ALKALOIDS OF CALABASH-CURARE AND STRYCHNOS SPECIES By A. R. BATTERSBY M.Sc. PH.D. and H. F. HODSON PH.D. (DEPARTMENT OF ORGANIC CHEMISTRY THE UNIVERSITY BRISTOL) NEWS of the dramatic paralysing effect of the South American Indian arrow and dart poisons known as curare was carried back by the early explorers of that continent and has held the interest of men of science and medicine since the sixteenth century. This interest was increased by the mystery surrounding the preparation of the poison and many fantastic stories became connected with it. These are included in an entertaining account of the history preparation and pharmacology of curare given in McIntyre’s excellent rnon0graph.l Recently a collection of authoritative accounts of all aspects of curare has been published2 as the proceedings of a UNESCO Symposium held in 1957; the present Review is desirable be- cause of the spectacular advances made in knowledge of the chemistry of curare since that time.As pointed out by McIntyre curare is a generic term which includes many types of arrow-head poison prepared in South America and the constitution of individual preparations varies according to geographical origin. All curares are powerful poisons which paralyse voluntary muscle2 and all are concentrated aqueous extracts of plant material. Boehm3 classified curares according to the type of container used to pack the final product; he writes of tube-curare packed in bamboo tubes pot- curare for which earthen pots were used and calabash-curare held in calabashes or gourds. Although the validity of this classification has been q~estioned,~ in particular the existence of pot-curare as a distinct group it is useful for this discussion to accept the broad division into tube-curare and calabash-curare.It was early realised3 that the active principles of curare are water- soluble quaternary alkaloids and in 1935 King5 isolated the now well- known curarising agent d-tubocurarine (1) from tube-curare ; d-tubocurar- McIntyre “Curare Its Natural History and Clinical Use” University of Chicago Press Chicago 1947. Bovet Bovet-Nitti and Marini-Bettolo (editors) “Curare and Curare-like Agents” Elsevier 1959; see also Craig “The Alkaloids” Vol. V ed. Manske Academic Press New York 1955 p. 265. Boehm Abhandl. KgZ. sachs. Ges. Wissensch. 1895 22 201; 1897 24 1. Lewin “Die Pfeilgifte” Barth Leipzig 1923; Gill Anesthesiol.1946 7 14. King J. 1935 1381. 77 78 QUARTERLY REVIEWS ine is an example of a quaternary bisbenzylisoquinoline alkaloid. Tube- curare is prepared mainly from the bark of Menispermaceous plants particularly the genus Chondrodendron and d-tubocurarine was later isolated by Wintersteiner and Dutcher from C. tomentosum6 Further work by King7 and others led to the isolation and structural elucidation of many more bisbenzylisoquinoline alkaloids from these sources. Calabash-curare originates in the northern parts of the South American continent particularly in the Amazon and Orinoco basins. It is con- siderably more active physiologically than tube- or pot-curare and has presented much more formidable chemical problems. Some of the major problems however have now been overcome and efforts have been increased owing not only to the rich and fascinating chemistry involved but also to the pharmacological interest of the curare alkaloids.d-Tubo- curarine chloride and synthetic curare agents with essentially the same action are now extensively used in surgery in conjunction with light anzesthesia. With their aid it is possible to achieve the required degree of muscular relaxation necessary for successful surgery without recourse to potentially dangerous deep anzsthesia. The interest attaching to the much more active calabash-curare alkaloids is thus obvious. Over a century ago Robert Schomburgke was able to see barks of Strychnos toxifera and other Strychnos species being used as important constituents of calabash-curare. This has been confirmed by later observa- tions and is also apparent from the general similarity in alkaloid content between calabash-curare and extracts from the bark of various Strychnos species.It has been shown particularly well by the extensive chromato- graphic studies of Marhi-Bettolo and his collaborator^.^ Thus chemical investigations of calabash-curare and of the barks of Strychnos species are mutually related topics and can bz conveniently considered together in this Review. An exhaustive survey of work in this field which has been largely carried out over the last decade has been published by Bernauer.1° Isolation of Pure Alkaloids.-Serious chemical work on calabash-curare was started by Boehm3 in 1897 and resulted in the isolation of a highly active amorphous principle ; much later (1 935) King" described the preparation of an equally active amorphous quaternary iodide from the bark of S.toxifera. However the first isolation of crystalline calabash- curare alkaloids was achieved by H. Wieland and his scho01.~~-~~ The Wintersteiner and Dutcher Science 1943 97 467. King J. 1948 265 and earlier papers. Penna Iorio Chiavarelli and Marini-Bettolo Gazzetta 1957 87 1 163 and earlier a See ref. 1 page 33. lo Bernauer Forschr. Chem. org. Naturstofe 1959 17 184. l1 King Nature 1935 135 469. le Wieland Konz and Sonderhoff Annalen 1937 527 160. l3 Wieland and Pistor Annalen 1938 536 68. Wieland Pistor and Bahr Annalen 1941 547 140. l6 Wieland Bahr and Witkop Annalen 1941 547 156. papers. BATTERSBY AND HODSON CURARE ALKALOIDS 79 German workers precipitated the quaternary alkaloids as a mixture of reineckate salts which was then fractionated by adsorption chromato- graphy on alumina; of the various reineckate fractions they obtained some yielded crystalline chlorides and picrates.C-Curarine I chloride* was the first calabash-curare alkaloid to be so isolated ; other well-characterised alkaloids isolated in this early work were C-calebassine and C-dihydrotoxi- ferine I from calabash-curare and toxiferine I and toxiferine I1 from the bark of S. toxifera. Using the same chromatographic technique Kinglo also isolated toxiferine I and toxiferine I1 from S. toxifera together with a series of new alkaloids all in small quantity designated toxiferine 111-XII. Recently it has been shown17 that several of these salts 111-XI1 are identical with well-characterised alkaloids described after King’s original paper; toxiferine V and toxiferine XI are identical with toxiferine I.Though chromatography of the alkaloidal reineckates was a major step forward in fractionation technique the method has its drawbacks. For example it has been found1’ that well-separated bands on the column can all contain the same quaternary alkaloid. It is now firmly established that the most satisfactory fractionation procedure in this field is partition chromatography on cellulose developed by the Zurich and Munich ~ c h o o l ~ . ~ ~ ~ ~ ~ Most of the present total of about seventy pure curare alkaloids have been isolated by this technique. Some idea of the com- plexity of the isolation problem can be gained from Schmid Kebrle and Karrer’s demonstration18 that at least forty-one alkaloids were present in a sample of calabash curare from the Amazon basin; even the single plant material Strychnos toxifera examined by Battersby Binks Hodson and Ye0we11,~~ contains at least thirty quaternary alkaloids.Extensive fractiona- tion involving repeated chromatography on cellulose and alumina is usually required before crystalline alkaloids can be obtained. The Table shows those alkaloids isolated from calabash-curare and Strychnos species which have been sufficiently studied to warrant their inclusion in this Review; little can be said at present about the other forty or so alkaloids and it will suffice to give their names and references to their isolation in the Appendix. Even during the early investigations by Wieland and King it became probable that the curare alkaloids are indole derivatives and with recent advances particularly from Karrer and Schmid’s group it is possible to correlate the ultraviolet spectra of many of the alkaloids with one or other of six related chromophores.These are the indoline (2) methyleneindoline * (a) The quaternary alkaloids are often isolated and handled as the chlorides and it is therefore convenient to use the name of the alkaloid as meaning alkaloid chloride. Thus in the sequel C-curarine I means C-curarine I chloride and other alkaloids will be treated in the same way. Anions other than chloride will be named. e.g. toxiferine I picrate (b) The letter C- denotes calabash. l6 King J. 1949 3263. l7 Battersby Binks Hodson and Yeowell J. in the press. lS Wieland and Merz Gem. Ber. 1952 85 731. Schmid Kebrle and Karrer Helv.Chim. Actu 1952,35 1864. 80 - YO. - 1 2 3 4 5 6 7 8 9 10 1 1 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 - QUARTERLY REVIEWS Alkaloids of Calabash-curare and Strychnos species Alkaloid ~~~ Toxiferine I C-Dihydrotoxiferine I C-Alkaloid H Nordihydrotoxiferine Caracurine VI E Caracurine I1 methochloride C-Alkaloid D Caracurine V Caracurine VII Hemitoxiferine I C-Calebassine C-Alkaloid A C-Alkaloid F C-Alkaloid Y C-Cararine I C-Alkaloid E C-Alkaloid G C-Curarine I11 E C- Flu or ocur arine Diaboline C- Mavacurine Melinonine A Melinonine B C-Alkaloid T Lochneram C-Fluorocurine Melinonine G Melinonine F ;ormula of cation or base 0.42 1.2 0.71 1.2 1.6 0.8 0.42 0.34 1.4 2.1 1.5 0.8 0.23 0.49 1.6 1-0 0.36 0.65 2.2 2.7 4.0 2.7 3.1 2.1 3.2 2.0 - Chromophore Meth yleneindoline 9 Y 9 9 Indoline Y Modifiez indoline Endoline "(a) as NH] Y Y Y9 Indoline carbinolamine Y9 Y 9 Y Y 3 Y 3 ) Unknown 9 9 See p.97 N- Acy lindoline Indole 9 9 3 93 Y $-Indoxy1 8- Carbolinium 8-Carbolinium The mobilities of the alkaloids on paper are referred to that of C-curarine I as in solvent "c".18 distance moved by toxiferine 1- distance moved by C-curarine I standard. Thus RC for toxiferine I = c These formula may still need revision particularly in respect of the hydrogen 2o Kebrle Schmid Waser and Karrer Helv. Chim. Acta 1953 36 102. 21 Asmis Waser Schmid and Karrer Helv. Chim. Acta 1955 38 1661. z2 Asmis Schmid and Karrer Helv. Chim. Acta 1954 37 1983. 23 Asmis Bachli Giesbrecht Kebrle Schmid and Karrer Helv. Chim. Acta 1954 content. 37 1968. BATTERSBY AND HODSON CURARE ALKALOIDS 81 - q0.- 1 2 3 4 5 6 7 8 9 10 1 1 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 - Alkaloids of Calabash-curare and Strychnos species-continued Clolour with ceric sulphate immediate-After 20 min. Red-violet-Colourless Blue-violet-Colourless Red-violet-Colourless Violet-Pale brown Purple-Brown Purple-Brown Violet Red-violet-Yellowish Purple-red-Brown Stable orange Stable orange Blue-violet-Carmine Blue-violet-Carmine Blue-violet-Carmine Red-violet-Olive green Blue-Chrome green Blue-Chrome green Blue-Chrome green Blue-green-Yellow-green Nil Carmine Nil Nil Very pale red - Red-violet-Brownish Nil Nil Physiological activityb 9 y 2o 30y 2 o 16y 2o Inactive 22 Inactive 22 Over 400 y l6 llOOy 2o Inactive 22 Inactive 22 1140y 21 240y 2o 7 0 y 2o 75y 2o 30y 2o 0 - 3 - 4 - 0 y 2o 0.6-5.0 y 2o 1800y 2 o *Inactive 24 ?Inactive lQ ?Inactive 25 tInactive 2s - - - - 4400y 2 o - - First isolation S.toxifera15 (Br. Guiana) Calabash15 Calabashla S. toxiferu21 (Venezuela) S. toxiferu22 (Venezuela) S. toxifera22 (Venezuela) S. t~xifera'~,'~ (Br. Guiana) Calabashla S. toxifera22 (Venezuela) S. toxifera22 (Venezuela) S. tq.xiferu17 (Br. Guiana) Calabash15 Calabashla and S. t o ~ i f e r d ~ ~ ' ~ Calabash18 Calabash2s Calabash" Calabashla Calabash1* Calabash14 S. diab01i~~ Calabash' S. melin~niana~~ S. melinoniana26 Calabash2s Calabash27 S. melin~niana~ S. melinoniana2 t~ The figures give the dose in y (pg,) per kg. for the head-drop assay on the mouse;s0 the figures marked * are for head-drop assay on the rabbit,ls and the results marked t are for toxicity test on the frog.1s$25 24 King J.1949 955. 25 Schlittler and Hohl Helv. Chim. Acta 1952 35 29. 26 Arnold von Philipsborn Schmid and Karrer Helv. Chim. Acta 1957 40 705. 27 Arnold Berlage Bernauer Schmid and Karrer Helv. Chirn. Acta 1958 41 1505. 28 Schmid and Karrer Helv. Chim. Acta 1947 30 2081. 2Q Bachli Vamvacas Schmid and Karrer Helv. Chirn. Acta 1957 40 1167. * O Waser Helv. Physiol. Pharmacol. Acta 1950 8 343. 82 QUARTERLY REVIEWS (3) indole (4) oxindole (5) or N-acylindoline (6) $-indoxy1 (7) and fi-carbolinium (8) systems. In some alkaloids simple modifications of these systems are involved for example the carbinolamine (9). The groups in the Table have been constructed on the basis of chromophore and it is worth drawing attention to the close connection between chromophore and colour reaction with ceric sulphate.C-Mavacurine C-FluGocurine and C-Alkaloid Y.-A first insight into the structures of alkaloids from calabash-curare came from a study of C-mavacurine and C-fluorocurine. The former was isolated by Wieland and Merz19 and the latter by Schmid and Karrer;28 both were subsequently found in S. tuxifera from Venezuela.22 Work on these alkaloids was greatly U+ OH + o-o- I 0-H \ \ 0 5) (14) (13) 0-1 _H+ 0-1 Q-1.S. ' Y"- N 6 ." (a) OH f +OH facilitated by the discovery by Karrer Schmid and their co-workers31 that they are structurally related. Thus the quaternary C-fluorocurine C20H2502N2+ which clearly showed the 2,2-disubstituted #-indoxy1 chromophore (1 0) was reduced to hydrofluorocurine (partial structure 1 1) by borohydride.Acid then catalysed the illustrated rearrangement which is of 2,2-disubstituted 3-hydroxyindolines (1 l) to yield a 2,3- disubstituted indole (partial structure 12). The indolic product was identical with natural C-mavacurine. Fritz Wieland and B e s ~ h ~ ~ were able to extend the relationships by making use of the extensive researches 31 Bickel Giesbrecht Kebrle Schmid and Karrer Helv. Chim. Acta 1954 37 553; Bickel Schmid and Karrer ibid. 1955 38 469. 32 Witkop J. Amer. Chem. SOC. 1950 72 614; Witkop and Patrick ibid. 1951 73 713. sa Fritz Wieland and Besch Annalen 1958,611 268. BATTERSBY AND HODSON CURARE ALKALOIDS a3 of Witkop and Patrick34 on simple tetrahydrocarbazoles. Oxidation of C-mavacurine (partial structure 12) catalytically with oxygen over platinum gave a product to which the partial structure (14) was assigned; this is no doubt formed by way of the peroxide (13).The product had the properties of a glycol and-what was important-was rearranged by acid as illus- trated to give C-fluorocurine (partial structure 10). Since the intermediate (14) was found to be identical with C-Alkaloid Y previously isolated from a calabash,23 a triad of related alkaloids was available ; structural informa- tion derived from one alkaloid can thus be used for the other two. C-Fluorocurine contains an ethylidene side chain (>=CHMe) an acetylatable hydroxyl group and one N-methyl group the last being attached to the quaternary N(b) These functions and the analytical evidence require that C-fluorocurine be pentacyclic and the foregoing correlations show that this also holds for C-mavacurine.A clue to the probable arrangement of three of these rings came from selenium de- hydrogenation of normavacurine the tertiary base derived by pyrolysis of the quaternary C-mavacurine. This degradation yielded a b-carboline derivative (1 6) which was not fully identified because of the minute amount t t Me \ I ;C-fj Me-CH-Et - Me-flH-Et * CO,H C ' I \ (20) 0 r) available a common difficulty in the curare field. However the spectral properties of the N(b)-methiodide of the degradation base particularly the stability of the chromophore towards alkali showed clearly that the indolic nitrogen N(a) is alkylated;31 it should be mentioned for comparison that the spectrum of the salt (17) is profoundly changed by alkali owing to the formation of the anhydro-base (1 8).Further catalytic hydrogena- tion of hydrofluorocurine (partial structure 11) resulted in Emde degrada- tion to a tertiary base (20) which unlike the starting material gave a-methylbutyric acid (21) onKuhn-Roth oxidation.31 This result establishes the presence in hydrofluorocurine of a quaternary allylamine system as in (1 9) which would undergo Emde degradation and reduction as illustrated. Witkop and Patrick Experientia 1950 6 183 and subsequent papers. 84 QUARTERLY REVIEWS Because of the cyclic set of changes shown in the partial formulae (10) +( 12)+( 14)+( lo) it is certain that the conversion of C-fluorocurine into C-mavacurine only involves the $-indoxy1 to indole change and so the same system (19) is also present in C-mavacurine. The foregoing evidence can now be combined in partial structure (22) for C-mavacurine and there remain two carbon atoms for construction of two rings free from C-methyl groups.These requirements can be met by the constitution (23) which Bickel Schmid and Karrer propose31 for the alkaloid ; no firm proposal is made for the site of the hydroxyl group though position 15 is favoured. un CHMe HO 07) The plausible biogenetjc arguments by which this structure is derived involve the indicated intramolecular cyclisation of the supposed inter- mediate (24) used in biogenetic theory for such alkaloids as corynantheine (25). Future developments here both structural and stereochemical will be of great interest for though the proposed constitution for C-mavacurine is not yet fully established it is clear that this alkaloid represents a new twist in the biogenetic pattern for the indole alkaloids.If formula (23) is correct for C-mavacurine it is certain that structure (26) represents C-fluorocurine ; C-Alkaloid Y then has structure (27). Alkaloids of Strychnos meZinoniana.-As a result of extensive fractionations by Schlittler and HohlZ5 and more recently by Bachli and his c o - ~ o r k e r s ~ ~ eleven new alkaloids have been isolated from this plant and named melinonine A and B and melinonine E to M inclusive. Structural studies have been described for four of these. BATTERSBY AND HODSON CURARE ALKALOIDS 85 The structure of melinonine F Cl3HI3N2+ was fairly clear29 on the basis of its ultraviolet absorption spectrum (P-carbolinium salt chromo- phore) and the alkaloid was shown by direct comparison to be the N(b)- metho-derivative (28) of harman.Melinonine F has the distinction of being the simplest quaternary alkaloid to be isolated from South American Strychnos species. Melinonine G C1,H,,N,+ is also notable for its low hydrogen content; in keeping with this its ultraviolet and infrared spectra are very similar to those of sempervirine salts (29) and moreover the spectra established U Et the absence of vinyl residues. On the basis of this and other evidence particularly the catalytic reduction of the alkaloid to an indolic product which contained a C-ethyl group Bachli and his c o - ~ o r k e r s ~ ~ propose structure (30) for melinonine G. The constitution (31) then follows for the indolic reduction product. Structure (30) is in fact the one rigidly established by Bejar et aZ.35 and by Hughes and R a p o p ~ r t ~ ~ for flavopereirine from Geissospermum species.There seems little doubt that melinonine G and flavopereirine are identical though no direct comparison has apparently been made. Melinonine G is a particularly interesting molecule because in terms of current biogenetic thinking it must be regarded as a degraded indole alkaloid. Thus it could be formed from the common intermediate (24) proposed for many indole alkaloids by the loss of the three carbon atoms attached to C(15) in a reversed Michael reaction. Aromatisation could then give the required system. The other two alkaloids melinonine A and melinonine B have rather more complex structures. A key degradation product in the chemistry of the former was obtained by Schlittler and H ~ h l .~ They found that melinonine A C,,H2,03N2+ could be pyrolysed to give the corresponding tertiary base normelinonine A C21H2,0,N2 which yielded alstyrine (32) on selenium dehydrogenation. Further chemical and spectroscopic evidence derived from normelinonine A established the presence of an indole system together with the group MeO,C-k=k-O- a common feature in 35 Bejar Goutarel Janot and Le Hir Compt. rend. 1957 244 2066. 56 Hughes and Rapoport J. Arner. Chem. SOC. 1958 80 1604. 86 QUARTERLY REVIEWS the indole alkaloid field. These results led directly to the gross formula (33) as a likely one for normelinonine A and direct comparison with tetra- hydroal~tonine~~ (33) established the identity of the two bases. Thus melinonine A is the A@)-metho-derivative of the base (33).The chemistry of the tetracyclic melinonine-B is not as yet so clear-cut though structure (34) is a likely one. The presence of an acetylatable hydroxyl group a vinyl group an indolic “3 and a quaternary / q X & ’ H2 ‘ Et (3 4) Y2 CH2-OH methylammonium group has been firmly established by Vamvacas et aL3* There is no C-methyl group in the alkaloid so that the hydroxyl group must be placed in the primary position as in formula (34) in order to accommodate the other evidence. Selenium dehydrogenation of dihydro- melinonine B gave an a-pyridylindole having the highly characteristic ultraviolet absorption of this system ; structural work on this degradation product was hampered by the very small amount available but there is reasonable evidence in favour of structure (35).Surprisingly dehydrogena- tion of melinonine B over palladium yielded yobyrine (36) which on the basis of structure (34) must have been formed by ring-closure of the side- chains. This result is rather disturbing in that it increases the difficulty of 87 Elderfield and Gray J. Org. Chem. 1951 16 506; Wenkert and Roychaudhuri J. Amer. Chem. SOC. 1957,79 1519. 38 Vamvacas von Philipsborn Schlittler Schmid and Karrer Helv. Chim. Acfu 1957,40 1793. BATTERSBY AND HODSON CURARE ALKALOIDS 87 degradative work in the indole field; no longer can the isolation of yobyrine be taken as proof of the presence of a carbocyclic ring E in the parent alkaloid (e.g. structure 37). It can be seen that the foregoing evidence is not sufficient to establish firmly structure (34) for melinonine B; it could be objected for example to constitution (34) that the side-chains on positions 15 and 20 might be interchanged as Vamvacas et al.have been careful to point The resulting structure would be less attractive biogenetically but clearly further investigation of melinonine B is desirable. However with the present knowledge of its structure we can add melinonine B to the growing number of a-indole alkaloids* which do not contain a carboxyl group. Until recently the presence of a carboxyl group seemed to be general in the a-indole series. Yohimbine (37) is just one example of the many bases which display this feature. C-Alkaloid T (0-Methyisarpagine) and Lochneram.-The non-phenolic C-Alkaloid T C20H2402N2 isolated in Zurich from a Brazilian calabash,26 is now known to be identical with O-methyl~arpagine~~ which in turn is identical with 10chnerine;~~ thus structural studies on this interesting base are available from several laboratories.There is rigid evidence for an isolated double bond which is present in part as ‘CH-CH=CH2 and in part as ‘C=CHMe since ozonolysis gives a mixture of formaldehyde and acetaldehyde. This means that “C- Alkaloid T” is really a difficultly separable mixture of vinyl and ethylidene isomers. It is also established that a 5-methoxyindole residue unsubstituted at N(a) three more rings and a C-methyl group are present. The primary nature of the hydroxyl group was shown very neatly26 by reducing the 0-tosyl derivative to the corresponding deoxy-derivative. This contained a new C-methyl group probably located as MeCH” or less probably as / / \C /c M e d - C since its Kuhn-Roth oxidation products did not contain \C propionic acid only acetic acid.All this evidence can be accommodated by the constitution (38) for * a-Indole alkaloids are those involving the p-carboline system (a) or some simple derivative of it e.g. (6). A (0) (b) sD Stoll and Hofmann HeIv. Chim. Acta 1953 36 1143; Stauffacher Hofmann and Seebeck ibid. 1957,40,508; Poisson Le Men and Janot Bull. SOC. chim. France 1957 610. O0 Mors Zaltzman Beereboom Pakrashi and Djerassi Chern. andInd. 1956,173. 88 QUARTERLY REVIEWS C-Alkaloid T which is based26 on an assumed relation between this alkaloid and ajmaline; ajmaline has been rigidly proved41 to have structure (39). However there are considerable gaps to be filled before structure (38) is proved to be valid.Lochneram was isolated by Arnold et aL2’ from the calabash which yielded C-Alkaloid T and is the A@)-metho-derivative of the latter. It seems though that lochneram is the pure ethylidene isomer on the basis of cleavage with ozone. Thus if the constitution (38) is correct for C- Alkaloid T structure (40) follows for lochneram. The C p o Alkaloids.-(a) General. As work on the alkaloids of calabash- curare and Strychnos species progressed it became apparent that they fall naturally into two more or less clearly defined groups. One group contains those alkaloids which have relatively high mobilities in the solvent systems used for partition chromatography and have little or no physiological activity; all the alkaloids discussed in detail so far in this Review belong to this group.The second group contains the alkaloids with high curare activity which account for all the physiological effects of the curares or bark extracts. These alkaloids all move slowly on paper chromatograms and on cellulose partition columns. As is the case with the “fast-running” alkaloids the “slow-running” alkaloids have two nitrogen atoms in a C1,-C21 unit. One nitrogen “a) is non-basic or only weakly basic and is involved in the indole or more usually modified indole chromophore; the second N(b) is the basic or quaternary basic centre. One of the most significant advances in the chemistry of the “slow- running” highly active alkaloids came with von Philipsborn Schmid and Karrer’s demonstration that they have molecular formulz based upon C38-C40 skeletons; that is two basic or quaternary nitrogen atoms are present in the molecule.Previously all alkaloids from calabash-curare and Strychnos species had been assigned formulre based upon C,,-C21 mole- cules. The method employed by the Swiss chemists is that of partial quaternisationp2 which in this series depends upon the fact (and incidentally 41 Woodward Angew. Chem. 1956 68 13; Robinson ibid. 1957 69 40. 4a von Philipsborn Schmid and Karrer Helv. Chim. Acta 1956,39,913; cf. Battersby - and Craig J. Amer. Chem. SOC. 1951 73 1887. BATTERSBY AND HODSON CURARE ALKALOIDS 89 provides evidence for this fact) that the two basic nitrogen atoms in the tertiary base prepared from the quaternary alkaloid are sufficiently far apart for protonation or quaternisation at one basic nitrogen atom not significantly to affect either of these processes at the second basic centre.The tertiary base nor-C-curarine I obtained by pyrolysis of the quatern- ary C-curarine I chloride was treated with half an equivalent of mineral acid to give an equilibrium mixture of the species (41) (42) and (43). (46) (45) ('44) Methylation then gave a mixture of nor-C-curarine I (44 =43) monometho- nor-C-curarine I (43 and C-curarine I (46). The monometho-derivative (45) was readily separated by partition chromatography and by further methylation was converted into C-curarine I. This preparation of a C-curarine I derivative with one A@)-methyl group in each C40 unit is decisive evidence that this alkaloid has a Cpo skeleton; the molecular formula was accordingly modified to C40H44-460N4-++. By the same method C-calebassine was shownp2 to be C40H4,-,o0,N4++ and C-dihydrotoxiferine I to be C40H46N4++ and it was suggested that all the highly active calabash-curare alkaloids have CqO molecules ; this is almost certainly true.However the converse that all C40 quaternary curare alkaloids are powerful curarising agents is certainly not true as is shown by the surprisingly low activity of caracurine V dimethochloride Now it is known that the active calabash-curare alkaloids contain two quaternary centres they fall satisfyingly into place with the other quatern- ary curarising agents. Thus the bisbenzylisoquinoline alkaloids such as d-tubocurarine (47) have two quaternary nitrogen atoms set some distance apart as do the synthetic preparations such as succinylcholine (48). Indeed it is interesting from the point of view of structure-activity relations that with the structure of toxiferine I known (p.94) it turns out that the two quaternary centres are very nearly the same distance apart as those in d-tubocurarine (approximately 14 A). (b) Toxiferine I C-dihydrotoxiferine I and diaboline. During their pioneer work (1947) in the curare field Wieland Bahr and Witkopf5 isolated the highly physiologically active alkaloid toxiferine I from S. toxifera grown in British Guiana. Schmid and Karrer43 later obtained the C40H460ZN4tt (P* g5) 43 Schmid and Karrer Helv. Chim. Acta 1947 30 1162. 90 QUARTERLY REVIEWS same quaternary alkaloid from a Venezuelan calabash. Wieland’s group15 also isolated from a Venezuelan calabash preparation what they took to be a relative of toxiferine I because of the close similarity between the known and the new alkaloid’s properties.The new material was named C-dihy- drotoxiferine I although no formal relation to toxiferine I was demon- strated and it is now known that the alkaloid has been misnamed; as will be seen later C-dihydrotoxiferine I is in fact C-deoxytoxiferine I. In 0 CH,CO -0 CH2CH,’NMe CH2CO.O.CH,CH,.NMe; addition nordihydrotoxiferine I the tertiary base corresponding to the quaternary C-dihydrotoxiferine I was found by Karrer and Schmid’s groupz1 in the tertiary bases from Venezuelan S. toxifera bark. Some ten years after its isolation little was known about C-dihydro- toxiferine I. In common with toxiferine I it had been shownz0 to contain the methyleneindoline chromophore (49). Its molecular formula had been + established as C40H4sN$+ with two N-methyl groupsz1 attached to the quaternary N(b)-nitrogen atoms and ozonolysis21 had furnished acetal-.dehyde. The usual attack by various dehydrogenating agents unfortunately gave only glimpses of the molecule. Thus dehydrogenation with sulphur and with zinc dust had yielded is~quinoline;~~ distillation with zinc dust gave a mixture of 3-methyl- and 3-ethyl-indole and palladium dehydro- genation of the corresponding nor-base gave traces of a 6-carboline derivative.21 As will be discussed in the sequel C-dihydrotoxiferine I can be transformed into C-calebassine and C-curarine I both known to contain the grouping (50). In view of the production of acetaldehyde by ozonolysis of C-dihydrotoxiferine I it was taken as probable that the same quaternary allylamine system (50) also occurs in this alkaloid.In 1957 still less was known about toxiferine I but against this lack of Wieland Witkop and Bghr Annalen 1947,558 144. BATTERSBY AND HODSON CURARE ALKALOIDS 91 progress must be set the great difficulty of obtaining even a hundred milligrams of the alkaloid. von Philipsborn Schmid and KarreF2 gave evidence for a Cl0 molecule and the earlier formulae were revised to C40H46-4802N2-f+. Moreover toxiferine I had been found to undergo a change of unknown nature when treated with dilute acid; C-dihydrotoxi- ferine I behaved similarly.45 The structure of C-dihydrotoxiferine I has recently been elucidated in an outstanding series of 47 by Karrer Schmid and their collabora- tors; in addition work at Zurich and at Bristol to the structure of toxiferine I.Papers on the chemistry of these and other C40 curare alkaloids have been appearing steadily from four different laboratories and it is impossible in the space available to pay the mass of information full justice. The following account is confined to the essentials necessary for a logical presentation. Asmis Schmid and Karrer22 in 1954 isolated nine tertiary alkaloids caracurine I-IX from Venezuelan S. toxfera and in a later ~aper-4~ it was shown that caracurine V is readily transformed by dilute mineral acid into an unstable methyleneindoline base named caracurine Va. After this initial stage further acid-catalysed changes occur to give a mixture of caracurine I1 and caracurine VII both of which are stable to dilute acid. Battersby and examined the very similar changes which occur when toxiferine I also a methyleneindoline is treated with dilute acid and isolated two crystalline products.One was identical with a new quaternary alkaloid provisionally called A8 which they had i ~ o l a t e d ~ ~ . ~ ~ from S. toxfera grown in British Guiana. This alkaloid showed indoline ultraviolet absorption and the indoline nitrogen atom was proved to be secondary. Since in earlier work they had shown that Alkaloid A8 is identical with caracurine VII methochloride it follows474g that the above acid-catalysed transformations are taking place on related tertiary and quaternary molecules. Thus caracurine Va must be nortoxiferine I and the second product from toxiferine I must be caracurine I1 methochloride which was by preparing this material from authentic caracurine 11.Toxiferine I was also prepared47 by methylating caracurine Va. The set of interconversions was completed by s h o ~ i n g ~ ~ - ~ ~ that alkaloid AS in acetic acid is converted into toxiferine I in moderate yield. These transformations are summarised in the annexed scheme which includes the demonstration50 that the formation of caracurine I1 metho- chloride from toxiferine I involves atmospheric oxygen. There can be no doubt that this also holds in the tertiary caracurine Va series. 45 Asmis Bachli Schmid and Karrer Helv. Chim. Acta 1954 37 1993. 46 Bernauer Schmid and Karrer Helv. Chim. Acta 1958,41 1408. 47 Bernauer Berlage von Philipsborn Schmid and Karrer Helv. Chim. Acta 1958 49 Battersby and Hodson J. 1960 736. 41 2293. Battersby and Hodson Proc.Chem. Soc. 1958 287. Battersby and Rao unpublished work. 92 QUARTERLY REVIEWS -1 Toxiferine I 7 i' / I HOAc / Caracurine I1 / methochloride I Alkaloid A8 1 Methylate " Methylate I 1 Caracurine 11 Caracurine VII ' Caracurine Va r H,O+ Caracurine V In addition Karrer Schmid and their collaborator^^^^^^ had shown that C-dihydrotoxiferine I C40H46N4++ is converted by dilute mineral acid into the Czo alkaloid hemidihydrotoxiferine which like Alkaloid A8 had indoline ultraviolet absorption and contained NH but in addition the infrared spectrum showed the presence of an aldehyde group. As with toxiferine I a second product is also formed identical with C-Alkaloid D isolated earlier18 in Zurich from a calabash. Moreover hemidihydrotoxi- ferine in aqueous acetic acid was back into C-dihydrotoxiferine I.When these facts are combined with the knowledge that hemidihydro- toxiferine has colour reactions and an ultraviolet spectrum identical with those of Alkaloid AS and caracurine VII they leave little doubt that the interconversions summarised in the scheme below are analogous to those outlined above for toxiferine I and its nor-derivative (caracurine Va). C-Alkaloid D f HsO+ / \\ C-Dihydrotoxiferine--- 7 Hemidihydrotoxiferine \-- L HOAc A complex of inter-related alkaloids had thus been established by the foregoing experiments when the Zurich group made the important identifi- BATTERSBY AND HODSON CURARE ALKALOIDS 93 cation51 of caracurine VII with the Wieland-Gumlich aldehyde (5 1). The latter had been known for over 25 years as a degradation product obtained during H.Wieland’s studies on the structure of strychnine (52). This identification was of great interest being the first reported natural occurrence of the intermediate proposed by Woodward in his seminal biogenetic scheme for strychnine.52 OH (53) (54) A very similar biogenetic interest attaches to the tertiary base diaboline and it is convenient to leave the main theme for a moment to consider this alkaloid. It was isolated by King24 from the bark of S. diaboli and Bader Schlittler and S c h w a r ~ ~ ~ showed that it is an N(a)-acetylindoline derivative. With the structure of caracurine VII known indications derived from colour reactions and from reduction of diaboline by lithium aluminium hydride to a glycol led Battersby and H o d ~ o n ~ ~ to compare deacetyldiaboline with the Wieland-Gumlich aldehyde (5 l) and the two were identical.Thus diaboline is N(a)-acetyl-Wieland-Gumlich aldehyde (53) and its reduction product is the glycol (54). Diaboline is intermediate in complexity between Wieland-Gumlich aldehyde (5 1) and strychnine (52) and its occurrence in S. diaboli gives further support to Woodward’s biogenetic proposals. We can now return to the way in which the identification of caracurine VII as the Wieland-Gumlich aldehyde leads when combined with the above results to the structures of toxiferine I and C-dihydrotoxiferine I. Since alkaloid A8 is caracurine VII methochloride (p. 92). the former is the Wieland-Gumlich aldehyde A@)-methochloride (55) and this was rigorously confirmed;49 A8 is now better named hemitoxiferine I. Bernauer Schmid and K a r r e ~ ~ ~ proposed structure (56) for hemidihydro- 51 Bernauer Pavanaram von Philipsborn Schmid and Karrer Helv.Chim. Acta 62 Woodward Nature 1948 162 155. 59 Bader Schlittler and Schwarz Helv. Chim. Acta 1953 36 1256. 6o Battersby and Hodson Proc. Chem. Soc. 1959 126. 1958 41 1405. 94 QUARTERLY REVIEWS toxiferine i.e. that of the deoxy-Wieland-Gumlich aldehyde metho- chloride and soon afterwards this was as follows. The degrada- tion of C-dihydrotoxiferine I can be duplicated on the corresponding tertiary base nordihydrotoxiferine I to give heminordihydrotoxiferine (tertiary base corresponding to 56) which was reduced at the aldehyde function by borohydride to the primary alcohol (57). This was identical with that prepared from the Wieland-Gumlich glycol (58; R = OH) by selective bromination of the reactive allylic hydroxyl group to give the base (58; R = Br) followed by reductive removal of the halogen.We must now consider the processes by which the oxygen-free C- dihydrotoxiferine I with a methyleneindoline chromophore is formed from two molecules of the aldehyde (56) with loss of water and loss of the N(a)-hydrogen atoms. Moreover the fact that this reaction is reversed in the presence of mineral acid must be explained. The only constitution for C-dihydrotoxiferine I which satisfactorily accommodates this evidence is (59; R = H) first proposed by Bernauer Schmid and K a ~ r e r . ~ ~ Structure (60) can be written46 as aformal representation of the intermediate in both dimerisation and fission with the postulate of a double prototropic shift to move the double bond into or out of the methyleneindoline position.In the same way by combination of all the evidence outlined above the structure (59; R = OH) must be given to toxiferine I.47-49 C-Dihydrotoxi- ferine I and toxiferine I are thus examples of a new type of natural product. It remains to fit caracurine V into the picture. This base has an indoline chromophore and is readily converted into caracurine Va by dilute mineral acids or dilute acetic acid. Karrer Schmid and their co-worker~~~ have assigned to it the amino-hemiacetal structure (61) although the published chemical evidence does not exclude the alternative arrangement (62). Either of these hemiacetals would be expected readily to yield caracurine Va (nor-base corresponding to 59 R =OH) under acidic conditions.Models of the molecules (61) and (62) show that whereas the former is BATTERSBY AND HODSON CURARE ALKALOIDS 95 relatively unstrained there is considerable steric compression in the latter and on this basis the former seems much more likely. We have already mentioned that dimerisation of Wieland-Gumlich aldehyde methochloride (hemitoxiferine I) (55) with hot acetic acid gives49 a reaction mixture from which only about 20 % of toxiferine I (59; R=OH) R can be isolated directly and a similar result is obtained in buffered solu- t i ~ n . ~ ~ Caracurine V dimethochloride [61 with both N(b)-nitrogen atoms methylated] is one of the side products but the major one is 00-diacetyl- toxiferine I dichloride (59; R =OAc). Battersby and Hod~on*~ have shown that dimerisation of hemitoxiferine I (55) by means of pivalic acid (tri- methylacetic acid) in which there is strong steric hindrance of acylation gives a product from which toxiferine I (59; R=OH) can be isolated by direct crystallisation in at least 70 % yield.In contrast dimerisation of the tertiary base Wieland-Gumlich aldehyde (5 1) both in buffered aqueous acetic acid46 and in pivalic acid49 gives mainly caracurine V (61) and only a little nortoxiferine I (tertiary base corresponding to 59 R=OH). The factors controlling the various possible cyclisations are thus delicately balanced. 55 Berlage Bernauer von Philipsborn Waser Schmid and Karrer Helv. Chim. Acta 1959,42,394. 96 QUARTERLY REVIEWS An interesting but no doubt very annoying difficulty arose when the synthesis of C-diliydrotoxiferine I was attempted by Bernauer Berlage von Philipsborn Schmid and Karre~-.~~ Deoxygenation of the Wieland- Gumlichaldehyde (5 1) was achieved when the allylic hydroxyl group formed by ring-opening was replaced by bromine and the product was reduced with zinc and acetic acid.The resulting aldehyde then dimerised as expected in buffered aqueous acetic acid but though the product had colour reactions ultraviolet spectrum Rc values infrared spectrum and rotation identical with those of C-dihydrotoxiferine I it gave a picrate melting more than 60” higher than that of the natural alkaloid; the two picrates were not interconvertible. By elimination the Swiss workers were forced to conclude that this product which they designate C-dihydrotoxiferine I* is stereoisomeric with C-dihydrotoxiferine I (59; R=H) about the 19,20- and 19’,20’-double bonds.However C-dihydrotoxiferine I identical with the natural material was prepared56 from caracurine V (61) obtained in turn from the Wieland-Gumlich aldehyde (51) (see above) by reaction with hydrogen bromide to give the allylic bromide (tertiary base cor- responding to 59; R=Br) followed by reduction with zinc and acetic acid. Conversion of the final base into the corresponding methochloride gave the desired material (59; R=H). (c) C-Fhorocurarine (C-curarine III). The yellow quaternary alkaloid C-fluorocurarine was first isolated by H. Wieland Pistor and Bahr14 from a calabash-curare preparation and more recently it has been identified57 chromatographically in extracts from the bark of S. mitscherlischii. Although it has been e~tablished~~ by the partial quaternisation method (p.88) that C-fluorocurarine is a Cz0 alkaloid it is discussed in this section because of its important relations to the C40 alkaloids of the C-dihydrotoxi- ferine I “family” (p. 98). Thus both B~ekelheide’s~~ and T. Wieland’sao group have shown that C-fluorocurarine is produced by the action of concentrated hydrochloric acid on C-curarine I and Volz and T. WielandG1 obtained it by treatment of C-calebassine with the mixed anhydride of formic and acetic acid. C-Fluorocurarine has one N-methyl group on the quaternary N(b)- nitrogen atom and one C-methyl group in an ethylidene side chain. N(a)-Acetylfluorocurarine can be prepared so that the N(a)-nitrogen atom is secondary. The most striking feature of the alkaloid is however its chromophore of a kind not hitherto encountered ; its ultraviolet spectrum has a long-wavelength peak at 358 mp which undergoes a reversible bathochromic shift in the presence of alkali.With dimethyl sulphate C-fluorocurarine gives the corresponding N(a)-methyl derivative having 56 Bernauer Berlage Schmid and Karrer Helv. Chirn. Acta 1959,42,201. 57 Kebrle Schmid Waser and Karrer Helv. Chim. Acta 1953,36 345. 58 von Philipsborn Meyer Schmid and Karrer Helv. Chim. Acta 1958 41 1257. 5s Zurcher Ceder and Boekelheide J. Arner. Chem. SOC. 1958 80 1500. 61 Volz and Wieland Annalen 1957 604 1. Fritz and Wieland Annalen 1958 611 277. BATTERSBY AND HODSON CURARE ALKALOIDS 97 ultraviolet absorption identical with that of C-fluorocurarine itself but which does not show the bathochromic shift in alkali.58 Thus this shift must involve the removal of a proton from the N(a)-nitrogen of C- fluorocurarine.As the result of a masterly study of the three products obtained by borohydride reduction of N(a)-methyl-C-fluorocurarine the Zurich group suggested58 that the assembly (63) must be the chromophore of the alkaloid. The presence of an aldehyde function was confirmed by the formation of an unstable oxime and by the infrared spectrum of C-fluorocurarine which shows bands characteristic of +unsaturated /3-amino-aldehyde~.~~ By studying C-fluorocurarine and also model compounds Fritz Besch and T. Wieland62 derived the same chromophore for the alkaloid although the aldehydic nature of the carbonyl group was not recognised. However all doubt about the chromophoric system was removed by Fritz’s syn- of the aldehyde (64) the simplest compound which contains the proposed C-fluorocurarine chromophore.The absorption spectrum of this material is closely similar to that of C-fluorocurarine in neutral and in alkaline solution. It is safe then to explain the bathochromic shift in alkali as being due to generation of the mesomeric anion (66). From a consideration of its relation to C-curarine I and C-calebassine and on biogenetic grounds the Swiss proposed (65) as a hypo- thetical structure for C-fluorocurarine and this was soon confirmed by Fritz Besch and Wieland.64 Hydrogenolysis of the allylic hydroxyl group in the Wieland-Gumlich glycol (58 ; R =OH) yielded the corresponding deoxy-derivative (58 ; R =H) which underwent Oppenauer oxidation to give norhemidihydrotoxiferine (67).No details are available for the next step which involves autoxidation of the aldehyde (67) to nor-C-fluoro- curarine (tertiary base corresponding to 65) ; N(b) methylation then gave C-fluorocurarine (65) identical with the natural material. The Zurich 62 Fritz Besch and Wieland Annalen 1958 617 166. 63 Fritz Chem. Ber. 1959 92 1809. 64 Fritz Besch and Wieland Angew. Chem. 1959 71 126. 911 QUARTERLY REVIEWS group provided independent proof65 from another angle; they reduced C-fluorocurarine to norhemidihydrotoxiferine (67) which was then dimerised as described earlier to yield C-dihydrotoxiferine I (59; R=H). Since the structure of C-dihydrotoxiferine I has been related chemically to that of the Wieland-Gumlich aldehyde C-fluorocurarine must have the constitution (65).Several important reactions have been described by various workers which show that many of the calabash- curare and Strychnos alkaloids can be grouped together in so-called “families” containing mutually related alkaloids. Thus when solid C- dihydrotoxiferine I is irradiated in the presence of oxygen it is converted into C-curarine I whereas C-calebassine is formed when the irradiation is carried out in solution containing eosin as a sensitiser. The formation of C-alkaloid D from C-dihydrotoxiferine I under acidic conditions was mentioned earlier (p. 92) as was the production of C-fluorocurarine by degradation of C-curarine and C-calebassine (p. 96). These reactions are summarised in the annexed chart. (d) The “jiarnilies” of alkaloids. C-Alkaloid D iH*o+ ref.48 Oz hv ref. 66 C-Dihydrotoxiferine I -+ C-CurarineI Conc. HC1 refs. 60 59 I 02 pyridine HoAc ref. 67 HCO.OAc ref. 61 C-Calebassine -- -+ C-Fluorocurarine Inter-relations in the C-dihydrotoxiferine I “family”. A similar scheme was established for alkaloids related to toxiferine I and there can be no doubt that the corresponding fundamental changes involved in the two families are strictly analogous. Thus C-Alkaloid E in the toxiferine I “family” corresponds to C-curarine I in the C-dihydro- toxiferine I “family” ; C-Alkaloid A corresponds to C-calebassine and caracurine I1 methochloride to C-Alkaloid D. In passing it is worth noting the recent work1’ on King’s alkaloidsls which has shown that caracurine I1 methochloride and C-Alkaloid A can be isolated from the bark of S.toxifera. The separation of C-Alkaloid A is the first example of 65 von Philipsborn Bernauer Schmid and Karrer Helv. Chim. Actu 1959,42 461. 66 Bernauer Schmid and Karrer Helv. Chim. Actu 1957,40 1999. 67 Asmis Schmid and Karrer Helv. Chim. Actu 1956 39 440. BATTERSBY AND HODSON CURARE ALKALOIDS 99 an alkaloid of the C-calebassine type to be isolated in substance from a plant; all previous isolations of this type of alkaloid and those of the C-curarine I type have been from calabash-curare preparations which have undergone largely unknown treatments by their native manufacturers.1° Caracurine I1 methochloride &O+ 0 2 refs. 49 50 01 hv -+ C-Alkaloid E C-Toxiferine I - ref. 68 1 0 2 pyr’dine pivalic HOAc (ref. 69a) acid or /%ti J. C-Alkaloid A Inter-relations in the toxiferine I “family”.The analogue of C-fluorocurarine in the toxiferine I family has not so far been isolated from natural sources. It would have the structure (68) or perhaps less probably the hemiacetal form (69); attempts to make it by oxidation of the Wieland-Gumlich aldehyde (51) both at Bristolsg and at Berlage Bernauer Schmid and Karrer have recently69a proved that C-Alkaloid H is the toxiferine-like mixed condensation product of Wieland-Gumlich aldehyde methochloride (hemitoxiferine I) and its 18- deoxy-derivative (hemidihydrotoxiferine) which gives constitution (70) for C-Alkaloid H. The tertiary caracurine VI is said to be the nor-base from this “hybrid” and it has been shownsga that C-Alkaloids F and G are respectively the C-calebassine and the C-curarine analogue in this ‘‘family”.have so far been unsuccessful. Bernauer Berlage Schmid and Karrer Helv. Chim. Acta 1958 41 1202. e9 Battersby and Hodson unpublished work. 60a Berlage Bernauer Schmid and Karrer Helv. Chim. Acta 1959,42 2650. Professor H. Schmid personal communication. 100 QUARTERLY REVIEWS The key position of the Wieland-Gumlich aldehyde (51) in this group of alkaloids is now obvious. From this compound its 18-deoxy-derivative and their N(b)-metho-derivatives can be derived no fewer than nineteen* of the tertiary and quaternary alkaloids isolated from calabash-curares and the barks of South American Strychnos species. (e) C-Curarine I and C-calebassine. No structures have yet been pro- posed for C-curarine I and C-calebassine which were the first properly characterised alkaloids to be isolated from calabash-curare.12* l3 The photo- oxidation of C-dihydrotoxiferine I C40H46N4++ (59; R=H) to give C- curarine I C40H44-460N4++ has already been mentioned (p.98) and the oxygen atom introduced is presumably in an ether linkage. Little help comes from the ultraviolet absorption of C-curarine I which was earlier a t t r i b ~ t e d ~ ~ ~ ~ ~ ~ ~ ~ to an indolenine chromophore (71) but is now thought to be of a unique type;1° however it is worth noting that the spectrum is very similar to that of the amino-hemiacetal caracurine V (61). The early chemical investigations showed that C-curarine I gives acetaldehyde on ozonolysis and that it contains the system (72). von Philipsborn Schmid and K a r ~ e r ~ ~ recognised that this undergoes an * Caracurine VII hemitoxiferine I C-fluorocurarine diaboline toxiferine I C- dihydrotoxiferine I nordihydrotoxiferine caracurine V C-Alkaloid A C-Alkaloid E caracurine I1 methochloride caracurine 11 C-calebassine C-curarine I C-Alkaloid D C-Alkaloid H C-Alkaloid F C-Alkaloid G.caracurine VI 71 Schmid and Karrer Angew. Chem. 1955,67 361. 72 Karrer Bull. SOC. chim. France 1958 99. 73 von Philipsborn Schmid and Karrer Helv. Chim. Acta 1955 38 1067. BATTERSBY AND HODSON CURARE ALKALOIDS 101 interesting vinylogous Hofmann elimination as indicated when it is attacked by alkali to give a bis-tertiary base containing two diene residues (partial structure 73). Hydrogenation of the latter gave the corresponding octahydro-derivative (partial structure 74) which yielded wmethylbutyric acid (75) when oxidised by the modified (micro-)Kuhn-Roth method;74 this was the first application in the alkaloid field of a technique which has since been widely used.A major clue concerning the structure of this alkal- oid comes from its degradation to C-fluorocurarine (65) by the action of concentrated hydrochloric acid,69* 6o but much more experimental informa- tion will be required before even reasonable working hypotheses can be formulated. That there may be considerable difficulties ahead is indicated by the experiments of Boekelheide et al.75 on the Hofmann degradation of C-curarine I which are considered by the authors to point to an unsym- metrical structure for the alkaloid. It looks at present as though C-curarine I the first calabash-curare alkaloid to be crystallised may well represent the most difficult structural problem in the field.C-Calebassine C40H4802N2++ also contains two of the systems (72) but unlike C-curarine I it does not undergo vinylogous Hofmann elimina- tion; however the +NMe(b)-CH bonds can be reductively cleaved over platin~m.'~ The alkaloid contains two oxygen atoms more than C- dihydrotoxiferine I and they are present as hydroxyl groups. Evidence for their carbinolamine nature [ >-&(OH)-] comes from their reductive removal by zinc and acetic acid to yield deoxy-C-~alebassine~~ which can be reconverted into C-calebassine by photo-oxidation.68 Bernauer Schmid and Karrer78 provided further support for two carbinolamine systems by showing that a dimethyl ether can be formed from C-calebassine by treat- ment with dry methanolic acid. This ether formation is readily reversed in dilute aqueous acid at room temperature.Tetrahydro-C-calebassine in which both ethylidene side chains of the two groupings (72) have been reduced forms an analogous dimethyl ether.78 It is certain that the hydroxyl groups are intimately connected with the chromophore of the alkaloid since the ultraviolet spectra of C-calebassine and of tetrahydro-C-calebassine undergo a shift of some 10 mp in alkaline solution that is not shown by the corresponding dimethyl ethers. Indeed the alkali-induced shift is held33 to be diagnostic for the 2-hydroxyindoline chromophore (as in 76). Further removal of the hydroxyl groups to form the deoxy-derivative increases the basicity of the N(a)-nitrogen atoms and perhaps in addition increases their steric accessibility such that a N(a) N(b),N(b)-trimethyl derivative can be formed by the action of methyl 74 Garbers Schmid and Karrer Helv.Chim. Acta 1954 37 1336. 75 Boekelheide Ceder Natsume and Zurcher J. Amer. Chem. SOC. 1959 81 2256. 76 Bernauer Schmid and Karrer Helv. Chim. Acta 1957,40 731. 77 Volz and Wieland Naturwiss. 1957 44 376. Bernauer Schmid and Karrer Helv. Chim. Acta 1958 41 673. 102 QUARTERLY REVIEW5 iodide on deoxy-C-calebassine at elevated temperature^.'^ Incidentally this gives further proof of the Cpo nature of this alkaloid. The non-forma- tion of a tetra-N-methyl derivative from deoxy-C-calebassine can readily be rationalised in terms of the field effect and probably steric effect also of what is presumably the closely placed quaternary N(a)-atom; this is in contrast to the relatively large distance between the N(b)-atoms noted earlier.2CI - - r C22H32 Thus it is possible to write79 the partial structure (76) for C-calebassine and illustrate the formation of the trimethyl derivative (78) from deoxy- C-calebassine (77). In lowacid the ultraviolet spectra of C-calebassine and its tetrahydro- derivative show peaks at ca. 320 mp indicating extended conjugation. This is interpreted by the Zurich in terms of the partial structure (79) for the alkaloid which in strong acid reversibly gives the mesomeric cation We have mentioned earlier (p. 98) the relation of C-calebassine to C-fluorocurarine and also the formation of C-calebassine by photo- oxidation of C-dihydrotoxiferine I. The latter oxidation can also be achieved68 under acid-base catalysis (pyridine-acetic acid) in presence of oxygen.With this knowledge of its close relation to C-dihydrotoxiferine I it is tempting to fit the partial structure (79) into the C-dihydrotoxiferine I skeleton to give constitution (8 1) for C-calebassine. However this is unlikely on a number of grounds particularly the stability of deoxy-C- calebassine towards acids ;lo the analogue in the toxiferine I “family” namely deoxy-C-Alkaloid A is also stable under strongly acidic condi- t i o n ~ . ~ ~ (80). 7 9 Bernauer Schmid and Karrer Helv. Chim. Acta 1958,41 26. BATTERSBY AND HODSON CUKARE ALKALOIDS 103 ItH’ The structures of C-curarine and C-calebassine and those of their hydroxy-analogues C-Alkaloid E and C-Alkaloid A represent the major challenge at present to chemists working in this field; no less interesting are the structures of C-Alkaloid D and caracurine I1 methochloride formed under such mild conditions from the methyleneindoline alkaloids.How- ever even when these problems are solved there remains much to do in this difficult field as the Appendix makes amply clear. APPENDIX The following alkaloids of unknown structure (not listed in the Table on p. 81) have been isolated from calabash-curare or Strychnos species C-Guianine;80 C-Alkaloids Q R and S C-Alkaloid X;28 C-curarine 11 ;13* 78 C-isodihydrotoxiferine ;15 C-Alkaloids B C I J UB and L;18 C-Alkaloids M 0 and P;23980 caracurine I 111 and IV;22 caracurine VIII and IX methochlorides;21 melinonine E H I K L and M;29 C-fluoro- curinine ;Is pseudofluorocurine xanthocurine fedamazine ;22 C-cale- bassinine ;I8 alkaloids 1 and 2;19 toxiferine 11 ;44 croceocurine ;82 macro- phylline-A ;83 kryptocurine ;82 toxiferine 111 VIII and XII ;16J7 macusine A and B.17 *O Giesbrecht Meyer Bachli Schmid and Karrer Helv. Chim. Acta 1954,37,1974. 81 Meyer Schmid and Karrer Helv. Chim. Acta 1956 39 1208. 83 Iorio Corvillon Alves and Marini-Bettolo Gazetta 1956 86 923. Meyer Schmid Waser and Karrer Helv. Chim. Acta 1956 39 1214.