April 5, 1961
COMMUNICATIONS TO THE EDITOR
Anal. Calcd. for CZgHz2S; C, 86.52; H, 5.51; S, 7.97. Found: C, 86.34, 86.83; H, 6.03, 5.86; S,8.25. Oxidation of IV with hydrogen peroxide in acetic acid converted this to its sulfone (VI), colorless crystals from ethanol, m.p. 192-193", showing very strong sulfone bands a t 7.70 and 8.82 p. Anal. Calcd. for C Z ~ H ~ ~ O C,~ 80.15; S: H, 5.10. Found: C, 80.04; H, 5.18. These transformations may be summarized by the scheme shown. Further work involving changing all four phenyl groups and also replacing sulfur by phosphorus is under way. DEPARTMENT OF CHEMISTRY GEORGESULD UNIVERSITY OF PENNSYLVANIA CHARLES C. PRICE PHILADELPHIA 4,PENNSYLVANIA RECEIVEDDECEMBER 30, 1960
R
~
3
Sir: The recent syntheses of aldosterone' and of 18nitriloprogesterone, oia photochemical rearrangement of nitrite esters a t positions C-11 and C-20 of the steroid nucleus, have demonstrated the power of this synthetic method. We wish to report here the preparation of a new class of nitrogen containing steroids, formed by photolytic rearrangement of C-17p-nitrite esters. Reaction of 5a-androstane-3a,l7~-diol Sa-acetate3 with nitrosyl chloride in pyridine gave the ~ (all 17P-nitrite (I; m.p. 177-180"; [ a ] -33" rotations in dioxane unless otherwise stated.) 4; A?:' 5.76, 6.10, 6.25, 7.94 p) which was photolyzed in benzene6 to give the hydroxamic acid (11, R = OH, R' = CHBCO; m.p. 229-233'; [:ID -2'; Ag:p' 3.28, 5.77, 6.12 p ; instant color with ferric chloride), isolated by direct crystallization. Acetylation of I1 with acetic anhydride gave the Nacetoxy derivative (11, R = CH3C02; R 1 = CH3CO; m.p. 173-176"; [ a l ~+IS0; A%?' 5.58, 5.75, 5.9, 8.05, 8.12 p ) , Our formulation of the photolysis product (11, R = OH) as 17a-aza-Dhomo-5a-androstane-3a,l7a-diol-17-one 3a-acetate is substantiated by the reduction of (11, R = OH) 3ato 17a-aza-D-homo-5a-androstan-3a-ol-17-one acetate (11, R = H ; R 1 = CH3CO; map. 280284"; [ a ] D + N o ; Xz:?' 3.12, 3.24, 5.76, 5.96, 6.22, 7.95 p ) with zinc and acetic acid,6 and to 17aaza-D-homo-5a-androstan-3a-ol-17-one. (11,R = R ' = H ; m.p. 341-346'; [ a ] D +25" (CH30H); A2zP' 3.05, 3.22, 6.05 p ) by reduction with hydrazine' in ethylene glycol. The lactam (11, (1) D. H. R. Barton, J. M. Beaton, L. E. Geller and M. M. Pechet, J . A m . Chem. S O L . ,8 2 , 2640 (1960); D. H. R . Barton and J. M. Beaton, ibid., 82, 2641 (1960). ( 2 ) A. L. Nussbaum, F. E. Carlon, E. P. Oliveto, E. Townley, P. Kabasakalian and D. H. R. Barton, ibid., 82, 2973 (1960). (3) K. Miescher, H. Kagi, C. Scholz, A. Wettstein and E. Tschopp, Biochem. Z.,294, 39 (1937). (4) Satisfactory analyses were obtained for all the compounds described in this communication. ( 5 ) We thank Mr. R. Armswood for his assistance with the photolyses. ( 6 ) Cf. J. D. Dutcher and 0. Wintersteiner, J . B i d . Chcm., 166, 359 (1944); see also J. A. Moore and J. Binkert, J . A m . Chem. Soc., 81, 6029 (1959). (7) Cf. D. W. C. Ramsay and P. S. Spring, J . Chem. Soc., 3409
(1950).
R
0
VIII
0-xo
0 I !I
H
/
E3
A
THE PHOTOLYSIS OF ORGANIC NITRITES. 11. SYNTHESIS OF STEROIDAL HYDROXAMIC ACIDS
1771
OH
;+Yo-
I&
E
c
0
,-'$px, tl
I .A
n
R = H , R 1 = CK,CO) was prepared independently from 17-oximino-5a-androstan-3a-ol-3-acetateE by Beckmann rearrangement, using thionyl chloride in d i o ~ a n eand , ~ authentic I1 (R = R' = H) then was secured by hydrolysis of the Beckmann product with methanolic potassium hydroxide. Similarly, testosterone 17p-nitrite (I11; m.p. cu. 100" dec; [ a ] +69O; ~ "::A: 239 mp (17,700); XEzp' 5.98, 6.05, 6.20 p ) and 19-nortestosterone 170-nitrite (IV; m.p. 83-87" dec; [ a ] D +go; 238 mp (17,800); A:,"?' 6.0, 6.14 p ) furnished, respectively, when photolyzed in benzene, 17a-aza-D-homo-4-androsten17a-ol-3,17-dione (V, R = OH R2 = CH,; m. . 220-223"; [ a ] D +67O; "::A: 238 mp (17,9007; A$:' 3.05, 5.98, 6.20 p ) and the 19-nor analog (V, R = OH, R2 = H ; m.p. 227-235', [ a ] D +21°, ";:A: 237 mp (17,900), A",?' 3.25, 6.04, 6.14 p ) . Acetylation gave the N-acetoxy derivatives (V, R = CH3C02, R2 = CH3; m.p. 172-174' [ a ]+Soo, ~ ",:A: 239 mp (17,000),A?:' 5.58, 5.95, 6.18, 8.45 p ) and (V, R = CH3C02, RZ = H ; m.p. 185-188", [ a ] +39O, ~ 237 mp (18,200); X~z~'5.60,5.90,6.04,6.18,8.45p). Reduction of V (R = OH, R2 = CH3) with zinc-acetic acid gave the knowng,lOtlllactam (V, R = H, R2 = CH3). Photolysis in benzene solution of estradiol 3benzoate 170-nitrite (VI; m.p. 165-167" dec. ; [a]D +So; 229 mp (21,700); A?:' 5.74, 6.1, 6.24, 6.30, 7.90 p) and estradiol 3-methyl ether 17p-nitrite (VII; m.p. 143-145' dec; [ a ] ~ -11'; "::A: 276 mp (2,000), 285 mp (1,800); A:?' 6.12, 6.18, 6.22, 8.0, 8.1 p) gave the hydroxamic acids (VIII, R = OH, R3 = CF,HSCO; m.p. 227-232' dec., [ a l ~$62'; A":, 229 mp ( 8 ) A. Butenandt and K. Txherning, 2. phyriol. Chem., 229, 188 (1934). (9) Cf. B. M. Regan and F. N. Hayes, J . A m . Chem. SOL.,7 8 , 639 ( 1956). (10) B. Kaufmann, ibid., 73, 1779 (1951). (11) We are indebted to Dr. B. M. Regan for his kindness in supplying an authentic specimen of the lactam V (R = H, R2 = CHs) for
comparison purposes.
1772
COMMUNICATIONS TO THE EDITOR
Vol. 83
(21,300); A?:' 3.25, 5.78, 6.12, 6.30, 6.70, 7.96, or Co-H stretching moti0n.~'5 If the former 8 . 2 4 ~ )and (VIII, R = OH, R 3 = CHs; m.p. were true, one would expect a t least one hydrogen 276 mp (2,000), vibration a t a frequency much greater than 703 186-191"; [ L Y ~ D+SO"; A$:"' 2S5 mp (1,800); A?:' 3.25, 6.10, 6.21, 6.35, 6.67, cm.-'. This question has been investigated pre7.99 p ) , respectively. Reduction of VI11 (R = viously by examining the infrared spectrum of OH, R 3 = C6H5CO) and VI11 (R = OH, R 3 = films of HCO(CO)~ and DCo(C0)k 011 a silver chloCH3) to the knowngJOlactams VI11 ( R = H, ride window a t -196°.3 It was reported that no R3 = CeH5CO) and VI11 ( R = H , R 3 = CHs) bands were observed in the region of -1400 em.-' was accomplished using zinc-acetic acid and hy- in the deuteride.3 Now, if there were no fundadrazine, respectively. mental in the deuteride a t -1400 cm.-l, there The photolysis mixtures from VI and VI1 also could be no "pure" Co-H stretching vibration gave, in each case, another hydroxamic acid. in the hydride near -1950 cm.-l, where bands These products (IX; m.p. 205-210" dec., [ L Y ~ D have been observed.2s8 -3"; A",: 230 mp (20,000); $A": 3,28, 5.75, This question now has been reexamined by study6.12, 6.24, 6.30, 6.68, 7.92 p ) and (X; m.p. 205- ing the infrared spectrum of gaseous HCo(C0)c 276 mp (2,000), 285 mp and DCO(CO)~ 207"; [LY]D 0"; a t pressures up to 20 mm. in cells (1,800);' ::A: 3.28, 6.12, 6.22, 6.34, 6.68, 8.0 p ) with effective path lengths as great as 5 meters. correspond, respectively, to isomers of (VIII, Special attention was paid to the region between R = OH, R 3 = CGH5CO)and (VIII, R = OH, 1000 and 2000 cni.-'. Among the bands observed R3= CHS). in this region for HCO(CO)~is one with easily This dieerence most probably is due to isomerism resolvable PQR structure a t 1934 cm.-l (previously a t C-13, rather than to the presence of the alterna- reported2J8)and a weaker, somewhat broad, band tive 17-aza-17a-keto system. l 2 a t 1400 cm.-l. The samples of DCo(C0)4 had A plausible mechanism (partial structures A-E) SO-SO% of the H atoms replaced by D atoms. The can be invoked to account for these photolytic hydride band a t 1934 cm.-l almost completely rearrangements. Initial fission of the 0-NO disappears in our deuteride spectra and the band bond can lead to the intermediate C-13 nitroso a t 1400 cm.-l is replaced by a sharper band a t 1396 intermediate(s) (C). The formation of the hy- cm.-l. In addition a number of new bands were droxamic acid (E) from (C) may then follow the found in the deuteride spectra including ones a t path C + D -+ E (or equivalent process). 1193, 1691, 1787 and 1850 cm.-'. The following interpretation is placed on these (12) The known 13-isoestrone 3-methyl ether (A. Butenandt, A. Wolff and P. Karlson, Bcr., 14, 1308 (1941)) has been converted, via results. The hydride band a t 1400 cm.-" is the the 17-oxime, t o a 13-isolactam identical with the lactam derived from overtone of 703 an.-'. The latter frequency shifts X. This evidence together with other corroborating data will be disto 600 cm.-l in the deuteride where its overtone cussed in a later, detailed publication. C. H. ROBINSON now gives rise to the band found a t 1193 cm.-'. 0. GNOJ The hydride band a t 1934 ern.-' is reasonably A. MITCHELL understood as the Co-H stretching frequency I SCHERING CORPORATION R. WAYNE The Co-D stretching vibration is the band a t BLOOMFIELD, NEWJERSEY E. TOWNLEY1396 cm.-l. The deuteride bands a t 1691, 1787 P. KABASAKALIAN E. P. OLIVETO and 1850 cm.-l arise from combinations of the 1396 cm.-l fundamental with the previously observed' RESEARCH INSTITUE FOR MEDICINEAND CHEMISTRY deuteride fundamentals a t 296, 393 and 458 cm.-'. D. H. R. BARTON 49 AMHERST STREET These results form a most convincing argument CAMBRIDGE, MASSACHUSETTS for the presence of a Co-H bond in HCo(C0)4. RECEIVED FEBRUARY 18,1961 They, together with the earlier are consistent with the C3v configuration previously proposed.2 They show that the C and/or 0 atoms THE HYDROGEN VIBRATIONS IN COBALT CARBONYL HYDRIDE. BONDING do izot participate significantly in the hydrogen CONSIDERATIONS.' stretching motion and hence the bonding of the Sir: hydrogen atom is like that recently found for The determination of the frequencies and the HMn(CO)~.4~s It is now possible to give a more character of the hydrogen vibrations in HCo(C0)4 advanced discussion of the molecular vibrations would be important in establishing the nature of and the frequency assignment in HCo(C0)c the bonding in this interesting molecule. Previous and DCO(CO)~. These, as well as the experimental work has shown that the infrared bands a t details of this and recent work4 on these molecules, 7032-4 and 3314 cm.-l arise from vibrations which will be given in a forthcoming paper which will involve substantial motion of the hydrogen atom. discuss their relation to the various models for the Since the C and/or 0 atoms also participate sub- bonding of the hydrogen atom. stantially in these modes in the hydride and/or The force constant for the Co-H bond is found deuteride,' a salient question is whether the 703 to be ca. 2.22 md./A. Thus this bond is stronger cm.-l vibration involves Co-H bending motion' than the Mn-H bond in HMn(CO)6' despite contrary expectations5 based on thermal and chemical (1) This work was supported by the Atomic Energy Commission under contract AT(ll-1)-164. stabilities. (2) W. F. Edgell, C. Magee and G. Gallup, J . Am. Chem. SOC.,1 8 , The fact that the hydrogen vibrations are similar 4185 (1956). in HCo(C0)d and HMn(C0)a strengthens the be(3) F. A. Cotton and G. Wilkinson, Chcm. and I n d . , 1305 (1956). (4)'W. F. Edgell, G. Asato, W. Wilson and C. Angell, J . Am. Chcm. SOC.,81, 2022 (1959).
(5) F. A. Cotton, J. L,Down and G. Wilkinson, J . Chcm. Soc., 838 (1959).
