Dehydtogenation of NDehydroestrone to Equilenin (VI) .-A solution of 2.5 g. of A*-dehydroestrone acetate in 70 cc. of glacial acetic acid was refluxed with 0.43 g. of freshly sublimed selenium dioxide for ten to fifteen minut& in a current of nitrogen. After filtration of seleni into water, there was obtained 2.4 g. of acetate with a reddish color, which was re tion in hexane-benzene ( 5 :1) solution through a short column of alumina. Recrystallization from nietbanol gave long prisms, with a slight pink tinge 111. p. 1A6-157' (COT,), [a]*% 4-72'; no depression wa iiescrved on admixture with the 3cetdt.e (nl. p. l.57-15R' prepared from the natural hor The uliravioli i i l l -
c , 73.16; 11, 6.33. Saponiticatiou of the acetirtc by boiling with .jc inethatiolic potassium hydroxide solution for twrn ty-five minutes, followed by sublimation t1z ~JUCUOand recrystdlli7ation from dilute ethanol, yielded equilenin ( V I ) , m . p . 256-258' (red melt), [a]WD +86.4" (dioxane) ; the m. p. was undepressed on mixing with a sample of the natural hormone (kindly supplied by Ayerst, McKenna and Harrisun, Ltd.) . The characteristic spectrum is given in Fig. 2 uith maxima a t 230, 270, 280, 292, 328 and 340 m,u a n d I I X S nearly indisti:~guisi~al~le from that of the natural hor1lione. 15
Summary The Wohl-Ziegler bromination of A1J4-androstadiene-3,17-dione (and of the corresponding 17acetoxy derivative) proceeds smoothly and after dehydrobromination of the intermediate A-bromo cornpouncl Ira leads t o A ' "-;tiidrostatriene-3,17dione jIIIa 1. .ironiatization in mineral oil solut i o i i of I11a at tiY)O" results iri the foriliation of A'>tlrhydroestrone ( A'-isoequilin) (IV) iii 40% yield, who5e structure provc'd by hydrogenatioii to t stroiie I i') arid by tlehytirogenation to equilenin (VI ) . This constitutes the first partial synthesis of equilenin from a steroid with an angular inethyl group a t C-10. and the last niaximum was observed a t 340 mp (in agreement with Dirscherl and Hanusch, Z physiol. C h ~ m . 258, , 13 (1935)) rather than at 34.5 ma (ref. 11). Our values were obtained on a Beckman Spectrophotometer which had been calibrated both in regard to W R V D l r i i g t h !benzrne) and extinction (potassium chromate in alkali) 1,) !hc 1rwiidiire of Kognr-is, P ! r ~ l . ,J , P h y s . Chcm., 41, 379 (1937).
131 Our log E values for both the natural and semi-svnthetie l u i p i i , ' r \ t r r atroi,r 0 ? ,init 1rint-r t h a n th0.c reported rarlirr irtf 11
Steroids. VI1.l
Contribution to the Bromination of A4-3-Ketosteroids and a New Partial Synthesis of the Natural Estrogens
BY CARLDJERASSI, G. ROSENKRANZ,J. KOMO,ST. KIUFMANN .\ND J. PATAKI
11,4-Dien-3-ones (XI) of the steroid series represent the key intermediates in the partial synthesis of the female sex hormones, estrone and estradiol from non-aromatic steroids possessing an angular methyl group at C- 10, since thermal treatment results in aromatization of ring A with fonnation of the desired phenol (XTVj.2*3This process is now used commercially and the required dienones have so far only been prepared by dibromination of saturated 3-ketoallostercfids followed by clehydrobromination of the resulting 2,4-dibrorno derivatives. The alternate approach involving as ii starting material a 2i4-3-ketosteroid (I), such as testosterone, appeared attractive and was studied in very considerable detail by Inhoffen? and by Rutenandt.-' Their study was limited to a model A4-3-ketosteroid, AJ-cho1esten-3-one (IC)and was concerned primarily with the behavior of this ketone toward bromine. These reactions, which a t times proved to be very involved, led these workers2,* to the conclusion that mono- and poly-broniinatiori (up to eight moles of bromine) of A*-3(1) Paper VI, Kautmann, Pataki, Rosenkranz, R o m o and Djerassi, T x s JOURNAL, 73, 4331 (1950). ( 2 ) Cf Inhoffru Angta'. Chcm., 63, 473 (19401, idid, 69, 207
(1047) (3) (ar Hershberg, Rubm and Schwenk, J. Qrg Chcm , 16, 292 (1960); (b) Wilds and Djerassi, Tms J o r a h n ~ ,68, 213B i1940)i Djerasoi snd Scholr, ibcd., 71, 3902 (1949) (1) Butenandt, Sfhramm and Hudnus, A n n , 6% 178 (1937).
ketosteroids (I) results in substitution at C-4 and in ring B and that, therefore, these ketones are useless as starting materials for the partial synthesis of the estrogens. Since a reinvestigation of this problem in this laboratory has provided a novel partial synthesis of all of the major natural estrogens and has thus thrown open to question the structure assignments of a considerable number of compound^,^^^ a detailed discussion of the brornination of A4-3-ketosteroids(I) is necessary. At the outset it should be noted that most of the work done on the bromination of A4-3-ketosteroids (I) was carried out in acetic acid, alone or diluted with another solvent, and the present discussion is limited to this type of solvent. It is stated2*4*5 that monobrominatwn of A4-cholesten%one (IC) in ether-acetic acid leads to the Abromo derivative (IIc). This substance (m. p. 132', u. v. maximum a t 248 rnp) is known and has been prepared by a number of methods6P6but it has never been isolated in the monobromination of A4-chofesten-3-one (IC) in acetic acid, because a difficultly separable mixture is formed,' and there(5) InhofXen, Ber., US, 2141 (1936). ( 6 ) Ruzicka, Bosshard, Fischer and Wirz, Xclw. Chim. Acta. 19, 1147 (1930); Dane, Wang and Schulte, 2. ghysiol. Chcm., MI, 80 (1938): Reich and Lardon, Heln. Chim. Acta, 49, 671 (1946); Mnrtens, Ann., W, iai (leis). (7) UnpubiiRhcd obaarvstioo; of. dao ref. 6, p. 9 1 4 6
Oct., 1950
THEBROMINA4TIONOF A4-3-KETOSTEROIDS
fore, although possible, i t is not proved that the 6bromo derivative I1 is the initial product in that r e a ~ t i s n . ~ * ' ,Dibrominatimt ~ of A4-cholesten-3one in etheracetic acid leads to a homogeneous productP*sm.p. 163O, u. v. maximum at 248 mp, to which the 4,6-dibromo-A4-cholesten-3-one (111) structure was a ~ c r i b e d . ~ The , ~ structure proof is based primarily on a series of reactions,6l9none of which exclude rearrangement of a bromine atom situated originally a t C-2. Finally, it is noteworthy that the introduction of a bromine atom on the double bond (111) results in no change in the spectrum (as compared to 11) and that the dibromo ketone I11 reacts readilyg with potassium acetate to yield an unsaturated enol acetate, originally (and most likely incorrectly) formulated by Inhoffens as IV, and more recently by Fieser, et ul.,sa as 3-acetoxy- A2kholestadien-4-one (IVa) . In marked contrast to these observations is the behavior of A'-2-bromocholesten-3-one and other A'-2-bromo-3-ketosteroids (V) whose structure has been established beyond doubtlo," and which proved to be extremely resistant toward potassium acetate.ll In that instance, the introduction of the bromine atom on the double bond resulted in a bathochromic shift of cu. 25 mp. The obvious conclusion must be drawn that the course of the dibromination of A4-3-ketosteroids (I) in general, and of A4-cholesten-3-one(IC) in particular, in acetic acid-ether has by no means been established. The further bromination with up to eight moles of bromine was carried out in Butenandt's laboratory4J2and need not be discussed in detail except to point out that structure VI (or its 8-14 double bond isomer) was proposed for this polybromination product. VI as well as its catalytically partially debrominated derivative VI112 were characterized by ultraviolet absorption maxima a t ca. 270 mpJ 295 mp and one below 240 inp which was not measured. Further removal of bromine" af(8) An isomer, m. p. 122O was obtained from the mother liquors and was believed to be a C-6 diastereoisomer, because it afforded the same product with potassium acetate as did the m. p. 163' isomer. Since no spectral data are given for this isomer, the point cannot be decided definitely. but the same situation may obtain as described in the experimental portion of this paper with testosterone acetate. (9) Inhoffen, Ber., 69, 1702 (1936). (9a) Fieser, Fieser and Rajagopalan, J. Org. Chem., 13,800 (1948), suggested this alternate structure (IVa) for the enol acetate as arising by a "special mechanism" from Inhoffen's 4,6-dibromo- A4cholesten-3-one (111) through initial rearrangement t o a hypothetical A'-4,4-dibromo-3-ketosteroid. Their alternate structure (IVa), though much more likely than IV, has not been proved uuequivocally and has no direct bearing on the structure assignment of the dibromination product(s) of A'-3-ketosteroids (I) (uide infra), since regardless of the structure of such bromoketones, a rearrangement must be involved in the enol acetate formation (Fieser, et ai.) and this rearrangement can of course occur equally well with any one of the four possible dibromination structures Xa-Xd. (10) Djerassi and Scholz, Tms JOURNAL, 69, 2404 (1947); for a discussion of absorption spectra, see p. 2406, footnote 5 , and Djerassi and Scholz, ibid., 70, 1911 (1948). (11) Inboffenend Zuehlsdorff. Ber.. 76, 283 (3948). (12) Werner Bukon, D. Sc. Diuat.tien, Danzig, 1988 (pub. lirhcd 1940).
4536 R
R
I3r
Ia, R = OCOCHl b,R=O C , R = CsH17
11~1, R = OC0CH.y
b,K = 0 C,
I
IZ
=
C8Hi7
I
t-
o I Va
d
q
I
Br Br 111
IV
forded an impure trienone, m. p. 76-SOo,believed to be VIII, which also showed these three characteristic peaks in the ultraviolet. I t is noteworthy (IX), whose synthethat A1J~5-cholestatrien-7-one sis seems ~nambiguous,'~ exhibits maxima a t 230 and 350 mp.la In that basis, the ultraviolet absorption spectra of compounds VI-VI11 appear to be incompatible with the structures assigned to them and this uncertainty is applicable, ipso facto, to their various precursors.* R GHi7 I
v
I
R Rt VI, R = R' = Br VII, R = Br; R' = H VIII, R = R' = H
IX
The above mentioned inconsistenciei led us to reinvestigate the dibromination of A4-3-ketosteroids, particularly of the androstane series, as a route to the partial synthesis of the estrogens. When carried out in acetic acid-ether solution, testosterone acetate (Ia) afforded a dibromo derivative in 85 to 90% yield on treatment with two moles of bromine a t 0'. This substance, m. p. 174' (dec.), [a]D +43', u. v. maximum a t 24% 250 mp was also obtained on dibromination with (18) K m t 8nd Ndk, E 6 h Chlm. Add, 8% 1892 (1849),
two moles of N-broniosuccinimide in carbon tetrachloride and on monobromination of 6-bromotestosterone acetate (IIa)lJ with bromine in etheracetic acid. Short boiling with collidine resulted in a smooth loss of two moles of hydrogen bromide and afforded in 45y0 over-all yield (based on la) A1~4*6-androstatrien-17-01-3-one 17-acetate (XII), identical with a specimen prepared by an unequivocal method,' which involved N-bromosuccinimide bromination of the corresponding A1$J-dienone iXTa) followed by collidine dehydrobromination. Occasionally in the dibrorninatioii of testosterone acetate (Ia) d dibrorno derivative was obtained with the same decomposition poiirt and approximately the same rotation as mentioned above but with an ultraviolet maximum a t ca. 240 nip. The same material was also obtained from the mother liquors of the 250 r n p product, but both substances afforded the A'1~*~-trien-3-0ne XIIa in essentially the same yield on dehydrobrominatioti
6-bromo- A4-androstene-3,17-dione (IIb), prepared from A4-androstene-3,17-dione (Ib) and N-bromosuccinimide, possesses a maximum a t 240 mp, in contrast to that at 248 mp exhibited by the corresponding 6-bromo derivatives (IIa and IIc) of testosterone acetate and A4-cholesten-3-one. The structure of all these 6-bromo derivatives was confirmed by conversion to the A4~6-dien-3-ones with their characteristic ultraviolet absorption maxima at 282 mg ; the alternate %bromo- A4-androstene3, 1'7-dione16structure was eliminated since under identical conditions it afforded the A1e4-dien-3-one (XI, R = 0 1 with its maximum a t 244 mp. This apparently anomalous spectral behavior of 6bromo- A4-androstene-3,17-dione (IIb) now finds a reasonable explanation in the light of Barton and Miller's observationL5"that 6a-bromo- A4-cholesten-3-one possesses a maximum a t 238 mp in contrast to the maximum a t 248 mp observed by InhoiTen6for the 60-bromo isomer. This would indicate that the configuration of the 6R bromine atom may have an unexpectedly large influence on the absorption spectrum. € 'h ' e supposition K , . ',/\, that 6-bromotestosterone acetate \ A\ I (Ha) possesses the 60 configuration I arid I1 --+ -+ and 6-bromo-A4-androstene-3,17-di, . niaximuni dt 248 mM (log E 4.15). I~ouiitl. C, Cl.S6, 11. T :;7 This q ~ ~ c i i n e \\as i i tlrhydrobrotninatett separately M ith To coilfirm the itructure ot the compound, it Wd:, de- collidine dud yielded 50% of the trienolone acetate XIIa. hydrobroniiridted with collidine (30 minutes of refluxing) to Monobromination of 6-Bromotestosterone Acetate (IIa) . yield A'~fi--a~idrostadie~i-17~l~~-one 17-acetate, ni, p. --hlonobromination of 0-bromotestosterone acetate in %",u. v maximum a t 284 mp (log ether-acetic acid afforded in 60% yield the dibromo de143-144". [el% E 4.53) Saponificatiou gave A4,6-androstadien-17-013- rivative with In. p. 169-173" (dec.), [ a ] * %+42", u. v. one, m. p 2U4-21)5", [ t r j W ~+76.6', maximum a t 254 mp maximum a t 248 in+ (log E 4.10), (Br, calcd.: 32.74; (log E 4.471.24 found: 33.033. Dehydrobromination similarly led to the 6-Bromo-A4-androstene-3,17-dione (IIb) .-The reactrienolone acetate SIIa. tion was carried out d-. dbove udiig A'-androstene-3,17Dibromination of AA-Androstene-3,17-dione (Ib) c corresponding &bromo derivadione (Ib)c~tndled t The dibrominatioii ~ 7 3 5carried out exactly as described tive in b i yield, i. 173-177" (dec., cor ) , [a]ZOo te5tosterone .tcct,tte except that a ~zispenszonof A?+ Log", u v $*iaxviii 240 it-1, flog K 1.33'1: reporter126 for .c1idrostcne-3,17-tlione (Ib) wa5 employed since the ketone in. , I . 171' 1% r i o t sufficiently soluble iii ether. Occasionally, detion WAS very slow :inti a red, unstable intermediate which suddenly diswlved x\-ith decolorization and 'The structure proof, necesbary because of the unusually vigorous evolution of hydrogel1 bromide. Usually, delorn position or the u. v. maximum, was provided by colli- colorization was rapid, the solid simultaneously going into dine dehydrobromination (91Yc yield of collidine hydro- solution, and it i5 conceivable that impurities in the ether bromide) which gave 5170 2f pure A4so-androstadiene- play a role in this reaction. On working up as above, ca. (Kofler), / O ~ ] ~+138", D 3,17-diour, in. p . iA8-170 AOyo of colorless crystals, in. p. 170--173', [a]% t l . l 2 " , maximum a t 282 mp (log E 4.52) ; reporte;lZGm. p. J X o . u . v. maximuin a t 240 mp was obtained and an additional The identical procedure with 2-hromo-A4-androstene-3, 17- 20-220I, OF equal purity from the mother liquors. Reof A*j4-androstadiene-3,17-dione crystallization from chloroform-ethanol yielded colorless dione'5 yielded ( X i , R = O)lGwith r n . p 13,5-138",t i . v. marimurn a t needles, in. p. 172- 17'5' (dec.), 157-160' (Kofler), [ a j r n ~ 214 nip (log E 4.23). -116O, u. v. maximuin at 240 mp (log E 4.19). Identical Dibromination of Testosterone Acetate (Ia) : A. With result 5 were obtained in four experiments. Bromine in Ether-Acetic Acid.-A solution of 19.85 p;. of .lnnl. Calcd. for C&-Iz4O2Br9: C, 51.37; H, 5.15; Br, testosterone acetate in 650 cc. of dry ether was cooled in :3.5.98. Found: C, .51 09; H, 5.69: Rr, 35.99. ire, a few drops of 4 A- hydrogen bromide in acetic acid was Monobromination of &Bromo- and 6-Bromo-A4-andro:~cldrdfollowed by a colution of 19 5 g . of bromine in 180 rc. OF acetic avid at the rate a t which the solution decolor- stene-3,17-dione. --The moiiobromiiiation of the 2\mmoi6 and 6-bromo iIIb, rtdc szipm) isomers had to be ized. Aftw ( a . hvc to ten minutes, the ether and part of carried out in :I suspension of 2 g. of ketone in 70 cc. of the acetic aciil nas d illed under reduced pressure at 15rther and 25 cr. o f glacial acetic acid in view of their in20' arid the product s filtered n t i d washed with ethanol. solubility, which w a s also responsible for the fact that comThe yield of colorless needlrs r%asusually
