Aug. 5, 1956
PREPARATION OF 16-OXYGENATED
T h e residue was dissolved in 50 ml. of hexane and chromatographed on 20 g. of silica gel. Elution with 5% ether9570 hexane yielded 226 mg. of 8-01 (m.p. 153-155"), 10% ether-90yo hexane yielded 140 mg. of a mixture of epimers and 20% ether-80% hexane yielded 120 mg. of a-01 (m.p. 177-179'). Rechromatography of the center fraction yielded 33 mg. of 8-01 (m.p. 153-155'), 39 mg. of a p-01s and 58 mg. of a-01 (m.p. 177-179'). T h e total recovery was 259 mg. (52%) of 8-01 and 178 mg. (377,) of 0-01. Recrystallization of the p-01 from methanol gave material melting 153.3-155.1' (lit.'3 154-155') and recrystallization of the a-01 from ether gave material melting 177.1-178.9" (lit.13 178-180 ') . ( b ) NaBH4: A solution of 0.2 g. (5.3 mmoles) of NaBH,, prepared by first dissolving the hydride in 5 ml. of H20 and then diluting with 20 ml. of MeOH, was added over a period of 20 minutes to a refluxing solution of 500 mg. of the ketone in 20 ml. of MeOH and 5 ml. of HzO. After heating for 3 hours, the reaction mixture was decomposed with 3 ml. of concd. HCl and then heated for a n additional hour. T h e mixture was cooled, diluted with HzO and processed as above; yield 488 mg. T h e crude product was chromatographed as before. T h e original hexane eluate gave 22 mg. of unreacted ketone and the yield of pure p-01 was 342 mg. (71%), m.p. 153-155", and of pure a-01 was 76 mg. (16%), m.p. 177-179'. Cholestan-3p-ol-7-one .-In our hands, the preparation of this compound by direct hydrogenation of 7-ketocholesteryl acetate followed by alkaline hydrolysis, as described by Buser,Ig was found to yield material containing some unsaturated ketone ( € 2 3 5 3000). T h e following indirect method was employed in the present work. 7-Ketocholesteryl acetate (15.0 g., 0.034 mole), prepared in 6570 yield from cholesteryl acetate by the procedure of Oppenauer and Oberrauchzs using t-butyl chromate, was dissolved in 200 ml. of ethyl acetate and hydrogenated using 1.0 g. of PtOz. After the uptake of hydrogen had ceased (1.2 moles), the catalyst was filtered and the solvent evaporated. T h e crude product had a m.p. of 134-140'. T h e crude material was then dissolved in 200 ml. of acetic acid and hydrogenated using 1.0 g. of PtOz until no further uptake of hydrogen occurred. T h e catalyst was filtered, the solvent evaporated and the residue recrystallized from MeOH to yield 7-hydroxycholestanyl acetate, m.p. 95-105", yield 9.20 g. (61%). T h e molar extinction a t 235 mp was 30. The above solid (9.20 g., 0.021 mole) was dissolved in 125 ml. of acetic acid and oxidized with a solution of CrOa (2.28 g., 0.029 mole) in 4.0 ml. of HzO and 150 ml. of acetic acid. T h e reaction was allowed to proceed for 16 hours a t 20". T h e solvents were removed under reduced pressure a t 2030". T h e residue was diluted with HzO, extracted 3 times with ether and the ethereal extracts washed with dilute HzS04, hTaHC03solution and HzO and then dried. After
+
(23) R. V. Oppenauer and H. Oberrauch, Anales asoc. puim. argen37,246 (1949).
&a,
[CONTRIBUTIOS FROM
THE
ETIANATES
3755
evaporation of the solvent, the material was recrystallized to yield 7-ketocholestanyl acetate, m.p. 146.3-147.4' (lit.24148-149"), yield 8.08 g. The over-all yield based on 7-ketocholesteryl acetate was 54%. The acetate (4.55 g., 0.01 mole) was saponified by refluxing for one hour with 125 ml. of EtOH containing 6.0 g. of KOH. T h e cooled reaction mixture was diluted with Hz0, filtered and recrystallized from aqueous EtOH to give 7-ketocholestanol, m.p. 156.3-158' (lit. 156-157" ,25 164-165" , 2 4 yield 3.89 g. (94%). Reduction of Cholestan-3p-01-7-one Acetate with LiAlH4. ( a ) Diethyl Ether a s Solvent.--A solution of 2.24 g. (5.03 mmoles) of the acetate in 50 ml. of dry ether was reduced in the usual manner with 100 ml. of an ethereal solution of LiXIHa containing 0.056 mmole per ml. The yield of epimeric 7-hydroxy compounds was 2.05 g. (10070), [ a I z 5 D +28.1" ( a 0.285, c 1.012). T h e infrared spectrum of this material showed no band in the carbonyl region. A sample (1.080 g.) of the crude mixture was absorbed on 30 g. of Woelm neutral alumina. Elution with benzene gave 3 mg. of non-crystalline material. Elution with redistilled ether gave 1.01 g. (93.5y0) of the isomeric 7-hyD , ( a +0.286, c 1.062). droxy compounds, [ ( Y ] ~ ~ +26.9 This rotation corresponds to a mixture of 57% a and 4370
+
8.26
(b) Tetrahydrofuran a s Solvent.--A mixture of 0.500 g. (1.13 mmoles) of the acetate and 0.171 g. (4.55 mmoles) of LihlH4 in 60 ml. of tetrahydrofuran was heated under reflux for 1 hour and processed a s above. A quantitative yield of crude product was obtained. Chromatography yielded 2 mg. of material with hexane. T h e mixture of epimeric 7-hydroxy compounds showed a n [ a ]25D +23.2' ( a +0.252, c 1.086) which corresponds to 657, a and 35% 8 . Reduction of Cholestan-3p-ol-7-one with IiaBH4.-To a solution of 0.822 g. (2.04 mmoles) of the compound in 40 ml. of MeOH and 10 ml. of ether was added a solution of 0.679 g. (18.4 mmoles) of NaBH4 in 5 ml. of HzO and 25 ml. of MeOH. T h e reaction mixture was stirred a t room temperature for 22 hours and then heated under reflux for 2 hours. After cooling, ether was added and the solution acidified with 57, HCl until acid to congo red. T h e layers were separated and the aqueous layer extracted 3 times with ether. The combined etheral solution was washed with 5% HC1, HzO and dried. Evaporation of the solvent gave 0.792 g. (96%) of crude product. Chromatography of 0.700 g. as described above yielded only 2 mg. of material with benzene. The mixture of isomeric 7-hydroxy compound ~ ( a +0.262, c 1.282) which corresponds had [ a I z 2+19.So to 73y0 a and 277, P . (24) 0. Wintersteiner and IM. Moore, THIS JOURNAL, 65, 1503 (1943). (25) A. Windaus and E. Kirchner, Ber., 59, 614 (120). (26) T h e [m]D used for t h e pure epimers was f51.0° for 78 and +8.5O for 7a.
BERKELEY 4, CALIFORNIA
DEPARTMEKT O F MEDICINE,WESTERN RESERVEUNIVERSITY,
A N D THE
LAKESIDEHOSPITAL]
The Preparation of 16-Oxygenated Etianates and their Relation to Gitoxigenin1 BY H. HIRSCHMANN AND FRIEDA B. HIRSCHMANN RECEIVEDFEBRUARY 13, 1956 The preparation and reduction of both 17-epimers of methyl 3p-acetoxy-16-oxoetianateis described. The principal reduction product of the 176-isomer is a cis-hydroxy ester which is shown t o be methyl 3P-acetoxy-16P-hydroxyetianate. I t gave a diacetate identical with a degradation product of gitoxigenin. The results strongly indicate that gitoxigenin is 16@-hydroxydigitoxigenin.
Gitoxigenin (I) has been obtained as the steroid who found i t to differ from digitoxigenin by an addimoiety from several cardiac glucosides. It has been tional hydroxyl group a t C-16. The sole structural intensively studied in many laboratories, notably features not yet fully established concern configuraby Jacobs and his co11aborators2 and by MeyerJ3 (1928); f b ) 82, 403 (1929); ( c ) 86, 199 (1930); (d) 88, 531 (1930); (1) This investigation was supported by grant C-1679 of the Sntional Institutes of Health, U. S. Public Health Service. ( 2 ) (a) LV. A . Jacobs and E. I.. Custos, J. Bioi. Chem., 79, 553
(e) W. A . Jacobs and R. C. Elderfield, ibid., 108, 4 9 i (l935), (3) ( a ) R. Meyer. Helv. Chiin. Acta, 29, 718 (1P10); !b) 29, 1350 (1946); ( c ) 89, 1908 (1946).
Vol. 7s
H. HIRSCHMANN AND FRIEDA B. HIRSCHMANN
3756
tions of ring D. The substituents a t C-14 and C-17 were shown to be cis to each other4 and their attachment, while not rigorously proven, is generally believed to be beta as it is in other cardiac aglycones. No such agreement exists with respect to C16. Many years ago Shoppees proposed the cisrelationship of all 3-substituents, since gitoxigenin forms a 16-21 oxide I1 on treatment with alkali,2d and was thought to form a 14-21 oxide upon oxidation of the 16-hydroxyl group.? Hydrolysis of I1 followed by oxidation gave a lactone I11 which 1;-C',
Ten years ago me ye^^^ reported a degradation of gitoxigenin to a methyl 3/3,16-diacetoxyetianate (VI) and thus obtained the only degradation product which retained the oxygen function a t C-16 and possessed a hydrogen with the normal ~ r i e n t a t i o n ~ ~ a t C-14. Although his sequence (I to VI) of reactions involved an allylic intermediate v, the conditions used seemed mild enough15to assure the C00bIe
Yo
I I J O
IV COOMc
/o-
COOMe
/o-
H
V
H
VI
retention of configuration a t C-16. Our developshowed extraordinary stability toward alkali.2a,b ment of a procedure for the introduction of a 16aThis compound therefore must be a cis-lactone since hydroxyl group into the steroid molecule16prompted great strain can be deduced for the epimeric struc- us to attempt the synthesis of compound iI with ture from the repeated failures to obtain trans-2-hy- the aim of determining its configuration. While droxycyclopentaneacetic lactone from its parent this work was in progress, Rloore5 reported a synacid.s &Lactones of this type show normal reac- thesis of the crystalline methyl 30,lea-dihydroxytivity toward The inertness of I11 can be etianate and of its amorphous diacetate. Since the explained, however, by steric hindrance if the lac- degradation product from gitoxigenin failed to intone ring is cis also to C-18 and to the 14-hydroxyl duce crystallization and possessed a different rotabut seems quite inexplicable if the ring were at- tion, it was concluded that the two diacetates were tached to the opposite (a)side of ring D. If the 14- not identical but most likely epimeric a t C-16. As hydroxyl group is removed, the lactone ring hy- Dr. Moore was unable to continue with this work, drolyzes a t a faster rate.2b Moreover, as pointed we have proceeded with our studies and have ohout by Turner,'O the conversion?CBe of the lactone to tained the isomer of methyl 30,lG-diacetoxyetianetianic acid demonstrates that I11 has the 0-con- ate that is identical with the degradation product figuration a t C-17. The same assignment, there- from gitoxigenin. The procedure used by Moore for the preparation fore, appears justified for the oxygen function a t of lea-hydroxyetianates consisted in the ozonizaC-16 in I11 and 11and probably also in I. The 16a-configuration was first proposed for gi- tion of the 21-benzylidene derivative of VIII, foltoxigenin by Reichstein" and was adopted by lowed by hydrolysis, hydrogenolysis and methylaCardwell and Smith.12 The main point in its favor tion of the resulting acids. Our reaction sequence is is the facile elimination of the 16-acetoxy group longer but gave better yields. The first steps were and the 17-hydrogen under conditions which nor- patterned after our preparation of 16a,21-diace16a-configuration is much more tenuous as i t is based not on observamally favor cis-elimination. l 3 tions of gitoxigenin b u t a companion substance, adynerigenin, which I1
111
(4) K. Meyer, Helu. Chim. A c f a , 32, 1993 (1949); see also t h e discussion by Moore., ( 5 ) J. A. Moore, ibid., 37,6.59 (1964). (6) C. W.Shoppee, A n n . Rev. Biochcm., 11, 125 (1942). (7) This interpretation failed t o account for t h e facile isomerization of t h e product and appears t o be incorrect (cf. C. W. Shoppee and E. Shoppee in E. H . Rodd, "Chemistry of Carbon Compounds," Vol. I I B , Elsevier Publishing Company, Hooston, Texas, 1953, p. 1012). (8) W. Hueckel and W. Gelmroth, A n n . , 614, 233 (1934); W. E . Grigsby. J. Hind, J. Chanley and F. H. Westheimer. THISJOURNAL, 64, 2606 (1942). (9) J . W. Corcoran and E€. Hirschmann. i b i d . , 78, 2323 (1956). (10) R. B. Turner, Chem. Rez's., 43, 7 (1948) (11) T . Reichstein, Angew. Chcm.. 63, 418 (19.51). (12) H. hl. E . Cardwell and S. Smith, J . Chem. Soc., 2012 (1054). (13) For references and discussion see Moorel and Cardwell and Smith." A second argument presented by t h e British workers for t h e
has been obtained from t h e same plant. Since t h e hydroxyl group which this compound is alleged t o have in t h e 16a-position could not be acetylated," i t is improbable t h a t t h e reactions of adynerigenin proceed as formulated by Cardwell and Smith. (14) R. Tschesche and G. Grimmer, Chem. Ber., 87, 418 (1954); R. Tschesche and G. Snatzke, i b i d . , 88, 511 (1955). (15) See, e.g , t h e retention of configuration of a n unstable (axial) 60acetoxy group under similar conditions of dehydration and hydrogenation ( D . N. Jones, J. R. Lewis, C. W . Shoppee and G. H . R.Summers, J . Chem. Soc., 2876 (19.55); V. A . Petrow, 0. Rosenheim and W.W. Starling, ibid., 677 (1938)). I n compound V the e-site is less crowded and if isomerization t o t h e more stable form occurs, one would expect i t t o be from B t o Q. Hence if V I , as will be shown, is IGB, this con6jiuration most likely was present already in gitoxigenin. (16) (a) H . Hirschmann, F. B. Hirschmann and &I. A. Daus, THIS J O U R N A L , 74, 539 (1952); (b) H . Hirschmann, F. B. Hirschmann and J. W.Corcoran, J . Or& Chcm., 10, 572 (1955).
Aug. 5, 1956
PREPARATION OF ~G-OXYGENATED ETIANATES
3757
toxyprogesterone. l7 3p - Acetoxy - A16- pregnen -20- atom possess a common enol, their ready interconone was hydroxylated a t C-16 (IX) by the benzyl version was to be anticipated. Partial epimerizaalcohol method16b and after acetylation (X) ace- tion of either isomer could in fact be observed spectoxylated a t C-21 with lead tetraacetate.l8 The 21- trographically on several occasions.2s Nevertheacetoxy group in XI was hydrolyzed with bicarbon- less the process must be considerably slower than ate to the keto1 which was degraded with m-perio- the enolization of simple P-ketoestersz6to have perdate.l9 The removal of unchanged compound X is mitted the fractionation achieved. Since the meltcarried out most readily a t this stage. Methylation ing ranges of the final products were rather wide i t of the acidic material gave chiefly methyl 3@,16a- is probable that epimerization interfered to some diacetoxyetianate (XII). This product could be extent with purification. Nevertheless this mutual hydrolyzed without appreciable elimination a t C-16 contamination could not have been very serious to yield methyl S~-acetoxy-16cr-hydroxyetianatejudging from the very good reproducibility of the (XIIIb) and the corresponding dihydroxy ester infrared curves and the results of the hydrogenaXIIIa. Their over-all yields from VI1 were 13.6 tions to be reported. The enol content of comand 3%, respectively. When the pure dihydroxy pound XIV, in 95% ethanol, was estimated specester was reacetylated, the product XII, like that of trographically as less than 1% and hence as subMoore, failed to crystallize. Non-identity with stantially lower than that of ethyl 2-oxocyclopenthe degradation product VI from g i t ~ x i g e n i nwas ~ ~ t a n e c a r b o ~ y l a t e . ~This ~ difference probably can be confirmed by infrared comparison. ascribed to the attachment of a chair form of a cyThe oxidation of the @-hydroxyester X I I I b with clohexane ring which should interfere with the cothe chromium trioxide-pyridine complex20 gave planarity of the ethenoid carbons and their substitvery little acidic material. Chromatography of uents, particularly if the fusion is trans.28 the neutral fraction removed some unchanged The conversion of 16-ketosteroids to 1GP-hystarting compound but otherwise caused little droxy compounds has been realized by a variety of fractionation. Nevertheless the main product procedures including hydrogenation with platicould be separated by crystallization from non-po- num in acetic acidz4or with nickel in a l ~ o h o lor~ ~ * ~ ~ lar solvents into two products with the composition reduction with sodium b ~ r o h y d r i d e . ~When these of a methyl acetoxyoxoetianate. They were recog- methods were applied to compound XIV, they all nized as distinct compounds by their rotations and gave the same hydroxy ester XVI although in their infrared spectra which differed widely in the greatly varying yields. When nickel was used in fingerprint region. The carbonyl peaks indicated large amounts, compound XVI was the sole prodthat both were 16-ketoetianates.*l Since only epi- uct that could be detected. I n contrast to the isomerism a t c-17 is consistent with the mode of for- meric structure XIIIb which possessed a hydroxyl mation and the spectra of these isomers, they have peak a t 2.78 and a single carbonyl peak a t 5.77 p , been assigned structures XIV and XV.22 Although the new hydroxy ester showed a hydroxyl peak a t vicinal interaction effects were expected, the ob- 2.86 and two clearly separated maxima a t 5.76 and served rotations agreed quite well with the calcu- 5.83 p. Since this spectral pattern persisted on lated values.23 Accordingly, the less levorotatory dilution, it was concluded that the hydroxyl group isomer has been assigned the 17-normal configura- is linked by hydrogen bonding to the ester carbonyl of the same molecule.31 This signified a cis relation. Since two @-ketoesters epimeric a t the a-carbon tionship between these groups.32 When the nickel reduction was applied to the 17-ko ester XV, the (17) H. Hirschmann, F. B. Hirschmann and G. L. Farrell, THIS reaction product, which failed to crystallize, posJ O U R N A L75, , 4862, 6363 (1953). sessed a spectrum very different from XVI. The (18) T. Reichstein and C. Montigel, Hclv. Ckim. Acta, 22, 1212 (1939). main constituent XVIIIa again showed hydrogen (19) S. A. Simpson, J. F. Tait, A. Wettstein, R. Neher, J. v. Euw, 0. Schindler and T. Reichstein, i b i d . , 87, 1200 (1954). (20) G. I. Poos, G. E. Arth, R. E. Beyler and L. H. Sarett, THIS JOURNAL, 75, 422 (1953). (21) N. J. Leonard, H. S. Gutowsky, W. J. Middleton and E. M. Petersen ( i b i d . , 74, 4070 (1952)) observed a frequency shift of both carbonyl peaks, if a keto group is in a 5-membered ring and @ t o the carboxylate. Ethyl 2-oxocyclopentanecarhoxylate (pure liquid) had maxima at 5.69 and 5.80 p. If the molecule contained a second more distant ester group, t h e 5.8 fi maximum a n d the normal ester peak were not resolved (Amax 5.76 or 5.77 p ) . Compound X I V showed maxima at 5.66 and 5.76 p , compound X V a t 5.69 and 5.78 p in carbon disulfide. (22) Recently a lecture was published in which Fried reported t h e conversion of biosynthetic 1 6 4 1-dihydroxyprogesterone t o methyl 3.16-dioxo-A4-etienate (J. Fried, R. W. Thoma, D. Perlman, J. E. Herz and A. Borman, Recent Pvogr. Hormmtr Research, 11, 149 (1955)). Isomerism of the keto ester was not mentioned. (23) [.bl]o calculated for X I V -334O, for X V - 6 0 6 O ; found -320O and -537'. The calculations are based on [ M I Dfor 3@,20@-diacetoxyallopregnan-16-one2~ -NE0, 3~,20@-diacetoxyallopregnane 125O (H. Hirschmann, M. A. Daus a n d F. B. Hirschmann, J. Biol. Chcm., 192, 115 (195111, methyl 38-acetoxyetianate (XVII) +197O (all in alcohol), methyl 38-acetoxy-17a-etianate - 105O (chloroform. M. Sorkin and T . Reichstein, Heiu. Chim. Acta, 29, 1209 (1946)), apChl. +30° (D. H. R. proximate solvent correction A[M]o EtOH Barton and W. Klyne, Ckcmisfry &Industry, 766 (1948)).
+
-
(24) H. Hirschmann, F. B. Hirschmann and M. A. Daus, J. B i d . Chcm., 178, 751 (1949). (25) I n view of this it was most startling t o find t h a t X I V gave a transitory purple color with ferric chloride, while X V did not unless i t had first been partially isomerized by exposure t o pyridine. (26) Ethyl acetoacetate is quite stable in either one of its tautomeric forms only if t h e preparation is highly purified. Polar solvents like alcohol which favor the keto form greatly accelerate equilibration. Non-polar solvents shift the equilibrium toward the enol h u t do not catalyze equilibration. However, trace impurities act a s very potent accelerators (P.Grossmann, 2. pkysik. Chcm.. 109, 305 (1924); F. 0. Rice and J. J. Sullivan, THISJOURNAL, 5 0 , 3048 (1928)). (27) W. Dieckmann, Ber., 55. 2470 (1922); Buu-Hoi a n d P. Cagniant. Bull. SOC.ckim. France, [ 5 ] 10, 251 (1943). (28) I n keeping with the usual behavior of (3-diketones, 16,20dioxosteroids enolize much more extensively, Nevertheless only a single crystalline species has been described in every case (G. E. Arth, G. I. Poos and L. H. Sarett, THISJOURNAL, 77, 3834 (1955); S. Bernstein, M. Heller and S. M. Stolar, ibid., 77, 5327 (1955)). This, however, appears not t o be a diketone b u t a chelated enol.Zlv29 (29) R. S. Rasmussen, D. D. Tunnicliff and R. R. Brattain, ibid., 71, 1068 (1949). (30) S. Kaufmann and G. Rosenkranz, i b i d . , 7 0 , 3502 (1948). (31) R. N. Jones, P. Humphries, F. Herling and K. Dobriner, i b i d . 74, 2820 (1952). (32) P . Hirsjaervi, Acto Chcm. Scond., 8 , 12 (1954).
3758
CHI
1 C=O
1-11
COOMe
XIIIa, R b,R
i
XVIIIa, R = H b, R = Ac
= =
COOMe
H AC
COOMe
XVIl
XI1
t
CrO3.pyr
XV
bonding between the hydroxy group and the ester carbonyl.33 The formation of two different cishydroxy esters from the isomeric keto esters XIV and X Y demonstrated that the hydrogenation could not have been preceded by an equilibration of the two keto esters. Moreover, the substrate must have been adsorbed on the catalyst as the ketone rather than the enol. If the reaction represents the hydrogenation of the carbonyl group, the hydroxy esters have the same configurations a t C-17 as the keto esters from which they were derived. However an alternative mechanism seems a t least conceivable, namely, that the keto ester enolized while attached to the catalytic surface and was reduced as such. In this case the configuration that results a t C-17 would depend entirely on the direction of the attachment of the catalyst to the steroid. If the catalyst were attached to the @-sideof compound XIV, inversion would result from this mechanism and the reduction product would then possess structure XVIIIa. The results of the hydrogenation of the 17-normal keto ester XIV with platinum, however, render (33) T h e wave length of t h e hydroxyl peak (2.83 p ) was significantly shorter t h a n in X V I , b u t t h e maximum a t 5.83 p provided full confirmation of hydrogen bonding. (In order not t o mistake a n a,@unsaturated for a hydrogen bonded carboxylate, t h e sample was acetylated. T h e peak a t 5.83 p disappeared.) T h e shorter wave length of the hydroxyl peak indicates a weaker (longer) hydrogen bond (L. P Kuhn, THISJ O U R N A L , 74, 2492 (1952)). The reawn for this difference between t h e two isomers i s not clear.
X1.I
XIV
this alternative forlnulation very unlikely. In this case the hydroxy ester XVI was accompanied by a hydrogenolysis product34XVII which was unequivocally identified as methyl 3P-acetoxyetianate. Since the hydrogenolysis of enol acetates is a common occurrence with this catalyst 35 we should expect that the postulated reduction through the enol would also be accompanied by hydrogenolysis. If this is the mode of formation of XVII. the a-attachment of the catalyst can be deduced from the configuration of the hydrogenolysis product a t C-17. This, however, is the opposite of that postulated for the alternative formulation of the hydroxy ester. The results obtained with sodium borohydride fully confirm that the reduction of XIV to XVI involved no inversion a t C-17. If it did, the epimerization would have had to occur prior to reduction, and the borohydride reduction of XV which already possesses the 17-is0 structure should then have given a t least the same amount of XVI as was obtained from the normal ester XIV. Examination (34) Carbon-oxygen hydrogenolysis has been reported also for acetoacetate when a platinum, but not when a nickel catalyst was used: (a) M. Faillebin, Aun. chim. ( P a v i s ) , (101, 4, 156 ( 1 9 2 3 ) ; ( b ) H. Adkins, “Reactions of Hydrogen with Organic Compounds over Copper-Chromium Oxide and Xickel Catalysts,” T h e University of Wisconsin Press, Madison, WIS , 1044, p 31 i c ) H Adkins and €1 R Billica. Trim J O U R N A L , 7 0 , 69: (1918) ( 3 6 ) H. H. Inhoffen, G. Stoeck. G . Korlliug and U. S t u c c k , A w 668, 52 (19.50), and references there cited; K. Hirschmantl, 1T I 3 n ) n u and X L. Wendler, THIS J O I ! R N A I . , 73, 5373 (1951). ~
Aug. 5, 1956
P R E P A R A T I O N OF 1 6 - O X Y G E N A T E D
ETIANATES
3759
3~-Acetoxy-l6~-benzyloxypregnan-2O-one (VIII).-The of the crude reaction product, however, failed to treatment of 1.33 g. of 3P-a~etoxy-A~~-pregnen-20-one (VII) give any indication of the formation of the hydroxy with 40 ml. of 3Y0 benzyl alcoholic potassium hydroxide was ester XVI from the 17-isoester XV. We conclude, carried out for 3.5 hr. a t 28" as described p r e v i o ~ s l y . ~ ~ therefore, that the hydroxy ester XVI has the same T h e crude reaction product was reacetylated with 8 ml. of p-configuration a t C-17 as compound XIV. Since pyridine and 4 ml. of acetic anhydride a t room temperature 17 hr. The mixture of acetates (1.6 9.) was chromatothe carboxylate was shown to be cis to the hydroxyl. for graphed. Most of the earlier eluates gave unreacted startcompound XVI is methyl 3@-acetoxy-16@-hydroxy- ing material, the later fractions furnished 3P-acetoxy-16aetianate. benzyloxypregnan-20-one which crystallized from methanol Comparison of the rotations of the two 16,17-czs in needles (798 mg., m.p. 133-136.5'). The analytical ~ ( c 0.8, chloroform) gave m.p. 135-137", [ a I z 6+24' methyl 3p, 16-diacetoxyetianates lends support to sample A,, 5.76, 5.86, 13.66 and 14.36 p (lasb two, benzyloxy).'B this assignment. Inversions of a carboxylate a t Anal. Calcd. for C30H4204: C, 77.21; H , 9.07. Found: C-17 from 6 to a and of an acetoxy group a t C-16 in C, 77.05; H , 9.23. Methyl 3P-Acetoxy-16a-hydroxyetianate (XIIIb).-The the same direction generally cause large negative finally adopted was carried out without purificashifts. The 16-acetate of XVI, therefore, should be procedure tion of intermediates. Infrared spectroscopy and, in the more dextrorotatory than XVIIIb. Although the case of step c, the tetrazolium reaction37 were used to check rotational difference was much smaller than esti- on the success of the reactions. (a) Reduction.-.% mixture of catalyst (303 mg. of 5% mated, assuming no vicinal effect, it has the correct palladium chloride-charcoal, pre-reduced and washed as sign. described before),'eb of 798 mg. of 30-acetoxy-16a-benzylAccording to our formulations compounds XVI oxypregnan-20-one (VIII) and of 180 ml. of 95% alcohol and X I I I b are epimeric a t C-16. The relative was stirred magnetically in an atmosphere of hydrogen. orientations of the substituents a t C-16 and C-17 in The gas uptake ceased after 37 minutes. After removal of the catalyst, the solution gave 646 mg. of crystalline residue. these compounds are consistent with the following Recrystallization of such material from acetone gave 3Pobservations : Whereas the trans-ester XI11 acetyl- acetoxy-16a-hydroxypregnan-20-one( I X ) (m.p. 165-1665', ates readily with pyridine and acetic anhydride a t [ O ~ ] ~ * 4-49' D (c 0.45)). Anal. Calcd. for CzrH&: C, room temperature under our, standard conditions, a 73.36; H,9.64. Found: C, 73.08; H,9.65. compound I X (646 mg.) in 5 complete reaction with XVI required much longer ml.( bof) Acetylation.-Crude pyridine and 3 ml. of acetic anhydride was kept a t periods. Furthermore, treatment of XI1 with car- room temperature for 14.5 hr. T h e reaction product (744 bonate caused little elimination a t C-16, whereas mg.) failed t o crystallize. I t s infrared spectrum showed no the acetate of XVI under these conditions gave hydroxyl peak nor any of the maxima assigned t o the benzyllarge amounts of the a$-unsaturated ester. Since oxy( c )group. Acetoxy1ation.-The crude acetate (744 mg.) was the 16-acetoxy group in XVI is trans to the hydro- distributed over 12 tubes. (This may not be necessary. gen a t C-17, the more facile elimination in an ionic It was done because in the A5-series17 yields were much better with small batches.) T o each was added 5.5 ml. of reaction is to be anticipated. warm solution of lead tetraacetate (6.72 g. in 92 ml. of The acetate of methyl 3@-acetoxy-16/3-hydroxy- aglacial acetic acid and 7.75 ml. of acetic anhydride). T h e etianate (XVI) was identical with the degradation sealed tubes were kept a t 68" for 46 hr. The excess anhyproduct VI from gitoxigenin. 3s Since the reac- dride was hydrolyzed with water after the addition of pyritions reported by Meyer could not have altered dine (1 ml. per tube). T h e combined reaction mixture gave of neutral product. the configuration a t C-IT, it follows that the side 785(d)mg. Hydrolysis and Keto1 Cleavage.-A mixture of this chain a t C-17 and, t h e r e f ~ r e the , ~ hydroxyl group material in 340 ml. of methanol and 8.25 ml. of 1 N aqueous a t C-14 of gitoxigenin likewise have the p-orienta- potassium bicarbonate was kept under nitrogen for 4.5 hr. tion. As pointed out above, the degradations most a t 24", acidified with 0.59 ml. of glacial acetic acid and concentrated i n ZJUCUO. Since the infrared spectrum of the neuprobably also preserved the configuration a t C-16. tral reaction product showed incomplete hydrolysis of the The identification of this degradation product a s a 21-acetoxy group, the hydrolysis was repeated (7 hr.). 16p-acetoxy steroid, therefore, lends strong support (The shorter reaction period is adequate if the 16-hydroxy to the view that gitoxigenin likewise possesses a p- group is formylated rather than acetylated.) A mixture of neutral fraction (713 mg.) in 100 ml. of methanol and oriented hydroxyl a t C-16. If the new evidence is the of 60 ml. of 27, sodium m-periodate in water was kept in the considered jointly with the observations of Ja- dark for 135 minutes and then concentrated i n F U C U O . The cobs, et a1.,2 there remains little ground to question product was extracted from the acidified solution with etherchloroform and separated with sodium carbonate into an the correctness of this assignment. acidic and neutral (363 mg.) fraction. T h e latter can be reworked by repeating steps b, c and d. (Sodium bicarbonate General Procedures.-All m.p.'s are corrected. Unless proved to be quite ineffective in extracting etianic acids.) when noted otherwise, rotations, ultraviolet spectra were ( e ) Methylation and Carbonate Hydrolysis.-The acidic measured in ethanol, infrared spectra in carbon di- fraction in 2 ml. of methanol was treated with an excess of sulfide with a Perkin-Elmer single beam spectrometer with diazomethane in ether (22 ml.). T h e product (341 mg.) was double pass. The infrared bands given in parentheses are mostly the diacetate XI1 of methyl 3P,16a-dihydroxyetianthose particularly useful for the differentiation from the ate, but contained some 3-monoacetate X I I I b owing to parother stereoisomers described in this paper, the others serve tial hydrolysis of a nuclear acetate group during the proto characterize functional groups. These include the com- longed bicarbonate treatment in step d . To complete the plex ester peak near 8 b (axial acetateI36 and the maxima hydrolysis, a mixture of this product in 25 ml. of methanol near 8.63 (acyl-oxygen stretching of methyl etianate~)~E and and 5 ml. of 0.5 N aqueous potassium carbonate was kept a t 9.80 p (alkyl-oxygen stretching of 3p-acetoxy-5@-steroids). 20" for 15 hr. The neutral reaction product (309 mg.) was Reaction products were isolated by ether extraction, chromatographed. T h e early eluates failed t o crystallize. washed free of basic or acidic reagents with hydrochloric The middle fractions (benzene 10Y0 ether) were freed of a acid or bicarbonate in the cold and then with water. I n sparingly soluble brown pigment by repeated extractions chromatography, a 2 : 1 mixture of silica gel and Celite (50 with ligroin-methanol (20: 1). Recrystallization from ligtimes the weight of the crude steroid) was used and preroin gave 198 mg. of methyl 3P-acetoxy-16a-hydroxyetianate washed as described .9 T h e eluent generally was benzene One specimen on reheating ( X I I I b ) (m.p. 122.5-124.5'). containing increasing amounts of ether. The ligroin used of the solidified melt gave m.p. 129.5'. T h e analytical had b.p. 90-96", the petroleum ether 60-70".
Experimental
+
( 3 6 ) R N. Jones and F. Herling, J . Org. Chem.. 19, 1252 (1954).
(37) W. J. LCader and R R . R u c k , Anal. Chrm , 24, GGG (19.52)
H. HIRSCHMANN AND FRIEDA B. HIRSCHMANN
3760
VOl. 7 8
sample (m.p. 123.5-124.5') had [ a I 2 ' D +20° ( c 0.6), A,,, and 13.02) p . Anal. Calcd. for C23H3606: C, 70.37; H, 2.78, 5.77, 8.00, 8.11, 8.63, 9.79 (9.50and 1 3 . 1 5 ) ~ . Anal. 9.25. Found: C, 70.18; H, 9.38. T h e infrared spectrum Calcd. for C23H310L: C, 70.37; H, 9.25. Found: C, 70.21; of the crude reduction product agreed closely with that of H, 9.36. pure XVI and showed none of the bands characteristic of The late eluates upon recrystallization from acetone gave X I I I b or XVIIIa. A non-crystalline preparation of this 40 mg. of methyl 3p,16ol-dihydroxyetianate (XIIIa). One compound was described by h f e ~ e r . 3 ~It probably conpreparation showed m.p. 204-205', another 190-192" and tained methyl 3p-acetoxy-A16-etienate. Reductions of XIV 204' on reheating of the solidified melt, [ C Y ] ~ ' D$22" (c 0.5, using only 30 mg. of catalyst came to a standstill before cornmethanol). Moore6 reported m.p. 190-192 and [ a ] ~pletion. 1-18.6'. Compound XVI (9.0 mg.) and its mother liquors were Alternative Preparation of Compound XII1.-At the out- each kept in 1 ml. of pyridine and 1 ml. of acet.ic anhydride for 44 hr. The reaction products had virtually identical set of the experiments i t was unknown whether the carbonate hydrolysis of step e would cause elimination. Therefore, infrared spectra and both gave on recrystallization from compound I X was formylated a t C-16 by the procedure de- dilute acetone and from petroleum ether methyl 3p,168diacetoxyetianate (VI) with m.p. 152.5-153.5', [a]25~ scribed.'eb Although the formate group was quantitatively hydrolyzed by the short hydrolysis of step d , keto1 +24' (c 0.6, 95% alcohol), +22" ( c 0.5, chloroform); A,:, cleavage and methylation gave a mixture of X I I I b and X I 1 5.75, 8.64. 9.80 (8.23, 8.40 and 13.01) p . Anal. Calcd. which were separated chromatographically. Evidently for C26H3806: C, 69.09; H, 8.81. Found: C, 68.99; H , 8.91. T h e infrared spectrum was identical in every respect step c had caused some ester exchange. This modification, with that of a specimen (m.p. 152.5-153.5') prepared by therefore, does not eliminate step e (hydrolysis). Prof. K . Meyer from gitoxin. There was no change in Methyl 3@,16a-Diacetoxyetianate (XII).-A solution of 10.3 mg. of methyl 3j3,lGa-dihydroxyetianate( X I I I a ) in 1 m.p. on admixture. T h e reported [ a ]is~ $30.9" (chloroml. of pyridine and 0.5 ml. of acetic anhydride was kept at form).3s When XVI was acetylated as described for XIIIa, room temperature for 15.5 hr. T h e neutral product (12.6 the hydroxyl and carbonyl peaks of the starting compound mg.), which failed to crystallize, showed an infrared spec- were still very distinct. B. With Platinum.-A mixture of 14 mg. of compound trum identical with t h a t of compound X I 1 obtained in the preparation of X I I I b by the alternative procedure. There X I V , of 15 mg. of platinum oxide (according to Adams and was no hydroxyl peak, A,, 5.75, 8.62, 9.78 (7.73, 8.34 and Shriner) and of 6 ml. of prereduced glacial acetic acid were shaken with hydrogen for 60 minutes. T h e product was 13.15) p . acetylated and chromatographed. T h e early eluates (benMethyl 3p-Acetoxy-16-oxoetianate (XIV).-A solution of zene+ 1 ether) yielded methyl 3p-acetoxyetianate (XVII). 195 mg. of chromium trioxide in 13 ml. of pyridine (for preI t s m.p. (127.5-125.5') was not depressed by admixture of a ~ ~ added ) to 194 mg. of methyl cautions see Poos, et ~ 1 . was 3B-acetoxy-16a-hydroxyetianate and kept in the dark a t reference sample, supplied by Prof. Reichstein, [ a ]2 5 ~ +52" ( c 0.3); A,,, 5.77, 7.99, 8.13, 8.64 and 9.80 p ; lit. 27' for 28 hr. The neutral reaction product (183 nig.) was freed by chromatography from unchanged starting material m.p. 126-128", [ a ] D +54' ( a c e t ~ n e ) . ~ aThe infrared spec(eluted with benzene and 15% ether). The earlier eluates trum was identical with t h a t of the reference sample. Anal. Calcd. for C23H38O4: C, 73.36; H, 9.64. Found: (153 mg. with benzene and 5y0 ether) were recrystallized C , 73.71; H , 9.75. from ligroin and gave 71.3 mg. of methyl 3j3-acetoxy-16T h e later eluates gave methyl 3p,lGj3-diacetoxyetianate oxoetianate ( X I V ) as needles melting a t 137-144". Continued recrystallization failed to sharpen the m.p.; [ a ] z e ~ (m.p. 152.5-153.5') identified b y mixture m.p. and infra-82" (c 0.6),ultraviolet A,, 259 ( e 78, enol), shoulder red comparison. All maxima of the crude unacetylated reduction product were present in the spectra of either XVI -293 m p ( E 34, ketone) (c 0.09); infrared A,, 5.66, 5.76, 8.00, 8.11, 8.64, 9.80 (7.44, 8.81, 10.42 and 12.41) p . or X V I I . C. With Sodium Borohydride.--A mixture of 8.4 mg. of Anal. Calcd. for C23H3406: C, 70.74; H , 8.78. Found: C, 70.82; EI, 9.04. T h e sample recovered from the rota- compound X I V in 0.5 ml. of 80% dioxane and of 19 mg. of sodium borohydride in 0.95 ml. of the same solvent was kept tion had practically the original spectrum, but others kept for similar periods in neutral polar solvents showed partial a t 22' for 6 hr. and then acidified with hydrochloric acid. The product showed hydroxyl peaks at 2.78 and 2.87 p . isomerization to compound XV. The twin maxima of XVI a t 9.24 and 9.40 p were easily seen, Reference data for estimating r$ enolz7: ethyl 2-oxocyclohut the very strong peak of XVIIIa a t 9.34 p was absent. pentanecarboxylate in absolute ethanol showed A,, -258 The spectrum indicated the presence of X I I I h in this mixmp, E 640 and bromometric enol content of 6.4%; in 95% ture, but only XVI was definitely identified by the isolatioii alcohol E 350. Methyl 3P-Acetoxy-16-0~0-17a-etianate(XV).-The first of VI after acetylation, chromatography and recrystallizamother liquor (70.7 mg.) of XIV was allowed to evaporate tion. Reductions of Methyl 3p-Acetoxy-l6-oxo-l7~~-etianate slowly. Heavy blocks separated which were removed mg. of compound XV was mechanically (43.7 tng.) and recrystallized from dilute (XV). A. With Nickel.-Ten acetone and from ligroin to yield 23.1 mg. of X V which had reduced with 350 mg. of nickel as described above. The product which failed to crystallize had A,,, 2.83, 5.76, 5.83,7.99, m . p . 150-159', [ a ] " D -137' (c 0.4); Amax 5.69, 5.78, 8.00, 8.11, 8.62, 9.79 (7.63, 8.71, 10.15 and 10.29) p . Anal. 8.09, 8.14, 8.57 (etianate?),9.79 and (9.34) p a n d a shoulder or weak maximum near 2.78 p . T h e maxima characteristic Calcd. for C23H3105: C, 70.71; H , 8.78. Found: C, 70.98; H , 9.01. A41though there were no manifestations of muta- of XVI were absent. .kcetylation for 45 hr. gave a product 5.75, 8.65, 9.79 and (10.30) p clearly different rotation, the sample recovered from this measurement on with A,, from VI. I t did not crystallize and was chromatographed. recrystallization gave m.p. 125-134" and infrared spectrum The middle fractions had [ a ] 2 5 D -12' (c 0.4). showing extensive isomerization to XIV. B. With Sodium Borohydride.-Compound X V (2.2 Isomerization of Keto Esters with Pyridine.-Compounds mg.) was reduced as described for XIV. The product XIV (2.5 mg.) and XV ( 2 . 5 mg.) were each kept in 1 ml. of pyridine for 28 hr., diluted with ether and washed with acid. showed two hydroxyl peaks (2.77, 2.83 p ) . .4lthough evidently less pure, it was spectrographically similar to the one The spectra nf the neutral residues were nearly identical, the minor differences indicating incomplete equilibration. obtained with nickel. The twin peaks at 9.24 and 9.40 p and other maxima characteristic of XVI were absent. The equilibrium appears to be in favor of the 17-normal Elimination Reactions.-Compound X I 1 (2.5 mg.) and epimer XI\'. Reductions of Methyl 3P-Acetoxy-16-oxoetianate(XIV). VI (2.5 mg.) after examination of their ultraviolet spectra A . With Nickel.-The catalyst for all experiments was W-5 mere each dissolved in 0.4 ml. of methanol and treated with 0.08 ml. of 0.54 N potassium carbonate a t 18' for 17 hr. of Xdkins and B i l l i ~ a ~ prepared '~ on 0.1 scale and washed The product derived from X I 1 showed a shoulder near 225 \vith more water (10 1.) than was needed to obtain a neutral mp with an increment in E of 500, the one from VI had a supernatant. I t was used within 3 days of preparation. This catalyst (300 mg.), 24.8 mg. of XIV and 6 ml. of 9
