April 5, 1962
21-AMINO-SUBSTITUTED SWFLUOROHYDROCORTISONES
This final fraction was converted to the barium salt, reconverted to the cyclohexylammonium salt, and then subjected to preparative paper chromatography t o remove Rr 0 material, as described for the a-anomer. The cyclohexyl-
[CONTRIBUTION FROM
THE
1265
Anal. Calcd. for C17H3,N207P (412.47): C, 49.50; H , 9.04; N, 6.79; P, 7.51. Found: C, 49.13; H, 9.30; N, 6.79; P, 7.11.
CHEMICAL PROCESS IMPROVEYENT DEPARTMEST, LEDERLE LABORATORIES, AMERICANCYANANID Co., PEARLRIVER, h'. U.]
16a-Hydroxy Steroids. X1I.l" 21-Amino Derivatives of 9a-Fluorohydrocortisone BY LELAND L. SMITH,^^ MICHAELMARX,HAROLD MENDELSOHN, THEODORE FOELL'~ AND JOSEPH
J. GOODMAN
RECEIVED AUGUST18, 1961 Fermentations of Sa-fluorohydrocortisone with Streptomyces roseochiomogenus afford as a minor product a steroidal amide, 21-acetylamino-9~-fluoro-ll~,l7~-dihydroxy-4-pregnene-3,2O-dione, whose isolation, characterization, structural elucidation, and synthesis from Sa-fluorohydrocortisone are described. A synthesis of the 21-amino analog of Sa-fluorohydrocortisone is also described.
The complex alteration of 9a-fluorohydrocortisone (I) by Streptomyces roseochromogenus includes 16a-hydroxylation, 2@-hydroxylation, 20-carbonyl r e d ~ c t i o ntogether ,~ with D-homoannulation of the 16a-hydroxylated products5 and conversion t o other non-reducing products. I n a continuing study of the bioconversions of this microorganism on Sa-fluorohydrocortisone, we have noted the regular occurrence in a variety of fermentation samples of one non-reducing component 11, located midway between the substrate I and the major fermentation product 16a-hydroxy-9a-fluorohydrocortisone (111) on standard paper chromatograms. Instrumental evaluation of the fluorescence of the isonicotinic acid hydrazones on paper chromatograms indicated that about 4 5 % of the substrate could be accounted for as the non-reducing component 11. The steroidal nature of I1 was supported by infrared spectra of papergram eluates. However, major concern for the structure of the compound was provoked by strong bands a t 6.5 p , implying that the steroid contained nitrogen. Isolation of I1 from very complex extract conccntrates from which 16a-hydroxy-9a-fluorohytlrocortisone had been removed by crystallization and from which other 16a,l7a-diols were removed as their water-soluble 16a,l7a-cyclo borates has been described3 The pure I1 was recognized as a neutral polyhydroxy-A4-3-ketosteroidal amide, C23H3206NF, from which a A1s4-3-ketoneanalog IV, C23Ha06NF1 was prepared microbiologically. Physical properties of the amides I1 and I V differentiated them from naturally occurring steroidal amides such as the toad poisons, bile acid conjugates, etc., and from the nitrogenous steroid prep(1) (a) Paper XI, J. J. Goodman and L. L. Smith, A p p . Microbiol., 9, 372 (1961); (h) Wyeth Laboratories, Philadelphia, Pa. (2) R. W. Thoma, J . Fried, S. Bonanno and P. Grabowich, J . A m . Chcm. SOC.,79, 4818 (1957). (3) L. L. Smith, H . Mendelsohn, T . Foell and J. J. Goodman, J . Org. Chem., 26, 2859 (1961). (4) L. L. Smith, T . Foell and J. J. Goodman, Biochemisfry, in press. ( 5 ) L.L.Smith, M . Marx, J. J. Garbarini, T. Foell. V. E. Origoni and J. J. Goodman, J . A m . Chem. Soc., 82, 4616 (1960).
arations of Voigt and Schroeder,6steroid-nucleotide/ purine complexes, steroid-polypeptide/protein complexes, and from various synthetic and derived steroid amine and amide c o m p o ~ n d s . ~ From the characteristic infrared absorption near 3, 6 and 6.5 p exhibited by both amides I1 and IV, a non-cyclic secondary amide was suggested. Absorption in these regions is characteristic for noncyclic secondary amides in general8 and for such steroidal amides in p a r t i c ~ l a r . ~Whereas some non-cyclic steroidal secondary amides have been reported without the amide I1 bands near 6.5 p, no steroidal secondary amide (cyclic) or tertiary amide is known to us with absorption near 6.5 p . Neither amide I1 nor IV was acetylated by acetic anhydride and pyridine a t room temperature,lo nor was a cyclic acetonide derivative formed between I1 or IV with acetone-perchloric acid. Mild acid hydrolysis of IV afforded a 16-dehydro amide V, recognized as such by loss of absorption a t 5.8 p and increased absorption a t 6.0 p and a t 238 mp. No hydrolysis occurred with base (as strong as 10 AT) a t room temperature, and no steroids could be isolated when hot alkali was used. Further experiments aimed a t amide hydrolysis were not fruitful.ll (6) K. D. Voigt atld W. Schroeder, Nafurc, 176, 509 (1955); W. Schroeder and K. D. Voigt, Acta Endocrinol.. 21, 343 (1956); 27, 110 (1958). (7) K . D. Voigt and G. Kallistratos. Endokrind., 85, 56 (1988). (8) L. J. Bellamy, "The Infra-red Spectra of Complex Molecules," 1st Edition, Methuen and Co., Ltd., London, 1954, pp, 175-196. (9) G. Rosenkrantz, 0. Mancera, F. Sondheimer and C. Djerassi, J . Org. Chcm., 21, 520 (1956); K. Heusler, P. Wieland and A. Wettstein, Hclo. Chim. Acta, 41, 997 (1958); Y.Sato, et al., J . Org. Chcm., 2 2 , 1496 (1957); 24, 893 (1959); 25, 783, 786, 789 (1960); V. cerng and P.Sorm, Coll. Czech. Chem. Commun., 24, 4015 (1959); 25, 2841 (1960); B. G. Ketchum and A. Taurins, Can. J . Chem., 88, 981 (1960): M. M. Janot, Q. Khuang-IIuu and R. Goutarel. Bull. SOC. chim. France. 1640 (1960). (10) Forced acetylation with acid catalysis gave complex products which did not absorb at 6.5 p , but exhibited 0-acetate bands. (11) 6 N hydrochloric acid-acetic acid gave complex products without 6.5 p absorption, which on reacetylation absorbed in the region of 0-acetates rather than near 6.5 c. Hydrolysis of steroidal amides has generally been a problem; cf. L. Ldhler and F . gorm, Coll. Czech. Chem. Commun., 25, 265 (1960); F. Ramirez and S. Stafiej, J. A m . Chem. Soc., 77, 134 (1955); R. H. Mazur, ibid., 81, 1454 (1959); J . Schmidt-Thome, Chem. Bcr., 88, 895 (1955); ref. 12.
1266
L. L. SMITH, M. MARX,H. MENDELSOHN, T. FOELL AND J. J. GOODM.4N
Vol. 84
from an internal standard of tetramethylsilane) were present. The multiplet was recognized as the AB-portion of the spectra of an ABX-grouping, tentatively associated with the -CH2-NH- feature of the m~lecule.'~Spectra of the A4-3-ketoamide I1 are presented in Fig. 1. The structures of I1 and I V thus induced were confirmed by synthesis of the indicated compounds. Copper acetate oxidation of the a-keto1 I led t o the 21-aldehyde IX, which was characterized as the aldehyde hydrate, so established by elemental analyses and by paper chromatographic mobility typical of a tetrahydroxy-diketone. X 21-oxime X was obtained selectively from reaction of IX with 1.1 equivalents of hydroxylamine. The oxime was more mobile on paper chromatograms Fig. 1.-Proton nuclear magnetic resonance spectra (80 than anticipated, and thus appeared t o be asmc.) of 21-acetj-lamino-90i-fluoro-l1~,17~~-dihydrosy-4-pregsociated under the chromatographic conditions. nene-3,2O-dione (11) in pyridine-deuteriochloroform (1:7). Reduction of the 21-oxime X was accomplished using the zinc dust-acetic acid-acetic anhydride The A1v4-3-ketoneIV was degraded by sodium borohydride reduction, followed by sodium bis- conditions of Treibs and Sutter.lG Catalytic and rnuthate oxidation t o a known 17-ketone 9a- sodium-alcohol reduction methods were rejected in fluoro - 110 - hydroxy - 1,4 - androstadiene - 3,17- view of the other reducible functional groups in the dione (VII).12 Preparation of VI1 from Sa-fluoro- molecule. The 21-acetylamino compound isolated prednisolone (VIII) by bismuthate oxidation and in 60% yield was identical in every respect with the amide I1 isolated from S. roseochromogenus ferfrom 9a-fluoro-ll~-hydroxy-4-androstene-3,17dione by selenium dioxide oxidation established its mentations on 9a-fluorohydrocortisone. Later work showed that the inclusion of the structure. A structure thus suggested for IV is that of 21- inorganic salts as catalysts in the reduction was acetylamino - 901 - fluoro - ll/3,17a - dihydroxy- unnecessary, and that the same reduction product 1,4-pregnadiene-3,20-dionej and for I1 that of 21- was formed when zinc-acetic acid-acetic anhydride acetylamino - 901 - fluoro - 11/3,1Sa- dihydroxy- was used alone. No unaltered oxime X was detected chromatographically nor were other reduc4-pregnene-3,20-dione. tion products found. Proton nuclear magnetic resonarice spectra supElimination of the acetic anhydride froiii the port in detail the suggested structures. Three lion-equivalent methyl group resonances are pres- reduction system resulted in the formation of an cnt in the spectra of both amides I1 and IV in amine, 2 1-amino-9a-fluoro-1I@,17a-dihydroxy-4pyridine. The two high-field resonances are as- pregnene-3,20-dione(XI), isolated as a hydrosociated with the angular CIS- and (219-methyl chloride. This synthesis marks the first reported groups. The third resonance is associated with a successful attempt a t preparation of 21-amino-21methyl group adjacent to a carbonyl group, thus an deoxy analogs in the corticosteroid series. Acetylation of the 21-amine XI with acetic acetyl group.13 Spectra in pyridine diluted with anhydride and pyridine gave the 2 1-acetylamino deuteriochloroform gave additional information. The three high-field methyl resonances, vinyl pro- derivative I1 identical with the amide samples preton resonances (A-ring) and a characteristic niulti- viously prepared. Acylation with phthalic anplet in the methylene region (for 11, 266 c.p.s. hydride and pyridine gave the 21-phthaliinido derivative %-fluoro- l 1/3,17a-dihydroxy-2l-phtlial(12) T h e intermediate 20a-dihydro amide V I was not isolated. 4 imido-4-pregnene-3,'O-dione ( X I ) previously prcsimilar 20-ketone reductiou of 21-nitrogenous 20-ketones has h e m pared in these laboratories by another route.'? reported, R . A. Micheli a n d C . K. Bradsher, J . A m . Chcm. Soc., 77, Fernientation of the 21-amine XI hydrochloridc ,4788 (1Q55). (13) Spectra run on pyridine solutions of t h e amides were recorded with S. roseochromogenus gave a single component, with a 40 mc. applied field, all frequencies being measured relative t o the amide 11, with no detectable amine remaining. benzene a s zero. Spectra of t h e amide I1 are: 243 (CIS-methyl), 229 (Clo-methyl), 212 c.p.s. (acetyl methyl); for amide IV: 235 (CIS- This result establishes yet another bioconversion of S. roseochromogenus, that of acetylation of primethyl), 223 (Cle-methyl), 209 C.P.S. (acetyl methyl). For 901fluorohydrocortisone 21-acetate: 243 (Cls-methyl), 229 (Cia-methyl), mary amines (microbiological acetylation of amines 215 C.P.S. (21-0-acetate); for 9e-fluoroprednisolone 21-acetate: 237 is a known process13s1gj.Slthough this finding (Cia-methyl), 223 (C19-methyl), 209 C.P.S. (21-0-acetate). i
Although pyridine has been used a s a solvent in nuclear magnetic rcsonance spectral studies," its use in this instance obscured t h e vinyl proton region and made interpretation of t h e methylene region unreliable. Dilution of pyridine solutions of the amides I1 and 1V with deuteriochloroform gave solutions whose spectra did ShIJw reliable features in the regions of interest. These spectra were recorded using a 60 mc. applied field, on pyridine solutions of t h e amides diluted with 9
