THE INDOLE ALKALOIDS EXCLUDING HARMINE AND STRYCHNINE By J. E. SAXTON B.Sc. M.A. DPHIL. (Oxox.) (RESEARCH FELLOW HARVARD UNIVERSITY CAMBRIDGE MASS. U.S.A.) ALKALOIDS containing the indole ring system occur widely in Nature and have so far been isolated from upwards of twenty genera of plants and trees. They include many important and widely used alkaloids such as the ergot bases valuable as oxytocic drugs in childbirth strychnine valuable as a general tonic and also employed as a vermin killer yohimbine used in veterinary medicine as an aphrodisiac and the extracts of RauwolJia serpentina Benth. used in India for several purposes chiefly as a sedative. In addition extracts of AZstonia barks have been used in the Far East in the treatment of malaria but pharmacological experiments have shown that this reputation is undeserved.The pharmacological properties of all these plant extracts have stimulated chemical investigation into the structures of the alkaloidal constituents and so far the structures of approximately fifty alkaloids have been completely elucidated. The object of the present Review is to sunimarise the present knowledge of these indole alkaloids. No attempt will be made to cover the field exhaustively particularly for those groups where work has almost ceased owing to the elucidation of tjhe structures of the alkaloids and their con- firination by synthesis. The alkaloids of Pegan rim harmala and those of the Xtrychnos species which will be reviewed elsewhere have been omitted entirely from the discussion; so have the alkaloids of the AmaryZZiducpx (e.g.lycorine) Erythrina (e.q. /3-erythroidine) and Cryptocarya bowiei (e.g. cryptowoline iodide) which although they niay be regarded as containing a hydroindole structure are not related biogenetically to the remaining indole alkaloids since they are probably formed in the plant froni dihy- droxyphenylalanine and not from tryptophan. Activity in certain fields for example the Yohim be a i d Rauz~~oZjla groups lias been intense during the last five years and substantial contributions have been made to our knowledge of the alkaloids occurring in the plants of the species and their structures. Hence emphasis will be given here to the recent developments in these series. Papers amilable up to July lst 1955 have been covered. A brief and authoritative account of the structural relations and probable course of the biosynthesis of these alkaloids has bcen given by Sir Robert Robinson in the recently puhlishecl Weizmnnn 1ectiires.l For a comprehen- Robinson ‘‘ The Structural 1Zc.lutions of Natural Products ” Oxford Univ.Press 1955. 108 sive accouiit of the earlier work on the iiidolc all~aloirls the reader is referred t o Volunie I1 of " The Alkaloids ".2 Simple Indole Derivatives.-Iiidole itself occurs in certain plants and has been obtained from the distilled oil of jasmine flowers and from the decaying wood of CeZtis reticulosa. However it is not a true alkaloid and probably arises by degra'dation of more complex indole derivatives e.g. tryptophan. Four comparatively simple inonosubstituted indole bases have beell isolated from natural sources. from the seeds of Abrus precatorius L.hypaphorine (11) from the seeds of various Erythrina species gramine (111) 5 from barley leaves and Arundo donax and donaxarine C13H1602N2,6 also from the leaves of A . donax. They are abrine (I) aTHzNMez (Ill) The relation of abrine and Gramine on the other hand was hypaphorine to tryptophan is evident. formerly regarded as arising by condensa- tion of indole with equivalents of dimethylainine and formaldehyde a re- action analogous to the usual laboratory preparation. Leete and Marion 7 have shown however that in common with the other indole alkaloids grarnine is produced from tryptophan. In their experiments tryptophan (IV) containing 14C in the 2- and the P-position was fed to sprouting barley. The grainine isolated after eleven days was shonm to possess activity in both labelled positions in the same ratio as in the original tryptophan.Thus it \li.ould seem that the /3-carbon bond in indole is not ruptured during the transformation. The structure of donaxariiie has not yet been elucidated and little is known about it beyond its formula. The Ergot Alkaloids.-The evidence leading to the structures of the ergot alkaloids and the attempts to syiithesise then? have been revieweds by 2 " The Alkaloicls " Vol. 2 cck. Manskc and Holmes Academic Press New York 1952. 3 Ghatak and KFLLL~ J . Itldicclz Cke??~. SOC. 1932 9 383 ; Hoshino AnTzalen 1935 520 31 ; Gordon and J&(*~soII d . Ukd. C ~ ~ e 9 ~ ~ . 1035 110 151 ; Miller and Robson J. 1938 1910. 4 References 11-21 in " The Alkaloids " Vol. 2 (see ref. 2) p. 481. 6 Euler and Hellstrom Z.pkpiol. Chenz. 1932 208 43 ; Orekhov Norkina and Maximova Ber. 1935 68 436 ; Wieland and Hsing A?znaEeiz 1936 526 188 ; Kuhn and Stein Ber. 1937 70 567. 6 Madinaveitia J . 1937 1927. 7 Leete and Marion Canad. J . Chem. 1953 31 1195. 8 Stoll " Progress in the Chemistry of Natural Products " Vol. IX Springer Verlag Vienna 1952 p. 114 ; Glenn Quart. Rev. 1954 8 192. 110 QUARTERLY REVIEWS CON H - I H Me C G . N H ~ H . C H ~ . OH @Me " (VI) BMe " (VII) Me t,Na BHqz$OC12,3,NaCN @- 4,Methanolysis s,Hydrol Ac H (+)- LyeerQic Acid SAXTON THE INDOLE ALKALOIDS 111 Stoll and by Glenn who sumnzarise the evidence available up to the end of 1953 by which time the general formulae for the alkaloids were firmly estab- lished as (V). Ergometrine the simplest alkaloid of the series.has structure Recent work in this field has been concentrated on lysergic acid (VII) the common structural unit of the alkaloids and has culminated in the elegant synthesis by Woodward and his collaborators of (&)-lysergic acid by the route shown below starting from N-benzoyl-3-2'-carboxyethyl- 2 3-dihydroindole (VIII).9 I n order to direct ring closure on to the 4- position of the indole nucleus a dihydroindole derivative was employed as starting material. The lysergic acid ring system was then synthesised by standard methods and the indole double bond introduced in the final stage by a catalytic dehydrogenation in neutral aqueous solution a t a deactivated Raney nickel catalyst Both the acid and the hydrazide were completely identified by comparison with naturally occurring samples.Since the hydrazide had already been resolved and converted into ergometrine lo this constitutes the first total synthesis of an ergot alkaloid. (VI). CO NH2 GNM~ CO NH2 O M e Cycl isn . co;! CO.NH2 H (Vll) Four schemes for the biosynthesis of lysergic acid have been proposed recently. The first proposed by van Tamelen,ll involves condensation of dihydronicotinic acid or its equivalent [e.g. dihydrotrigonelline or (X)] with a didehydro-5-hydroxytryptophan (IX) the doubly activated 5-position Kornfeld Fornefeld Kline Mann Jones and Woodward J. Amer. Chem. SOC. 1954 76 5256. 10 Stoll and Hofmann Helv. Chim. Acta 1943 26 922 944. l1 Van Tamelen Experientia 1953 9 457. 112 QUARTERLY REVIEWS of the former coupling with the 4-position of the latter. Subsequent stages include the removal of the oxygen from the indole ring transamination oxidation of the product to an indolylacetic acid derivative cyclisation and isomerisation to lysergic acid (see previous page).This scheme is evidently not rigid and many variants can be envisaged for example the use of 5-hydroxy-3-indolylacetic acid as starting material. Since trigonelline and nicotinic acid [and so presumably dihydrotrigonelline and (X)] can arise in biological systems from tryptophan and since both tryptophan and 5-hydroxytryptophan are possibly formed from tyrosine the whole biosynthesis can be achieved by starting from tyrosine ; so it is noteworthy that tyrosine has been found t o be associated with the ergot alkaloids. l2 Sir Robert Robinson l2 commenting on this biosynthesis pointed out that although all the stages are feasible no 5-hydroxy-indole analogues of lysergic acid have been found in Nature as would be expected if this scheme were correct.I n addition quinones of type (IX) are not known although this does not mean that they cannot be formed transiently in vivo. Harley- Mason’s suggestion l 3 suffers from these same disadvantages. Since lysergic acid contains a tryptamine residue it is reasonable to assume that the two nitrogen atoms of the tryptophan precursor are retained and are the two nitrogen atoms of lysergic acid. This is the fundamental postulate of the other proposed biosyntheses. Harley-Mason’s scheme consists of condeiisation of tetradehydro-5- hydroxytryptophan with acetonedicarboxylic acid and formaldehyde by the standard Mannich reaction l3 The starting material could presumably also be didehydro-5-hydroxytrypto- phan (IX).The third mechanism proposed by Wendler l4 starts from tryptophan itself which is presumed to undergo ring closure to a tricyclic amino-ketone. Addition of citric acid dehydration and further ring closure then give l2 Beadle Mitchell and Nye Proc. Nut. Acad. Sci. U.S.A. 1947 33 155 ; Mitchell and Nye ibid. 1948 34 1 ; Dalgliesh Quart. Rev. 1951 5 227 ; Robinson Chew. and I n d . 1952 358 ; Fraiikel and Rainer Biochem. Z. 1916 74 167. l3 Harley-Mason Chem. and I n d . 1954 251. l4 Wendler Experientia 1954 10 338. SAXTON THE INDOLE ALKALOIDS 113 lysergic acid. The first stage which assumes anionoid reactivity of the &position of the indole nucleus is supported by the results of nitration of 2-methylindole and dihydropentindole which are nitrated in the 4-position,15 NH2 HO,C-C/H 0 NHR H H H -r GR -+I) H and is also a feature of the most recent and most elegant scheme proposed by Robinson.1 The initial stage in this biosynthesis involves condensation of t,ryptophan and succinic acid to give the keto-acid intermediate (XI) which then gives the tricyclic intermediate (XII) by ring closure and dehydration.C02 H $02 H /CH2 /CH2 HBC NHMe "ZC\ / cw &O?H co - c,H NHMe COz H ,CH2 CHzo - GYM' - (VII) i X l l ! " I londensation with formaldehyde (or its equivalent) completes the synthesis. I'his ketonisation to give the hypothetical intermediate (XII) is analogous to i he proved formation of aminolzvulic acid from glycine and succinic acid :16 NH,*CH,CO,H + HO,CCH,*CH,*CO,H + NH,*CH,.CO*CH,*CH,*CO,H 15 Plant J.1929 2493 ; Mathur and Robinson J . 1934 1415. l6 Sherniri arid Russell J . Amer. Chew?,. SOC. 1953 75 4873. H 114 QUARTERLY RBVIERS These four biosyntheses postulate dissection of the lysergic acid molecule according t o the various modes annexed. While it is not possible a t present to appraise the relative merits of these theories they should stimulate investigations involving the application of tracer-element techniques. CO H COZ H COe H &->NMe 4 7 i i e &'=;Me dS 673 byf H H H Harley - Mason Wendler van Tamelen and Robinson Alkaloids of Evodia rut zcarpa and Crypt olepis spp.-Condensation of tryptamine with its biological degradation product anthranilic acid and formaldehyde or formic acid leads to the alkaloids evodiamine (XIII) and rutacarpine (XIV) which occur in the Chinese drug wu chu yu the dried fruit of Evodia rutmarpa.These formula? were proposed by Asahina and have been amply confirmed by synthesis.17 -+ / J MeN (XIII) 1 Cryptolepine Cl6HI2N2 was originally isolated by Clinquart from Cryptolepis trianguzaris N. E. Br. and differs from all known alkaloids in that it is dark violet giving rise to yellow salts. The constitution was elucidated by Gellert Raymond-Hamet and Schlittler,lgb who obtained it from C. sanguinoZenta. Dehydrogenation by selenium gave a colourless Asahina and Kashiwaki J . Pharm. SOC. Japan 1915 405 1293 ; Asahina &id. 1924 503 1 ; Asahina Irie a i d Ohta ibid. 1927 543 51 ; Asahina Manske and Robinson J. 1927 1708 ; Asahina and Ohta J . Pharm. Soc. Japan 1928 48 313 ; Ber.1928 61 319 ; Ohta J. Pharrn. Soc. Pormosa 1938 51 2 ; J . Pharnz. Xoc. Japan 1940 60 311 ; Schopf and Steuer Annulen 1947 558 124. (u) Clinquart B u l l . Acad. m e d . belges 1929 9 627 ; (b) Gellert Raymond-Hamet and Schlittler Helv. Chirn. Acta 1951 34 642. SAXTON THE TNDOLE ALKAT,OIT)S 115 base C 15H10N2 which was converted into cryptolepine hydriodide by reaction with methyl iodide showing that the degradntion simply involved removal of a methyl group froin a nitrogen atom. The parent ring system was identified as quindoline (XV) which was already known. Hence cryptolepine hydriodide is quindoline methiodide and cryptolepine is (XVI) . \ Me I - It is tempting to explain the origin of this alkaloid in the plant by con- densation of tryptophan with anthranilic acid or o-aminobenzaldehyde followed by loss of the ethanamine side chain but there is little or no evidence a t present to indicate that such a course of reaction is possible.There is no doubt that the mechanism of formation is much more complex and it may involve preliminary condensation of the aldehyde a t the 3-position of the indole nucleus followed by rearrangement and cyclisation with elimina- tion of the ethanamine side chain aJE:~;;o+.c H~.CH(NH$CGH C&CH( NH~CGH MeH H H OH Alkaloids of Australian Rufacez.-The Aust<ralian rain-forest tree Penfaceras australis Hook has yielded three alkaloids containing a ring system not hitherto found in Nature.lg The alkaloids have been shown by Price and his co-workers to be 6-oxocanthine (XVII) and its 5-methoxy- (XVIII) and 4-methylthio-derivative (XIX).The first two alkaloids occur in the leaves of the plant whereas the first and the last are present in the wood. l9 Haynes Nelson and Price Austrulian J . Sci. Res. 1952 5 A 387 ; Nelson and Price ibid. pp. 563 768. 116 QUARTERLY REVIEWS Oxidation of 6-oxocanthine gave P-carboline-1-carboxylic acid (XX) which was readily identified and hydrolysis gave the acrylic acid derivative QTP 0 6 5 a3 SMe Me jxvrr) (XVIII) (XIXI (XXI) which was obtained in both cis- and trans-forms. As expected only the former of these was reconvertible into the alkaloid. The second alkaloid 5-methoxy-6-oxocanthine also gave p-carboline-1- carboxylic acid on oxidation and a P-carbolinylacrylic acid derivative fJJENK_Mno- (XVII) EtOH aTQ H CH // H02C.CH H C02H (xx) (XXO (XXII) on hydrolysis which was easily reconverted into the alkaloid.Hence the methoxyl group must be at position 4 or 5. Hydrogen bromide demethyl- ated the alkaloid to anenol (XXIII) which readily condensed with o-phenyl- enediamine to give the quinoxaline (XXIV) showing that the methoxyl group is a t position 5. CH // HO2C.C.OMe (xx I I) 03 0 (XXIV) The third alkaloid isolated from this plant is unique in that it contains sulphur. It is slowly attacked by alcoholic alkali to give an acidic substance (XXV) and methanethiol. The acidic substance which gave p-carboline-l- carboxylic acid on oxidation did not condense with o-phenylenediamine and hence the second oxygen atom and the methylthio-group in the alkaloid must be at position 4. This was proved by synthesis of the derivative SAXTON THE INDOLE ALKALOIDS 117 W N // CVvle SMe HOZC'CH I NaOH-)( EtOH (xx) -*-- (E t 0 C) t H (XXV) and the alkaloid itself from p-carboline-1-carboxylic acid by the annexed react ions.Biogenetically these alkaloids could arise by condensation of tryptamine with glutamic or hydroxyglutamic acid. The former condensation leads to 118 QUARTERLY REVIEWS hexaliydro-6-oxocanthine which is then oxidised to 6-oxocanthine or to 5-methoxy-6-oxocanthine (via XXVI) which on methylation gives the alkaloid 5 -methoxy - 6 - oxocant hine. The second condensation leads directly to 4-hydroxy-6-oxocanthine (XXVII) which can be converted by oxidation and substitution of a methylthio-group for a hydroxyl group into the third alkaloid 4-methylthio-6-oxocanthine. Alternatively a tlhio-analogue of hydroxyglutamic acid may be used as starting material.Sir Robert Robinson1 prefers condensation of tryptamine or tryptophan and aspartic acid to give 6-oxocanthine directly. 5-Hydroxy-6-oxocanthine is then derived as above and the third alkaloid via the hydroxy-ketone (XXVII) the product of hydration of 6-oxocanthine. Alkaloids of Calycant hacez-The alkaloid calycanthine has been isolated from various species of Calycanthacez e.g. C. glaucus Willd. C. floridus L. C. occidentalis Hook and Am. and Meratia przcox Rehder and Wilson. 2o A second alkaloid calycanthidine has been isolated from C. ghucus Willd. Recently a third alkaloid folicanthine has been isolated by Eiter and Svierak from C. j?oridus.20 Calycanthine C,,H2,N4 is a diacidic base which contains two methylimino-groups.Its ultraviolet spectrum shows that it is a true dihydroindole and the infrared spectrum shows the presence of an NH group. Quantitative coupling indicates that the molecule contains two reactive para-positions and oxidation with potassium nitro- sodisulphonate shows that there are two NH groups attached to benzene rings.21 Benzoylation and oxidation of calycant'hine afford N-benzoyl-N- methyltryptamine (XXVIII) which was identified by synthesis.22 When heated with phthalic anhydride calycanthine yields a substance identical Calycanthine and calycanthidine. H (XXVIII) with that obtained from tryptamine and phthalic anhydride which has been formulated as (XXIX) .23 Benzoylcalycanthine when treated with soda lime affords 2-phenylindole and quinoline but on the other hand calycan- thine itself gives N-methyltryptamine and a base which is probably a methyl-/3-carboline.24 Degradation with lead copper oxide sulphur or selenium produces calycanine C,GH,ON,. Selenium also produces @-carbo- 2o Eccles Proc. Amer. Pharm. ASSOC. 1888 84 382 ; Manske J . Amer Chem. SOC, 1929,51,1836 ; Manske and Marion Canad. J . Res. 1939,17 B 293 ; Barger Jacobs and Madinaveitia Rec. Trav. chim. 1938 57 548 ; Eiter and Svierak Monatsh. 1951 82 186; 1952 83 1453. 2 1 Robinson and Teuber Cl~em. and Ind. 1954 783. 2 2 Manske Canad. J . Res. 1931 4 275. 28 Marion and Maimke ibid. 1938 16 B 432. 2 4 Barger Madinaveitia and Streuli J . 1939 810. 119 line skatole 3-ethylindole and lepidine. These confusing results were explained in various ways by different investigators but none of the earlier formuh was completely satisfactory.Robinson has proposed a sym- metrical oxidatively coupled N-methyltryptamine dimer structure (XXX) for calycanthine. 21 There are other possible formulz involving aa- or PP-coupling of the two indole nuclei but this one is preferred. The degrada- SAXT'ON THE INDOLE ALKALOIDS tion product calycanine can be explained as arising from fission of the MeN-C-N system followed by rearrangement of the bisindolenine inter- mediate (XXXI) loss of the two ethanamine chains and aromatisation. Calycanine (XXXII) is thus formulated as quinolino(4' 3'-3 4)quinoline ; this has now been confirmed by ~ynthesis.~4~ Oxidation of calycanthine with silver acetate produces a pyrrolo- quinoline identified by synthesis as 1'-methylpyrrolo(2' 3'-3 4)quinoline (XXXIV).25 This could arise from a hexahydro-@-carboline by oxidation and ring closure a reaction reminiscent of Witkop and Goodwin's ozonolysis experiments in the yohimbine series,26 but its formation can also be explained on the basis of formula (XXX) for calycanthine. Loss of an ethanamine chain from the intermediate (XXXI) followed by oxidative coupling of the other ethanamine chain t o the /3-position of the indole nucleus would give the hypothetical intermediate (XXXIII). Rearrangement to give a quinoline ring system analogous to the above rearrangement of (XXXI) followed by loss of aniline and aromatisat'ion would give the required pyrroloquinoline. Hence the formation of the substance (XXXIV) does not necessarily imply the presence of a P-carboline or pyrroloquinoline ring system in calycanthine.24a Clark and Woodward personal communication. 25 SpBth Stroh Lederer and Eiter Monatsk. 1948 79 11 17 ; Eiter and Nagy 26 Witkop and Goodwin J . Amer. Chem. SOC. 1953 75 3371. ibid. 1949 80 607. 120 QUARTERLY REVIEWS Calycanthidine C,,H,,N, the second alkaloid obtained from C. glaucus Willd. is a dihydroindole containing a NH group linked directly to a benzene ring It affords a quinone on oxidation and couples with diazonium salts.21 Zinc dust distillation gave norharman and since the alkaloid contains no C-methyl group and no double bond it must be represented by a bridged-ring formula e.g. (XXXV). The position of the bridge and the indole structure require confirmation. H (xxxv) Me (XXXVI) (XXXVII) (xxxv II I ) FoEicanthine. The constitution (XXXVI) has been suggested for foli- canthine C18H2,N3 by Eiter and Svierak from the following evidence.,O It was degraded by hydrogen chloride to a base C,,H,,N, which was formulated as 2-3'-aminopropyl- 1 -methylindole (XXXVII) although it was not unequivocally identified.This indole base was also obtained by acetylation of the alkaloid followed by hydrolysis. I n contrast to this folicanthine itself was unaffected by alkalis. Hofmann degradation of folicanthine methiodide gave a base C,,H,,N, identical with that obtained by methylation of the base (XXXVII) and oxidation of the alkaloid by silver acetate gave dehydrofolicanthine formulated as (XXXVIII). Eiter and Svierak's formula requires this alkaloid to be an indole derivative con- taining one N-methyl group but the evidence recorded by these authors would seem to indicate a dihydroindole structure (ultraviolet spectrum) containing two N-methyl groups.The Yohimbine Group of Alkaloids.-By far the majority of the indole alkaloids can be regarded as originating in the plant from tryptophan dihydroxyphenylalanine and formaldehyde or their biochemical equivalents. SAXTON THE INDOLE ALKALOIDS 121 U Sempervirine Cinchonine OH Yohimbine,etc. OH I i / S-Yoh imbi ne etc Alstonine J. serpentine) (s) +Deserpidine reserpine and rescinnamine in methoxylated series) 0 Meltnonine - A ( 122 QUARTERLY REVIEWS The primary product of this condensation could be either of the dihydric phenols (XXXIX) or (XL) depending on whether reaction occurs at the 2- or the 3-position of the indole nucleus. The former possibility leads to the yohimbine series of alkaloids (called by Robinson the a-series) and may even lead to such apparently unrelated alkaloids as cinchonine in which the indole moiety has been converted into a quinoline derivative.The principal transformations of the primary condensation product (XXXIX) are given in the accompanying scheme. Occasionally mono- or di-meth- oxylated derivatives occur in Nature alongside the unsubstituted alkaloids but their biogenesis offers no difficulty since we can equally well start with a mono- or a di-methoxytryptamine. It is frequently possible to explain the formation of an alkaloid by alternative routes and these have in some cases been indicated. One point of unusual interest is the origin of the methoxycarbonyl group in yohimbine and its congeners.It could arise by simple condensation of the intermediate (XXXIX) with formal- dehyde followed by oxidation and methylation. An alternative suggestion by Robinson postulates conversion of the intermediate (XXXIX) into a tropolone by introduction of a single carbon atom. Reduction and a benzilic acid type of rearrangement then lead to yohimbine. It should be noted that in contrast to many natural processes this series of transformations appears not to be stereospecific. Thus for example nine stereoisomers of yohimbine are known and these differ not simply in the orientation of ring substituents but also in the stereochemistry of the C-D and D-E ring junctions. The formation of structure (XL) by the alternative condensation of dihydroxyphenylalanine with tryptophan and formaldehyde was postulated by Woodward to account for the biogenesis of the Xtrychnos alkaloid^.^' (4 Fission of the aromatic ring between the hydroxyl groups gives an inter- mediate which on condensation with an acetic acid equivalent leads to strychnine directly.The probability that this convincing and revolutionary theory was essentially correct was increased when it was applied by Robinson with marked success to derive the structure of emetir~e.~' It was still further increased when laboratory analogies for the formation of the com- pound (XL) were realised. 28 These two reactions namely the /I-condensation and the fission of the aromatic ring are probably the key stages in the formation of the other 27 Woodward Nature 1948 162 155 ; Robinson ibid. p. 524. 2* Robinson and Saxton J .1953,2596 ; Woodward Cava Ollis Hunger Daeniker and Schenker J . Amer. Chem. SOC. 1954 76 4749. SAXTON THE INDOLE ALKALOIDS 123 dihydroindole alkaloids. Unfortunately the structures of many of them e.g. gelsemine akuammine and aspidospermine are still unknown. In these examples Woodward’s theory is used as an invaluable aid in deriving the structures of the alkaloids to act as a basis for discussion and investiga- tion until further evidence becomes available. The features of both types of alkaloids may be combined in ajmaline (see below) which results from an internal #?-condensation in an intermediate (XLI) of the %-series. It is necessary to emphasise here that although biogenetic schemes are frequently written with definite chemical substances as intermediates it is not intended to convey the impression that these specific entities are involved to the exclusion of all others.For example the useof dihydroxyphenyl- alanine in the early stages implies the participation of this or of any biochemical equivalent e.g. dihydroxyphenylacetaldehyde. Similarly the use of formaldehyde implies the intervention of formaldehyde or any possible equivalent or progenitor e.g. glycine. The theories are justified by the structural relations observed between the alkaloids and by the (at present) limited evidence from experiments in vivo (e.g. the conversion of tryptophan into gramine in barley). The results of laboratory analogies in this respect (e.g. Robinson’s synthesis of tropinone 28a) are frequently encouraging although the greatest caution must be exercised in their interpretation and application to reactions in vivo.Further these speculations do not even presume an exact order in which the various stages occur particularly the simpler ones of methylation or acetylation which can be expected to proceed a t any convenient stage in the biosynthesis. Thus as Robinson points out,28b all the alkaloids in this series contain the “ berberine bridge ” carbon atom provided by the formaldehyde equivalent. This may mean the initial condensation of dihydroxyphenylalanine with formaldehyde to give an intermediate of type (XLB) which subsequently condenses with tryptophan to yield (XXXIX). The Alkaloids of Yohimbehe Bark.-Yohimbehe bark is themain source of yohimbine and its congeners and is obtained from a tree (Pausinystalia yohimba Pierre; syn.corynanthe yohimbe K. Schum.) found in the Came- roons and the French Congo. Included among the yohimbehe alkaloids are those obtained from Pseudocinchona africuna A. Chev. Of the thirteen alkaloids isolated from these sources nine are stereoisomers of yohimbine and by 1950 their structure was firmly established as (XLII). The final link in the chain of evidence leading to this structure namely the proof of the position of the hydroxyl group was provided by Swan who achieved the total synthesis of yohimbone (XLIII).29 The nuclear structures of the alkaloids having been established attention was directed towards the configurations of the asymmetric centres. Witkop was the first to adopt this approach and obtained evidence relating to the stereochemistry of the D-E ring junction by degrading yohimbic acid to an optically active trans-decahydro-N-methylisoquinoline (XLIV) and 3-vinyl- 286 Ref.1 p. 124. 20 Swan J. 1950 1534. 2*a Ref. 1 p. 63. 124 QUARTERLY REVIEWS indole (not isolated). Hence in yohimbine the rings D and E must be truns-fused provided that no change in stereochemistry has occurred during the degradation.30 Utilising this result and the methods of conformational analysis Janot and Goutarel and their co-workers have derived the stereochemistry of yohimbine and several of its isomers.31 Thus since yohimbine and y-yohim- bine give the same tetradehydroyohimbine (XLV) with lead tetra-acetate these two bases differ only in the configuration a t position 3. Catalytic reduction of this tetradehydroyohimbine regenerated yohimbine which is presumably the more stable isomer and therefore contains the greater number of equatorial carbon-carbon bonds as shown in (XLVI).The configurations of the hydroxyl and methoxycarbonyl groups remain to be determined. Since corynanthine can be converted into yohimbine by alkaline hydrolysis and re-esterification these two substances are identical with the exception of the configuration at position 16. The hydroxyl and the methoxycarbonyl group are therefore cis in one isomer and trans in the other. The different behaviour of the hydrogen sulphates of yohimbine and corynanthine towards dilute alkali allows the conformations to be deter- mined yohimbine hydrogen sulphate (XLVII ; R = S03H) gives an unsaturated acid whereas corynanthine hydrogen sulphate (XLVIII ; R = S03H) gives an unsaturated hydrocarbon by simultaneous decarboxy- lation.Since this reaction proceeds by elimination of axial groups it may be deduced that the hydroxyl group and the 16-hydrogen atom are in the axial positions in yohimbine and that the hydroxyl and the methoxycarbonyl group are in the axial positions in corynanthine. This conclusion is con- firmed by the readier hydrolysis of the equatorial ester group i.e. yohimbine 30 Witkop J. Amer. Chem. SOC. 1949 71 2559. 31 Janot Goutarel Le Hir Amin and Prelog Bull. SOC. chim. France 1952 1085 ; Le Hir Janot and Goutarel ibid. 1953 1027 ; Le Hir and Goutarel ibid. p. 1023 ; Bader Dickel Huebner Lucas and Schlittler J. Amer. Chem. Soc. 1955 77 3547. SAXTON THE INDOLE ALKALOIDS 125 should be more easily hydrolysed than corynanthine as found by experi- ment.Hence yohimbine is (XLVII ; R = H) and corynanthine is (XLVIII ; R = H). Cookson and Klyne have arrived at the same relative configurations and the latter has also presented evidence that (XLVII) represents the absolute configuration of yohimbine. 32 Similar reasoning allows the conformations of p-yohimbine (XLIX) y -yohimbine (L) alloyohimbine (LI) a-yohimbine (corynanthidine) (LI) and 3-epi-a-yohimbine [C,, epimer of (LI)] to be established with the exception of the methoxycarbonyl group in the three last-named alkal0ids.~1 Thus derivatives of all four possible yohimbanes have been found in Nature. OR (XLVl I I) c ":p H02C Two alkaloids of RauwolJa serpentina Benth.-isorauhimbine 33 and aerpine 34-have also been formulated as stereoisomers of yohinzbine. It has been suggested that the latter differs from v-yohimbine only in the configuration of the methoxycarbonyl group.The elucidation of the stereochemistry of these alkaloids stimulated attempts at stereospecific syntheses and a certain amount of success has c 32 Cookson Chem. and Ind. 1953 337 ; Klyne ibid. p. 1032. 33 Hofmann Helv. Chim. Acta 1954 37 314. 34 Chatterjee and Bose Experientia 1964 10 246. 126 QUARTERLY REVIEWS already been achieved. Att'ention was naturally paid initially to the syn- t'hesis of the four yohimbanes. aZEoYohiiiibnne (LIII) the first isomer to be synthesised had already been obtained by direct hydrogenation of semper- virine (LII),35 and it was prepared later together with its stereoisomer 3-epialloyohimbane (LIIIA) by synthesis from cis-perhydroindan-2-one (LIV).36 An analogous synthesis starting from trans-perhydroindan-2-0ne~ led to (-+)-yohimbane [epimeric with (LIII) at C,,? Comparison of synthetic aZZoyohimbane and yohimbane with material obtained from the related alkaloids demonstrated that the derived stereochemistry of the C-D and D-E ring junctions in these alkaloids was correct.* (LIV) Corynantheine b-yohimbine and their derivatives. Introduction of a carbon atom into the aromatic ring E of the intermediate (XXXIX) followed by conversion into an ester and Woodward-type fission between the hydroxyl groups leads to a hypothetical intermediate written formally as (LV) which may be the precursor of corynantheine (LVI) and many alkaloids in which ring E is heterocyclic. Corynantheine the only tetracyclic alkaloid of this series may thus be regarded as the biogenetic lirlk between the yohimbine isomers on the one hand and the d-yohimbine type with a heterocyclic ring E and the cinchonine type in which further complex transformations have occurred on the other.Corynantheine C22H'&O,"& obtained from Pseudocinchona africana is a tertiary indole base which gives alstyrine (LVII) on dehydrogenation by selenium. A whole series of transformations proved conclusively the 35 Le Hir Janot and Goutarel Bull. SOC. chim. (Prance) 1952 1091. 36 Stork and Hill J . Amer. Chern. Soc. 1954 76 949 ; van Tamelen and Shamma ibid. p. 950 ; Janot Goutarel Le Hir Tsatsas and Prelog Helu. Chim. Acta 1955 38 1073. * The assumption that hydrogenation of sempervirine gives a syn-cis-product and hence the accepted stereochemistry a t C(3j in alloyohimbine a-yohimbine 3-epi-u- yohimbine reserpine deserpidine and rescinnamine have been challenged by Janot Goutarel Le Hir Tsatsas and Prelog who deduced from a study of the molecular rotation changes observed during dehydrogenation of yohimbane and aZZoyohimbene that these two substances possess the same configuration at C(3).36 Hence these authors believe that all the above-mentioned alkaloids are epimers at C(3) of the generally accepted formulations. I n the absence of further evidence t8he earlier structures are used here. . SAXTON THE INDOLE ALKALOTDS 127 presence of the grouping Me0,C.C CH*OMe and when the base was finally obtlained analytically pure the presence of a vinyl group was also demon- strated. Many early specimens of the alkaloid were contaminated with dihydrocorynantheine which accompanies it in the plant.The mixture therefore yielded formaldehyde on ozonolysis and acetic acid in the Kuhn- Roth determination. The constitution (LVI) for corynantheine based on these results by Janot Goutarel and P r e l ~ g ~ ' was confirmed by degrada- tion of t,he alkaloid to 3-ethyl-4-isopropylpyridine and de-ethylalstyrine (LVIII) which were identified by synthesis. 37 Et (LVII) (LVIII) d-Yohimbine C21H2403N2 was first isolated from the mother-liquors of yohimbine preparations and has more recently been obtained from RauwoZ$a serpentina and named ajmalicine and raubasine. 38 Its ultraviolet spectrum is typical of an indole containing the system Me0,C.C C-OR with maxima a t 230 and 290 mp and an inflexion near 250 mp.The infrared spectrum L'x) Me0,C U' 37 Prelog Karrer and Enslin Helv. Chint. Acta 1949 32 1390 ; Chattcrjee arid Karrer ibid. 1950 33 802 ; Janot and Goutarel Bull. SOC. chinh. (France) 1961 588 ; Janot Goutarel and Prelog Helv. Chim. Acta 1951 34 1207 ; Karrer and St. Mainoni ibid. 1953 36 127 ; Prelog Janot Goutarel and Mirza ibid. p. 337 ; Karrer Schwyzer and Flam ibid. 1952 35 851 ; Janot and Goutarel Compt. rend. 1944 218 852 ; Karrer Blumenthal and Eugster Helv. Ghiwz. Acta 1964 37 787 ; Janot Goutarel and Chabasse-Massoiineau Bull. SOC. chim (France) 1953 1033 ; Anderson Clemo and Swan J. 1954 2962. 38 Siddiqui and Siddiqui J . Indian Chent,. Soc. 1931 8 (67 ; Heinemann Ber. 1934 67 15 ; Raymond-Hamet and Goutarel Conapt. rend. 1931 233 431 ; Goutarel and Le Hir Bull. SOC.chim. (France) 1951 909 ; Klohs Drap?r Keller Malesh and Petracek J. Amer. Chem. SOC. 1954 76 1332 ; Popelak Spingler anti Kaiser Natctr- wiss. 1953 40 625. 128 QUARTERLY REVIEWS confirms the presence of this grouping with characteristic twin peaks a t 5-89 and 6-21 ,LL. Selenium dehydrogenation gave alstyrine (LVII). Since the alkaloid contains C-Me but no isolated double bonds it must be penta- cyclic and was formulated by Goutarel and Le Hir as (LIX).38 The pro- duction of alstyrine on degradation is characteristic of all the alkaloids in this series in which ring E is opened or heterocyclic and is in striking contrast to the behaviour of yohimbine and its isomers which on similar treatment yield a mixture of yobyrine (LX) tetrahydroyobyrine (LXI) and keto- yobyrine (LXII).The alkaloids mayumbine from Pseudocinchona mayumbensis and akuammigine from Picralima nitida are formulated as stereoisoniers of d- yohimbine . 39 Alkaloids of Rauwolfia Species.-These alkaloids are obtained from the various species of RauwoZJia and in particular from R. serpentina Benth. indigenous t o the Dehra Dun valley or the Bihar district of India and from R. canascens Linn. Extracts of R. serpentina have been used medicinally for centuries in India. The drug has been prescribed for various disorders e.g. as a febrifuge as a cure for dysentery and as a hypnotic and sedative. It is also recommended for insomnia hypochondria and some forms of insanity but its most important action consists in its ability to reduce the blood pressure. The plant extracts vary somewhat in pharmacological activity depending on their origin those collected from the Dehra Dun valley being more active as a sedative and less active in the treatment of insanity than those obtained from the state of Bihar.This indicated the presence of several active principles in varying proportions in the different specimens and stimulated the chemical and pharmacological investigations which have been intense during the last few years. No less than 24 alkaloids have been isolated from this species alone although as yet some of them are not well known or characterised. The Table opposite gives a list of alkaloids isolated from RauwoZJia species complete up to October lst 1965. This group of alkaloids has recently been reviewed by Schlitt'ler Schneider and Pluninzer and by ChatterjeegO The versatility of the Rauwolufia species in respect of their pharmacological properties is paralleled by their ability as biosynthetical agents.In contrast to yohiinbehe bark which contains only alkaloids of the a-series extracts of RauwoZJia yield alkaloids of both the a- and the @-series together with mono- and di-methoxylated derivatives of the parent alkaloids. The presence of the unrelated alkaloids thebaine and papaverine has even been reported but this may prove to be due to contamination of the samples by opium Hofmann and Chatterjee and Talapatra report t,hat these two alkaloids were not present in their plant extracts.41 39 Raymond-Hamet Conapt. rend. 1951 232 2354 ; Janot Goutarel and Masson- neau ibid. 1952 234 860 ; Robinson and Thomas J . 1954 3479. 40 Schlittler Schneider and Plummer Angew.Chem. 1954 66 386 ; Chatterjee " Progress in the Chemistry of Natural Products " Vol. X Springer Verlag Vienna 1953 p. 390. 41 Hofmann Helv. Chirn. Acta 1954 37 849 ; Chatterjee and Talapatra Nuturwiss. 1955 42 182. SAXTON THE INDOLE ALKALOIDS Alkaloids of Rauwolfia species 129 Ajmalicine (6-yohimbine raubasine) . . . . Ajmaline . . . . . isoAjmaline . . . . neoAjinaline . . . . Ajmalinine. . . . . Alstonine . . . . . Aricine . . . . . . Corynanthine (rauhimbine) Deserpidine (canescine) . Methyl reserpate . . . Papeverine . . . . Perakenine. . . . . isoRauhimbine . . . Raumitorine . . . . Rauwolfinine . . . . Rauwolscine (a-yohimbine) Rescinnamine . . . . Reserpiline. . . . . isoReserpiline . . . . Reserpine . . . . . Reserpinine (raubasinine) isoReserpinine.. . . Sarpagine (raupine) . . Semperflorine . . . . Seredine . . . . . Serpentine . . . . . Serpentinine . . . . Serpine. . . . . . Serpinine . . . . t Tetraphyllicine ** Tetraphylline . . Thebaine . . . Yohimbine. . . alloYohimbine . . 3 -epi - ct -Y ohimbine p-Yohimbine . . y-Yohimbine . . $-Yohimbine . . Unnamed . . . . . . . . . . . . . . . . . . . . . . . Formula 21H 2 4 O 3N2 C20H 2 6 O ZN2 C20H2602N2 ,OH 2 6 O ZN2 C20H2603N2 21H 20' BN 2 2ZH 26' 4N 2 C21H 2 6 O 3N 2 c3 ZH 2 2sH3Oo 5N 2 C211H2104N C21H2603N2 2 2H 2 6 O aPIT 2 C1ilH260 ZN2 21H260 3N2 35H4209N 2 C23H 2 1 3 ~ 5 ~ 2 2sHm0 gN2 C33K4009N2 (= 2ZH260 4N2 2 ZH 26O 4N 2 21H 2 6 0 N 2 23H SO0 gN 2 21H 20° 3N2 C21H 22' 3N2 C21H2603N2 (=ZOH24ON2 C 1 9 H 2 2 O 2 N 2 DT C2,H,,ON c 2 0 H 2 4 0 N 2 (= 22H 2 6 O bN 2 C19H2103N 21H 26°3N2 C21H2603N2 21H2603N2 ,1H2603NB ,IH 2 1 3 ~ 3 ~ 2 C21H 26°3N2 B1.p 250-252" * 158-160 264-266 * 205-207 180-181 254 * 188 * 218-225 228-232 * 244-245 147 236 * 225-228 138 235-236 * 231-232 * 2 24-2 2 6 Amorphous 211-212 * 263 238-239 225-226 * 325 * 295 * 291 158 213 315 * 263-265 3 20-322 220-223 * 2 35-2 3 7 135-136 181-1 83 246-249 2 5 8-2 5 9 265-278 195 323 Rotation t - 58.1" (C) + '72.8 (E) + 128 (C) - 97 (C) - 58.3 (E) - 82 (P) - 137 (C) - 106 (P) 0 - 104 (P) - 34.7 (E) - 40 (E) - 38 (E) - 82 (P) - 117 (C) + 60 (C) - 98 (C) + 21 (PI - 73 (C) + 105 (P) - 90 (P) + 27 (PI - 279 (P) - 72.7 (P) - 48 (P) - 28.3 (P) Source 2 * With decomp.t Solvents A aqueous acetic acid ; C = chloroform ; E = ethanol ; M = meth- anol ; P = pyridine 2 c = R. canescens. ht = R.hirsuta. s = R. serpentina. cf = R. caffra. m = R. micrantha. d = R. densijlora. o = R. obscura. t = R. tetraphylla. se = R. semperylorens. h = R. heterophylla. p = €2. perakensis. v = R. vomit~ia;. ** This is now known to be didehydrodeoxyajmaline ; reuvomitine a recently isolated alkaloid is probably its tri-0-methylgalloyl derivative (Djerassi Gorman Pakrashi and Woodward personal communication). Several of tb. 2 Sllkaloids e.g. yohimbine alloyohimbine and Q-yohim- bine have already been discussed and hence need no further comment. The monomethoxy-derivatives of 6- yohimbine are represented by reserpinine I 130 QUARTERLY REVIEWS (from R. serpentina) and isoreserpinine (from R. canascens) which are formulated as stereoisomers of 11 -met.hoxy- 6-yohimbine (LXIII ; R = H R’ = OMe).38 41 42 It is noteworthy that reserpinine has also been “-Q R :lQ-$ N CH ct %; 0 O-CHE ~LXIV):R- R ~ = H (LXIVB):RIH,R=OM~ ( LX IVA) R-H R L O M ~ ( LX IVC) R- R’ = o M= (LXIII) Meo2C obtained from Vinca major L.collected in Normandy by Janot and Le Men 43 ; this is the first reported occurrence of an alkaloid of the yohimbine series in a plant indigenous to Europe. Aricine (10-methoxy-6-yohimbine) (LXIII R = OMe R’ = H) occurs in R. canascens and Cinchona pelletierana Wedd.,42 44 and its stereoisomer raumitorine in R. ~ o r n i t o r i a . ~ ~ The series is completed by the extraction of reserpiline (from R. serpentina) and isoreserpiline (from R. canascens) which are stereoisomers of 10 ll-dime- thoxy-6-yohimbine (LXIII R = R’ = OMe).429 46 Thus we find a quartet of alkaloids in the RauwoZJia species completely analogous to the Xtrychnos alkaloids strychnine (LSIV) a- and P-colubrine (LXIVA and B respectively) and brucine (LXIVC).As with the yohimbines the stereoisomerisni within the series illustrates the fact that the biogenetic pathway is not stereospecific. The Xtrychnos alkaloids on the other hand in common with most alkaloids of the #I-series must be produced by a much more selective mechanism since such stereoisomerism is not encountered. Taking cognisance of the fact that dehydrogenation reactions frequently participate in natural processes (cf. the biosynthesis of nicotine and papa- verine) it would be surprising if none of the alkaloids of this group existed in a more highly oxidised state than that represented by yohimbine or d-yohimbine.Several such alkaloids are known which are coloured and belong to the class of anhydronium bases. Their true constitution is inter- mediate between the zwitterion structure and the alternative quinonoid form. Sempervirine (LII) is the simplest of these alkaloids and occurs in Gelsemiurn sernpervirens and Mostuea buchh0lxii.4~ In view of the compre- 4 2 Weisenborn Moore and Diassi Chem. and Ind. 1954 375 ; Schlittler Saner and Muller Experientia 1954 10 133 ; Stoll Hofmann and Brunner Helv. Chiin. Acta 1955 38 270. 4 3 Janot and Le Men Compt. rend. 1954 238 2550 ; 1955 240 909. 44 Pelletier and Corriol J . Pharm. 1829,15 565 ; Goutarel Janot Le Hir Corrodi and Prelog Helv. Chirn. Acta 1954 37 1805 ; Raymond-Hamet Compt. rend. 1945 221 307. 45 autarel Le Hir Poisson and Janot Bull.SOC. chim. (France) 1954 1481. 46 Klohs Draper Keller and Malesh Chem. and Ind. 1954 1264. 47 Forsyth Marrian and Stevens J . 1945 579 ; Goutarel Janot and Prelog Experientia 1948,4,24 ; Prelog Helv. Chirn. Acta 1948,31,588 ; Bentley and Stevens Nature 1949 164 141 ; Woodward and Witkop J . Arner. Chem. SOC. 1949 ’71 379 ; Woodward and MacLamore ibid. p. 379 ; Gellgrt and Schwarz Helv. Chim. Acta 1951 34 779. SAXTON THE INDOLE ALKALOIDS 131 hensive range of alkaloids isolated from Rauwolfia species it is perhaps surprising that this comparatively simple one does not occur there. Its structure was derived from a study of its dehydrogenation products (LX- LXII) which are typical of the pentacyclic yohimbane system and its spectra and was confirmed by Woodward’s synthesis of sempervirine methochloride and sempervirine itself.(LXV) ( LXV A ) Serpentine serpentinine and alstonine. The coloured anhydronium base alkaloids are represented in the RauwolJia series by serpentine serpentinine and alstonine. The last -named was originally found in Alstonia constricta F. Muell. in which it is the principal alkaloid. It is of interest that alstonine has not been found in any of the remaining ten Alstonia species which have been investigated. The structure of serpentinine is at present obscure although it is evidently closely related to serpentine and alstonine (LXV) which are stereo- is0mers.~8 48 The constitutions of these two alkaloids were derived in- dependently and indeed it was only very recently realised that they are isomeric since serpentine was initially believed to contain two hydrogen atoms more than alstonine.The isolation of alstyrine (LVII) on dehydrogenation of serpentine by selenium and the close similarity of the spectra of the alkaloid and tetra- dehydroyohimbine (XLV) indicated the presence of the partial structure (LXVA). Proof of the presence of an ester group an ether link C-Me and finally a double bond indicated the formula (LXV) which was proved by identification of py-tetrahydroserpentine with d-y~himbine.~*> 49 The reputation of Alstonia barks in the Far East as an antimalarial and febrifuge stimulated investigation of the alkaloidal constituents in the search for possible quinine substitutes. However reliable pharmacological experi- ments have failed to substantiate this claim and AZstonia extracts are no 48 Schlittler Huber Bader and Zahnd Helv.Chim. Acta 1954 37 1912. 49 Schlittler and Schwarz ibid. 1950 33 1463 ; Bader and Schwarz ibid. 1952 35 1594. I* 132 QUARTERLY REVIEWS longer included in the official British Pharmacopceia. These investigations resulted in the isolation of several minor alkaloids but of these only alstoniline has been studied in any detail. The presence of the partial structure (LXVA) in alstonine was readily proved but the constitution of ring E presented more difficulty until it was realised that the anomalous behaviour of tetrahydroalstonol (obtained by reduction to a tetrahydrocarboline and the change (C0,Me -+ CH,*OH) was due to the presence of the grouping *O*C C*CO,Me which on reduction gave a labile ally1 alcohol derivative which rearranged in acid solution and yielded ethers with comparative ease.Comparison of the spectra and reactions of tetrahydroalstonine with those of the model compounds (LXVI) and (LXVII) by Bader proved conclusively the position of the double bond and the structure of alstonine was firmly established as (LXV). 50 Me4 Me Me Oy%Oyb MeOzC (LXVI) (LXVII) ( LXVI I I) (LXIX) Reaction of tetrahydroalstonine with methyl chloride yields the quater- nary salt melinonine-A which occurs in the bark of South American Xtrychnos melinoniana Baillon. 51 The alkaloid alstoniline C,,H,,O 3Nz*OH is characterised by the brilliant red colour of its salts the hydrochloride being obtained directly from the bark without previous addition of acid. It thus appears that alstoniline exists in Nature as the chloride a very rare occurrence in the alkaloid series.In spite of the fact that in alstoniline chloride the ring system is in a very highly dehydrogenated state this compound does not contain the chromo- phore of alstonine since the spectra of the two alkaloids are quite different. On the other hand the spectra of alstoniline and tetrahydroalstoniline resemble those of ketoyobyrine (LXII) and 6-methoxyindole respectively. Surprisingly dehydrogenation by selenium gave no identifiable products but potash fusion gave 2-methylisophthalic acid. Elderfield and Wythe’s tenta- tive formula (LXVIII) for alstoniline chloride is supported by comparison of its spectrum with that of 3-(6-methoxy-3-methyl-2-indolyl)-2-methyl- isoquinolinium iodide (LXIX) .52 Hence in the formation of alstoniline chloride from an intermediate of type (XXXIX) ring c has remained hydro- aromatic and ring E aromatic (although this may involve reduction and 6o Sharp J.1934 287 ; Sharp J. 1938 1353 ; Leonard and Elderfield J. Org. Chem. 1942 ‘7 556 ; Elderfield and Gray ibid. 1951 16 506 ; Bader Helv. Cizim. Acta 1953 36 215. 61 Schlittler and Hohl Helv. Chirn. Acta 1952 35 29. 62 Hawkins and Elderfield J. Org. Chem. 1942 ‘7 573 ; Elderfield and Wythe ibid. 1954 19 683 693. SAXTON THE INDOLE ALKALOIDS 133 re-aromatisation stages) while ring D has suffered dehydrogenation. The fully aromatic system in which ring c is also aromatic has not so far been found in Nature. Reserpine rescinnumine and deserpidine. Retention of all three sub- stituents in ring E of (XXXIXA) followed by further obvious transforma- tions leads to the important drug deserpidine (canescine) (LXX) found in Rauwolfia canuscens.The methoxylated derivative reserpine (LXXI) ocours in R. serpentina and several other RauwoZjia species while an analogous derivative of trimethoxycinnamic acid rescinnamine (LXXII) is also present in R. serpentina and R. v0rnitoria.5~ All three alkaloids are extremely valuable hypotensive and sedative agents and the limited amounts available have led to the prohibition of their export from the Indian sub-continent. In this connection the isolation of reserpine from Australian Alstonia con- stricta is important since it may provide an alternative source of this alkaloid for medical use.54 Reserpine C33H4009N2 the first of these three alkaloids to be isolated was easily shown to be an ester alkaloid by hydrolysis to reserpic acid (LXXIII) and trimethoxybenzoic acid.Rescinnamine similarly gave reserpic acid and trimethoxycinnamic acid. Reserpic acid was shown to be a derivative of yohimbane by degradation to 4-methoxy-N-oxalyl- anthranilic acid 5-hydroxyisophthalic acid and yobyrine (LX) and by its colour reactions which were characteristic of a tetrahydro-p-carboline. The relative positions of the hydroxyl and the carbonyl group were indicated by formation of a y-lactone and by isolation of 5-hydroxyisophthalic acid. Gs Muller Schlittler and Bein Exper;entia 1952 8 338 ; Haack Popelak Spingler and Kaiser Natumiss. 1954 41 214 ; Klohs Draper and Keller J . Amer. Chem. SOC. 1954 76 2843 ; Schlittler Ulshafer Pandow Hunt and Dorfman Experientia 1955 11 64 ; Stoll and Hofmann J .Amer. Chem. SOC. 1955 7'7 820. 5 4 Report from C.S.I.R.O. Melbourne quoted in the London Times May 26th 1955. 134 QUARTERLY REVIEWS These results enabled Schlittler and his co-workers and Neuss Boaz and Forbes to propose formula (LXXI) for reserpine.55 The methoxyl group was placed at position 17 for purely biogenetic reasons but was supported by conversion of methyl 0-toluene-p-sulphonylreserpate (LXXIV) into a derivative (LXXV) containing the chromophore Me0,C.C C-OMe which by acid hydrolysis and decarboxylation yielded reserpone (LXXVI). The position of the carboxyl group and hence the positions of the other substituents was proved by degradation of reserpinol (LXXIII ; CH,*OH in place of C0,H) to 7-hydroxymethylyobyrine (LXXVII) which was identified by synthesis of its methyl ether.56 OW? OMe OMe ( LXXV) (LXXV I II 1 (LXXVII) 5 H" 20 l5 I9 R - tri - 0- methyl -galloyl By analogy with this deserpidine which exhibits very similar pharmaco- logical properties to reserpine was formulated as (LXX).55 The stereo- chemistry of the ring system was elucidated by conversion of deserpidine into a-yohimbine a derivative of uZbyohimbane (LIII).Since the alkaloid is known to have the less stable configuration at position 3 and since one of the stages in this series of transformations involves epimerisation at this centre deserpidine and hence reserpine can be formulated as derivatives of 3-epiaZbyohimbane. The configurations of the substituents in ring E remain to be determined. Since reserpic acid readily gives a lactone it can be assumed that the carboxyl and the hydroxyl group are in the cis-position relative to one another.These conclusions were supported by Diassi 5 5 Furlenmeier Lucas MacPhillamy Muller and Schlittler Experientia 1953 9 331 ; Dorfman Huebner MacPhillamy Schlittler and St. Andre ibid. p. 368 ; Dorf- man Furlenmeier Huebner Lucas MacPhillamy Muller Schlittler Schwyzer and St. And-6 Helu. Chim. Acta 1954 37 59 ; Neuss Boaz and Forbes J . Amer. Chem. SOC. 1954 76 2463 ; Schlittler MacPhillamy Dorfman Huebner and St. Andre ibid. 1955 77 1071. 66 Huebner MacPhillamy St. Andre and Schlittler J . Amer. Chem. SOC. 1955 77 472 ; Schlittler MacPhillamy Dorfman Huebner and St. Andre ibid. p. 1071 ; Diassi Weisenborn Dylion and Wintersteiner ibid. p. 2028. SAXTON THE MDOLE ALEALOIDS 135 Weisenborn Dylion and Wintersteiner who obtained the quaternary salt (LXXVIII) in addition to the unsaturated ester by removal of the toluene- p - sulp hony 1 group from met hy 1 0 -toluene -p - sulph on ylr eserpat e .Since quaternary salt formation was assumed to involve inversion a t C(ls) the 18- oxygen bond must be cis with respect to the 15- and 20-hydrogen atoms. Finally if trans-elimination occurs in formation of the unsaturated ester (LXXV) then the 17-methoxyl group must also be cis with respect to the hydroxyl and the carboxyl group. Reserpine was therefore completely represented by formula (LXXIA) .56 More recent studies however have shown that these deductions were not entirely correct. The stereochemistry of the D-E ring junction has been confirmed by Huebner's synthesis of (A)- reserpane [l l-methoxy-3-epiaZZoyohimbane (LXXVI) with CO -+ CH,],56° but the spontaneous quaternisation which occurred on treatment of reser- pinol (LXXVIIIA ) or 3-isoreserpinol with toluene-p-sulphonyl chloride to give the mixed salt (LXXVIIIB ; X- = C1- or OTs-) indicated that the C(16)-C(22) bond must be trans with respect to the 15- and 20-hydrogen atoms instead of cis as a t first Since reserpine can be hydrolysed and methyl reserpate can be treated with sodium methoxide in boiling methanol without inversion the methoxycarbonyl group must be equatorial.Application of Hudson's lactone rule to reserpic lactone indicates that the 18-hydroxyl group has the ,&configuration and therefore reserpic acid is correctly represented by (LXXVIIIC) with reservations with regard to the configuration of the methoxyl group.According to van Tamelen and Hance this methoxyl group is probably trans with 18-substituents and the formation of (LXXV) respect to the 16- and and (LXXVIII) from participation. Quater- (LXXIV) probably occurs by neighbouring-group L nisation therefore proceeds with double inversion and retention of 56aHuebner Chem. and Ind. 1955 1186. 66b Huebner and Wenkert J . Amer. Chem. SOC. 1955,77,4180 ; Diassi Weisenborn Dylion and Wintersteiner ibid. p. 4687 ; van Tamelen and Hance ibid. p. 4692. 136 QUUTERLY REVIEWS configuration at C(ls). The all-trans-arrangement also explains the ready isomerisation of reserpine and its analogues to alloyohimbane derivatives. 56c Dihydroindole Alkaloids of the RauwoZfla Species.-Ajmaline. The principal alkaloid 38 of RauwoZjia serpentina and one of the first to be isolated is ajmaline C2,H,,02N2.Its ultraviolet spectrum and colour reactions the production of id-N-methylharman on distillation with zinc dust and the oxidation of ajmaline with potassium permanganate in acetone to 3-acetonyl-2 3-dihydro-3-hydroxy-1 -methyl-2-oxoindole (LXXIX) in- dicate clearly that the alkaloid is a derivative of 2 S-dihydro-l-methyl- indole. 57 Robinson and his co-workers have provided abundant evidence that the second nitrogen atom of ajmaline is contained in a carbinolamine grouping. However it' does not behave as a normal carbinolamine e.g. y-strychnine (LXXX). Thus the alkaloid shows reducing properties and Me (LXXIX) (LXXXI) gives an oxime which on dehydration and reduction of the nitrile with lithium aluminium hydride regenerates ajmaline.It can also be reduced by the Huang-Minlon procedure and on treatment with Raney nickel at 130" loses carbon monoxide [N*CH(OH*& *&He H 6 + CO] the pro- ducts from both reactions being seconda;y bases. On the other hand the basic strength of ajmaline (pK 8-15) is much higher than expected (cf. y-strychnine pK 5-60) and it gives 0-acetyl derivatives without fission of the C-N bond and is not reduced by zinc and acid. These results are interpreted by assuming that in ajmaline the carbinolamine grouping is so situated in a bridged-ring system that participation of the ammonium form (-+NH :CH-) is imp0ssible.~7 Deoxydihydroajmaline gives ethyl methyl ketone on oxidation while similar treatment of decarbonoajmaline gives butyric acid (and traces of the lower homologues).These results can be accommodated in the following partial formdz I I Me*CO*Et + -NH Me-CHEt +- -N-CH(OH)-CHEt I I c 4 - c c-c-c Deoxydihydr oa j maline A j maline C3H,*C0,H I CH3*C0,H c-c-c C,H,.CO,H +- -" + CO + CH,Et I Decarbonoajmaline '13~ Wenkertc and Liu Experiefitia 1955,11 302 ; Huebner MacPhillamy Schlittler and St. Andr6 ibid. p. 303. SAXTON THE INDOLE ALKALOIDS 137 The second hydroxyl group in ajmaline was assumed to be tertiary since although it can be acetylated it is comparatively inert towards oxidation. The failure of dehydration and replacement reactions indicates that the hydroxyl group may be situated at the apex of a bridged-ring system as it is in apocamphan-1-01 (LXXXI) which behaves similarly. Dehydrogenation of deoxydihydroajnialine and deoxyajmaline with palladised charcoal at 326" yields bases of the alstyrine type but their structures have not yet been completely elucidated.So far it is the only dihydroindole alkaloid known which gives alstyrine-like degradation products. The above results require that this alkaloid possess the partial formula (LXXXII) . By postulating an internal condensation at position 3 of the indole nucleus with one of the fragments of the ruptured benzene ring Robinson has expanded this to the alternatives (LXXXIII) and (LXXXIV).57 The structure of ajmaline poses a fascinating problem. OH E t Me Me ( LXXX I II) 'C The constitution of ajmaline was finally established in an elegant series of degradations by Schenker and Woodward who conclusively proved that it possesses the structure (LXXXIVA).5* Oxidation of deoxyajmaline with lead tetra-acetate led to the rapid formation of an indole-aldehyde (LXXXIVB) which suggested that one of the carbon atoms attached to position 2 or 3 of the dihydroindole moiety contains the inert hydroxyl group which must moreover be secondary since the product is an aldehyde and not a ketone.This was confirmed by isolation of a dihydroindole- ketone by the prolonged oxidation of deoxyajmaline with benzophenone and potassium tert.-butoxide. Reduction of this ketone by sodium boro- hydride gave epideoxyajmaline which gave the same aldehyde (LXXXIVB) as deoxyajmaline with lead tetra-acetate. The infrared absorption of the dihydroindole-ketone (carbonyl band at 5.74,~) suggested that the carbonyl group was present in a five-membered ring.These facts together with biogenetic considerations led to the formula (LXXXIVA) for ajmaline 67 Mukherji Robinson and Schlittler Experientia 1949 5 215 ; Raymond-Hamet Compt. rend. 1949 229 1165 ; An& Robinson Chakravarti and Schlittler J. 1954 1242 ; Robinson (with Anet Finch and Hobson) Chem. and Ind. 1955 285 ; Finch Hobson Robinson and Schlittler ibid. p. 653. m Schenker and Woodward personal communication. 138 QUARTERLY REVIEWS which was soon confirmed by a study of the dehydrogenation of deoxydi- hydroajmaline (LXXXIVC) by palladised charcoal at 250". The four pro- ducts isolated included N(a,-methylharman ajarmine (LXXXIVD ; as racemate) ajmyrine (LXXXIVE) and a base C20H24N2 probably identical WEt (LXXXIV A) Et \yEt (LXXXIV B) with one of Robinson's dehydrogenation products which had a spectrum almost superimposable on that of AT(,,-methylalstyrine.Final confirmation of this structure was obtained by synthesis of ajarmine (LXXXIVD) from N(,,-methylharman. The isolation of ajarmine and ajmyrine whose struc- ture (LXXXIVE) has now been confirmed by synthesis 58a demonstrates conclusively that N(b) in deoxydihydroajmaline is common to two six- membered rings and that in ajmaline it must be common to three such rings. The formulation of this alkaloid as a product of both 01,- and ,&type biogenetic condensations is thus confirmed. Me (LXXXIV E> R. serpentina also contains isoajmaline and a consideration of the chemical properties which are identical with those of ajmaline shows that this compound is simply a stereoisomer of the latter.571 59 In accordance with this conception isoajmaline can be produced from ajmaline by the action of heat or alkali and since the two alkaloids give the same decarbono- ajmaline they differ only in the configuration of the carbon atom adjacent to the carbinolamine grouping (denoted by an asterisk in LXXXIV).se Woodward and Yang personal communication. sn Siddiqui and Siddiqui J . Indian Chem. SOC. 1935 12 37. SAXTON THE INDOLE ALKALOIDS 139 The existence of neoajmaline reported by Siddiqui is now regarded as doubtful. An examination of Siddiqui’s specimen by the Oxford workers suggests that it is very probably ajmaline.60 Little is known about the remaining dihydroindole alkaloids which have been isolated from 3. serpentina. Both rauwolfinine C19H2602N2 and serpinine C,oH,,ON or C,,H,,ON, show weak reducing properties which may indicate the presence of a potential aldehyde group.Rauwolfinine shows certain other superficial resemblances to ajmaline hence it may possess a similar constitution and it is significant that in some specimens of R. serpentina ajmaline is replaced by rauwolfinine and the latter is the major alkaloid of the plant growing in North-Western India. It is therefore surprising that it was not isolated earlier.61 The Quebracho Alkaloids.-The quebracho alkaloids are obtained from various species of Aspidosperma and VaZZesia which are hard-wood trees found in South America. Several alkaloids have been isolated from Chese sources the four principal ones being aspidospermine vallesine quebrachine (yohimbine) and quebrachamine.62 The remaining alkaloids are aspido- samine haslerine and quirandine which were extracted from Aspidosperma quirandy Hassler by F l ~ r i a n i ~ ~ and aspidospermicine aspidospermatine and hypoquebrachine obtained from Aspidosperma quebrachoblanco Schlecht.62 Very little is known about these alkaloids and indeed the status of two of them as true alkaloids is open to question.Ewins has suggested that aspidosamine and hypoquebrachine are simply decomposition products of aspidospermine.63 Aspidospermine C22H3002N2 is the principal alkaloid of the group and occurs along with quebrachine in A. quebracholdanco with vallesine in Vallesia glabra and in several related species. It is a monoacidic base containing methoxyl methylimino- and N-acetyl groups. Hydriodic acid removes the N-acetyl group and ruptures the ether link to give aspidosine C19H260N2 which behaves in all respects as an aminophenol.The position of the phenolic hydroxyl is shown by a study of the ultraviolet and infrared spectra of the demethylated alkaloid which show a marked resemblance to those of vomicine (LXXXV). I n particular there is no hydroxyl band in the infrared spectrum owing to hydrogen bonding with the acetyl group. The presence of a dihydroindole group is proved by isolation of a mixture of indoles on dehydrogenation with zinc dust or palladium. The first formula (LXXXVI) proposed for aspidospermine explained readily all these experimental results and was easy to justify biogenetically but could not account for the occurrence of 3 5-diethylpyridine among the products of zinc dust distillation or for the result of Kuhn-Roth determination which 6o Siddiqui J.Indian Chem. Soc. 1939 16 421. 61 Bose ibid. 1954 31 47 691 ; Naturwiss. 1955 42 71. 6 2 Fraude Ber. 1878 11 2189 ; 1879 12 1560 ; Schlittler and Rottenberg Helv. Chim. Acta 1948 31 446 ; Hesse Annalen 1882 211 249 ; Fourneau and Page Bull. Sci. p h r m . 1914 21 7. 6s Floriani Rev. centro Estud. farm. bioquim. 1935 25 373 423 (Chem. Abs. 1936 30 1415); Ewins J . 1914 105 2738. 140 QUARTERLY REVIEWS indicates only one C-Me for deacetylaspidospermine. Indeed the isolation of 3 5-diethylpyridine is unexpected since all the alkaloids discussed above whether they belong to the a- or the p-series are 3 4-disubstituted pyridine (LrnV) (LXXXVI) derivatives. Witkop's alternative formula (LXXXVII) for aspidospermine explains readily its known behaviour but its possible biogenesis is not so apparent.Openshaw's ingenious suggestion is that it is formed from a methoxy-aldehyde (LXXXVII) analogous to the Wieland-Gumlich aldehyde (XLA ) which represents an intermediate stage in Woodward's biogenesis of strychnine. Reduction of the aldehyde followed by oxidation in the /&position to the carbonyl group (i.e. position 1 of the dihydroxyphenyl- alanine precursor) yields an aldol which on fission recyclisation reduction and acetylation gives aspidospermine (LXXXVIII) .64 (LXXXVII) 1 (LXXXVIII) Vallesine the second alkaloid isolated from these sources is probably deacetyl-N-formylaspidospermine. Schlittler and Rottenberg 62 have shown that deformylvallesine and deacetylaspidospermine are identical in chemical and physical properties but in spite of this and the identity of Rontgen powder diagrams and infrared spectra these authors formulate deacetyl- aspidospermine as a homologue of deformylvallesine for reasons which are not clear.The only other alkaloid known which gives a 3 5-disubstituted pyridine on degradation is ibogaine C20H260N2 obtained from Tabernanthe iboga Baillon. Extracts of the root bark of the Witkop states that the last-named substances are i d e n t i ~ a l . ~ ~ AZkaloids of Tabernanthe iboga. 64 Raymond-Hamet Compt. Tend. 1948 226 2154 ; Witkop J . Amer. Chem. SOC. 1948 70 3712 ; Witkop and Patrick ibid. 1954 76 5604 ; Openshaw Smith and Chalmers 13th Internat. Congr. Pure Appl. Chem. 1953 Abs. p. 223. SAXTON THE INDOLE ALKALOIDS 141 shrub are used by West African natives to increase resistance to fatigue and so this species has attracted wide attention from pharmacologists.Its chemical study however is in a much more elementary stage and although four alkaloids have been isolated and characterised none of their structures is known.65 Ibogaine is a monoacid base containing methoxyl C-methyl and imino-groups. Its ultraviolet spectrum and colour reactions disclose the presence of an indole nucleus and this is confirmed by the production of 5-methoxy-N-oxalylanthranilic acid on oxidation. Alkali fusion gives 3-ethyl-5-hydroxy- 1 2-dimethylindole (LXXXIX) which was identified by synthesis and 3-ethyl-5-methylpyridine.66 Assuming from the isolation of the latter a relation to aspidospermine Robinson derives the constitution (XC) for a dihydroibogaine by fission of the bond between the carbon atoms marked with an asterisk.Ibogaine must then contain another ring since there is no evidence for the presence of a double bond. Alkaloids of Picralima nit ida.-The dihydroindole alkaloids discussed above show few structural analogies with those of the a-series with the possible exception. of ajmaline. Aspidospermine is presumed to be formed by fission of the aromatic ring of the dihydroxyphenylalanine precursor followed by various reactions involving rupture and formation of further rings. The comparatively simple recombination of the aromatic ring frag- ments to give a six-membered ring similar to ring E of 6-yohimbine might also be expected and this has been postulated by Robinson to account for the behaviour of two alkaloids akuammine and y-akuammigine isolated from Picralirna nitidu Stapf.extracts of which are used by West African natives in the treatment of malaria and as an antipyretic. This reputation led Henry and Sharp to investigate the alkaloidal constituents but the investigations were not pursued when it was shown that the extracts are inactive in avian malaria. More recently however Raymond-Hamet has demonstrated that the principal alkaloid akuammine has a local anaesthetic action almost equal to that of cocaine and the investigations have been resumed by Robinson and his co-workers.67 Of the eight alkaloids so far isolated only akuammigine has been studied in addition to akuammine 6 5 Dybowski and Landrin Compt. rend. 1901 133 748 ; Haller and Heckel ibid. p.850 ; Delourme-Houd6 Ann. pharm. frang. 1946 4 30 ; Burckhardt Goutarel Janot and Schlittler Helv. Chim. Acta 1952 35 642 ; Goutarel and Janot Ann pharm. frang. 1953 11 272. 66 Janot Goutarel and Sneeden Helv. Chim. Acta 1951 34 1205 ; Raymond- Hamet Compt. rend. 1949 229 1359 ; Goutarel Janot Mrtthys and Prelog ibid. 1953 237 1718 ; Schlittler Burckhardt and GellBrt HeZv. Chim. Acta 1953 36 1337. 67 Henry and Sharp J. 1927 1950 ; Goodson Henry and MacFie Biochem. J. 1930 24 874; Raymond-Hamet Arch. mp. Pathol. Phamakol. 1942 199 399; Janot Le Men Aghoramurthy and Robinson Experientia 1955 11 343. 142 QUARTERLY REVIEWS and y-akuammigine. Its constitution as a stereoisomeride of d-yohimbine has already been mentioned. Akuammine,* CB2HB6O4N2 is a ditertiary dihydroindole base containing hydroxyl C-methyl N-methyl and methoxyl groups which dissolves readily in methanolic alkali giving a solution from which it can be regenerated.Its colour reactions are strongly reminiscent of those of p-methylamino- phenol and the presence of a similar grouping (based on 2 3-dihydro-5- hydroxy- 1-methylindole) also explains the relative instability of the alkaloid which has been known to decompose during attempted recrystallisation. Zinc dust distillation of the product gave 3-ethylpyridine and probably skatole. Hydrogenation was inconclusive and requires further study but presence of a double bond is considered probable by Robinson and Thomas who have combined these results in the provisional formula (XCI; R = OH) for akuammine.68 The second alkaloid y-akuammigine is believed to be deoxyakuammine (XCI ; R = H).It contains similar functional groups (with the exception of the phenolic hydroxyl group) but its basic strength and colour reactions are anomalous. Thus its basic strength is closer to that of strychnine than to that of strychnidine. Further ferric chloride gives a colour only on warming and diazo- coupling occurs reluctantly and only in buffered acetic acid. However nitration and nitrosation proceed normally. On the other hand the lithium aluminium hydride reduction product y-akuammigol behaves like strychnidine in all respects. These peculiar properties are interpreted in formula (XCI ; R = H) by assuming that the proximity of the dihydroindole nitrogen atom and the ester group results in deactivation of the potentially basic group across space by the carbonyl group.This proximity apparent from a study of molecular models and the tertiary nature of the ester group also explain the difficulty of hydrolysis of y-akuam- migine which was unaffected by vigorous treatment with alcoholic alkali.68 Alkaloids of Gelsemium Species.-The North American plant Gelsemiurn sempervirens has been known since 1870 to contain alkaloids ; Wormley isolated an amorphous base later obtained crystalline by Gerrard,Ggb and named gelsemine. Although a number of bases has been isolated in addition to gelsemine only four of these namely sempervirine gelsemicine gelsedine Henry J. 1932 2759 ; Millson Robinson and Thomas Experientia 1953 9 89 ; Robinson and Thomas J. 1954 3522. 6s (a) Wormley Amer. J . Pharm. 1870 42 1 ; ( b ) Gerrard Pharm.J. 1883 13 641 ; Jahresber. 1883 1353 ; Chou Chinese J . Physiol. 1931 5 345 (Chem. Abs. 1932 26 806) ; Chou Wang and Cheng ibid. 1936 10 79 (Chem. Abs. 1936 30 4270). * This alkaloid also known as vincamajoridine has recently been isolated together with reserpinine from Virzca major L.67 SAXTON THE INDOLE ALKALOIDS 143 and gelseverine are well established. Chinese Gelsemiurn species also contain several alkaloids but with the exception of gelsemine obtained from G . elegans (Gardn.) Benth. they do not appear to have been related to those extracted from the North American varieties.69 Gelsemine C20H2202N2 is the principal alkaloid of this series and has been studied in recent years by several groups of workers. One of its nitrogen atoms is contained in a tertiary basic group while the other is inert.Reduction of gelsemine with lithium aluminium hydride results in elimination of one oxygen atom with the formation of a dihydroindole base. This behaviour is characteristic of 3 3-disubstituted oxindoles and the presence of this grouping was confirmed by comparison of the ultraviolet spectra of gelsemine and 3 3-dimethylindo1ey and the colour reactions of gelsemine and strychnine. The second inert oxygen atom is presumed to be present in an ether link and the alkaloid was readily proved to contain a terminal double bond. The formula (XCII) was deduced from these results by H ( xcv) (xcvr) (X=C.NorO; n = 0. I or 2) Gibson and Robinson but the alternative formulae (XCIII) and (XCIV) were preferred by Goutarel Janot Prelog Sneeden and Taylor who interpreted the results of a series of transformations on gelsemine as indicat- ing the presence of the grouping (XCV).For example reaction with bromine led to a cyclic derivative formulated as (XCVI) and the reverse process was achieved by reduction with zinc and acid.70 These conclusions of Goutarel et al. were challenged by Jones and Stevens,714 who doubted 7O Chu and Chou J. Arner. Chem. SOC. 1940 62 1955 ; Marion Canad. J. Res. 1943 21 B 247 ; Kates and Marion J. Amer. Chem. SOC. 1950 72 2308 ; Canad. J . Chem. 1951 29 37 ; Gibson and Robinson Chem. and Ind. 1951 93 ; Goutarel Janot Prelog Sneeden and Taylor Helv. Chim. Acta 1951 34 1139 ; Goutarel Janot Prelog and Sneeden ibid. p. 1962 ; Prelog Patrick and Witkop ibid. 1952 35 640 ; Habgood Marion and Schwarz ibid. p. 638. Jones and Stevens J .1953 2344. 144 QUARTERLY REVIEWS whether gelsemine is an allylamine since Hofmann degradation of dihydro- gelsemine did not proceed normally as expected for a substance possessing the partial structure (*NMe*CH*CHMe*). The first positive evidence against the formulz (XCII-XCIV) was provided by Habgood and Marion,71b who demonstrated that gelsemine contains a methylene group adjacent to the basic nitrogen atom. Reaction of dihydrogelsemine with diethylazodi- carboxylate gave a product which behaved as a carbinolamine. Oxidation of this gave an amide the infrared spectrum of which suggested that it probably contained a five-membered lactam ring. Habgood and Marion interpreted this evidence by the following partial formula? and proposed for gelsemine the from tryptamine constitution (XCVII) which can be derived biogenetically and tyrosine.(XCVII) O;p? H The minor alkaloids of G. sempervirens are probably also oxindole deriva- tives but little is known about them. Oxindole alkaloids also occur in other botanical species e.g. mitraphylline formosanine and rhynchophylline in Ourouparia formosanu Matsumura et Hayata.72 Although the formula of gelsemine is not known with certainty the presence of the 3 3-disubstituted oxindole chromophore is well established. This grouping presumably arises in vivo by oxidation and rearrangement of indole derivatives a process which is also feasible in vitro e.g. the conversion (XCVIII) of 9-acetylhexahydro-10 1 l-dihydroxycarbazole by means of acid into a spirocyclic oxindole derivative (XCVIII). The reverse process was attempted with gelsemine by Witkop and Patrick since the tetrahydro- carboline obtained should be easier to study than gelsemine but so far without SUCC~SS.~ 71b Habgood and Marion Canad.J. Chern. 1955 33 604. 73 Chou Chinese J. Physiol. 1931 5 131 (Chern. Ah. 1931 25 4085) ; Raymond- Hamet Cornpt. rend. 1950 230 1405 ; Janot Goutarel and Friedrich Ann. pharm. franp. 1951 9 305 ; Schwarz and Marion Canad. J. Chem. 1953 31 958 ; J . Amer. Chem. SOC. 1953 75 4372 ; Cook Gailey and Loudon Chem. and Ind. 1953 640. 73 Plant and Robinson Nature 1950 165 36 ; Witkop J . Amer. Chem. SOC. 1950 72 614 ; Perkin and Plant ; J. 1923 -123 676. SAXTON THE INDOLE ALKALOIDS 145 Alkaloids of Calabash Curare.-The highly toxic constituents of gourd or calabash curare from various sources in South America have been investi- gated by Wieland and his collaborators and more extensively in recent years by Karrer and his associates.More than thirty alkaloids have so far been isolated by precipitation and separation of the reineckates or by chromato- graphic techniq~es.~~ They are all quaternary salts and a study of their ultraviolet spectra has shown that they all contain the indole ring system either as such or in various reduced or oxidised-reduced forms.75 Except with fluorocurine and mavacurine also isolated from Venezuelan Strychnos toxifera very little progress has been made in the elucidation of the structures of these alkaloids. Chemical studies have been hampered by the difficulty of isolating the pure alkaloids in large quantity. No doubt the problem would be simplified if the samples of curare consisted of extracts from one botanical species instead of a mixture of several species.Another dis- advantage is that the composition of the curare varies according to its source. Included in this group are the alkaloids of Xtrychnos toxifera and 8. mitscher- Eichii which are possible constituents of some samples of curare. Fluorocurine C20H2602N2+ contains one methylimino- one C-methyl and one acetylatable hydroxyl group and there is a double bond in the molecule. The ultraviolet spectrum closely resembles that of a 1 2 2- trisubstituted indoxyl and it is probable that the alkaloid contains this grouping. Reduction of fluorocurine with lithium aluminium hydride yields hydrofluorocurine which is converted by sulphuric acid into the trisubstituted indole mavacurine C20H250N2+ by loss of water.The quaternary nitrogen atom is the one which bears the methyl group since thermal decomposition of fluorocurine chloride gives norfluorocurine which is reconverted into the alkaloid by methyl chloride. Similarly normava- curine yields mavacurine. The indoxyl-nitrogen atom is probably bound in another ring since there is no imino-group in the molecule and the position of the C-methyl group excludes the possibility of an N-ethyl group. Selenium dehydrogenation of mavacurine gave a base which was not completely identified but was very probably an N-substituted /3-carboline derivative. The position of the double bond in these alkaloids was demonstrated by ozonolysis and Kuhn-Roth oxidation which gave respectively acetaldehyde and acetic acid.Further chromic acid oxidation of dihydrofluorocurine gave a significant amount of propionic acid confirming the presence of an ethylidene group CH,*CH :. Catalytic reduction of hydrofluorocurine chloride (C20H2,02N2Cl) yielded a tertiary base C20Hso02N, which contained two C-methyl groups and probably arose by reduction of the double bond and simultaneous Emde reduction. This product gave a mixture of acetic 7 4 Karrer and Schmid Helw. Chirn. Acta 1946 29 1853 ; 1947 30 1162 2081 ; 1950 33 512 ; Schmid Kebrle and Karrer ibid. 1952 35 1864 ; Kebrle Schmid Waser and Karrer ibid. 1953 36 345 ; Asmis Bkhli Giesbrecht Kebrle Schmid and Karrer ibid. 1954 37 1968 ; Giesbrecht Meyer Biichli Schmid and Karrer ibid. p. 1974 ; Asmis Schmid and Karrer ibid. p. 1983 ; Wieland Bahr and Witkop Annalen 1941 54'7 173; Wieland and Merz Chem.Ber. 1952 85 731. 75 Kebrle Schmid Waser and Karrer Helw. Chim. Acta 1953 36 102. 146 QUARTERLY REVIEWS and a-methylbutyric acid on oxidation. can be accommodated in the annexed partial formuh Hence the abovet ransformations LiAl Hi- CHZ$:CHMe C QYJbMe Hydrofluorocurine + Fluorocurine Mavacurine Emde base Since the infrared spectrum suggests the presence of an ortho-disubstituted benzene derivative it is clear that the two remaining carbon atoms and the hydroxyl group must be situated in the alicyclic portion of the molecule and further since the indoxyl group and the double bond are the only centres of unsaturation the alkaloids must be pentacyclic. Mavacurine and melinonine-A occur together in Xtrychnos melinoniana and hence it is possible that they are related biogenetically.The formula for mavacurine can therefore be expanded to (XCIX) and it only remains to justify the position i Melinonine - A c- of the hydroxyl group. Mavacurine is not a carbinolamine and Karrer and his co-workers are of the opinion that the most likely position is a t C(15). If this formula represents mavacurine fluorocurine must be (C) and may be formed in vivo from the former by oxidation and rea~~angement.7~ This 76 Bickel Giesbrecht Kebrle Schmid and Karrer Helv. Chim. Acta 1954 37 553 ; Bickel Schmid and Karrer ibid. 1955 38 649 ; €or summaries see Karrer Nature 1955 176 277; Karrer and Schmid Angew. Chem. 1955 67 361. SAXTON THE INDOLE ALKALOIDS 147 parallels the known rearrangement of 9-acetylhexahydro- 10 11 -dihydroxy- carbazole in alkaline solution to a spirocyclic iiidoxyl (CI).73 The author thanks Professor Sir Robert Robinson O.M.F.R.S. for his interest and advice Professor R. B. Woodward for permission to include some unpublished results and Miss J. C. Clark B.A. for reading the manuscript.