3120
J. A. ZDERIC,D. C. LIXON,H. J. RINGOLD AND C. DJERASSI
Anal. Calcd. for CS1HS206:C, 73.76; H, 10.38. Found: C, 74.32; H, 10.57. Reduction of 4cu,5-diacetoxycholestan-3p-olwith lithium aluminum hydride in the usual nianncr gave cholestan-3@,4a,5-triol, m.p. 212-214", undepressed by triol prepared by direct reduction of the cu-diolone 111. Reduction of 4p,5-Diacetoxycoprostan-3-one with Sodium Borohydride.-In the manner described above 0.85 g. of this diacetoxy ketone was reduced with sodium borohy(tentative asdride to give 4~,5-diacetoxycoprostan-3~-ol signmentIg at 3), m.p. 163-166". Anal. Calcd. for C31Hj10S:C, i3.76; H, 10.38. Found: C, 73.56; H , 10.47. Reduction of 4~,5-diacetoxycoprostan-3~-ol with lithium aluminum hydride in the usual manner gave coprostan-3a,48,5-triol, m.p. 81-82', undepressed by triol prepared by direct reduction of the p-diolone 11.
1'01. 81
Reduction of 4~-Acetoxy-5-hydroxycholestan-3-one with Sodium Borohydride.-A suspension of 1.80 g. of this ketone in a solution of 0.5 g. of sodium borohydride in 45 ml. of methanol and 5 ml. of water was stirred for seven minutes and then diluted with dilute hydrochloric acid. After the evolution of hl-drogen ceased the acidic mixture was extracted with ether, which was in turn washed with aqueous sodium bicarbonate, dried and evaporated. The residue was recrystallized from ethanol to give 4cu-acetoxycholestanJp,S-diol (tentative assignment'g a t 3), m.p. 207-208". Anal. Calcd. for CzsHjoOc: C, 75.28; H, 10.89. Found: C, 75.29; H , 10.82.
Acknowledgment.-This investigation was aided by the Texas Company, which furnished G. B. I f . a graduate fellowship for the year 1936-1957. KNOXVILLE, TENN.
[COSTRIBUTION FROM THE RESEARCH LABORATORIES O F SYSTEX, S.X.]
Steroids. BY
CXII. JOHN
Cycloethylene Ketal Formation of 19-Nor-a4-3-keto Steroids
A. ZDERIC, DINORAH CHAVEZ LIMON,H. J. RINGOLD .iND CARLDJERASSI RECEIVED OCTOBER 30, 1958
Cycloethylene ketal formation in the 19-nor-A4-3-keto steroid series has been shown to produce a mixture of isomeric The structure proofs for these double bond isomers were accomplished via epoxidation followed by fission of the epoxides with boron trifluoride. The resulting ketal fluorohydrins were then degraded to known or readily identified products. The results not only permit the unamibiguous location of the double bonds but also provide a complete description for the stereochemistry of the intermediate epoxides. A6 and A3(10)-3-cycloethylene ketals.
In connection with another problem, we have had occasion to investigate the formation of 3-cycloethylene ketals in the 19-nor steroid series and now report some of our findings. The fact that the formation of 3-cycloethylene ketal derivatives in the 10-methyl-A4-3-ketonesystem results in rearrangement of the double bond to the A6-position has been known2 for some time. More recently the mechanism of the reaction has received attention3 and the conclusion reached was that the isomerization resulted from the intermediate formation of a A3,5-enolether type compound. While the isomerization of the double bond was expected to occur in the 19-nor steroid series i t was evident t h a t such an isomerization could now lead to Ab- and/or A5(10)-cycloethylene ketal isomers. Several considerations must be taken into account in predicting the preferred position for the double bond in such cases. For example, i t is known4 that bromination of 3-keto all0 steroids I leads to 2bromo compounds, thus indicating that in the A/R trans series the enolic double bond is more stable in the Az-position 11. On the other hand, bromination and sulfonation evidence shows that in A/R
JOURNAL, 81, (1) Paper CXI, A. Bowers and H. J. Ringold, THIS 1264 (1959). (2) (a) F. Fernholz and H. E. Stavely, abstracts of t h e 102nd Meeting of t h e American Chemical Society, Atlantic City, hT,J . , 1941, p. 39 M : see also E. Fernholz, U. S. Patents 2,356,154 and 2,378,918; (b) R. Antonucci, S. Bernstein, R. Littell, K. Sax a n d J. H. Williams, J . Org. Chem., 17,1341 (1952); (c) G. I. Poos, G. E.Arth, R. E. Beyler and L. H. Sarett, THIS JOURNAL, 76, 422 (1953). (3) C. Djerassi and 11. Gorman, abid., 76, 3704 (1953). (4) A. Butenandt and A. Wolff, Ber., 68B, 2091 (1935).
cis steroids (111)the 3-4 position of the enolic double bond (IV) is favored.
0
H
I11
H
IV
It follows that the nature of the ring junction has an important bearing on the position for the double bond. If these observations are now applied to the 19nor steroid series and the B/C trans ring junction is equated to the A/B trans decalone type junction I, i t might be expected that a double bond would be more stable in the A5-position V rather than in the A5(Io'-positionVI.
Mediating against this conclusion are two factors, the first of which is that the nor steroid V does not accurately approximate the model I, since i t is reasonable to expect that ring A will conformationally modify the double bond stability a t A5. Secondly, it is known t h a t octalin-type double bonds are more stable when in the tetrasubstituted A9-p0sition, as may be witnessed by an interesting reports quantitatively describing the equilibria involved. As will be seen in the sequel, both of the possible positional isomers were obtained during the ketalization reaction. ( 5 ) W. G. Dauben, E C. Martin and G. J Ponken, J. Ovg Chem 23, 1205 (1958).
June 20, 1959
CYCLOETHYLENE
VIIIa, R = CGCH b, R=CH3 o&C=CH C, R = H
VIIa, R = CECH b, R=CH3 @ C EC,C R H= H
KETALF O R R I A T I O N
i";
3121
IN STEROIDS
IXa, R = C r C H b, R=CH3 o& C, R = H
Xa, R = C=CH b, R = CH3
,,,/
IC.
R=H
OH 1 ,,R
c -
HOF
\
0
\OH6F
XIVa, R = C r C H b, R = H
XVII
o& & p R
0 F XXa, R = H b, R = A c
t
XXI
XVI
-
E XVa, R = C - C H b, R = H
y,C=CH
?H.C=CH
HO&-Ho& o**
XIX
XVIII
Thus when the 19-nor-A4-3-ketones, VIIa,6" b,6ac , were ~ ~ treated under the usual conditions for ethylene ketal formation, the total crude reaction products were always obtained noncrystalline and on the basis of their low ultraviolet absorption maxima appeared to be 90-95% converted to the 3cycloethylene ketals. From these gums by careful chromatography in each of the three cases, only a single pure compound VI11 could be isolated and i t has been necessary to express its formula either as a As- or A5(10'-isomersince the small amounts of the pure compounds available precluded the possibility of structure proofs. It was, however, readily apparent t h a t either of the structural forms was possible, since on epoxidation of the total crude ketalization products there always resulted two isomeric epoxides I X and X. At this point i t may be pertinent to mention t h a t the formation of the 3-cycloethylene ketal of 1701methyl-19-nortestosterone (VIIb) proceeded without dehydration at C-17, since the epoxides I X b and X b were obtained in fair over-all yields based on the starting ketone VIIb. This point becomes of interest in view of the report' that attempts to form ( 6 ) (a) C. Djerassi, L. Miramontes. G. Rosenkranz and F. Sondheimer, THISJOURNAL, 76, 4092 (1954); (b) A. J. Birch, J . Chem. Soc., 367 (1950); A. L. Wilds and N. A. Nelson, THISJ O U R N A L , 76, 5366 (1953). (7) G . Cooley, B. Ellis, D. N. Kirk and V. Petrow, J Chem. Soc.,
.
the 3-cycloethylene ketal of 17a-methyltestosterone resulted in dehydration of the 17P-hydroxy group. In order to prove the structures of the epoxides I X and X the main effort was concentrated on the 19-nortestosterone VIIcGband 17a-ethynyl-19-nortestosterone (VIIa)Ga cases, since in these series most of the final products of the proposed structure proofs were already known. Thus in the 19-nortestosterone case i t was found that the epoxide IXc on treatment first with perchloric acid to effect epoxide fission, followed by strong alkaline dehydration and rearrangement,s led to a product whose analytical data and infrared spectrum indicated the structure of 3,6-diketo-19-norandrostane-17@-01 (XVII). Indeed, the agreement of the physical constants with those in the literatureg leaves little doubt that structure XVII is correctly assigned. It now logically followed that XVII could only arise from a 5,C-epoxide although the stereochemistry of the epoxide had not yet been established. With regard to the isoineric epoxide Xc, i t was allowed to react with boron trifliioride under condi4112 (1957); however, J. 4.Campbell, J. C. Babcock and J. A. Hogg, THISJOURNAL, 6 0 , 4717 (19581, report a 757, yield of the ketal of 17amethyltestosterone. (8) B. Ellis and V. A. Petrow, J . C h e m Sor , 1078 (1939). (9) R. L. Pederson, J. A. Campbell, J. C. Babcock, S. H. Eppstein. H. C. Murray, A. Weintraub, R. C. Meeks, P. D. Meister, L. .\I. Reineke and D. H. Peterson, THISJ O U R N A L 78, , 1512 (1956).
3122
J. A. ZDERIC,D. C. LIMON,H. J . KIXGOLDAND C. I.)JEKASSI
Vol. b l
tions which are already knownlO to give the fluoro- tal epoxide IXa, it io1lo.c~~ that IXa is correctly hydrins of 5,B- and 3(10)-epoxides. By this means represented as 5a,6 a-oxido- 17a-ethynyl-19-norailthere was obtained the fluorohydrin XIVb which drostane-1~3-ol-3-onc3-cycloethylene ketal. upon treatment with perchloric acid in tetrahydroTt-hile the stereochemistry of the 5a,Gcr-oxidofuranzc was smoothly converted to the Sa-fluoro- 19-nortestosterone 3-cycloethylene ketal (IXc) was 106-hydroxy-3-ketone (X1.b) . This same com- not proved rigorously as in the preceding case IXa. pound was recently prepared in this Laboratory'"' there appears to be no reason to assume that the by an unambiguous route and a comparison of the epoxidation of its prccursor would difler from that samples showed them to be completely identical. of the corresponding 1 7a-ethynyl analog IXa. Furthermore, since the structure of XI*b TWS Xs shown in the Experimental section, the 56,proved1oCby base-catalyzed dehydrofluorination to 103-epoxides X were less polar than the 5a,Ba-epox1OB-hydroxy-19-nortestosterone(XVI), and since ides I X in both the 19-nortestosterone and 17athis same 103-hydroxy compound has now been ob- ethynyl-19-nortestosterone series. -Applying this tained directly from the 5,10-epoxy-3-ketal X c by observation to the 17cr-meth~l-19-nortestosterone treatment with perchloric acid followed by dilute derivatives. one may- assign to the less polar epoxide alkali! there can be no question that the epoxide Xc the structure SP,lOt3-oxido-l7a-methyl-l9-noranis correctly formulated as $3, 106-oxido-19-noran- drostane-l7,5'-01-3-one 3-cycloethylene ketal (Xb) drostane-lT~-ol-3-one3-cy-cloethylene ketal. and to the more polar epoxide that of Sa,Ga-oxidoThere still remained the problem of stereocheini- 17a-niethyl-19-tiorandrost:Lne-l;~-ol-~-one 3-cyclocal assignment to the 5,B-epoxides and this was ac- ethylene ketal (IXb). complished by studying the isomeric epoxides I X a It is interesting to note that in the j,B-position, and Xa in the 17a-ethynyl series. the epoxide appears t o be almost completely in thc ItThen the epoxide X a was treated with boron tri- a-configuration whereas, in contrast, a t the 3,lOfluoride it provided the ketal fluorohydrin XIVa position the $-epoxide is the sole product isolated.Ioc which was then hydrolyzed a t C-3 to yield the Considering that in the 19-nor steroids the relaknown1nc 5a-fluoro-17a-ethynyl- 19-norandrostane- tively bulky 103-angular methyl group has been 108,1/,Bjdiol-3-one (XVa). The isolation of this replaced by hydrogen, it is perhaps surprising that compound characterized the epoxide X a as 5p, 108- a considerable quantity of the 58,GO-epoxide is not oxido- 17a -ethynyl- 19-norandrostane- 17~-01-3-one formed. The possibility that a small amount of 3-cycloethylene ketal. the Sp,G,&epoxide is contained in the epoxides The positionally isomeric epoxide IXa, when mother liquors cannot be completely- discounted similarly treated, led to the fluorohydrin XI and since our recovery of the two pure epoxides IXa upon hydrolysis of its ethylene ketal group it pro- and X a was only 6iFT. A suitable rationale for vided the fluorohydrin-3-one XII, whose constitu- the predominant formation of the 513,lOp-epoxides rather than the epimeric 5a,lOcr-epoxides has altion was demonstrated as follows : Starting with 17a-ethynyl-Aj-19-norandrostene-ready been presented in an earlier paperlGCfrom 3,B,17P-diol (XVIII)l1 and treating with monoper- these laboratories. On the basis of the yields encountered for the phthalic acid, there was isolated in iO-SO%l, yield a single epoxide (XIX) which upon treatment with isomeric epoxides IX and X, and assuming no boron trifluoride was converted to 6p-fluoro- 17a- Sb',B,O-epoxideformation, it can be concluded that ethynyl-l9-norandrbstane-3,8,5a,l7~-triol (XXa) . the ratio of A5(10) and aj,6-3-cycloethyleneketals T h e stereochemistry of XXa and hence of XIX, is in excess of two to one. This in turn suggests was clearly indicated by the fact that upon acetyla- that the governing factor in the stability relationtion X X a provided only a 3-monoacetate (XXb) as ship is the higher degree of substitution as found shown by elemental analysis and acetyl determina- in the A5(10)-isomer. tion. Had X X a been a 3/3,6&epoxide, i t would E~perimental'~ have led to a 5a-fluoro-3fl,G,8-dioland i t has already 3-Cycloethylene Ketal of 17ru-Ethynyl-19-nortestosterone been shownlGathat compounds of this type readily ('VIIIa).--To 150 ml. of lxnzenc, 15 1111.of ethylciie glycol forin diacetates. :inti 2.00 g. of 17u-eth\-n?l-ln-nortestosterone (TII:i),6,L ;idded 0.10 g. of p-toluenesulfonic acid. T h e rcsultitig Additional confirination for the Sa-hydroxpt u r e was then stirred untler reflux for 14 hours using :I ($3-fluoro structure was obtained by oxidatioii o f 5v-atc.r separator. rlfter this tiiiii. the solution w:l\ \vasl~e(l thc fluorohydrin X X a to the 3-keto-fluorohydriii first with dilute aqucous sodium bicnrbonate, t h v n \ \ ' a t v ~ XI1 which could then be dehydrated t o provide 6p- and finally dried over sodium sulfate and cv;tpor:iteti. 244 mp, log e 2.77, fluoro-17cr-ethynyl-19-nortestosterone (XXI). The There remained 1.8 g. of gum, structure of this latter compound was indicated by which was chromatographed on 35 g. of neutral alutniiia. with henzene-ether provided 0.3 g. o f cryst:lls its elemental analysis, ultraviolet spectrum and Elution (n1.p. 140-150") which were repeatedly recrystallized from characteristic12rotatory dispersion curve. ether-hexane t o furnish the analytical sample, m.p. 163-Since XI1 has been obtained both from the au- 1 6 5 O , [ u ] D -27". Amd. Calcd. for C 2 2 H 3 ~ 0 3C,: 77.15; H , 8.83. Found: thentic 5a,Ga-epoxide XIX as well as from the ke(10) (a) 11. B. Henbest and T. I. &'rigley, J . C h e m SOL.,4763 (I!).;i); (b) A. Bowers a n d € I . J. Ringold, THIS J O U R X A L , 80, 412d ( 1 9 3 ) ; fc) J. Perez Kuelas, J Iriarte, F. K i n d and C. Djerassi, J . OYK. f h r ~ ,i 23, 1744 ( l 0 S X ) . I I I) . I Iriierx-.,i . ~ n ( l1 I I l:inKnlri, ? ' H I S J c i r ~ n s z r . ,81,
r
l>jrra,ii, n,,? & f l , l < m , l e.;tnstrr
