CR 00 3455

Synthesis of Strychnine Josep Bonjoch* and Daniel Sole´ Laboratory of Organic Chemistry, Faculty of Pharmacy University of Barcelona, 08028 Barcelona, Spain Received March 13, 2000 Contents I. Introduction 3455 II. An Overview of the Synthetic Strategies 3456 III. Total Syntheses 3458 A. Woodward’s Synthesis 3458 B. Magnus’ Synthesis by Transannular Cyclization of a Stemmadenine-type Derivative 3462 C. Overman’s Synthesis through the Tandem Cationic Aza-Cope Rearrangement−Mannich Cyclization Reaction 3466 D. Kuehne’s Syntheses Using a Tandem Sequence To Form Pyrrolo[2,3-d]carbazole Intermediates 3469 E. Stork’s Synthesis via Intramolecular Conjugate Addition of a Vinyllithium Intermediate 3472 F. Rawal’s Synthesis through Intramolecular Diels−Alder and Heck Reactions 3473 G. The Bonjoch−Bosch Synthesis Using a 3a-(2-Nitrophenyl)hexahydroindole as a Building Block 3475 H. Martin’s Formal Synthesis: A Biomimetic Approach 3477 IV. Concluding Remarks 3478 V. Tabular Survey of Total Syntheses of Other Strychnos Alkaloids 3479 VI. Note Added in Proof 3480 VII. References and Notes 3480 I. Introduction The historic total synthesis of strychnine by Wood- ward1 in 1954 represented a milestone in the field of organic synthesis.2 Strychnine (C21H22N2O2) ranks as one of the most complex natural products of its size, inasmuch as it incorporates six contigous asymmetric centers (five of which are in the core cyclohexane ring) and contains a mere 24 skeletal atoms com- pactly arranged in seven rings. Given its intricate architecture, coupled with its pharmacological and extremely toxic properties,3 strychnine has always fascinated organic chemists.4 It is a notorious poison (�50 mg is lethal for an adult human), which blocks postsynaptic inhibition in the spinal cord where it antagonizes the transmitter glycine.5 This property has made strychnine very useful as a tool in experi- mental pharmacology. Strychnine was first isolated as far back as 1818 from the seeds and bark of Strychnos nux vomica by Josep Bonjoch Sese´ was born in Barcelona (Catalonia, Spain) in 1952. He received his BSc (1974) and Ph.D. (1979) degrees from the University of Barcelona, Faculty of Chemistry. His Ph.D. Thesis work was conducted under the direction of Professor Joan Bosch. He then moved to the Faculty of Pharmacy at the University of Barcelona, where he was promoted to Associated Professor (1984) and subsequently became Full Professor of Organic Chemistry in 1992. His main research involves the synthesis of azapolicyclic natural products and the development of synthetic methods using radical and organometallic species. Daniel Sole´ Arjo´ was born in 1965 in Vielha, Spain. He studied Pharmacy at the University of Barcelona, where he received his Ph.D. in 1992. After postdoctoral studies at the Spanish research council (CSIC) with Prof. Josep MaMoreto´, he returned to the Faculty of Pharmacy at the University of Barcelona, where he became Associate Professor in Organic Chemistry in 1997. His research is centered on the use of transition metals in the synthesis of complex organic molecules. 3455Chem. Rev. 2000, 100, 3455−3482 10.1021/cr9902547 CCC: $35.00 © 2000 American Chemical Society Published on Web 08/08/2000 Pelletier and Caventou6 and its elemental composi- tion was established some 20 years later by Reg- nault.7 Strychnine was the subject of a very large number of degradative studies before the advent of modern spectroscopic techniques, and the elucidation of its constitutional structure represented one of the major achievements of classical organic chemistry. Degradative work started in the 1880s, and the finishing touches were published in 1948 by Wood- ward and Brehm,8 the major contributions being made by Leuchs and his school and by Robinson and his collaborators.9 An exhaustive and excellent review covering a century and a half of historical accounts of the work on the chemistry of strychnine was written in 1964 by Smith.10 The relative configuration of strychnine was provided via two independent X-ray crystal analyses done by Robertson and Bevers, and Bijvoet.11 The absolute stereochemistry of strychnine was established by Peerdeman12 with X-ray crystal- lography and was later confirmed by Schmid and his collaborators13 using a chemical method. The exten- sive NMR data available for strychnine14 have been used for collecting information on conformational and configurational assignment of this and other related alkaloids or precursors. Strychnine is the flagship compound of the family of Strychnos alkaloids,15 one of the most populous classes of indole alkaloids. Its biogenetic pathway involves, in the initial steps, the enzymatically catalyzed Pictet-Spengler condensation of tryptamine with secologanin to provide strictosidine. Next to be formed is geissoschizine, the common biogenetic intermediate for all monoterpenoid indole alkaloids (Scheme 1). After an oxidative cyclization involving C-16, followed by a skeletal rearrangement, the characteristic framework of Strychnos alkaloids ap- pears with dehydropreakuammicine. The unrear- ranged monoterpenoid unit characteristic of the Corynanthe skeleton (depicted in boldface in geissos- chizine, Scheme 1), originally attached to the indole R-carbon (C-2), is now bonded to the â-position (C- 7), and a new bonding between the rearrangable unit (C-16/C-17/C-22) and C-2 is in place.16 The next step involves the loss of the methoxycarbonyl group from dehydropreakuammicine to give norfluorocurarine, which, upon hydroxylation and reduction, could lead to the Wieland-Gumlich aldehyde, a biogenetic precursor of the heptacyclic base strychnine, as shown by Heimberger and Scott17 in 1973. To com- plete the strychnidine backbone,18 two additional carbons are required. Robinson’s suggestion that they come from acetate was proven by Schlatter in 1969,19 and probably occurs through prestrychnine, formed by an aldol condensation involving acetyl-CoA. The numbering system and ring labeling used throughout this review is based on the biogenetic interrelationship of indole alkaloids, as proposed by Le Men and Taylor.20 To avoid confusion, it is worth mentioning that this numbering differs from Wood- ward’s system,8 which is used in some papers in the strychnine field as well as in the strychnidine ste- reoparent nomenclature of Chemical Abstracts (Fig- ure 1). II. An Overview of the Synthetic Strategies Woodward’s strychnine synthesis1 remained the sole approach for a long time despite the monumental work in indole alkaloid synthesis developed in the second half of the twentieth century. After a dormant period of more than 30 years, interest in the chem- istry of strychnine revived, and in the 1990s several groups succeeded in synthesizing this fascinating molecule.21-28 Three syntheses22,25,27 culminated in the enantioselective total synthesis of the natural enantiomer, (-)-strychnine,29 Overman’s route also leading to the dextro isomer, ent-strychnine.22b In an overview of strychnine syntheses, as outlined in Table 1, the first noteworthy feature is that all approaches are directed to isostrychnine or the Wieland-Gumlich aldehyde, whose synthetic conver- sion to strychnine was reported while Woodward’s first total synthesis of strychnine was in progress (Scheme 2). Isostrychnine, which is the product of a base- or acid-induced retro-Michael addition with double-bond migration obtained from strychnine,30 Scheme 1. Biosynthesis of Strychnine Figure 1. On the left, strychnine structure with number- ing and ring labeling proposed by Woodward and used in Chemical Abstracts. On the right, strychnine showing the biogenetic numbering and ring labeling used in this review. 3456 Chemical Reviews, 2000, Vol. 100, No. 9 Bonjoch and Sole´ was converted back to strychnine in 20% yield when treated with alcoholic potassium hydroxide.31 The Wieland-Gumlich aldehyde is another degradation product isolated in the course of strychnine chemical investigations.32,33 Its conversion back to strychnine was achieved in 68% yield when treated with a mixture of malonic acid, sodium acetate, and acetic anhydride in acetic acid.34,35 After strychnine had been chemically correlated with isostrychnine and the Wieland-Gumlich alde- hyde, both compounds were identified as Strychnos alkaloids, the former in 1973,17 when it was isolated from a natural source, and the latter when it was found to be the same as the already known caracu- rine VII.36 The synthetic strategies developed to reach strych- nine deserve a brief general comment. The major stumbling blocks in the synthesis of the target alkaloid are the following: (i) the generation of the spirocenter at C-7; (ii) the assembling of the bridged framework of the alkaloid (CDE core ring); (iii) the elaboration of the hydroxyethylidene substituent. The crucial spirocenter at C-7 has been constructed by either taking advantage of the indole reactivity or elaborating this quaternary center without the use of indole derivatives (Table 1 and Chart 1). Wood- ward, Magnus, and Kuehne all use the electrophilic attack of an iminium ion upon a 2,3-disubstituted indole to generate the C-7 spirocenter, but they undertake this crucial step at different stages of the synthesis. Thus, Woodward constructs the quater- nary center early on in the synthesis (ABC ring fragment), while Magnus and Kuehne elaborate the C-7 spirocenter at more advanced stages of the process: the former to assemble the pentacyclic curan skeleton (ABCDE rings) and the latter to construct the pyrrolo[2,3-d]carbazole fragment (ABCE rings). On the other hand, both Stork and Martin generate a 3-chloroindolenine to promote the formation of the key quaternary center by means of a skeletal rear- Table 1. Main Features of Strychnine Syntheses a See Chart 1. b See Scheme 3. c The synthesis of (+)-strychnine was published in 1995. d Desymmetrization of cis-3,5- diacetoxycyclopentene. e Personal communication (1999). Scheme 2 Synthesis of Strychnine Chemical Reviews, 2000, Vol. 100, No. 9 3457 rangement that leads to a pyrrolo[2,3-d]carbazole intermediate (ABCE rings) and to a pentacyclic curan derivative (ABCDE rings), respectively. In contrast, Overman, Rawal, and our team worked with intermediates incorporating a functionalized phenyl ring that does not participate in the elabora- tion of the spirocenter. Overman used a tandem aza- Cope/Mannich rearrangement, Rawal an intramo- lecular Diels-Alder reaction, and our team a classical Claisen rearrangement to build up the quaternary C-7 center. While the formation of the quaternary center in the Rawal synthesis also involves the closure of the indoline ring, in Overman’s and our approach the substituted phenyl ring remains as a latent form of the indole nucleus until an advanced stage of the synthesis. The second key step in the synthetic approaches to strychnine is the assembling of the bridged CDE ring fragment. The synthetic strategies adopted for its construction are outlined in Scheme 3. In the majority of the synthetic approaches the bridge framework is assembled once the quaternary C-7 center has already been constructed. In all of these cases the closure of the piperidine D ring, either by formation of the N4-C21 bond (Woodward; Kuehne) or by formation of the C15-C20 bond (Rawal; Stork; Bonjoch and Bosch), is used at this crucial step. In the former syntheses, the process involves a reaction of a nitrogen atom upon an oxygenated carbon (carbonyl, epoxide, or tosylate), whereas in the latter, the ring closure is accomplished by the addition of a vinyl organometallic species to a double bond. In the other approaches, in which the piperidine ring has already been constructed, the bridged ring fragment and the C-7 spirocenter are assembled simultaneously, either by the transannular cycliza- tion of a stemmadenine-type compound (Magnus, forming C3-C7 bond) or by multistep sequence processes, such as the cationic aza-Cope rearrange- ment/Mannich cyclization (Overman, C5-C6 and C3-C7 bonds formed) or the skeletal rearrangement of a 3-chloroindolenine (Martin, C3-C7 and C2-C16 bonds formed). The last key operation in the synthetic routes to strychnine is the elaboration of the hydroxyeth- ylidene side chain at C-20. Woodward, Magnus, and Kuehne took advantage of a ketone carbonyl at C-20 to introduce the hydroxyethylidene substituent in the last steps of the synthesis by either an allylic rear- rangement (Woodward) or a Wittig olefination pro- cess (Magnus and Kuehne). On the other hand, both Overman and Martin constructed the hydroxyeth- ylidene-bearing piperidine ring early on by means of â-elimination reactions that stereoselectively intro- duce the E-configured double bond. Finally, in the other approaches (Rawal; Stork; Bonjoch and Bosch) the stereoselective incorporation of the (E)-hydroxy- ethylidene double bond is accomplished during the closure of the piperidine D ring by means of intramo- lecular coupling reactions of vinyl halides with alk- enes. III.Total Syntheses A. Woodward’s Synthesis The total synthesis of strychnine achieved by Woodward in 1954, only 6 years after the elucidation of its structure, is a historical landmark in organic synthesis. Considering the complexity of the strych- nine molecule it is admirable that Woodward was able to undertake, let alone satisfactorily complete, its total synthesis with the resources at his disposal. It is noteworthy that, when he devised a way to synthesize strychnine, Woodward was strongly in- fluenced by contemporary ideas about the biogenesis of the indole alkaloids and especially by his own hypothesis about the biogenesis of strychnine itself,37 which although finally shown to be not essentially correct, proved to be very fruitful. Woodward took advantage of two of his own biogenetic proposals in his synthetic work: (a) the nucleophilic character of the â-position of the indole nucleus and (b) the oxidative cleavage of an aromatic ring and the subsequent recombination of the fragments to build up the skeleton of the alkaloid. The general features of the synthesis are shown, in retrosynthetic form, in Scheme 4. At the end of the 1940s, it was known that isostrychnine could be converted to strychnine by the action of a base.31 Therefore, it is not surprising that Woodward chose this process for the last step (closure of ring F) when planning the synthesis of strychnine. Dehydrostrych- ninone (2) was envisioned as a suitable precursor of isostrychnine. Two crucial transformations needed to be done from 2: (i) the introduction of the hydroxy- ethylidene side chain, which could be easily elabo- Chart 1. Generation of C-7 Spirocenter of Strychnine 3458 Chemical Reviews, 2000, Vol. 100, No. 9 Bonjoch and Sole´ rated from the corresponding carbinol through an allylic rearrangement reaction, and (ii) the reduction of the aromatic R-pyridone ring to the dihydro level. The closure of the piperidine ring (ring D) was planned by oxidative cyclization of methyl ketone 3, a compound that should be readily available from â-keto ester 4. The disassembly of ring E at 4 by a retro-Dieckmann condensation would led to diester 5, which, in turn, could be derived from 6 by selective cleavage of the veratryl protecting group and recom- bination of some of the carbon atoms to build up ring G. Finally, disassembly of intermediate 6 by a retro- Scheme 3. Construction of the Bridged Framework of Strychnine Scheme 4. Woodward’s Retrosynthetic Analysis of Strychnine Synthesis of Strychnine Chemical Reviews, 2000, Vol. 100, No. 9 3459 Pictet-Spengler reaction (ring C formation) would lead to 2-veratryltryptamine (7) and ethyl glyoxylate. The starting material for the synthesis was the 2-veratrylindole (9), which was readily prepared by Fischer indole synthesis from acetoveratrone (8) (Scheme 5). The first steps in the synthesis involved the introduction of the 2-aminoethyl chain at the â-position of 2-veratrylindole (9). Thus, reaction of 9 with the iminium salt derived from formaldehyde and dimethylamine afforded a gramine derivative, which by treatment with methyl iodide was converted to the ammonium salt 10. Reaction of the latter with sodium cyanide followed by reduction of the resulting nitrile with lithium aluminum hydride gave tryp- tamine 7. Having obtained 2-veratryltryptamine (7), Wood- ward undertook the first key step of the synthesis, the elaboration of the C-7 quaternary center, which was accomplished by the construction of a spiro ABC derivative. Although condensation of 7 with ethyl glyoxylate afforded the corresponding Schiff base, all attempts to promote the desired cyclization of this intermediate by means of acid catalysts resulted in failure. To drive forward the desired process, it was necessary both to increase the electrophilic character of the iminium moiety and to stabilize the cyclization product. In fact, when the Schiff base was treated with tosyl chloride in pyridine the indolenine 11 was obtained as the sole product.38 Reduction of indole- nine 11 with NaBH4 occurred with complete stereo- selection since an attack by the borohydride ion occurs from the more accessible â-face to give an indoline, which on subsequent treatment with acetic anhydride provided the N-acetyl derivative 6. Once the veratryl group had been used to block the R-carbon of the indole nucleus and direct the attack of the electrophilic species to the â position, Wood- ward wondered whether this protecting group could be used for the elaboration of the other rings of strychnine. When 6 was treated with ozone in aque- ous acetic acid the veratryl group was selectively cleaved at the bond between the two methoxy groups to give muconic ester 12, which on heating in metha- nolic hydrogen chloride directly afforded pyridone 13. This transformation, which leads to the formation of ring G of strychnine, brought about the cleavage of the N-acetyl group, formation of the six-membered lactam, and isomerization of the exocyclic double bond to the stable aromatic R-pyridone. Diester 13 contains all of the carbon atoms and the functionality necessary to undertake the construction of ring E by means of a Dieckmann condensation. Nevertheless, when 13 was treated with a base, the leaving group behavior of the tosyl group changed the expected course of the process, and consequently had to be removed prior to the condensation reaction. This removal was accomplished by treating 13 with hot hydriodic acid in the presence of red phosphorus (Scheme 6). These reagents also cleaved the two ethyl ester groups of the starting material to give an amino diacid intermediate, which on sequential N-acetyla- tion and esterification with diazomethane afforded the dimethyl ester 14. Treatment of 14 with sodium methoxide in methanol resulted in the epimerization of the stereogenic center at C-3 and the subsequent Dieckmann cyclization to give 4. Scheme 5. Closure of Rings C and G: Synthesis of Intermediate 13 Scheme 6. Closure of Ring E: Synthesis of Intermediate 17 3460 Chemical Reviews, 2000, Vol. 100, No. 9 Bonjoch and Sole´ The â-keto ester 4 exists as a stable enol, thus preventing classical methods for carbonyl group reduction from being used to remove the oxygen atom at C-14. Fortunately, this oxygen atom could be removed by indirect methodology. Thus, reaction of 4 with tosyl chloride in pyridine afforded the corre- sponding O-tosyl derivative, which on treatment with sodium benzylmercaptide, gave sulfide 15 in an addition-elimination process. Desulfuration of 15 using deactivated Raney nickel, followed by hydro- genation of the resulting unsaturated ester furnished cis saturated ester 16, together with a small amount of the trans isomer. The obtention of cis saturated ester 16 as the major isomer in the hydrogenation reaction is the result of the addition of hydrogen from the less-hindered R-face of the