1540 FINDLAY VOL.

24
[CONTRIBUTION INSTITUTE
FROM THE NATIONAL OF ARTHRITIS
AND METABOLIC
DISEASES,
NATIONALINSTITUTES
OF HEALTH,
PUBLIC HEALTHSERVICE, DEPARTMENT EDUCATION
O F HEALTH, AND WELFARE]

The Three-Dimensional Structures of the Cocaines. 11.

Racemic Allococaine and Racemic Allopseudococaine
STEPHEN P. FINDLAY

Received M a y $6,1969

Catalytic hydrogenation of racemic Zcarbomethoxy-tropinone in acetic acid yields racemic alloecgonine methyl ester,
which can be transformed to the racemates of alloecgonine, allococaine, allopseudoecgonine, allopseudoecgonine methyl
ester, and allopseudococaine. Some limitations of a generalization concerning the course of the catalytic hydrogenation of
cyclic ketones as i t applies to certain keto derivatives of the tropane and morphine alkaloids are noted. The three-dimensional
structures of the new cocaines are tentatively assigned. The possible utility of molecular rotation data in ascertaining the
absolute configuration of transformation products of the 2-carbomethoxy derivatives of both tropinone and N-methyl-
granatonine is indicated. Some other possible methods of syntheAizing the new cocaine isomers and the drawbacks thereof
are mentioned.

Having established the molecular structure of the The maturation of stereochemistry which has
naturally occurring and medicinally important occurred meanwhile makes possible now the de-
alkaloid, E-cocaine (I) Willstatter and his col- termination of such questions in natural products
laborators remarked that three other stereoisomers generally, and, relative to the cocaine problem,
having the same sequence of atomic linkages has allowed the unequivocal assignment of three-
should exist.lt2vaOf these he and his associates ob- dimensional structures to the known isomers,
tained, both by the transformation of 1-cocaine and cocaine and pseudococaine (I11 and IV, respec-
also by total synthesis, one other stereoisomer, tively, R = CHs R'= COC6H6).6-g By the process
pseudococaine (11) .4.6 Although they recorded of elimination the two unknown isomers must be
represented by the structures, V and VI (R= CH3,
* ,COzCH3 /H R'=COCsH6 in both); and this information per-
CH,-(H-C;..H CHz-CH-C;*. CO,CH,
mits a more rational approach to their selective
synthesis.
NCH, NCHj
I\ Ii
I 11
the availability of a third14neither of the two re-
maining cocaine stereoisomers had been isolated
when the problem was abandoned some thirty-
five years ago.
The isolation of the four possible modifications
at that time, while it would have substantiated
their predictions based on the fundamental struc-
tural theory of organic chemistry, would have left
unanswered the related and equally important
question as to the steric or three-dimensional re-
lationship of the functional groups in each of them.
Although Willstatter had, to be sure, not over- b H
looked this problem, stereochemistry was a t that OR' OR'
time in too rudimentary a state to permit a certain V 1'I
conclusion; and his opinion concerning the relation The isolation of these two unknown cocainee
of cocaine to pseudococaine2 has indeed recently would be noteworthy as completing one of the
been disproved.6 classical topics of alkaloid chemistry. It would
also afford ready access t o potentially valuable
(1) R. Willstatter and W. Muller, Ber., 31, 2655 (1898).
(2) R. Willstatter and M. Bommer, Ann., 422, 15 derivatives of the medicinal, atropine ; and indeed
(1921).
( 3 ) R. Willstatter and A. Bode, Ann., 326, 42 (1903). (7) S. P. Findlay, J. Am. Chem. SOC.,7 5 , 1033 (1953).
(4) R. Willstatter, 0.Wolfes, and R. Mader, Ann., 434, (8)G. Fodor, Nature, 170, 278 (1952); G. Fodor and
111 (1923);Cf.,E.Merck, German Patent 389,359. 6. KovAcs, J . Chem. Soc., 724 (1953).
(5) Cj., (a) A. Einhorn and A. Marquardt, Ber., 23, 468 (9)According to E. Hardegger and H. Ott these struc-
(1890); (b) C. Liebermann and F. Giesel, Ber., 23, 508 tural formulae rather than their mirror images represent
(1890). the absolute three-dimensional likeness of the naturally
(6) S. P. Findlay, J. Am. Chem. Soc., 7 5 , 4624 (1953); occurring alkaloid and its derivatives [Helv. Chim. Acta, 38,
76,2855 (1954). 312 (1955)l.
OCTOBER 1959 RACEMIC ALLOCOCAINE AND ALLOPSEUDOCOCAINE 1541

there are recorded few, if any, examples of tropines applied to the keto ester, converts it in excellent
bearing substituents at the C2-position. yield to alloecgonine methyl ester (V, R=CH,,
The three-dimensional structures of the related R’=H) .g
substances, pseudotropine (VII) and tropine (IX), Since, unlike tropinone, 2-carbomethoxytropin-
have also been established.’O Reduced with sodium one exists largely as the enol (Xa)14 in solutions
and alcohol, sodium amalgam, or sodium and most favorable to its catalytic hydrogenation, an
moist ether, the ketone, tropinone (VIII), affords adequate explanation of the steric outcome of this
pseudotropine (VII)11; and on catalytic reduction process must take cognizance of the possibility that
this ketone yields the isomeric amino alcohol, an enolic rather than a carbon-oxygen double
tropine (IX). l2,I3These transformations are, among bond is reduced. The results of the reduction of
the tropane alkaloids, remarkably stereospecific, some basic ketones carried out in connection with
no evidence of the formation of more than one this investigation together with information of a
isomer having been found in either the sodium and similar kind drawn from the chemistry of the mor-
alcohol” or the catalytic reduction of tropinone12*13 phine alkaloids permits one to account satisfactorily
(see experimental section). By the action of sodium for the product obtained, regardless of the tauto-
amalgam on the 2-carbomethoxy derivative of this meric form actually reduced.
ketone, both ecgonine and pseudoecgonine (I11 According to the nomenclature now customary
and IV, respectively, R = R‘ = H) 2 - 4 result in which for specifying the location of ring substituents, l5
compounds the C3-hydroxyl group thus created the C3-oxygenis equatorial in molecules having the
has the same steric relation to the nitrogen structures, 111, IV, and VII, and axial in the struc-
as the similarly produced C3-hydroxyl group of tures, V, VI, and IX. As a rule, the reduction of
pseudotropine. l1 Hence, the catalytic hydrogena- ketones by means of such reagents as sodium and
alcohol affords predominantly the equatorial con-
figuration of the hydroxyl group,15 and the well
nigh exclusive formation of pseudotropine (VII)
(see experimental section) and the equally one-
sided production of ecgonine and p~eudoecgonine~
by similar methods apparently provides additional
corroboration of this generalization.
In general the catalytic hydrogenation of both
hindered and unhindered cyclic ketones in strongly
acidic media is rapid and affords the axial configura-
tion of the hydroxyl group formed, while in basic
and in neutral media this process is slow and leads
to the axial configuration only when the ketone is
strongly hindered.15 I n this investigation the plati-
num-catalyzed hydrogenation of tropinone which
x o \I” is not notably hindered was observed to occur a t
b H the same rate in alcohol as in aqueous acetic acid,
OH and in both instances the product appeared to con-
sist entirely of tropine (IX) (which has the axial
configuration of the C3-hydroxyl group). It was ob-
served also that in aqueous hydrochloric acid the
platinum-catalyzed hydrogenation of dihydro-
codeinone (XI) proceeds quite slowly and gives
”---”---$
H
only a mixture of products from which no
pure component could be isolated, while in
Xb 0
the relatively weakly acidic medium, aqueous
acetic acid, this process took place about ten
times more rapidly and yields largely dihydro-
tion of the keto tautomer (X) of this substance
codeine (X1I).l6Finally, it has been reported that
would be an obvious way of attempting the synthesis
8-hydroxydihydrocodeinone (XIV) is reduced much
of the unknown isomers (V and VI): and indeed,
more rapidly in alcohol than in aqueous hydro-
as described in more detail hereinafter, this process,
chloric acid and to the same alcohol, presumably
(10) (a) G. Fodor and K. NAdor, Nature, 169, 462
(1952); (b) J. Chem. Soc., 721 (1953); (c) L. F. Fieser and (14) S. P. Findlay, J. Org. Chem., 22, 1385 (1957).
A. Nickon, J . Am. Chem. Soc., 74, 5566 (1952). (15) For an informative discussion of ring conformations
(11) R. Willstatter, Ber., 29, 936 (1896). and the appertaining nomenclature, consult D. H. R.
(12) J. van de Kamp and M. Sletainger, Chem. Abstr., Barton, J. Chem. SOC.,1027 (1953).
39, 2080 (1945). (16) I n basic media (pyridine) dihydrocodeinone is also
(13) L. C. Keagle and W. H. Hartung, J. Am. Chem. readily hydrogenated to dihydrocodeine. K. Goto and T.
Soc., 68, 1608 (1946). Ami, Ann., 547, 194 (1941).
15$8 FINDLAY VOL. 24

8-hydroxydihydrocodeine (XV) .l7 Quite evidently that are essentially flat and, furthermore, contain
the foregoing generalization, of undeniable use- no nitrogen which is accessible from only one side of
fulness among most alicyclic compounds, fails the molecule. I n such instances, interaction between
in its application to such complex basic ketones as either side of the molecule and the catalyst is rela-
tropinone and dihydrocodeinone; the cause of this tively probable; therefore, ring conformation may
failure suggested below has the double advantage be decisive in determining the configuration of the
of accounting satisfactorily for the catalytic hy- substituent produced by the catalytic hydrogena-
drogenation of 2-~arbomethoxytropinone,whatever tion. On the other hand, among bridged ring com-
the tautomeric form in which it is reduced, and of pounds such as the complex morphine alkaloids and
indicating the nature of the exceptions to this rule the bicyclic terpenes of the camphor type, one side
which may be anticipated.18 of the molecule is much more convenient to the
catalyst than any other-i.e., steric accommodation
and/or coordinating power become factors to be
considered in the steric course of the hydrogenation
of ketones. If these two factors exceed the confor-
mational one in importance, any adherence of such
bridged ring compounds t o the foregoing generali-
zation arises from the fortuitous circumstance
that in the examples so far studied these three
H XI1 influences have usually acted in concert. Only by
an examination of instances in which these in-
fluences operate in opposing directions can one
form some estimate of their relative importance and
thus predict intelligently the probable steric
course of such hydrogenations generally.
The catalytic hydrogenation of camphor (XVI)
and of dihydrocodeinone indicate the possible
results of such an examination. I n acetic acid the
XIV XV platinum-catalyzed hydrogenation of camphor
Possibly what may be referred to as steric ax- results in isoborneol (XVII)19120in which the hy-
commodation between the catalyst and the molecule droxyl group is equatorial.'S From X-ray crystal-
to be reduced (particularly those molecules con- lographic studies it appears that Ring C of mor-
taining two or more basic atoms), in conjunction phine and many of its derivatives may have the
with a coordinating power of metallic catalysts for boat or semi-boat conformation.21*22Hence, the
basic atoms like nitrogen and oxygen, outweighs in platinum-catalyzed hydrogenation in acidic media
importance the tendency of a ring to assume a of dihydrocodeinone (XI) should, according to the
particular conformation and hence to yield a partic- foregoing generalization, convert this ketone largely
ular configuration of an attached substituent. to dihydroisocodeine (XIII) in which the c6-
Many organic compounds, such as certain of the hydroxyl would have the axial configuration for the
terpenes and sterols, have molecular structures boat conformation.22Such an outcome is, however,
contrary to that expected from considerations of
(17) S. P. Findlay and L. F. Small, J . A m . Chem. Soc., coordinating tendencies and of steric accommoda-
73, 4001 (1951).
(18) The supposition that catalytic hydrogenation of tion, which generally produce among the morphine
unsymmetrical cyclic ketones in acidic media (which leads alkaloids a C6-hydroxyl group trans to the nitrogen
to the alcoholic groups with the axial configuration) is rapid16 bridge ;23 as noted above, dihydrocodeine (XII)
appears to be based on the usually lesser thermodynamic having presumably an equatorial C6-hydroxyl
stability of an alcohol with an axial hydroxyl group re-
lative to the epimeric alcohol in which this group is equatorial.
Hence, assuming the rate of formation of the latter to be (19) G. Vavon and P. Peignier, Conzpt. rend., 181, 183
the same in acidic as in basic and in neutral media, the (1925).
predominant production of the former in acidic media must (20) M. Lipp, E. Oeckinghaus, and C. L. Conze, Ber.,
be due to a marked favouring by acid of the hydrogenation 74, 6 (1941).
mechanism whereby the former is produced. It is neverthe- (21) M. Mackay and D. C. Hodgkin, J. Chem. SOC.,
less conceivable that the axial configuration might pre- 3261 (1955).
dominate under such conditions by the reverse process- (22) K. W. Bentley and H. M. E. Cardwell, J . Chem.
namely, the inhibition by strong acid of the hydrogenation Soc., 3252 (1955).
mechanism whereby the equatorial configuration is formed; (23)Aoonsiderationof thesteric factorin theratalytic hydro-
i t has been pointed out that, in a t least one or two instances genation of morphine Csketones enabled Fieser to predict
(coprostanone and possibly 2-methylcyclohexanone), the correctly the configuration of the C6-hydroxyl so produced
rate of platinum-catalyzed formation of the alcohol having in this class of compounds. L. F. Fieser and M. Fieser,
the axial Configuration of the hydroxyl group in strongly Natural Products Related to Phenanthrene, Third Ed. , Rein-
acid media is slower than the rate of formation of the hold Publishing Corporation, New York, 1949, pp. 24-25
equatorial epimer in neutral media (S. P. Findlay, Archives [Cf., H. Rapoport and G. B. Payne, J. Orq. Chem., 15,1093
of The Chemical Society, Paper 8/1044). (1951); J . Am. Chem. Soc., 74, 2630 (1952)l.
OCTOBER 1959 RACEMIC ALLOCOCAINE AND ALLOPSEUDOCOCAINE 1543

group (trans, however, to the nitrogen bridge)

is in fact produced.24
That coordinating power and steric accommoda-
tion are highly significant in other, similar types of
catalytic hydrogenation makes itself evident from
the reduction of the ethylenic and enolic double
bonds of the complex alkaloid, thebaine. The nu-
merous products thus ~ b t a i n e d ~ result ~ - ~ ' from the
addition of one, two, or three molecules of hydrogen
to the side of the molecule bearing the nitrogen
bridge;28v29 no such products arising from addition
to the other side appear ever to have been found.2a
Among the tropane alkaloids, tropinone consti-
tutes another example wherein these factors may
either cooperate or conflict. Because in this in-
stance an acidic medium (which favors the forma-
tion of the axial configuration) obviously rein-
forces the effects of coordination and of steric
fit, the conversion of tropinone (VIII) in acetic
acid to tropine (IX) by catalytic hydrogen is not
surprising. That the substitution of alcohol as
solvent-which puts the conformational factor in
opposition-has no appreciable influence either
on the rate of reduction or on the configurational H
XXII
outcome indicates that for this molecule, as for
dihydrocodeinone, the steric and coordination should produce the tropine configuration also.
factors predominate and hence bring about the Hence, the conversion of this p-keto ester to al-
addition of hydrogen somewhat as illustrated by loecgonine methyl ester (V, R=CHs, R=H)
the structure, XVIII, regardless of the relative appears to harmonize very well with pertinent
acidity of the solvent. If so, this process applied available information concerning such processes.
to the isomeric and isosteric substances, tropanone-2 The platinum-catalyzed hydrogenation of race-
(XIX and its mirror image) and isotropinone (XXI mic 2-carbomethoxytropinone (X or Xa and its
and its mirror image), may afford substantial antipode) in aqueous acetic acid gives the 0-
quantities of tropanol-2 (XX and its mirror hydroxy amino ester, racemic alloecgonine methyl
image) and isopseudotropine (XXII and its mirror ester (V and its antipode, R=CH3, R'=H) in
image), respectively, wherein the newly created high yield (ca. 80%). As described in the experi-
hydroxyl groups are equatorial, the relative acidity mental section this process was attempted under
of the solvent notwithstanding. a variety of conditions. Of particular note are the
Hydrogenated in the keto form, 2-cnrbomethoxy- observations that methanol and benzene are poor
tropinone should, as already noted, resemble or unsatisfactory solvents for the hydrogenation
tropinone in yielding the tropine configuration of and that hydrogen chloride markedly retards the
the Ca-hydroxyl group. If it is hydrogenated in the reductioixa0
enol form, the conformational factor is eliminated; Racemic alloecgonine methyl ester, the only
steric accommodation and coordination power, reduction product found, can be readily isolated
operating by analogy with the catalytic reduction and purified as its acetate. The free ester, C10H17N03,
of the enolic double bond of thebaine to neopine which melts a t 81.5-83.5", resembles the long
methyl etherz7 and to t e t r a h y d r ~ t h e b a i n e ~ known
~ , ~ ~ methyl esters of ecgonine and pseudoec-
gonine in physical and chemical properties. Sub-
(24) If, as suggested,21 the CG-hydroxyl group of di-
limation or distillation in vacuo except a t low tem-
hydrocodeine and its near relatives is equatorial, the retard- peratures causes a pronounced decrease in its
ing influence of hydrochloric acid on the catalytic process melting point; and, when hydrolyzed, it affords
leading to these substances is explicable.'E racemic allopseudoecgonine (VI and its antipode,
(25) C. Schopf and L. Winterhalder, Ann., 452, 232 R=R'=H), CgHI5NO3,melting a t 243", in about
(1927).
(26) L. F. Small and G. L. Browning, J . Org. Chem., 3, the same quantity as racemic alloecgonine (V
618 (1939). and its antipode, R=R'=H), CgHl6NO3,melting
(27) L. F. Small, J . Org. Chem., 20, 953 (1955). at 241". The former amino acid gives upon Fischer
(28) G. Stork, The Alkaloids, R. H. F. Manske and esterification the expected Cz-epimer, racemic
H. L. Holmes, Eds., Vol. 11, Academic Press, New York,
1952, pp. 175 and 199. (30) Aqueous hydrochloric acid appears to be quite
(29) Cf.,K. W. Bentley, The Chemistry of the Morphine satisfactory as a medium for hydrogenating 3-carbomethoxy-
Alkaloids, Oxford a t the Clarendon Press, London, 1954, 1,2,6-trimethylpiperidone-4. C. Mannich, Arch. Pharm.,
pp. 197-203. 272, 323 (1934).
1544 FINDLAY VOL. 24

allopseudcecgonine methyl ester (VI and its anti- lished, ** the three-dimensional structures assigned
pode, R=CI-IB, R'=H), Cl0HI7SO3,which melts throughout this discussion follow.
a t 80". Since the Cs-hydroxyl groups of the two new
Although ttle melting points of the two new ecgonine methyl esters have the axial configuration
ecgonine methyl esters lie close together, mixtures and are otherwise hindered, a greater resistance to
of the two melt much lower; and their distinctive benzoylation might be anticipated ;benzoyl chloride
salts and i:ifrared spectra leave no doubt con- in pyridine, which was the most effective technique
cerning their noli-identity. Likewise, a mixture of employed, did give a low yield (ca. 40%) of racemic
racemic ulloecgonine and the allopseudo isomer allococaine, C17HzlN04(m.p. 82-84"), and hardly
melts much lower than either component, and their more than a trace of racemic allopseudococaine,
hydrochlorides (melting a t 231-233" and at 213" C17H21N04 (m.p. 93-95'). As methyl benzoate was
respectively) and differing capacity for hydrate liberated during the benzoylation of both esters
formation make their dissimilarity indubitable. in noticeable degree, the pyridine-catalyzed de-
From the large-scale reduction of racemic 2- composition of the new cocaines contributes to the
carbomethoxytropinone with sodium amalgam, difficulty of realizing a satisfactory yield. This
Willstatter and his collaborators obtained, besides propensity to transesterification is such that
racemic ecgonine and racemic pseudoecgonine, a the pure bases appear to autocatalyze their de-
small quantity of a third substance, which he called composition, the odorless crystals of each of the
the 'drittes racemisches E k g ~ n i n ' . This
~ melted new racemic cocaines gradually changing to a
a t 229' (corr.) and afforded a hydrochloride melt- brown oil and emitting the unmistakable aroma of
ing a t 231". Assuming the reliability of the data methyl benzoate. The physical and chemical prop-
noted above, one may reasonably infer that their erties of the new racemic cocaines do not depart
hydrochloride mas pure or nearly pure racemic noticeably from those of the known isomers. Some
alloecgonine hydrochloride and that their free impression of the similarities and differences
'drittes racemisches Ekgonin', because of the ready between them may be obtained from Table I.
epimerization of alloecgonine methyl ester, con- The extensibility of the foregoing procedures to
tained both the allo and allopseudo ecgonines. the acquisition of the optically active antipodes of
the all0 and allopseudo series was established in
The three-dimensional structures assigned above the ready conversion of d- and 1-(2-~arbomethoxy-
to the allo and allopseudo series derive primarily tropinone) to the corresponding optically active
from the reactivities of the methyl esters toward alloecgonine methyl esters. No evidence of racemi-
methyl iodide. The variety of products got from zation or other unexpected phenomena like that
the reaction of ecgonine methyl ester and this reported by Mannich30 was noticed. In methanol
alkiodide4,6 arises from the trans relation of the d-alloecgonine methyl ester is weakly dextrorota-
Cz-hydrogen and the nitrogen atom, Hofmann
tory (ag +0.15 =k0.03) and affords a strongly
Degradation being thus possible by the relatively
facile trans elimination process and further pro- laevorotatory ecgonine (a?. ca -47" in water).
moted by the electrophilic C2-carbomethoxygroup.6 Reduced with lithium aluminum hydride in tetra-
Only the expected methiodide results from the hydrofuran this ester (XXIII) was converted to a
combination of methyl iodide with pseudoecgonine substance, presumably alloecgoninol (XXIV), CS-
methyl ester4$6in which the (&hydrogen is in the
cis-position relative to the nitrogen atom. Racemic NCHj NCH3
/I / 'i
alloecgonine methyl ester yields readily a pure
methiodide, C11H201N03,melting a t 106-197",
both from methanol and from acetone, while the
allopseudo isomer affords, under the same circum-
stances, a mixture of products from methanol
from which no methiodide was obtained and only OH
FH
OH
XXIII SXIV
a poorly defined methiodide from acetone. One
may conclude teutatively that allopseudoecgonine (31) Evidence of an indirect character supports this con-
methyl ester and its derivatives have the ecgonine clusion. If 2<arbomethoxytropinone is hydrogenated in the
methyl ester configuration of Cz in which the enol form,14 hydrogen should add, for reasons already given,
hydrogen a i d carbomethoxy groups are attached in a cis manner from the side having the nitrogen bridge
trans and cis, respectively, to the nitrogen atom, creating the structure, V ( k C H 3 , R'=H); if in the keto
form, of the two possible and interchangeable configurations
while the pseudo and the all0 series are alike in a t CS (X and Xb), that (X) in which the carbomethoxy
having the Cz-hydrogen and -carbomethoxy groups group is lrans to the nitrogen bridge and, presumably,
attached cis and trans, respectively, to the nitrogen equatorial, appears to be the stabler one, and this would
atom. 31 The spatial location of the C3-hydroxyl be frozen by the reduction of the keto group.
Also, racemic allopseudoecgonine combines with meth-
having been ascertained above and the configura- anolic methyl iodide, the corresponding methyl ester result-
tional relation of the optically active 2-carbometh- ing [C'f., ref. 6; also F. G. Novy, Am. Chem. J., 10, 145
oxytropiiiones to 1-cocaine having been estab- (1888) 1.
OCTOBER 1959 RACEMIC ALLOCOCAINE A N D ALLOPSEUDOCOCAISE 1545

TABLE I
POINTS
MELTING OF THE RACEMIC A N D SOMEOF THEIRDERIVATIVES
COCAINES
Cocaine Derived Racemic Ecgonine
Modifications Racemic Cocaine Methyl Ester Derived Racemic Ecgonine
Naturally occurring 79-80 '' Liq.' 21204
hydrochloride 187'' 19504 24704
Pseudo 81.503 12803 251 O 3
hydrochloride 205. 503 - 193-194'
All0 82-84 '32 81.5-83. 5'g2 241°32
hydrochloride 201. 5°32 - 233°32
Allopseudo 93-95' 32 80'32 243°32
hydrochloride - 192°32 213°32

H17N02,melting a t 201.5-202.8", which was also Still another variation of this kind of attack,
optically active (& -6.3, in water). which has been attempted elsewhere, 3 4 involves
The foregoing synthetic scheme is simple and the use of ecgoninol (XXVIII), the lithium alumi-
unambiguous; and, were it not for the benzoylating num hydride reduction product of cocaine. Thionyl
reaction which can no doubt be improved, the over- chloride is said to convert this diol to the chloro-
all yields would be high. Moreover, as it leads to hydrin (XXIX) which heat transforms to an ether,
both antipodes of each of the new cocaines, it con- possibly XXX.34By the action of hydroxide the
stitutes a complete solution of the problem, while ether reverts in part to ecgoninol (XXVIII) and
a scheme based on the transformation of 1-cocaine gives in part a new isomeric diol t o which the
can afford but one antipode of each. Several ap- structure, XXXI, is assigned34;this is isomerized
proaches to a limited solution of the latter kind were with sodium amylate to a product which, if the
considered, but investigation showed them to be preceding part of the scheme is correctly repre-
unpromising. For example, the oxidation of the sented, should be alloecgoninol (XXIV). However,
methyl esters of ecgonine and pseudoecgonine to the reported melting points of this substance and
d-(2-~arbomethyoxytropinone)in substantial quan- its hydrochloride differ appreciably from those of
tity proved impracticable;'* while an SN2 displace- the alloecgoninol and its hydrochloride described
ment reaction involving the tosyl derivatives of above. In view of the reduction of cocaine to ecgon-
these esters might allow the desired inversion of the inol with lithium aluminum hydride in the ex-
configuration at C3, the preparation of the requisite pected manner,s the similar reduction of alloec-
tosylates is difficult.33 gonine methyl ester permits no alternative to the
structure, XXIV, for the product so obtained in
(32) See experimental section. this investigation. It is therefore reasonable to
(33) The mutarotation of salts of anhydroecgonine suppose that the foregoing chlorohydrin is a rear-
(XXV) in aqueous solution (A. W. K. de Jong, Rec. trav.
chim., 42,996-7 (1923)), which is no doubt due to hydration
rangement product and that an alternative se-
of the a,p-ethylenic linkage, suggests another avenue of quence, perhaps XXVIII+XXXII+XXXIII4
access. This phenomenon which, among simple aliphatic XXXIV+XXXV, is the correct one. The selective
compounds, has been studied extensively by Lucas and his oxidation of alloecgoninol and of allopseudoecgon-
collaborators [see, e.g., D. Pressman and H. J. Lucas, inol would give optically active alloecgonine and
J. Am. Chem. Soc., 64, 1953 (1942)l now finds exemplifica-
tion among the a,p-unsaturated ketones derived from mor- allopseudoecgonine, r e s p e ~ t i v e l y , which
~~ could,
phine" [U. Weiss, J. Org. Chem., 22, 1505 (1957)l. An of course, be transformed to the hitherto unknown
equilibrium involving ecgonine (XXVI) and alloecgonine cocaines in the manner indicated above. The suc-
(XXVII) in the salt form is thus quite conceivable: cessful oxidation in this manner of the two diols
obtained by the foregoing reaction sequence has
apparently not yet been realized.35
Although the rotatory powers of 1-cocaine and its
+ 2H20 - derivatives have not been measured in a uniform
manner, the chemical literature contains enough
usable data of this kind to warrant the conclusion,
xxv (34) 6. KOV&CS, I. Weisz, P. Zoller, and G. Fodor,
+ Helv. Chzm. Acta, 39, 99 (1956).
(35) It has recently been reported also (K. Zeile and
W. Schulz, Chem. Ber., 89, 678 (1956)) that a third racemic
cocaine, C1~HPINOI.l/~H~O, m.p. 156-157', is obtainable by
converting Willstatter's 'drittes racemisches Ekgonin' to
the corresponding methyl ester, CIOH17N03.1/~H~0, m.p.
203.5', and benzoylating this. No information concerning
, II
the purity of the starting material, the removability of the
hydrated water, the properties of appropriate derivatives,
XXVII Ho or the three-dimensional structure was furnished.
1546 FINDLAY VOL. 24

NCH3 NCH, NCHI NCH,

C H zOH

H
soc12

H
I
OH
-HC1
:o
H
-HzO(0H-I
OH OH
H

XXVIII XXIX xxx XXXI XXIV

+
XXVIII

d3H3 "C1

\c OH
H H HO H
XXXII XXXIII XXXIV xxxv
+
XXVIII

suggested earlier,I4 that the simpler derivatives of sumption [1100 ml. (S.T.P.), loti%] of gas had nearly
the naturally occurring base have a molar rotation ceased. The catalyst-free solution was concentrated in
vacuo on a warm water bath to a viscous, nearly colourless
more positive (or less negative) than that of the cor- gum which was treated cautiously with water (15 ml.)
responding salts (Table 11),and vice versa for the and saturated aqueous potassium carbonate (50 ml.), the
derivatives of the unnatural or antipodal base. resulting white mixture then being extracted with ether
Among the pseudo derivatives some exceptions to (4 X 100 ml.). Unreduced keto ester was removed by wash-
the rule occur, but for asymmetric derivatives of ing with saturated aqueous potassium carbonate (20 ml.)
mixed with 3N aqueous potassium hydroxide ( 4 ml.), the
such compounds as 2-carbomethoxytropinone and potassium salt separating in a yellow solution between the
2-carbomethoxy-N-methylgranatonine the deter- ethereal and the colourless aqueous phases (omission of
mination of the sign of this difference for several this step did not noticeably interfere with the isolation of
closely related base-salt pairs is sufficient to decide the pure reduction product). Concentration of the dried
(sodium sulfate) extracts on a water bath (ca. 50°), first a t
the absolute configuration or three-dimensional atmospheric pressure and then in vacuo, afforded a nearly
structure of the substance in question. colourless viscous oil which crystallized when seeded with
The synthesis of the racemic forms of allococaine racemic alloecgonine methyl ester: 8.9 g. (98%). As the
and allopseudococaine together with the ready ester is always impure a t this stage, the oil was as a rule
accessibility of d- and I-alloecgonine methyl ester taken up at once in acetone (25 ml.) and ether (50 ml.) and
the solution treated with glacial acetic acid (2.55 ml.) in
reduce the preparation of the optically active iso- ether (20 ml.). Left for several hours, the yellowish solution
mers of the all0 and allopseudo series as yet un- deposited white crystalline warts of the hydroacetate which
known to a routine laboratory assignment; to- were collected: 7.5 g. (64%), m.p. 10&110". By concentrat-
gether with confirmatory evidence for the three- ing the mother liquors in vacuo and taking up the residue
dimensional structures tentatively proposed herein, in a little acetone and ether, a second (ca. 1.5 g.) and a small
third crop of the salt were also obtained, the total yield
this preparation will complete a chapter of classical of acetate in several such experiments varying between 77
alkaloid chemistry begun seventy-five years ago. and 83%. Dissolved in water (50 ml.), treated with satu-
rated aqueous potassium carbonate (50 ml.), and processed
as above described for the crude ester, the acetate (9.15 g.)
EXPERIMENTALa7 yielded colourless racemic alloecgonine methyl ester as a
Materials. The racemic and optically active forms of 2- thick oil which crystallized spontaneously on keeping: 6.6
carbomethoxytropinone employed herein were obtained as g. (94%), m.p. 81.5-83.5'.
described e1~ewhere.l~ The platinum oxide came from one Conducted in the foregoing manner, the rate of catalytic
batch and was a product of the American Platinum Works. hydrogenation of racemic 2-carbomethoxytropinone was in
The catalytic hydrogenation of racemic B-carbomethoxy- three identical preparations remarkably constant, even a t
t r o p i m e . I n a typical experiment anhydrous racemic 2- the beginning when it is most rapid. Furthermore, the scale
carbomethoxy tropinone (9.00 g., 0.0456 mole), dissolved of operations scarcely affects either the rate or the yield.
in glacial acetic acid (195 ml.) and water (30 ml.), and plati- Several variations of this method were also tried with the
num oxide (0.75 g.) were shaken with hydrogen a t 1.3 same hydrogenation apparatus and gas pressure. The use of
atmospheres for 48 hr., a t the end of which period the con- glacial acetic acid as solvent did not appreciably alter the
rate. As expected, the rate was approximately proportional
(36) C. Lieberman, Ber., 21, 2342 (1888). to the relative quantity of catalyst employed. I n methanol
(37) The melting points recorded herein are corrected the rate was about half that in aqueous acetic acid and in
and were observed in boro-silicate glass capillaries. Unless benzene it was very low: in these experiments the reduc-
noted otherwise, rotations are for solutions in absolute tion product was not isolated. I n either methanol or glacial
methanol. acetic acid the hydrogenation of 2-carbomethoxytropinone
OCTOBER 1959 RACEMIC ALLOCOCAINE A N D ALLOPSEUDOCOCAINE 1547

TABLE I1
MOLECULAR
ROTATION
DIFFERENCES
OF BASE-SALTPAIRS
AMONG
THE COCAINE
ALKALOIDS
Solvent [a1'," [MI '," A[M]',"
1-Cocaine
Hydrochloride
Methanol
Methanol
-29.g3'
-66. 432
- 9,060
-22,500) +13,400
Ecgonine Water - 45.536 - 8,420 +2,100
Hydrochloride Water ' -47.436 - 10,500)
Pseudoecgonine Water +22.76 + 4,2001
Hydrochloride Water +216 + 4,6001
- 400
Ecgonine methyl ester
Hydrochloride
Methanol
Methanol
-12.36
- 506 274501
- 11,800
- +9,350

+ 4J5401
Pseudoecgonine methyl ester Water +22. 86
-990
Hydrochloride Water $23, 46 + 5,530
Anhydroecgonine Water -84. 633 -14,100
Water - 73.533 -14,900) +800
Hydrochloride
d-( 2-Carbomethoxytropinone) Water +36.714
Ammonium ion Water + 7.8314
+5,680
Pseudoecgoninol Water +58.3*
Hydrochloride Water +46.38 + 9,600 +400
Alloecgoninol Water - 6.3'' +150
Hydrochloride Water - 5.932 - 1,230
+
d-Alloecgonine methyl ester
Hydrochloride
Methanol
Methanol -
0.1P
2.132 +
- 2g1
495
+524

hydrochloride was extremely slow. The optically active Pseudotropine. Tropinone (10.0 g.) was reduced with
forms of the keto ester as the bitartrate were readily reduced sodium and alcohol according to the directions of Will-
in aqueous solution, but the yield of alloecgonine methyl ester stiitter." The reduction mixture was mixed with water (100
was much inferior to that obtained when aqueous acetic ml.), the liberated alcohol removed in vacuo, and the residue
acid was used. Ruthenium on charcoal, which is advertised extracted with ether (6 x 100 ml.). The dried (KzCOa)
as a powerful catalyst for reducing keto groups, had little extracts furnished a brown oil which readily crystallized
or no activity in this reaction either in benzene, in methanol with the evolution of considerable heat. This wa8 freed of
or in aqueous acetic acid. Also, rhodium (5% on alumina) brown gummy by-products by sublimation in vacuo: 9.3 g.
had no hydrogenating activity for solutions of the keto ester (93%). The product was further purified by dissolving the
in methanol. sublimate in benzene (12 ml.) and adding hot ligroin (60-
The catalytic hydrogenation of tropinone in alcohol and in 71') (16 ml.): white prisms, m.p. 106-108'. I n three such
aqueous acetic acid. Tropinone (1.40 g., 0.0100 mole), dis- experiments the yields of pure product were 78,80, and 82%.
solved in 95% alcohol (37 ml.), was mixed with platinum The brown residue from the sublimation which partially
oxide (0.12 g.) and shaken with hydrogen (at ca. 1.3 atm.). crystallizes on keeping gave, upon chromatographing on
After 6 hr. the uptake of hydrogen had become negligible, alumina, as the only crystalline component, a small addi-
the total consumption being then 233 ml. (S.T.P.) (104%). tional quantity of pseudotropine, m.p. 108-109.5'; meth-
The catalyst-free, colourless solution, which had originally iodide, m.p. 323" (dec.).
been brownish, was mixed with picric acid (2.30 g., 0.0100 The catalytic hydrogenation of dihydrocodeimne. ( a ) I n
mole) in hot alcohol (30 ml.). Tropine picrate separated a t hydrochloric acid. Dihydrocodeinone (3.00 g., 0.0100 mole),
once and was filtered from the cold solution: 3.1 g., m.p. m.p. 196-198.5', dissolved in 3.1N hydrochloric acid (13.2
295O.I' An additional quantity (0.6 g., m.p. 293') was ml.) and water (16.8 ml.) was shaken with platinum oxide
isolated from the mother liquors, the total yield of tropine (0.10 9.) and hydrogen ca. a t 1.3 atmospheres for 36 hr.
thus being quantitative. (uptake: ca. 170 ml., S.T.P.). The old catalyst was then
Exact repetition of the foregoing experiment in aqueous replaced with fresh and the shaking continued for a like
acetic acid (32 ml. of glacial acetic acid and 5 ml. of water) interval. The total uptake of hydrogen by the base was
gave similar results. The plot of gas consumption as a func- about 350 ml. (S.T.P.) (78% of the theoretical quantity for
tion of time was colinear with that of the previous prepara- the consumption of 2 equivalents). The catalyst-free solution
tion for the first 150 minutes, and after 5.5 hr. the total contained much free phenolic material as indicated by a
uptake amounted to 105'% of the theoretical. Concentrated strongly positive diazosulfanilic acid reaction. Treated with
in uacuo, the catalyst-free solution yielded a nearly colour- some sodium hydrosulfite and made basic with potassium
less oil which was mixed with water (10 ml.) and saturated bicarbonate, the mixture liberated an oily mixture extract-
aqueous potassium carbonate (20 ml.). Extraction of the able with chloroform. The recovered oil (3.05 g.) did not
liberated bases with ether (5 x 25 ml.) and removal of the crystallize on long keeping; and, when treated with picric
solvent from the dried (NagSOd) extracts furnished a colour- acid, gave a liquid picrate which only partially crystallized
less liquid which could not be induced to crystallize by seed- and could not readily be resolved into its component salts.
ing either with tropine or with pseudotropine. ( b ) I n aqueous acetic acid. Dihydrocodeinone (3.00 g.,
The base was heated briefly a t 100' with potassium 0.0100 mole), m.p. 196-198.5', dissolved in a mixture of
hydroxide (1.0 g.) in water (11 ml.), which caused the forma- glacial acetic acid (32 ml.) and water (7.0 ml.), was shaken
tion of a black product also obtainable by similarly treat- with platinum oxide (0.12 g.) and hydrogen as before. A t
ing tropinone. After mixing with some potassium carbonate the beginning the uptake of gas was quite rapid, and after
sesquihydrate, the base was recovered by ether extraction 3 hr. it had nearly ceased: 232 ml. (S.T.P.) (103'% of the
as before: 1.25 g., which crystallized when seeded with theory for 1 equivalent). The catalyst-free solution gave a
tropine. Recrystallized from ligroin it melted at 54-62'. negative diazosulfanilic acid response, Recovered as noted
The base was then converted entirely to the picrate as de- above for tropine in this solvent except that chloroform
scribed above: 3.0 g. (81%), m.p. 293-295". ( 3 X 25 ml.) was used, the amber gum did not crystallize on
1548 FIR'DLAY VOL. 24

keeping. Dissolved in ethyl acetate containing a little water, opalescence to the hot mixture. Concentration of such a solu-
it gave no crystals on scratching, but largely crystallized tion to incipient turbidity and cooling afforded racemic
when seeded with authentic dihydrocodeine hydrate: 2.3 alloecognine as a finely divided precipitate melting a t 237'
g. (73%). (dec.). After recrystallization from 95% alcohol it melted
Racemic alloecgonine methyl ester hydroacetate. Prepared at 241.5' (dec.); i t separated from alcohol (ea. 90%) as
as described above, this salt was purified from acetone-ether lustrous, thin plates, melting a t 240-241 ' (dec.).
from which it separated as aggregates of stout prisms, m.p. Anal. Calcd. for CgH16NOs.1H20:C, 53.18; H, 8.43.
110-110.5'. It is extremely soluble in most polar solvents, Found: C, 52.90; H, 8.24.
and even small quantities of acetic acid greatly retard both Drying 5 hr. a t 100' in uacuo alto removed about half the
the inception and the rate of its crystallization. water of crystallization, and drying 9.5 hr. a t 117" in
Anal. Calcd. for C12H21NOB: C, 55.58; H, 8.16. Found: vacuo alto removed all of it. Recovered from aqueous solution
C, 55.66; H, 8.20. it did not melt below 125' and thus appeared to be dif-
Racemic alloecgonine methyl ester. Isolated from the hydro- ferent from Willstatter's 'drittes racemisches Ekgonin'
acetate as described above for the hydrogenation mixture, which, as the hydrate, was reported to melt at l l O o , and in
the oily pure ester crystallized spontaneously and slowly the anhydrous condition a t 229' ( c o ~ . ) . ~
to a white cake. It was recrystallized both from acetone and Anal. Calcd. for CgH1QNO3:C, 58.35; H, 7.75. Found:
from ligroin (60-71') and separated from the former as C, 58.17; H, 8.02.
irregular crystals and parallelopipeda and from the latter Racemic Allopseudoecgonine. From the hot absolute alco-
aa striated square tablets, m.p. 81.5-83.5'. It is quite soluble holic leaching of racemic alloecgonine (described above),
even in the cold in most polar and non-polar solvents. When needles melting a t 241' separated as the solution cooled.
evaporatively distilled or sublimed in vacuo a t higher tem- These were recrystallized to constant melting point from
peratures, the ester appeared to be partially epimerized as absolute alcohol for analysis: feathery needles, m.p. 243'
indicated by a decline of the melting point to as low as (dec.).
70-73 '. Anal. Calcd. for C9HlaO3: C, 58.35; H, 7.75. Found:
Anal. Calcd. for CIOHITNOs: C, 60.28; H, 8.60. Found: C, 58.14; H, 7.95.
C, 60.44; H, 8.82. This isomer appeared not to crystallize with water. The
The pure ester (0.060 9.) in methanol was mixed with hydrochloride, obtained by evaporating a solution of the
picric acid (0.040 9.) in methanol and heated with enough amino acid in the equivalent amount of hydrochloric acid,
more solvent to complete solution; shiny yellow flakes, was purified readily from absolute alcohol: m.p. 213'.
m.p. ca. 185', separated. Recrystallized from methanol, the Anal. Calcd. for C9HlsC1NO8:C, 48.76; H, 7.28; C1, 16.00.
picrate melted at 194-197', then crystallized partially as Found: C, 48.64; H, 7.33; C1, 15.86.
the temperature was raised and remelted completely a t Racemic allopseudoecgonine methyl ester. Allopseudo-
203.5'. Recrystallized again, this salt melted a t 195-196' ecgonine hydrochloride (1.45 g.) was dissolved in absolute
(softening at 193'). By lowering the bath temperature to methanol (75 ml.) containing dry hydrogen chloride (6.2
185' the salt solidified completely and then remelted only g.); the mixture, protected from moisture, refluxed 4 hr.
a t 203-203.5". The solvent was removed in vacuo, the residue treated with
Anal. Calcd. for ClaHzoN~Olo: N, 13.08; Found: N, 13.11. saturated aqueous potassium carbonate (20 ml.) and water
The neutral oxalate was prepared in methanol-ether and (10 ml.), and the resulting mixture extracted with ether
purified from methanol. It dissolved rather slowly in (3 X 100 ml.). Recovered from the dried (Na&O,) extracts
methanol from which i t separated readily only after seed- in the same manner as the epimeric ester, the colourless oil
ing: warts of feathery, slender prisms, the melting point obtained crystallized almost a t once: 0.9 g. After one re-
of which appeared to depend somewhat on the rate of heat- crystallization from ligroin ( 60-71') these crystals (0.85 g.)
ing but usually occurred about 200.5'. melted a t 79-80' and after a second a t 80-80.5". This mate-
Anal. Calcd. for C ~ Z H ~ ~ NC,Z 54.08;
O ~ ~ :H, 7.43. Found: rial consisting of small, colourless, short prisms was sub-
C, 53.84, 53.77; H, 7.31, 7.46. limed about 60°/1 mm. for analysis. The melting point of
The hydrochloride was not obtained crystalline either its mixture with the racemic allo isomer was 60-70'.
from methanol or from methanol-ether. Anal. Calcd. for CI0H1,NO3:C, 60.28; H, 8.60; N, 7.03.
Racemic alloecgonine. Hydrolyzed in the same manner as Found: C, 60.03; H, 8.61; N, 7.05.
pseudoecgonine methyl ester: racemic alloecgonine methyl The ester (0.100 g.) and oxalic acid dihydrate (0.0317 g.)
ester afforded a white crystalline residue which was con- were dissolved in methanol ( 2 0 . 2 ml.). No crystals were
taminated with a purplish impurity when the starting obtained either by keeping the solution or after mixing it
material had not been freed of small amounts of unhydro- with an equal additional quantity of oxalic acid.
genated 2-carbomethoxytropinone. Leaching the dried The ester (0.10 g.) and picric acid (0.12 9.) were mixed
hydrolysis product with small amounts of absolute alcohol in acetone. No crystals separated. Removal of the solvent
removed one of its two principal components. The remainder left a gum which gradually solidified. Recrystallized from
(map. ca. 235') was then evaporated at room temperature acetone to constant melting point ( 2 X ), the picrate was
with the stoichiometric amount of hydrochloric acid and obtained as yellow, feathery tufts of minute, slender prisms
the recovered salt recrystallized from absolute alcohol from melting a t 135-136.3'.
which it is readily purified: colourless crystals of alloecgonine Anal. Calcd. for ClaHzoN4010: C, 44.86; H, 4.71; N, 13.08.
hydrochloride, m.p. 231.5-233.5'. The hydrochloride of Will- Found: C, 44.80; H, 4.72; N, 12.57, 12.74.
stiitter's 'drittes racemisches Ekgonin' is reported to melt As a by-product of the action of benzoyl chloride on the
at 231'.* ester in pyridine, allopseudoecgonine methyl ester hydrochloride
Anal. Calcd. for C9HlsClX03: C, 48.76; H, 7.28. Found: was obtained as a salt insoluble in pyridine. This was re-
C, 49.02: H, 7.18. crystallized to constant melting point ( 2 ) by ~ dissolving
The hydrochloride (0.27 g.) in water solution (10 ml.) in methanol and adding ether: irregular, stout prisms m.p.
was shaken with silver carbonate (0.4 g.), the mixture 191.5-192O.
filtered, excess silver removed with hydrogen sulfide, and Anal. Calcd. for CloH18C1N03:C, 50.95; H, 7.70. Found:
the &&-free solution evaporated to dryness in vacuo. The C, 50.60, 50.76; H, 7.57, 7.53.
residue dissolved readily in hot absolute alcohol and the Racemic alloeocaine. T o a stirred mixture of racemic
solution suddenly deposited a fine crystalline precipitate: alloecgonine methyl ester (4.0 g., 0.0200 mole) in dry pyri-
m.p. 239'. This material, presumably the anhydrous modi- dine (10 ml.) a t '
0 was added benzoyl chloride (1.0 ml.).
fication, is much less soluble in hot absolute alcohol, re- The mixture which reddened at once was kept 15 minutes
quires about 300 times its own weight of this solvent for a t 0' and then mixed with more chloride (1.5 ml.). Originally
complete solution and, until dissolved, imparts a noticeable transparent orange red, the benzoylation mixture gradually
OCTOBER 1959 RACEMIC ALLOCOCAINE b N D ALLOPSEUDOCOCAINE 1549
acquired a brownish red opacity. After 12 hr. a t room tem- mixed with aqueous sodium bicarbonate, and the liberated
perature, the mixture contained a few crystals and after two oily bases extracted with ether. From the dried ether ex-
days had become semi-solid. The mixture, which smelled tracts was recovered a small quantity of greenish brown oil
strongly of methyl benzoate, was rubbed with methanol- which was induced to crystallize by scratching a t dry ice
ether (10 ml. of each) and the solid material was collected temperature. By several recrystallizations from ligroin (60-
at the pump and recrystallized from methanol to constant 71") and manual separation of the crystals from gummy
melting point: 1.2 g., m.p. 201.5' of allococaine hydrochloride, impurities, pure racemic allopseudococaine was obtained:
consisting of aggregates of minute prisms. colorless, irregularly shaped crystals melting a t 92.5-98.4'
Anal. Calcd. for Ct7H21N0~HC1:C, 60.08; H, 6.53; C1, and, after grinding, a t 93-95O. Like the all0 isomer, these
10.43. Found: C, 59.88; H, 6.57; C1, 10.37. crystals liquefy slowly on standing and acquire the aroma
The mother liquors from the purification of the hydro- of methyl benzoate.
chloride furnished additional alkaloid through basification, Anal. Calcd. for C17H~1N04:C, 67.31; H, 6.98. Found:
extraction, and conversion to the neutral oxalate described C, 67.45; H, 6.70.
below: 1.1 g. The picrate was prepared in and purified from methanol:
The pyridine mother liquors were concentrated i n vacuo, m.p. 161-162'.
treated with water and potassium carbonate, and extracted Anal. Calcd. for C Z ~ H ~ ~ C, O ~ ~H,: 4.54. Found:
N ~51.88;
with ether which removed both allococaine and unreacted C, 51.61; H, 4.10.
alloecgonine methyl ester. These were separable by taking A comparison of the reactivities of alloecgonine methyl ester
advantage of the much greater solubility of the latter in and allopseudoecgonine methyl ester toward methyl iodide.
weakly alkaline aqueous solutions: ca. 0.6 g. of each ester. Methyl iodide (15 ml.) was diluted to 100 ml. with absolute
The total yield of the cocaine isolated as the hydrochloride methanol and racemic alloecgonine methyl ester (0.20 g.),
and the binoxalate was about 40%. m.p. 81-83", dissolved in some of this solution (0.50 ml.),
The feasibility of benzoylation by the Schotten-Baumann Kept at 0' overnight the mixture deposited a compact mass
method, by Willstatter's method utilizing benzoic anhydride of crystals, and an additional quantity was obtained by
in benzene', by the action of benzoyl chloride in inert adding ether to the mother liquor: m.p. 198'. Recrystallized
neutral solvents, by benzoyl chloride alone, and by trans- from methanol the first crop of alloecgonine methyl ester
esterification with methyl benzoate was examined. None methiodide gave colorless crystals melting at 196-197'.
of these methods appears to give results as satisfactory as ~ : 38.72; H, 5.91; I, 37.2.
Anal. Calcd. for C I I H ~ I N OC,
benzoyl chloride in pyridine. Found: C, 38.88; H, 5.69; I, 37.0.
Recovered from the pure hydrochloride or the pure Treated in exactly the same way allopseudoecgonine
binoxalate by basification and ether extraction in the usual methyl ester, m.p. SO', gave no crystalline precipitate even
manner, racemic allococaine was obtained as an initially after 44 hr. By diluting with ether to near turbidity and
brownish oil which readily crystallized. It could conveniently scratching, crystals were obtained: m.p. 175-177'. After
be recrystallized from ligroin from which it separated as one recrystallization from methanol these melted at 208-
striated, square tablets melting at 82-84'. During several 209'; after two, at 212'. The analysis of this material
weeks this alkaloid gradually melts to a viscous brown oil indicated that it was still impure (Anal. Found: C, 40.36;
with a liberation of methyl benzoate. H, 6.14; I, 41.7).
Anal. Calcd. for C17H~lN04:C, 67.31; H, 6.98. Found: Dissolved in acetone (1.0 ml.) and treated with methyl
C,67.30; H, 6.85. iodide (0.20 ml.) a t 0", alloecgonine methyl ester (0.185 g.),
Prepared in and purified from methanol, the binozalate m.p. 81-83', gave almost immediately a crystalline pre-
was obtained as long, slender prisms melting a t 177.5- cipitate melting a t 198' and at 196-197" after mixing with
178.5'. the preparation obtained from methanol.
Anal. Calcd. for C1~H21NOcH&~04: C, 58.01; H, 5.89. When allopseudoecgonine methyl ester (0.20 g.), m.p
Found: C, 57.88; H, 5.67. 80", dissolved in acetone (1.0 ml.) was mixed at 0' with
Its picrate was prepared in and purified from methanol: methyl iodide (0.20 ml.) and the solution allowed to warm,
stellate clusters of short, slender prisms, m.p. 178.5-180'. crystals soon precipitated. Two different samples from this
Anal. Calcd. for C~H24N*011:C, 51.88; H, 4.54. For preparation melted at 160-165' and 164-167'. After one
CZ,HZ,N~O~I.~/~CH,OH: C, 51.46; H, 4.78. Found: C, 51.42, recrystallization from methanol the product melted at 166-
51.38; H, 4.41, 4.49. 167'; after two, at 164165'. Although apparently inhomo-
Prepared in methanol (0.35 ml.) from the base (0.30 g.) geneous, i t has the composition of allopseudoecgonine methyl
and Gtartaric acid, the lrbitartrate separated as dense ester methwdide.
warts, m.p. 144-148'. Recrystallized from methanol, the Anal. Calcd. for CtiH20INO~:C, 38.72; H, 5.91; I, 37.2
crystals obtained melted at 145.5-148'. The degree of Found: C, 38.33; H, 5.77; I, 37.26.
resolution was not investigated. The reaction of allopseudoecgonine with methyl iodide.
Anal. Calcd. for C21Hz,N010:C, 55.62; H, 6.00. Found: Allopseudoecgonine (0.20 g.), m.p. 243", was dissolved in
C, 55.16, 55.22; H, 6.09, 6.13. hot methanol (10 ml.), the solution cooled somewhat, and
The dibenzoyl-bbitartrate waa obtained by mixing the methyl iodide (Malliinckrodt A.R., 2.5, ml.) added. The
base (0.30 g.) with the acid monohydrate (0.37 9.) in mixture was refluxed for 3 hr. and the volatile solvents
methanol. The salt which precipitated was recrystallized removed i n vacuo. The residual gum gradually crystallized
three times from methanol, the melting point being raised 0.22 g., m.p. 165.5-170'. The crystals, which are extremely
successively: 161' to 166" to 168-168.5". The degree of soluble in methanol, were recrystalliied by dissolving them
resolution thus effected was not determined. in a small amount of this solvent and keeping the solution
Anal. Calcd. for C85Ha6N012: C, 63.53; H, 5.33. Found: a t 0': colourless crystals, m.p. 182.5-185 (Anal. Found
C, 63.68; H, 5.34. C, 40.83; H, 5.83). This may largely be the methyl betaine
Racemic Allopseudocacaine. Racemic allopseudoecgonine hydriodide which like the pseudo isomera seemingly loses
methyl ester (0.50 g.) was treated at 0' with a solution hydrogen iodide readily.
(1.4 ml.) made by diluting benzoyl chloride (2.0 ml.) to The mother liquors from the crystallization were concen-
10.0 ml. with pyridine. This mixture waa stirred at 0" for trated in vacuo to a gum which was taken up in aqueous
30 minutes and kept overnight a t room temperature. Re- potassium carbonate and the resulting solution extracted
moval of the solvent in vacuo left a reddish semi-solid with ether. The oily base (0.030 g.), recovered in the usual
(having an odor reminiscent of methyl benzoate), which manner from the extracts, did not crystallize when seeded
waa mixed with methanol-ether and a crystalline precipitate, with racemic alloecgonine methyl ester (m.p. 81-83'), but
aUopseudoecgonine methyl ester hydrochloride (see above), fil- did so immediately when treated with a trace of racemic
tered off. The filtrate was concentrated i n vacw), the residue allopseudoecgonine methyl ester.
1550 FINDLAY VOL. 24
d-Alloecgonine methyl ester. Hydrogenated in a manner boiling the mixture 1 hr., the excess reducing agent was
essentially the same as for the racemic modification, d-(2- decomposed with methyl acetate. The mixture was suc-
carbomethoxytropinone) yielded the d-antipode of allo- cessively treated with 3N aqueous sulfuric acid, tartaric
ecgonine methyl ester.14 Hydrogenated in aqueous acetic acid acid (2.0 g.), and potassium carbonate, and the whole con-
as the bitartrate sa1t,14 the &keto ester gave a yield (67%) tinuously waa extracted with chloroform for 3 days. Re-
of pure acetate lower than was obtained from the racemic moval of the chloroform gave a light brown oil (0.3 g.)
@-ketoester, and this salt in water alone gave a lower yield which was readily purified from methanol: colourless, tri-
still. The remainder of the hydrogenation product appeared angularly shaped crystals, m.p. 201.5-202.8', a? -6.3'
to be non-crystalline. Recovered from the pure hydroacetate (C, 1, water).
as indicated above for the racemic isomer, the &ester was Anal. Calcd. for CQH17N02:C, 63.12; H, 10.01. Found:
recrystallized from ligroin (60-71'): stout, colourless prisms, C, 63.45; H, 10.09.
m.p 79-82.3', [a]: +0.15' f 0.03 (C, 2). Treated with the stoichiometric amount of methanolic
Prepared from the stoichiometric amount of methanolic hydrogen chloride the base was converted to the hydro-
hydrogen chloride and recrystallized from methanol-ether chloride, which was recrystallized from methanol-ether:
the hydrochloride was isolated as colourless, hygroscopic, minute, stout prisms, m.p. 236.5-238', : a -5.9' (C, 1,
hexagonal, and irregularly formed stout prisms: a? -2.1' water).
IC. 21.38
\ - I -,- Anal. Calcd. for C9Hl&lN02: C, 52.04; H, 8.73. Found:
Anal. Calcd. for CloH&1N03: C, 50.95; H, 7.70. Found: C, 51.78; H, 8.86.
C, 51.30; H, 7.80. The foregoing is presumably what has been called 2a-
'Prepared 'in methanol and purified from methanol-ether hydroxymethyl-3a-tropanol for which the melting point and
and from methanol acetone, the binozalate, as small stout that of ita hydrochloride are reported to be 165-166' and
prisms, melted at 164-165'. It was dried in vacuo a t room 172', re~pectively.3~ An isomeric substance, said to be
temperature for analysis. 2~hydroxymethyl-3a-tropanol(XXXI), has been reported
Anal. Calcd. for ClzH1gN07: C, 49.82; H, 6.62. Found: to melt at 139-140' and its hydrochloride a t 258-260°.a4
C, 49.87; H, 6.63. Infrared measurements. All measurements were made with
The picrate was prepared in and recrystallized from a Perkin-Elmer Model 21 double beam spectrometer with
methanol: m.p. 194-201". sodium chloride optics. In carbon tetrachloride solution
Anal. Calcd. for C16H20N4010: C, 44.86; H, 4.71. Found: the methyl ester of ecgonine, of pseudoecgonine, of racemic
C, 44.81; H, 4.58. alloecgonine, and of racemic allopseudoecgonine each had
1-Alloecgonine methyl ester. Likewise prepared from 1- unmistakable bands in the regions both of hydroxyl group
( 1-carbomethoxytropinone), the bantipode crystallized from absorption and of carbomethoxy group absorption. For
ligroin (60-71') as stout rectangular prisms, m.p. 78-81', the foregoing esters, respectively, these had the following
which in vacuo alto at low temperature gave a sublimate of locations: 2.83,~and 5.79,~,2.79,~
and 5.80p, 2.83,~and 5.81y,
stout prisms melting a t 79.5-81.8', :a -0.11' f 0.03 and 2.76,~and 5.75,~.
(cJ 2). .iViscelhneous rotations. An apparently pure sample of
Anal. Calcd. for C1OH1,NOs: C, 60.28; H, 8.60. Found: cocaine [a? -15.9 (C, 4, chloroform)," [reported -15.9
C, 60.18; H, 8.67. (C, 4, chloroform39] was found to have in methanol :a
Hydrolysis of d-alloecgonine methyl ester. The ester (5.0 g.) -29.9' (C, l ) n and -31.0' (C, 2)" Mallinckrodt cocaine
dissolved in water (15 ml.) was heated a t 70' until the pH hydrochloride in methanol was found to have -66.4' (C,
of the solution had been reduced to 8. Evaporation of the 1)3*and -65.6' (C, Z)."
solution a t low temperature afforded a colourless syrup Ecgonine methyl ester which had originally n'," 1.48866
which when kept in vacuo over potassium hydroxide largely and CY: -12.3' (C, 2)e had, after keeping in the dark in a
crystallized. Taken up in hot absolute alcohol (10 ml.) desiccator over CaCl? in excess of four years, acquired a
the syrup gave colourless stout prisms which separated brownish colouration. This was removed by evaporative
during several hours: m.p. 214-216'. Repeated recrystal- distillation at 95"/1 mm. The distillate appeared to be
lization from absolute alcohol furnished a pure or nearly pure: n: 1.4887, a g -12.75' (C, 5, methanol).
pure substance isomeric with Z-ecgonine: m.p. 221.5-222'
(dec.), -47.4' (C, 2, water)." Acknowledgments. The analytical data recorded
Anal. Calcd. for C Q H I ~ N OC,~ : 58.35; H, 7.75. Found: herein were by the Institute's Microanalytical
C, 58.48; H, 8.13. Laboratory directed by Dr. William C. Alford.
Lithium aluminum hydride reduction of d-alloecgonine
methyl ester. Lithium aluminum hydride (ca. 50 mg. alto- Mr. Harold K. Miller of this Institute determined
gether) was added in small portions to pursed tetrahydro- the infrared absorption spectra.
furan (15 ml.) until effervescence ceased and then a further
quantity (0.60 g.). To the gently refluxing solution of the NATIONALINSTITUTEOF ARTHRITIS
AND METABOLICDISEASES
reducing agent another of d-alloecgonine methyl ester ( 1.0
g.) in tetrahydrofuran (10 ml.) was admitted dropwise. After BETHESDA14, MD.

(38) Rotation determined by Mrs. Evelyn G. Peake of (39) I. Heilbron, Dictionary of Organic Compounds, Rev.
this Inetitut,e. Ed., Vol. I, Eyre and Spottiswood, London, 1953, p. 59.1.