'E. E. V A N
621.1-
T.&hfELEN,
[CONTRIBUTION FROM TIIE
P.E. A L D R I C I I
DEPARTMENT OF
ASD
1. n. HFSTER,J R .
Vol.
s1
CHEMISTRY OF TIIE UNIVERSITY OF i ~ I S C O X S I N 1
The Stereochemistry of the Ipecac Alkaloids BY EUGENE E. VAN TAMELEN, PAUL E. ALDRICHAND
JACKSON
B. HESTER, JR.
RECEIVED MAY18, 1959 T h e results of stereospecific synthesis, supplemented by other definitive findings, are used to derive the first complete stereochemical formula ( X X I X ) for emetine and related I p c c x dkaloidq.
Interest in the Ipecac alkaloid group, accentuated by the importance of emetine (I) in the treatment of amoebic dysentery, has resulted in degradative investigations sufficient to establish complete gross structures for all the known members of the family. In addition, the total synthesis of emetine has
+
HaC?OOCCI-I2CII=CHCOOC~H~ CNCH2COOC2Hj --+ (HaC?OOCCH?)zCHCH(Cr\')COOC2Ha( 11) +
(EIsC~OOCCH~)~CHC(CN)C?H~COOC~H~ +
CHTO CH10 h, R = C H , c,R=H
been accomplished, the first success belonging to Preobrazhensky and co-w~rkers.*-~It is remarkable that, in contrast to the usual case of programs dealing with natural product structures, none of the information acquired in these investigations lends itself to stereochemical interpretation, and thus the interrelation of the four asymmetric centers in emetine remained unknown. The solution of this stereochemical problem is the subject of this publication. Because of a particular foundation of experience already established in this a seemingly expedient approach to positions 10 and 11 of emetine involved proving, through independent rational synthesis, the stereochemistry of a key intermediate in the Russian approach to emetine.2 The route described by Preobrazhensky, et al., involves addition of cyanoacetic ester to glutaconic ester; the resulting cyanotriester I1 was ethylated and decarboethoxylated to give the cyanodiester 111. Reductive coupling with 3,4-dimethoxy-Pphenethylamine, carried out under hydrogenation conditions in the presence of nickel catalyst, led to the lactam ester IVa. In the generation of this intermediate, an opportunity for diastereoisomerism develops, and it was reported that suhstantial amounts of both isomers were formed. (1) First reported in Communications t o the Editor. TIIISJOURNAL, (a) 79, 4817 (1957); (b) 61, 507 (1959). (2) (a) R. P. Evstigneeva, R. S. Livshits, L. I. Zakharkin, IT.S. Bainova a n d N. A. Preobrazhensky, Doklady A k a d . S a u k . , U.S.S.R., 76, 539 (19.50): (b) .'IP A. Preobrazhensky, R. P. Evstigneeva, T. S. Levchenko a n d K. M. Fedyshkina, ibid., 81, 421 (10.51); (c) R. P. Evstigneeva a n d N. A. Preobrazhensky, Tetrahedron, 4 , 223 (1058). (3) h l . Barash and J. M. Osbond, Chemistry fPI n d x s f r y , 490 (19.55). (1) A. R . Battershy a n d J. C. Turner, ibid., 1324 (1958). (.5) A. TV. Burgstahler and Z. J. Bithos, T m S J O U R N A L , 81, 503 (19.59). (0) E . E. v a n Tamelen and hI. S h a m m a , ibid., 76, i2.50 ( 1 9 Z 4 ) ; cf. C.. Stork and R. K. Hill, ibid.. 76, 949 (1954). (7) E. E. van Tamelen, A I . Shamma and 1'. E:. Aldrich, ibid., 78, 41'28 (IMG). (8) E. E. v a n Tamelen, P. E. Aldrich a n d 1'.J. K a t z . Chamisfry c?' I n d i ~ s t r y 703 , (1956); THISJOURNAL, 79, Gt2G (1057).
One of these piperidones, without elucidation of its stereochemistry, was carried through the steps IV-VI1 by the Russian workers in order to com-
-+
dl-emetine
plete the synthesis.2a Alternatively, the piperidone IVa could be first cyclized (VIII) and reduced to the tricyclic ester I X ; the latter was reported also
to yield emetine, after conversion to the corresponding 3,i-dimethoxyphenethylamide (X), followed by a second cyclization-reduction sequence.2b Clearly the operations proceeding from IV cannot affect the stereochemical relationship already established in this intermediate, and therefore assignment of its stereochemistry would define also the corresponding spatial configurations in emetine itself. The necessity of securing intermediate I V provided the occasion for repeating the procedures of Preobrazhensky and co-workers. The steps leading to the cyanodiester I11 were reproduced, although on occasion some difficulty was encountered in effecting the last step, viz., the selective hydrolysis and decarboxylation of the cyano-
Dec. 5, 1959
STEREOCHEMISTRY OF IPECAC ALKALOIDS
acetic ester m ~ i e t y . The ~ non-crystalline isomers (IVa) produced by reductive alkylation of 3,4dimethoxy-P-phenethylamine with the cyanodiester were characterized by the Russian group only as “toluene-soluble” (“isomer-A”) and “toluene-insoluble” (“isomer-B”) ; the “toluene-insoluble” material was used in the conversion to emetine. Although we could not confirm the reported solubility behavior, we were able to separate, by means of chromatography on silicic acid, the mixture of piperidonecarboxylic acids (IVc) obtained by mild saponification of the crude ester mixture. There were isolated two pure isomers, m.p.’s 151.2-152.5’ and 155.2-156.3’,’O the infrared and ultraviolet spectra of which were completely consistent with the structure IV. Because there was no way for us to associate either of these piperidonecarboxylic acids with the material used by the Russians in carrying through the total synthesis, we also were required to demonstrate which acid is convertible to the alkaloid. Thus, the methyl ester IVb of each of the acids was transformed by heating with 3,4-dimethoxy-Pphenethylamine to the lactam amide V, which was cyclized by treatment with phosphorus oxychloride. The bis-quaternary salt VI obtained from the higher-melting acid, on catalytic hydrogenation over platinum, gave product which, after purification by crystallization in the form of the hydrochloride salt, was identified as dl-emetine, contaminated with some dl-isoemetine (the C-1 epimer of emetine). On the other hand, the lower melting piperidone acid gave material a t the emetine stage which did not easily form crystalline salts; infrared analysis showed the absence of dlemetine. The plan evolved for proving the stereochemistry of the piperidone acids included (i) transformation of one of these isomers, through routes which cannot alter the stereochemistry, to the lactam -threoX I of dl-N-(3’,4’-dimethoxy-p-phenethyl)
3,4 - diethyl - 5 - aminovaleric acid, and (ii) stereorational synthesis of this reference compound through N-alkylation of 3,4-dimethoxy-P-phenethylamine with ethyl dl-threo-3,4-diethyl-5-bromovalerate (XII). The synthesis and stereochemistry of this bromoester were known from earlier work carried out in this Laboratory.8 The conversion of XI1 to XI proceeded smoothly; as expected, the intermediary alkylation product, the aminoester (9) In order t o circumvent this last pair of steps, benzyl-rather t h a n ethyl-cyanoacetate was utilized in t h e Michael addition t o diethyl glutaconate. Following t h e alkylation of t h e resulting monobenzyl, diethyl ester, removal of t h e benzyl group b y palladium-catalyzed hydrogenolysis was accompanied b y decarboxylation, thereby affording cyanodiester identical with material provided b y t h e original route. This modification was not, however, entirely free of practical difficulties, a n d offers no a d v a n t a g e over t h e original procedure. (10) Preparation of these acids also has been reported b y Barash a n d Osbonda a n d b y B a t t e r s b y a n d Turner.‘
6215
X I I I , cyclized under the conditions of the reaction to the desired trans-lactam X I , a high-boiling liquid. In order to accomplish the correlation between the emetine synthesis intermediate IV and the diethyl lactam stereoisomer XI, the methyl ester of the higher melting piperidone acid was reduced selectively through the use of lithium borohydride under carefully defined conditions to the lactam alcohol XIV, m.p. 115.8-118.0°.11 Preparation of the corresponding tosylate X V was followed by conversion to the isothiouronium salt XVI ; reductive desulfurization with Raney nickel then gave rise to a good yield of product identical
(
CH3
C HlSC ‘-Iu
I I1
I:.T
X v I NH,
with the lactam X I of known stereochemistry.12 The identification of the lactarn IV as a member of the trans series was substantiated by phosphorus oxychloride cyclization and catalytic hydrogenation of the lactam X I from both sources; the two specimens of crystalline hydrochlorides, as well as the (11) Reduction of t h e ester group with maintenance of t h e l a c t a m function appeared t o be necessary for t h e successful prosecution of t h e scheme outlined above. Thus, although total reduction with lithium aluminum hydride is attractive because of its simplicity, tosylation of t h e resulting alcohol i would very likely be followed b y cyclization
I
%H3CH20Ts
ii (ii) t o t h e quinuclidine quaternary salt (iii); because of t h e s y m m e t r y properties of t h e latter, t h e stereochemical relationship characteristic of t h e preceding intermediates would be absent in this substance. (12) Because our supply of t h e Irans-lactam acid I V was limited, t h e conditions for its conversion t o t h e frans-diethyl compound XI were developed in a model series. 8-Phenethyl tosylate was converted t o t h e isothiouronium tosylate, which, on nickel desulfurization (E. Hardegger a n d R. M. Montavon, Helv. Chim. Acta, 29, 1199 (1946)) provided a 95% yield of ethylbenzene.
E. E.
6210:
VAN
TAMELEN, P. E.
A
CH 40 XI+ CHYO
CH
4
tetrahydroisoquinoline free bases XVIII, were indistinguishable. In this way, the t r a m stereochemical relationship of C-10 and C-11in emetine was established.l39l4 All of the available evidence indicates that the hydrogen at C-1' in emetine is cis oriented with respect to the hydrogen at C-10.The fact that emetine, rather than a diastereoisomer, is produced from the A"J'-unsaturated intermediate VI by catalytic reduction suggests that this more stable configuration, i.e., C-1' axial hydrogen (XIX), is generated. Since similarly constituted substances,
c -N+,
~ AND ~ J.
R.~HESTER, ~ ~JK.
~
Vol. 81
i
amount of its epimer.'j The reduction product must have, therefore, the more stable configuration, and the stereochemistry already portrayed in structure XVIII follows. Catalytic reduction of the einetine synthesis intermediate VI is closely akin to that of XVII, and one would expect the same stereochemical outcome at C-1'. Finally, a correlation independent of certain assumptions implicit in the above arguments, has been achieved : the conversion, by means of an unambiguous method, of the emetine synthesis intermediate XXI15 to the tricyclic reference compound XVITI. Lithium borohydride reduction followed by tosylation provided the lactain X X I I I ; lithium aluminum hydride reduction of the latter then afforded the base XTVII, of established trans stereochemistry.16
H CH,C H
e.g., Aj-dehydroyohiiiibarie (XX) and -corynaii-
theane (XXI) (both D, E frans) are reduced to members of the C-3, C- G - c i s similar
XX
l
During the course of the work described above, and after our preliminary results were announced, l a the assignment of relative stereochemistry a t C-1' was questioned by Brossi, Cohen, Osbond, Plattner, Schnider and \.T'ickens.'' These authors alleged that (i) contrary to the report of Preobrazhensky et al., the tricyclic interinediate IX gave rise to-riot emetine-but two diastereoisomers of the natural base (-lal and Abz); (ii) again in conflict with the Russian investigators, catalytic reduction of the tetradehydroenietine VI afforded SOf/h of tlie emetine isomer Ab2; and (iii) the intermediate XXV on hydrogenation led to only a 20% yield
XXI
behavior might be anticipated in the case under scrutiny. Support for this tentative conclusion was acquired in several ways. First, lithium aluminun1 hydride reduction of the cyclized intermediate VI yielded the same products formed by catalytic reduction. This imine salt does not present any exceptional steric barrier to reduction and, in the absence of such a factor, hydride reduction to the more stable product would be expected. Second, sodium and alcohol reduction of one of the reference compounds, the imine salt XVII, gave rise to the same tetrahydroisoquinoline isomer produced by catalytic hydrogenation, with no detectable
of einetinc, in addition to *Yb2. Ijrussi, ct ul.,
accepted the previous findingsLa which required that the reduction processes described generate the and therefore more stable configuration at C-l', were obliged to entertain a new stereochemical expression (XXVI) for emetine and related alkaloids, in which the assignment at C-1' was the reverse of that initially proposed. More recently, (I;) In addition to the normal reduction product, there was isolated secoiid substance, which elemental, methoxyl and spectral analysis s h o w e d
( 1 8 ) A synthesis of XI, identical w i t h the one described a h o v e , c o i i ~ stltutes the unarnbiguoua portion of the evidence cited in support o f t h e IO,il-lvu?zs relatioilship b y Battersby, et ut. (Cl:einisfi,y I%fiustry, '383, 983 (1057)), whv, without reference to our previous work,8 de-
L:
scribed the same bromo ester intermediate S I I , anti synthesis thereof, i i n t iiied in another connection in this 1,aboratory. (1.1) l'he 10.11-traiis assignment t o emetine w a s supported by -1. Urossi, A . Cohen, J. R I . Osbond, P I . A . Plattner, 0. Schnider and J. C . IVickens (Chemisir-v & I?idusfvy, 491 (193S)), who related the cis-piperidoneacetic acid (IV) t o ethyl cincholoiponate, the stereochemistry of which had been established prei-iously ( V . Prelog :mil E Zalnri, H e l r ~ .Chilli. . I c l n , 27, 5 3 3 (1914).
reduction is precedented b y tlie case of O-methJ-ll,sychotrirle (XXVIT, ' CH.4 I< = C H s j (P. I. P p m a n , J . C'i:e717. Soc., 111, -110 (1017)) a n d structure iv
