyclobutanes, 208 utanium enolate, 284 utanium methylidene, 208 titanium tetrachloride, 110 titanium(IV) enolates, 18 utanium-promoted regioselective opening, 79 titanocene alkylidene complexes, 445 TMEDA, 160 TMS ether hydrolysis, 192 TMS iodide, 276 TMS triflate, 52 TMSBr, 92 TMS-CF~, 321 TMSCN, 319 TMSI, 486 TMSOK-promoted coupling, 9 TMSOTf, 200 TMSOTf, Et~SiH, 201 TNE 122 TNFe, 220 toluenesulfonyl hydrazone, 53 tomentanol, 59 tomentol, 59 torsional strain, 212, 481 tosyl chloride, 106 tosylhydrazone, 73 TPAP, 199-200, 203, 212 4,4'-tppcp, 375

trans-l,2-bis(4-pyridyl)ethylene (4,4'bpe), 373 trans-conformation of secondary amide, 353 trans-crotyl chloride, 282 trans-2,5-dimethyl-pyrrolidine, 473 trans-diaxial couplings, 42 trans-[3.3.0]octane, 425 trans-stilbene, 271 tra ns, tra ns- l , 4-b is ( 4-p yri dyl ) - l ,

3-butadiene, 377 trans, trans, trans- 1,6-bis (4-pyridyl)-1,3,

5-hexatriene (1,6-bpht), 377 transamination, 272 transannular cyclizations, 252 transannular olefin-enone cyclization, 255 transannular reactions, 246, 264 transition metal-catalyzed cyclocarbonylation and cycloisomerization reactions, 335, 361 transition metal-catalyzed reactions, 332, 334 transition state, 21,252, 473 transmetalation, 85 1,2-transposition, 76 triacetoxyborohydride, 24 triallylaluminum, 174 triallylborane, 174 triazoline, 416 2,4,6-trichlorobenzoyl chloride, 5, 28 trichlorosilyl enolate, 15, 17, 20 tncycloillicinone, 463 triethyl orthoformate, 284 triethylamine, 103, 106 triethylphosphite, 28 triethylphosphonoacetate, 94 triflate, 194 triflic acid, 438 trifluoroacetic acid, 51 trimethyl orthoformate, 284, 311 trimethylaluminium, 68 trimethylphosphonoacetate, 64 ~trimethylsilyl acrolein, 11 triphenylphosphine, 61, 68, 336-337 2-triphenylphosphoranylidenebutanedioate, 68 triphenylsilanol, 81 triquinanes, 428 tris (dibenzylide neacetone)-dipalladium (0)-chloroform, 27 trypticene, 116 tuberostemoninol, 324 m-tyrosine, 291

506

0~,[3-unsaturated amide, 355 unsaturated amide, 355 unnatural amino acids, 269 ureas, 17

valium, 428 vanadium-catalyzed Sharpless epoxidation, 255 vanadyl bis-acetoacetate [VO(acac)2], 62 van-der-Waals or dispersion forces, 99 vibsane, 245 Vilsmeier reaction, 40 vinyl epoxide, 78, 81 vinyl magnesium bromide, 187 vinyl triflate, 396 vinyldiazoacetate, 393 vinyldiazoacetates, 388, 391 vinylogous aldol addition, 11-12 vinylogous ester, 477, 480, 484-485 vinylstannane, 81

W

Wacker's conditions, 81 Weinreb amide, 13, 94-95, 431,447 Wesley oxidation, 464 Wieland-Miescher ketone, 73, 78 Wilkinson catalyst, 479 Williamson cyclization, 78 Williamson reaction, 78

INDEX Wittig olefination, 16, 182, 261 Wittig reaction, 45, 321,399, 402 Wittig reagent, 90 Wolff-Kishner reaction, 53

X-ray analysis, 471 X-ray crystal structure, 355 X-ray crystal structure analysis, 236 X-ray crystallography, 44, 308, 348, 369, 453 X-ray diffraction, 247, 260, 264 X-ray diffraction analyses, 256

Yamaguchi method, 5 Yamaguchi protocol, 204 ylide, 306, 308

zetekitoxin, 324 Zimmerman-Traxler chair transition state, 297, 302 zinc, 110, 255 zinc chloride, 275 Zn, 160 Zn (Cu) amalgam, 215 Zn(N3) 2, 441 Zn(OTf) 2, 214 ZnC12, 336, 338, 340-341,486 (+)-zoapatanol, 59

Strategies and Tactics in Organic Synthesis, Volume 4 (Strategies and Tactics in Organic Synthesis)

Strategies and Tactics in Organic Synthesis, Volume 4 (Strategies and Tactics in Organic Synthesis)

Strategies and Tactics in Organic Synthesis, Volume 6 (Strategies and Tactics in Organic Synthesis)

Strategies and Tactics in Organic Synthesis, Volume 6 (Strategies and Tactics in Organic Synthesis)

Strategies and Tactics in Organic Synthesis, Vol. 6

Strategies and Tactics in Organic Synthesis, Vol. 6

Strategies for Organic Drug Synthesis

Strategies for Organic Drug Synthesis

Strategies for Organic Drug Synthesis and Design

Strategies for Organic Drug Synthesis and Design

Palladium in Organic Synthesis

Palladium in Organic Synthesis

Tellurium in Organic Synthesis

Tellurium in Organic Synthesis

Ruthenium in Organic Synthesis

Ruthenium in Organic Synthesis

Introduction to Strategies for Organic Synthesis

Introduction to Strategies for Organic Synthesis

Ruthenium in Organic Synthesis

Ruthenium in Organic Synthesis

Radicals in Organic Synthesis

Radicals in Organic Synthesis

Microwaves in Organic Synthesis

Microwaves in Organic Synthesis

Palladium In Organic Synthesis

Palladium In Organic Synthesis

Microwaves in organic synthesis

Microwaves in organic synthesis

Organic Synthesis

Organic Synthesis

Photochemistry and Organic Synthesis

Photochemistry and Organic Synthesis

Organic Synthesis

Organic Synthesis

Organic Synthesis

Organic Synthesis

Reagents for Organic Synthesis, Volume 7

Reagents for Organic Synthesis, Volume 7

Reagents for Organic Synthesis, Volume 7

Reagents for Organic Synthesis, Volume 7

Nitrile Oxides, Nitrones and Nitronates in Organic Synthesis: Novel Strategies in Synthesis

Nitrile Oxides, Nitrones and Nitronates in Organic Synthesis: Novel Strategies in Synthesis

Hydrogen Bonding in Organic Synthesis

Hydrogen Bonding in Organic Synthesis

Domino Reactions in Organic Synthesis

Domino Reactions in Organic Synthesis

Cycloaddition Reactions in Organic Synthesis

Cycloaddition Reactions in Organic Synthesis

Cycloaddition Reactions in Organic Synthesis

Cycloaddition Reactions in Organic Synthesis

Iridium Complexes in Organic Synthesis

Iridium Complexes in Organic Synthesis

Unit Processes In Organic Synthesis

Unit Processes In Organic Synthesis

Domino Reactions in Organic Synthesis

Domino Reactions in Organic Synthesis

Microwave Methods in Organic Synthesis

Microwave Methods in Organic Synthesis

Metal Carbenes in Organic Synthesis

Metal Carbenes in Organic Synthesis

Our partners will collect data and use cookies for ad personalization and measurement. Learn how we and our ad partner Google, collect and use data. Agree & close