3206
COMMUNICATIONS TO THE EDITOR
n.m.r. at 4.57 broad (>C=C(
H); 7.73 (OH), isomer, I 1
1
8.35 (CHs-C=C-),
I
1
( J 6 cs., CH3-CH),
doublet centered a t 9.0 9.05, 9.147 (CH3-c!!-CHa).
Hydroboration-boxidation led to the diol (14) (45%), m.p. 112-113', [a]D -2' subsequently transformed to the liquid acetate (15) (87%), [ a ]$39' ~ with acetic anhydride in hot pyridine. The position of the double bond in 15 was revealed by its n.m.r. spectrum: doublet a t 5.4 (lH, J 6 cs.), 8.15 (3H), 8.55 (3H), doublet at 8.85 (3H, J 6 cs.), 8.97, 9.20 7 (6H) with no olefinic protons. Catalytic reduction of 15 afforded the liquid ace~ which with lithium alutate (16) (98Oj0), [ a ]-3' minum hydride gave the alcohol (17) (87%) m,.p. 97-97.5', [a]D-7' identical in every respect with an alcohol obtained from a-patchoulene (20) by hydroboration-oxidation. Catalytic reduction of the diol (14) over platinum in acetic acid containing some perchloric acid yielded the alcohol (17)in one operation.
-C=C-H)
[ a ] +51' ~
Vol. 84 had n.m.r. a t 4.95 (broad,
1
; 8.37 (CHI-C=C-)
tered a t 9.10 ( J 6 cs.) 9.05;
1
; doublet cen-
9.11 r (CH-d-.
I
CH,) and was spectroscopically (infrared and n.m.r.), vapor chromatographically and polarimetrically identical with authentic a-patchoulene (20). Preferential formation of 20 agrees with the postulate of double bond character in the transition state of the pyrolytic elimination8because the other isomer is destabilized by 1,3-methyl interaction. We have previously described' a conversion of a-patchoulene (20) to patchouli alcohol (1) and since total syntheses of (+)-camphor have been accomplished the transformationsg described constitute a total synthesis. Financial support by the National Institutes of Health (RG9186) and by Firmenich and Cie, Geneva, is gratefully acknowledged. (8) C. H. DePuy and R. W. King, Chem. Reus., 60, 431 (1960). (9) G.Komppa, Bey., 36, 4332 (1903); A n n . , 370, 209 (19091. (10) National Institutes of Health Predoctoral Fellow 1960-1962.
DEPARTMENT OF CHEMISTRY G. BUCHI MASSACHUSETTS INSTITUTE OF TECHNOLOGY CAMBRIDGE, MASSACUSETTSWILLIAMD. MACLEOD, JR.%O RECEIVED JULY 12, 1962 STEROIDS. CC.1 SPECTRA AND STEREOCHEMISTRY, PART 111.' STEROIDAL 5.6-EPOXIDES
13
J. @OH
-
@OAC
14
\
J
Sir : Epoxidation of As-steroids frequently produces both the 5a,6a- and 5PI6fi-epoxides, the former stereochemistry normally being assigned to that isomer which preponderates. Proof of structure has hitherto rested on further chemical reactions and molecular rotation differences.a An examination of Dreiding models of the stereoisomeric 5,6-epoxides suggested that the angles (+) subtended by the epoxidic C-H bonds with the C-H bonds a t C, (Table I) are sufficiently different to permit a differentiation between the a- and 8-epoxides to be made through a study of JHH values, providing that the Karplus correlato 4 is valid for these cases.6 tion,'l relating JHH (1) Steroids. CXCIX, L. H. Knox, E. Velarde, S. Berger and D. Cuadriello, Chcm. and I n d . , in press, 1962. (2) Part 11, A. D. Cross and P. W. Landis, J. A m . Ckcm. SOC.,84,
16, R = A c 17, R = H
t 18, RI=CHI; Rz=H 19, R I = H ; Rz=CHs
The synthesis of patchoulione (19) was completed by chromic acid oxidation of 17 to the ketone (18) (97%), m.p. 25-26', [ a ] +13' ~ followed by epimerization with alkali to 19 (60%), m.p. 48-50° identical (infrared, mixture m.p. and rotation) with an authentic sample of patchoulione (19). Pyrolysis of 16 a t 350' afforded a mixture of two liquid olefins in a ratio of 4 ; 1. The major
1736 (1962). (3) A survey of the literature and collected data4 reveals that the &epoxide is more dextrorotatory than the Q. However, this requires both compounds for the comparison. An apparent exception concerns acetate methyl the epoxides of l7,9-carboxy-A~~~bandrostadiene-3@-ol ester, where the correct specific rotation of the @-epoxide,-7' (ref. 4c, p. 1167),is erroneously given as -70° (ref. 4c, p. 1171) and repeated elsewhere as the incorrect value (ref. 4a, p. 78). (4) (a) A. Petit and J. P. Mathieu, "Tables de Constantes et D o o d e s Numkiques. 6. Constantes S6lectionn6es. Pouvoir Rotatoire Naturel. I. St&oldes," Masson et Cie., Paris, 1956; (h) A. Bowers, L. C. IbBEez and H. J. Ringold, Tcfrahcdron, 7 , 138 (1959); (c) L. Ruzicka, E. Hardegger and C. Kauter, Hclu. Chim. Acta, 27,1164
(1944). ( 5 ) M. Karplus, J. Chcm. Pkys., 80, 11 (1959). (6) It has been shown recently' that in the rigid steroid molecule, the constants of the Karplus equations need to be much larger, at least for the steroidal 2- and 4-acetoxy-3-ketones examined. Table I also includes therefore calculated J values according t o these modified equations.' (7) K. L. Williamson and W. S. Johnson, J. A m . Chem. Soc., 83,4623
(1961).
Aug. 20, 1962
COMMUNICATIONS TO THE
EDITOR
3207
TABLE I. Noticeable differences in the chemical shifts of the 19-methyl protons for a- and P-epoxides were CALCULATED AND OBSERVED COUPLING CONSTANTS FOR 6 also anticipated? PROTONS OF STEROID ~,~-EPoxIDES~~~~ Eight 5a,6a-epoxides and seven 5P,BP-epoxides, 5a,6aObserved all having either an ethylene ketal or a 30-acetoxy 60 JK, c i s P,CIS J , CIS epoxide substituent a t C-3, were e ~ a r n i n e d . ' ~ , 'In ~ no 68H-7aH 94 f 4' 0.28- 0.1 0.0-0.27 . .... compound were there C=C or C = O functions 6pH-7BH 2 8 f 4' 5.8 6.8 7.2-8.3 3.3-4.1 which could give rise to long-range shielding of 5888either the 19-protons or the epoxidic proton. epoxide For the a-epoxide-3-ketals the 6P-proton resonance 6cuH-7cuH 75 f 4' 0.03-0.62 0.36-1.06 . . ... always appeared within the range 169.5-171.5 6aH-7pH 49 f 4' 2.8 -4.0 3.6 -5.0 2.1-2.7 c./s. from the TMS reference as a doublet, J , a These values of 4 are measured from models. Twelve 3.3-4.1 c./s., while the 6a-proton resonance of the measurements were made in each case and the majority fell 8-epoxide-3-ketals appeared a t 183.0-185.4 c./s. well within the limiting range of f 4 ' . from TMS,15 J , 2.1-2.7 c./s. Replacement of the 2.5 c./s. for 3-ketal by 38-OAc raises the epoxidic proton ab- of 4- 15.0 c./s. for the a-epoxide, 2 c./s. for the 3-ketal of the sorption range by ca. 5 c./s. for the a-epoxides and the @-epoxide,and by ca. 2 c./s. for the &epoxides. A comparison P-epoxide.18 Thus, after epoxidation of a series of of observed and calculated J values is given in A5-steroid 3-ethylene ketals the 19-protons resoTable I. It is of note that these epoxides constitute nated a t 64.4-65.6 c./s. for five a-epoxides and a t very rigid systems, yet the observed J values for 60.2-60.7 c./s. for four @-epoxides. Agreement of the epoxidic protons are lower than either set of calculated and observed 19-proton frequencies is calculated values. The agreement of the ob- a further criterion for the epoxide stereochemistry. served values and those calculated from the Moreover, if the epoxide bears a 38-hydroxy or Karplus equations5 is however acceptable. The 3P-acyloxy function, the 3a-proton is shifted 25range of J values is conveniently narrow in each 30 c./s. away from the TMS frequency in the a-epoxcase and coupled with the value of the chemical ides 0nly.'9 We consider i t important to note that additivity shift completely defines the epoxide stereochemistry. That each proton absorption appears as a doublet values for steroid methyl proton frequency shifts agrees well with the implications5J that for 4 = are applicable only when the introduction of further 70-llOo, J is quite small. In both epoxides only substituents into the steroid nucleus does not cause serious alterations of the relative positions and one 6,7-proton coupling is observed. From the positions of the 19-proton resonances orientations of the methyls and substituents capable of long-range shielding. This is particularly imit was possible to calculate additivity valuesg~10~17 portant for unsaturated and keto steroids. Thus, (8) The chemical shift of the methyl protons a t C1o or CISin steroids additivity generally holds for steroids containing a is determined by the over.111 shielding experienced, which is a net effect A9(11) double bond or a 58,68-epoxide, but when of t h e position and orientation of all neighboring single and double both groups are present this is no longer true since bonds, plus long-range shielding effects due to magnetic anisotropy uf more distant groups of electrons. Changes i n srcreochcmistry can lead now ring B is forced into a half-boat form and the l o considerable changes i n long-range shielding effects.'-*-lo Failure t o position of the 19-methyl relative to both groups consider fully these effects has led to errors in interpretation in reis changed. For example, 5P,GP-epo~y-A~(~~)cently published work.lo*ll A more detailed discussion of some new pregnene-3,20-dione-l7a,21-diol diacetate 3aspects will appear shortly.', (9) A. D. Cross and P. W.Landis, unpublished results. ethylene ketal has calculated and observed 19(10) R. F. Zurcher, Hclo. C h i n . Acta, 44, 1380 (1961). proton resonances of 70.0 and 73.7 c./s. respectively, (11) J. Jacquesy, J. Lehn and J. Levisalles, Bull. SOC.chim. France, a disparity considerably larger than the agreement 2444 (1961). normally observable. g,lO,ll (12) A. D. Cross, forthcoming publication. (13) N.m.r. spectra were obtained for purified chloroform solutions We thank the Universidad Nacional Aut6noma a t 60 Mc. using tetramethylsilane (TMS) as an internal reference. A de Mexico and Prof. A. Sandoval for time on the Varian A-60 spectrometer was employed, but final calibration is against A-60 spectrometer. spectra run on a Varian HR-60 instrument, suitably equipped for cali-
-
+
+
bration by the standard side-band technique. Accuracy limits are of (18) Positive shifts are away from TMS, for the 19-proton resonance the order of 1 1 c./s. for chemical shifts and 1 0 . 3 c./s. for J values. frequency, due t o the extra deshielding induced by these functional (14) The a-epoxides were: 5a,6a-epoxy-androstan-3B,l78-diol group*, relative t o Sa- or 5p-androstane, according to stereochemistry diacetate, 5a,6a-e~xy-androstan-l7-one-3R-olacetate, Ba,Ba-epoxyat CI. androstan-3-0ne-17@-01-17-acetate 3-ethylene ketal, 5a,6a-epoxy(19) Shifts of the 3 a - H resonance have been noted for the other preman-320-dione 320-diethylene ketal, 5a,6a-epoxy-pregnan3,20- steroids with a highly polar 5a bond., dime-21-01 320-diethylene ketal, 5lr.6a-epoxy-pregnan-3,20-dione3RESEARCH LABORATORIES ethylene ketal, 5a,6a-epoxy-pregnan-3,2O-dione-2l-ol3-ethylene S.A. ALEXANDER D. CROSS ketal 21-acetate, and 5a.6a-epoxy-pregnan-3,20-dione-17a,2l-diol SYNTEX, POSTAL2679 APARTADO 3-ethylene ketal Iln,Zl-diacetate The @-epoxides were: 50,68epoxy-pre~nan-3,20-dione-21-ol3-ethylene ketal 21-acetate, 58,6@- MEXICO,D. F. epoxy-pregnan-3,20-dione 3,20-diethylene ketal, 56,6B-epoxy-pregnanRECEIVED MARCH19, 1962 3,20-dione-17a.21-diol 3.20-diethylene ketal 17a,Zl-diacetate, 58,6& epoxy-androstan-17-one-a5-01 acetate, 5,8,6&epoxy-17a-ethinylandros. tan-3B,17B-diol 38-acetate. 5B,68-epoxy-androstan-3~,17~-diol diTHE ANODIC OXIDATION OF TRIPHENYLMETHANE acetate, and 5~,6~-epoxy-androstan-3-one-l7B-ol 17-acetate 3-ethylene DYES ketal. Sir: (15) On t h e r scalen the a-epoxide 6B-proton is ca. 7.2 and the 8epoxide 6a-proton is ca. 6.9. We wish to report an unusual electrochemical (16) G . V. D. Tiers, J . P h y s . Chem., 6 8 , 1151 (1958). (17) J. N. bhoolery and M. T. Rogers, J. Am. Chem. Soc., 80, reaction, which is exhibited in the anodic oxidation 5121 (1958) of crystal violet and related triphenylmethane dyes.
