46%
E. E. VAN TAMELEN, PI. SHAMMAAND P. XLDRICH
nnol containing approximately 2 mg. of sodium methoxide. After refluxing 12 hr., the solution was neutralized and concentrated t o a small volume. A solid product remained, which, after being recrystallized from ethanol, melted a t 139.5-140.5' (34% yie!d.). The analysis and infrared absorption spectrum (pyridinium ringI4 and nitrile absorption) indicated the substance t o be the 1:2 (picoline methiodide: acrylonitrile) condensation product. Anal. Calcd. for C13H16N31:C, 45.73; H, 4.73. Found: C, 46.15; H, 4.78. Addition of a-Picolyl Sodium to 1-Cyanocyclohexene .To an ether suspension of sodium amide freshly prepared from 2 . 3 g. of sodium and liquid ammonia there was added (11) l f r . John Baran of this Laboratory has shown t h a t t h e Nalkyl pyridinium ring exhibits bands a t 6.18, (3.36 and 6 61 p ir. t h e infrared, whereas the pyridine system absorbs a t 6.29, 6.38 and 6.78 p.
Vol. 78
dropwise 10.2 g. of 1-cyanocyclohexene dissolved in 20 ml. of dry ether. The addition was carried out over a period of 1 hr. in a system protected from atmospheric moisture. After being refluxed for 18 hr., the mixture was cooled t o room temperature, poured into an ice-water mixture and neutralized with dilute hydrochloric acid. Ether extraction, followed by a sodium bicarbonate wash, drying and removal of ether yielded a residue, the fractionation of which gave a forerun of recovered a-picoline and 1.0 g. of product, b.p. 150-162" (0.1 mm.), n% 1.5100. The boiling point range and analysis indicated a mixture of isomers, which was not separated but which appeared to consist of 1:2 (picoline: 1cyanocyclohexene) condensation products. Anal. Calcd. for C20H2jN3:C, 78.13; H , 8.20. Found: C, 78.26; H,8.20. &$ADISON, Li'ISCONSIN
[ COSTRIBUTIOS FROM THE DEPARTMENT OF CHEMISTRY, USIVERSITY O F WISCONSIS]
The Stereospecific Synthesis of &Yohimbane. BY EUGEKE E. V A N TAMELEN, MAURICE
Stereochemistry of Yohimbine
SHAMMA A N D P A U L
ALDRICH
RECEIVED APRIL 26, 1956 Proceeding from dl-trans-2-hydrindanone, the stereospecific synthesis of dl-yohimbane establishes thc D / E ring juncture of the parent natural product yohimbine (I),2as trans.
The stereochemical relationships expressed in the formula I for yohimbine are based3 upon (1) certain eliminations and epimerizations bearing configurational significance which the alkaloid or closely related substances ~nderg0,~a-C and (2) the Hofmann degradation of dihydrochanodesoxyyohimbo1 (11) to d-N-methyl-trans-decahydroisoquino(1) First reported in a Communication t o the Editor, THISJOURNAL,
76, 950 (1954).
line, a result which was takensd to indicate a trans
( 2 ) I n t h e interests of rational nomenclature we wish t o propose t h e generic term yohimbane for t h e ring system (i) and t h a t substances
11 1
possessing, this 3,15,20-cis,trans-pentacyclicsystem be named a s derivatives of yohimbane. W e suggest further t h a t t h e system featuring t h e 3,15,20-all-cis arrangement (ii) be designated as alhyohimbane and t h a t t h e epimers a t Cs of (i) and (ii) be termed, respectively, cpiyohimbane (iii) and cpialloyohimbane (iv). Finally, i t seems worthwhile t o continue application (6. C. F. Huebner, H. B. MacPhil-
lamy, E. Srhlittler and A. F.St. Andre, Ezperientia. 11, 303 (1955)) of t h e configuratiunal conventions used commonly in t h e steroids series, v i e . , a-,signifying a substituent placed below t h e plane of the fused system (dotted line); a n d D-,denoting a substituent attached above t h e plane (solid line). Formula (i) features a n ar-oriented Cihydtogen, a choice in keeping v i t h t h e absolute stereochemical assignment made b y Klyne, Chem. and I d . ,14, 1032 (1954). T h u s yohimbine becomes, systematically, 16m-carhomethoxy-17arhydroxyyohimbane; and reserpine, ll,lia-dimethoxy-16P-carbomethoxy- 188-(3',4',5'-trimethoxybenzoyloxy) -cpialloyohimbane.
v
D/E ring juncture for yohimbine itself. The base I1 results from the catalytic hydrogenation of chanodesoxyyohimbol, an unusual product obtained by heating yohimbic acid a t 280" in vacuo with either thallous oxide or carbonate.3d NOW the position of the double bond in chanodesoxyyohimbol remains uncertain; ar,d i t is clear that its termination a t C15 (or Czo)-a distinct possibility in view of the drastic treatment required for its production-invalidates any configurational deductioris which can be made regarding the chano base and therefore the D/E juncture of yohimbine, since the stereochemical outcome of the reduction to 11 would not be clear, Because of the ambip i t i e s 4 involved in assigning, on the basis of rela(3) (a) G. Stork, a s quoted b y B. Witkop a n d S. Goodwin, THIS 75, 3371 (1953); (b) M..h'l. Janol, R. Goutarel, A. LeHir, M. Armin and V . Prelog, B d 1 . S O L . chim., 1085 (1952); (c) R. Cookson, Chem. and I n d . , 16, 337 (1953); (d) B. Witkop, THISJ O U R N A L 71, 2559 (1849). (4) See, for example, W. G . Dauben, R. C. Tweit and C . hfannerskantz, ibid., 76, 4420 (1954).
JOURXAL,
Sept. 20, 1956
STEREOSPECIFIC SYNTHESIS O F
dl-YOHIMBANE
4629
tive stabilities, configurations to substituents at- solves itself into the search for a method of inserttached to a cis-bicyclo [4,4,0] system, the assigneda ing the sidechain nitrogen of VI into the ketonic stereochemical relationships of Ca and C16 to CIS ring of IV, followed by cyclization-reduction to or C20in yohimbine are uncertain to a degree which form ring C. trans-2-Hydrindanone was conveniently prepared parallels the probability that the D/E ring juncture is cis. Thus to the extent that the objections in quantity, as previously described,' by pyrolysis above are justified, the formulation I is in jeopardy of the diacid V, the nitric acid oxidation product of in all points, with the exception of the Cl6-C17 trans-2-decalone, The latter ketone, in contrast to the cis isomer, was not readily accessible when assignment. Because of these considerations, we felt com- this research was initiated, and of the possible synpelled to secure incontrovertible evidence regard- thetic routes which were investigated, that ifivolving the terminus a quo in the stereochemical argu- ing the lithium-liquid ammonia reduction* of ments: the nature of the D/E ring juncture in A1v9-2-octalone was followed. Perbenzoic acid oxidation of the bicyclic ketone yohimbine. The degradation product yohimbane (111) appeared ideally suited for stereochemical proceeded smoothly a t room temperature and afforded in 76% yield, the lactone (VII) of trans-2hydroxymethylhexahydrophenylacetic acid, which was secured, after distillation, in a crystalline state (m.p. 38.5'). In order to attain a system which v'; would allow attachment of the tryptamine moiety, the lactone was subjected to the action of anhydrous ethanolic hydrogen bromide and thereby was IE I converted to the ethyl ester IX of trans-2-bromo-
fiqQ
yJ
I11
study in that i t is derived6from yohimbine through operations (Oppenauer oxidation-accompanied by loss of the carbomethoxyl function-to yield I +
.n.ri7 E
VI1
--f
0
111
yohimbone, followed by Wolff-Kishner reduction) which can hardly affect asymmetric centers 15 and 20. A synthesis of 111, for example, which, by its very nature, defines the stereochemical relationship of these two centers would also, therefore, establish the relationship of these same centers in the parent alkaloid. The "stereorational" synthesis which was developed (1) utilizes a starting material, dl-trans-2-hydrindanone (IV), the stereochemistry of which had been proved unequivocally by resolution of the diacid V from which it is preCHzCOOH HOOCCH~,J
!I v
IX
methylhexahydrophenylacetic acid (VIII). The desired ring-opening of VI1 is not readily accomplished; of the various conditions tried, only several days mechanical shaking a t room temperature of the lactone and ethanol, which had been saturated with gaseous hydrogen bromide a t - 5 , gave material with an acceptable halogen analysis. The bromoester, a liquid boiling a t 129" (14 mm.), appears to be stable on standing for some months at room temperature. On the basis of the detailed course of peracid oxidation of ketonesg opportunity for epimerization a t the bridge carbon of I11 could not arise, and the intermediate VI1 can be safely regarded as the trans isomer. I t is possible to visualize a mechanism for the lactone ring opening which would H O
0
H
0
v
pared6 and (2) involves transformations which do not affect either of the asymmetric carbon atoms present in the starting ketone. I t is obvious that between tryptamine (VI) and the hydrindanone, H 0
0
\/ !
there are just the number of carbon and nitrogen atoms required for yohimbane and the problem re( 5 ) J. Jost, Hrlu. Chim. Acta, 34, 1301 (1949). (6) W. HLickel, M . Sarhs, J . Yantschirelewitsch and F.Nerdel, 618, 155 (1935).
Airit.
CH,
OC:H,
0
\/ I
CHI
(7) R . S. Thakur, J . Chem. Soc., 2147 (1932). (8) E. E. van Tamelen and W. C. Proost, Jr., THISJOURNAL, 1 6 , 3632 (1954). (9) (a) W. von E. ]>orring and E. Dc,rfrn.in, {bitl,, 75. : i s 3 i l D 5 X ) ; (b) S I.. l;riess, ibid., 71, 2871 ( l Y 4 Y ) .
iiivolve ;I disturbance a t the y-position. However, such a circumstance was not realized since, contrary to fact, a yohimbane skeleton could not then result from the operations described below. Introduction of the indolic system was accomplished through the monoalkylation of tryptamine by the bromoester, a reaction which was carried out by refluxing with potassium iodide in ethanolic solution in the presence of potassium carbonate. Either attack by the primary amino group at the ester function, yielding a bromoamide X, or attack .it the bromomethyl group, yielding an aminoester
comparison with the base obtained by the treatment of dl-trans-decahydroisoquinoline with @-3'indolylethyl bromide. The immediate product of the alkylation was a crystalline hydrobromitle, m.p. 240-242", the base derived from which did not depress the melting point of the reduction product secured from the lactam. The racemate IIa corresponds in structure to that derived3d for dihydrochanodesoxyyohimbol (II),although a comparison between the two substances was not made. With the structure and stereochemistry of thc intermediate lactam secure, attempts to effect ring closure and reduction to dl-yohimbane were made. After several abortive trials a t p a r f i d lithium aluminum hydride reduction, which was expccted"
XIIi I
v
XI, iiiight appear reasonable as a11 initial operation. On the basis of comparison with similar cases, l o however, the bromoamide might be anticipated to cyclize by displacement of halogen by oxygen,
affortliiig ai1 iminolactone X I I . Since the lactam X I I I is the product which results from the alkyla-
d
xr11
pr ,'
H
\y/ ir
--+ Ila I€Ur
-
1
to lead to the alkanolamine XIV and then to dl-yohimbane (or the C3-epimer) via the imine cation XIr, we turned to the Bischler-Napieralsky type ring closure. On occasion, refluxing phosphorus oxychloride-benzene served to convert X I I I , after careful crystallization. to an unstable yellow substance, m.p. 196-198", the analysis of which suggests its formulation as the dichlorophosphate of dl-A3(4)-dehydroyohi~nbane (XVI).l 2
'A
@;$, I{
1 /'
I
+ \I1
'0
XVI
tion, we believe that the process must follow largely the pattern I X + X I ; a t the same time, the presence of the iminolactone in the reaction product has not been excluded. The lactam, a high-melting (244-245") solid which crystallized nicely from hot ethanol, was non-basic, stable to aqueous mineral acid and exhibited a strong absorption band a t 6.20 p, which corresponded well with the carbonyl band displayed by N-methyl-2-piperidone. Reduction of the lactam XI11 with lithium a h minum hydride afforded dl-N-@-3'-indolyIethyl)trans-decahydroisoquinoline (IIa), m.p. 150-151°, the structure of which was confirmed readily by
v
xv
XI v
x
1 XLlL
Ordinarily, however, work-up of the cyclizatioii mixture led to the hydrochloride, m.p 300" dec., of, presumably, the same dehydro base. We have gathered no direct evidence excluding the enamine structure XVII for the salts, although the yellow of the hydrochloride and the observation by Leonard and GashI4 that enamines nor.. H + + >C=C--N< I
I
--+
>C-C=N< I I
H
!
1
(11) Cf.F. Galinovsky and 12. Weiser, Experienlia, 6, 377 (1950). (12) T h e dichlorophosphate anion has been previously observed t o result from dehydration-cyclization reactions brought about by phosphorus oxychloride, K. Gleu, S. Nietzche and A. Schubert, B e y . , 73, 1093 (1939). (13) Diacetylallocinchonamine ( v ) is reported t o be colorless ( K .
Sept. 30, 1956
STEREOSPECIFIC Syivrmsrs OF dl-Yoilmruaiw
4031
layers had separated, shaking was commenced and continued for 3 days. The tube then was opened, and the volatile components were removed a t the water pump. The strongly acidic residue was neutralized with sodium bicarbonate; the resulting solution was extracted with ether; the ether layer was washed with saturated salt solution, dried over magnesium sulfate and finally evaporated to an oil. By distilling the products and recycling several times the fractions analytically low in bromine content, there was obtained 10.6 g. bromoester (45% yield), analyzing for 29.1% bromine. Distillation of this material gave an analytically pure fraction, b.p. 9G96" (0.1 mm.), n% 1.4882. Anal. Calcd. for CllHIQBrOz: Br, 30.36. Found: Br, 30.2. Lactam of dl-trans-N-(~-3'-Indolylethyl)-2-aminomethylhexahydrophenylacetic Acid (XIII).-Tryptamine (248 mg.), bromoester IX (284 mg.), anhydrous potassium carbonate (75 mg.) and a crystal of potassium iodide were added t o 15 ml. of anhydrous ethanol, and the solution was refluxed 3 days under nitrogen. The ethanol was evaporated off a t reduced pressure, water was added and the mixture was extracted several times with chloroform. The combined chloroform extracts were washed several times with 1 N hydrochloric acid and finally with saturated salt solution. The chloroform layer was dried over magnesium sulfate and evaporated. Crystallization of the residue from et$anol A gave 217 mg. of lactam (60% yield), m.p. 239-240 sample was recrystallized to the constant melting point 242-243"; under the microscope the substance was observed t o undergo a crystalline change a t about 210-220". Anal. Calcd. for C19H21N20: C, 77.00; H, 8.16. Found: C, 77.19; H, 7.94. Salts of dl-A3(4)-Dehydroyohimbane(XVI). Hydrochloride.-To 20 ml. of dry thiophene-free benzene was added Lactone of dl-trans-2-Hydroxymethylhexahydrophenylace- 214 mg. of lactam XI11 and 0.5 ml. of freshly distilled phostic Acid (VII).-To a solution of 17.2 g. of perbenzoic acid phorus oxychloride. The mixture was refluxed gently 3 to 4 hr. and allowed to stand overnight a t room temperature; in 245 ml. of chloroform was added 14.1 g. of dl-trans-hydrindan-2-one,T b.p. 88-89' (9-10 mm.), %*OD 1.4756. The the mixture became yellow-orange in the early stages of resolution was allowed to stand in the dark a t room tempera- action. The precipitate was removed by filtration, washed ture for 40 hr. At the end2f this time, the chloroform was several times with benzene and dried. Several ml. of purievaporated in vucuu a t 100 , and the lactone was then dis- fied dioxane were added to the product, followed by suffitilled a t 102-104O (0.1 mm.) a s a colorless liquid (nP5~ cient absolute ethanol to effect solution. Then the excess 1.4905), which solidified on short standing to give crystals, ethanol was gently boiled off on the steam-bath until the first crystals just appeared; the solution was allowed t o cool, m.p. 38.5'; yield 12 g. (76%). and the crystals were filtered off, washed with benzene and Anal. Calcd. for CQHlaOn: c , 70.10; H , 9.15. Found: dried; yield 178 mg. of bright yellow crystals, m.p. 300" C, 70.35; H, 9.11. dec. The product gave a rapid silver nitrate test in dilute Ethyl dl-trans-2-Bromomethylhexahydrophenylacetate nitric acid and a negative phosphate testlQ; it appeared to (IX).-To 12.5 g. of hydrindanone was added 15.8 g. of be a solvate of indeterminate constitution. A sample dried perbenzoic acid in 120 ml. of chloroform, and the solution to constant weight a t 173" weighed 160 mg. (70% yield). was allowed to stand a t room temperature in the dark for 3 Anal. Calcd. for ClQHZ4C1&: C, 72.21; H , 7.11. days. After evaporation of the solvent, the residue was Found: C, 72.71; H , 7.11. taken up in 40 ml. of absolute ethanol, and the solution was Dichloroph0sphate.-At the end of the cyclization reacpoured into a glass tube and saturated with dry hydrogen The tube was sealed and al- tion the benzene was distilled off under reduced pressure, bromide (ca. 70 g.) a t -5'. and the crystalline reddish-orange precipitate left was dislowed t o stand overnight a t room temperature. When two solved in absolute ethanol and then fractionally recrystallized from the same solvent. This procedure finally yielded, (16) Viewing the hydrogenation as proceeding by way of simultaneon occasion, a crystalline yellow solid, m.p. 196-196", which ous, cis addition of hydrogen t o t h e carbon-nitrogen double bond does gave a positive phosphate test with ammonium molybdate; not appear t o account for the stereospecificity observed in the reducyield 21%. These yellow crystals slowly assumed a green tion, since (assuming the essential equivalence of > C = C C and color on standing. and H H ) the approaches > C = N + < as well as t h a t of H H I / I I Anal. Calcd. for C10H23C15K;202P: C , 55.23; H , 5.61. >c-c< >C--N< Found: C, 55.16; H, 5.46. from the two sides of the planar imine salt XVI are sterically equivalent dl-Yohimbane (IIIa).-A3(4'-Dehydroyohimbane hydroa n d since the immediate products of hydrogenation are, as judged by W. S. Johnson, Expevientio, 7 , 316 (1951)), chloride was hydrogenated over platinum (59 mg. platinum conformational analysis (6. oxide) in ethanol. Uptake of hydrogen was loa% of theory. energetically equivalent. On the other hand, the concept of an indeThe hydrogenation mixture was filtered, 150 mg. of powpendent (of h-r) attachment of hydrogen a t CS accommodates the dered potassium hydroxide was added and the mixture preference for the C r C l a cis product, since-assuming t h a t a t this shaken for ten minutes. The solvent was evaporated off, point in the reaction Nd is either essentially trigonal or tetrahedral the residue taken up in chloroform; the solution was with a n axial hydrogen attached-the stability of such a product is washed free of alkali, dried and evaporated down. Cryshigher t h a n t h e C r C l a trans type, and consequently the transition tallization of the residue from ethanol gave 367 mg. of dls t a t e energy leading t o I I I a would probably be sensibly lower t h a n yohimbane, m.p. 181.5-183.0". Two more recrystallizat h a t involved in going t o dl-epiyohimbane. tions raised the melting point to 182-183'. The mother (18) Spectrum supplied b y Dr. R. K. Hill. liquors were chromatographed on alumina to give an addi(17) T h e synthesis of dl-alloyohimbane, reported by G. Stork a n d tional 29 mg. of material, which gave no mixed melting R. K. Hill, THISJ O U R N A L , 7 6 , 949 (1954), and proceeding along the point depressiori with yohimbane as obtained above; same liues as outlined above, serves t o independently confirm the total yield of product 396 mg. (91% of theory). There was stereochemistly
