[CONTRIBUTION FROM
THE
MEDICAL RESEARCH LABORATORIES, CHAS.PFIZER AND Co., INC.,GROTON, CONN. 1
Chemistry and Stereochemistry of 16-Substituted 17,20 ;20,21-Bismethylenedioxy Steroids1 BY R:.T. XORELAND, D. P. CAMERON, R.G. BERG.INI) C.E. ?\I,ZXWELL, I11 RECEIVED J A N T J A R Y 31, 1962 The preparation of isomeric 1~,20;20,21-bismetl1~-lenedio~~ (BML)) derivatives from 16p-broin0-17a,2l-dih~droxypregn4-ene-3,20-dione is reported. Evidence indicating that these represent 20K- and 20s-forms of the B M D group is presented and tentative assignment of the %OK-configuration to the familiar levorotatory B M D species is made. Displacements of 160-bromine from the 20R-isomer are described which provide a. nieans for introducing new 160-substituents.
In the course of a study of corticosteroid strucHydrolysis of either I or I1 in aqueous acetic acid ture-activity relationships, it became desirable to in the presence of a small amount of perchloric find new means for introducing substituents in acid furnished the starting l(iP-bromodio1which was position 16.3 h survey of possible methods sug- converted without isolation to the known 16a,17agested that protection of the side-chain as the epoxy-21-hydroxypregn-4-ene-3,20-dione acetate.$ 17,20;20,21-bismethylenedioxy (BMD) derivative4 3-Ketones K 3-Ketals in a steroid having bromine a t C-16 might provide I -Br Ia a useful intermediate for this purpose. It appeared I11 -H IIIa that if displacements of halogen could be ac1’1 -OH 1% complished a t this stage, a series of new derivatives VI1 -0COCH3 VIIa might become available which could be converted to potentially active corticoids5 in relatively few IS -OCH( CH3)z steps. Accordingly, investigation of the preparation s (20) Xa and properties of 1GP-bromo-17,20;20,21-bis-methylenedioxypregn-4-ene-3-one(I)5 was undertaken. ?,-Ketones R 3-Ketals Treatment of 16P-bromo-17a,21-dihydroxypregn4-ene-3,20-dione with aqueous formaldehyde and hydrochloric acid in chloroform or methylene chloride as originally described&did not give promising results, but when the organic phase was changed to a mixture of benzene and hexane, a product crystallized directly from the reaction mixture in 50% yield. This material had the high melting point, papergram mobility and negative 0 rotation expected4for the desired 16P-bromo-BMD (I). Careful examination of the mother liquors The levorotatory isomer I could be related to revealed a second product which was isolated in previously encountered BMD’s by hydrogenolytic fair yield under slightly modified conditions. This removal of bromine from its 3-ethylene ketal proved to be an isomeric, dextrorotatory product Ia and selective hydrolysis of the resulting halogen(11) believed to be an example of the hitherto free ketal I I I a with aqueous methanolic sulfuric acid. The product of this hydrolysis (111) was undetected 20-epimeric BMD’s. identical with the B M D prepared directly from 17a,2 l-dihydroxypregn-4-ene-3,20-di0ne.~~~~~ In a (1) Portions of this work have been reported in earlier communications, W.T. Moreland, R . G . Berg and D. P. Cameion, J . A m . Chem. similar sequence, the dextrorotatory isomer I1 SOL., 82, 504 (1960); W. T. Moreland, R. G. Berg, D . P. Cameron, was converted to the ketal IIa and then to a C . E. IhIaxwell, 111, J. S. Buckley and G. D. Laubach, Cheinislvy b bromine-free ketal IVa. In this series, however, Indzislvy, 1084 (1960). selective hydrolysis of the ketal was not realized, ( 2 ) (a) Background literature is discussed in I.. F. Fieser and A f . Fieser, “Steroids,” Reinhold Publishing Corp., l i e w York, N. Y . , and IVa was completely hydrolyzed to 17a,211959, p p , 682-701. (b) More recent summaries of corticoid structuredihydroxypregn-4-ene-3,20-dione under conditions E . Dulin, F. I,. Schmidt and S. C . Lyster, Pmc. activity include which smoothly transformed I I I a into 111. A SOC.Erptl. Bid. M i d . , 104, 343 ( l 9 f i O ) ; I. Ringler, S. Mauer and E. comparison of molecular rotations for the two Heyder, i b i d , 107,4 5 1 (1961). ( 3 ) Since completion of this work there have been additional reports series of isomers is shown in Table I. o f : (a) 16p-duorine, D . E. Ayer and W.P. Schneider, J . A m . Chem. Halogen-free displacement products were readily Soc., 82, 1249 (1960); (b) 16a-fluorine, B. J. Magerlein, R . D. Birkenobtained from the levorotatory isomer I provided meyer and F. Kagan, ibid.. 8 2 , 1262 (1960); (c) 16a-methoxy, S.Bernelectrophilic assistance from silver ions was furstein, M Heller and S. M. Stolar, Chemislvy & Indtrslvy, 516 (1961). (4) R . E. Beyler, R. 11. Xoriarty, F. Hoffman and L. H . Sarett, nished. Although brief warming of I in methanolic J . A m . Chem. SOL.,8 0 , 1.517 (1938); J . Ovg. Chrin., 26, 2421 (1961). silver nitrate was sufficient to precipitate silver ( 3 ) I n spite of the essentially equivalent activities of 16a- and 160bromide, examination of the steroidal product methyl derivatives,%i t was apparent from the deleterious effects of suggested a mixture of nitrate ester and methoxy 160-chlorines and 160-acetoxyl7 thLt configuration of t h e new substituents might be of considerable importance and t h a t it would be compound which proved impractical to separate. most desirable t o obtain members in t h e 16a-series. Substitution of silver perchlorate, however, avoided
u
f0) R . 13. Beyler and I? Hoffman, J . Org. Chem., 21, 572 ,(I!X6). ( 7 ) S Bernstein, M. Heller and S . h l . Stolar, J . A m . C h e m S o r . ,
81, 1256 (19a9). (8) Komenclature where possible conforms with the “IUPAC Definitive Rules for Nomenclature of Steroids,” i b i i l . , 8 2 , 3377 (1960).
( 9 ) P. I, Julian, E. \%’RIeyer, \T. J . Karpel and I. R . XYaller, i b i d . , 72, 314.7 (19.50). ( I O ) R . E . Beyler and L . H . Sitrett, U. S . Patent 2,888,457 ( M a y 26, 1E l 9 ) .
metric center a t C-204and should lead to the epimeric 20R- and 20S-f0rms.~~Models show that the environment of the ring which incorporates C-IT and C-20 should be little changed in passing from one isomer to the other. The terminal ring TABLE I bridging C-20 and (2-21 determines isomerism and ISOMERIC B M D DERIVATIVES is so located that its plane is roughly parallel with Levorotatory series Dextrorotatory series the axial 13(18)-bond. Rotation a t C-20 therefore Substituents ComCom316pound MD pound MD brings 20-oxygen in the 20R-isomer or the C-21 Ketone Br I -210 I1 +464 methylene group in the 2OS-form in close proximity Ketal Br Ia -471 IIa f313 to C-18. The interactions occurring between 18Ketone H I11 - 74 .. ... and 21-hydrogens in the latter case appear to be Ketal H IIIa -512 IV-3. + 38 exceptionally severe. Contrasting relationships with the smaller 1GP-hydrogen on the other side refluxing acetic acid gave the 16P-acetoxy compound appear less important. VI1 and silver fluoride in 2-propanol furnished the In the previously reported examples of BMD 16P-fluoro-BMD VI11 accompanied by a small formation only one isomer has been encountered amount of the isopropoxy derivative IX. By con- and these have all been characterized by a large trast, the dextrorotatory isomer I1 was completely negative shift in molecular rotation relative to the unreactive under similar conditions. When treat- original ~ t e r o i d . ~ The major product (I) of BMD ment sufficiently vigorous to remove bromine was formation in the presence of 16P-bromine was shown used, there was disruption of the molecule to un- by direct chemical means to belong to the familiar recognizable products. levorotatory series, but in addition it was possible The products V, VI and VI1 were shown to have identical stereochemistry by interconversion; i.e., to isolate a second, dextrorotatory isomer 11. Comacetylation of the alcohol VI furnished the acetate parison of the latter material and the products derived from it with counterparts in the isomeric VII, while methylation’l converted VI to the methoxy compound V. The fluoro and isopropoxy series (see Table I) shows the substantial difference analogs VIII and I X were assumed to have the in optical rotation which is the most obvious distinguishing feature. Other physical properties same configuration. l 2 Evidence for the orientation such as high melting points, solubility characterisa t C-16 was first sought by lithium aluminum tics, and infrared absorption16 which have been hydride reduction of the protected 16-ketone X a previously described4 for BMD’s are grossly simi(prepared by oxidation of the alcohol VI and selective ketal formation a t the 3-ketone) which would lar. These relationships and the fact that chemibe expected to give the more hindered 166-alcoh01.l~ cal evidence limits isomerism to the side-chain make i t clear that the bromo compounds I1 and Acetylation of the product from this reduction IIa and the halogen-free product IVa are members gave material identical with the 3-ketal VIIa of a new series of dextrorotatory BMD derivatives prepared from the silver acetate product VII. The 16/3-configuration was therefore indicated for inverted a t C-20 with respect to the familiar levothe substituents introduced in the displacement rotatory products. The striking differences in reaction. Subsequent conversion of the acetate molecular rotation are comparable to rotational ascribed to an analogous spiroketal inVI1 to the known 16P,17a,21-trihydroxypregn-4-changes ene-3,20-dione 1G,21-diacetate14 confirmed this version a t C-22 in sapogenitis.ls Evidence for the absolute configuration a t C-20 assignment . in the BMD’s is provided by the following obserRemoval of the BMD protection with formic acid4 and subsequent treatment with meth- vation. Infrared absorption of the 16P-hydroxyanolic sulfuric acid to cleave formates furnished IGP- B l l D VI in 0.0026 Ai carbon tetrachloride solution methoxy- and 1G,R-fluoro-17,21-dihydroxypregn-4-shows the single sharp band a t 3620 cm.-’ typical for an unbonded secondary alcohol. l 9 Measureene-3,20-dione, respectively, from V and VIII. ments from models of VI in the 20s-form indicate a The Ag(l1)-ana1ogof I was prepared similarly from 16P-bromo-17a,2l-dihydroxypregna-4,9( 11) - d‘iene(15) T h e system of R and S nomenclature [R.S. Cahn, C . K. 3,20-dione and is also indicated to be in the levoro- Ingold and V. Prelog, Expevienlia, 12, 81 (1956)l is preferred over t h e tatory series. Treatment of this product with silver more cumbersome c q and PF usage which can be applied a s follows. perchlorate in methanol furnished the expected Bridging to l7u-oxygen fixes one of t h e 20-oxygens in t h e u-configurain either isomer. T h e remaining oxygen a t C-20 can then be Ag(ll)-16P-methoxy-BhlD which could be converted tion unambiguously designated from Fischer projections referred t o termias above t o 16/3-methoxy-17a-21-dihydroxypregna-nal-ring-opened forms as proposed in “Tentative Recommendations of t h e Steroid Somenclsture Suh-Committee.” I C P A C Informalion 4,9(1l)-diene-3,20-dione. Bullelin S o . 21, Munich, 1959, p. 50. I n this system, the 20R- and Discussion 20s-forms become, respectively, 1 7 a , 2 0 a ; 2 0 a ~ , 2 1and - 17u,20a;Z0PF,21Formation of BMD derivatives from 17a,21- bismethylenedioxy. (16) T h e infrared spectra in both series are characterized by a numdihydroxy-20-keto steroids introduces a new asym-
such mixtures and this salt in refluxing methanol or aqueous acetone smoothly converted I to, respectively, the 16P-methoxy-BMD V or the 16Phydroxy analog VI. Similarly, silver acetate in
(11) >Neeman, ‘I. hf. C . Caserio, J. D. Roberts and W. S. Johnson.
Telrahedron, 6 , 36 (1959). (12) This was subsequently verified for t h e Auoro compound VI11 by t h e similarity of properties and activities of derived products’ t o l6,5’-fluoro derivatives prepared in another way.@ (13) Reference 2a, p. 268; cf. S. Bernstein, M. Heller and S . M. Stolar, J . A m . Chem. Soc., 77, 5327 (1955). (14) K. Heusler and A . Wettstein, C h e m . B e y . , 87, 1301 (1954).
ber of strong bands in t h e C - 0 stretching region between 9.0 and 11.3 Y. It is of interest t h a t the sapogenins, which similarly contain a spiroketal function, also show distinctive absorption in this region.17 (17) C. R, Eddy, &I. E. Wall and h I K . Scott, Anal. Ckem., 26, 266 (19,53); R. N. Jones, E. Katzenellenbogen and K . Dobriner, J . A m . C h e m . Soc., 76, 168 (19.53). (IS) M. E. Wall, Erpevienl;a, 11, 340 (1953); ref. 2a, pp. 826 ff (19) S. A. Barker, J. S. Brimacombe, A . B. Foster, D. H. Whiffen and G. Zweifel. Tetrahedron, 7 , 10 (1959).
minimum distance of approximately 1 k . between hydrogen of the 16P-hydroxyl and 20-oxygen in the terminal ring. As these groups appc:ti- itleally
and hydrolysis mediated to an unknown degree by solubility effects in the two-phase system. The familiar Icvorotatory (20R)-BMD's have bcen characterized by remarkable stability to hydrolysis by mineral acids4 whereas, if the single 1G-unsubstituted dextrorotatory example (IVa) is representative of the series, the 20s-isomers are readily susceptible to attack by acids.23 Thus, the usual predominance of levorotatory isomer in BMD formation could be the result of greater stability under the strongly acidic conditions of preparation. This difference in ease of hydrolysis is compatible J with the assignment of 20R-configuration (having less severe strain near C-18) to the more resistant 20 R 20 form. I n an extension of this, i t is reasonable that situated for intramolecular hydrogen bonding20 VI bromine a t C-1G should impose more strain in the would be expected to show strongly bonded 20R-isomer in which i t is near c-21 than it would hydroxyl absorption were it in the 20S-series. in the 20s-isomer where i t is near 20-oxygen. Consideration of VI in its 20R-fol-mshows the oxy- The additional stress introduced by the halogen gen nearest 16p-hydroxyl to be that attached at might therefore act to reduce the net difference C-21. Although the corresponding distance f s r between isomers and permit isolation of both under bonding in this isomer appears to be less than 2 A., the altered conditions found t o be favorable in this the model suggests effective shielding between the series. The Present data, however, do not disether oxygen and the 1G-alcohol by c-21 methyl- tinguish between possible effects of the organicene. The observed absence of intramolecular phase solvent and factors concerned with the ease hydrogen bonding in VI therefore Seems to be Of acid-catalyzed formation or cleavage SO that accommodated only in the 20R-fom ; accordingly, precise evaluation of reasons for the isolation Of one VI must be 1GB-hydroxy,1701,20;20,21-bismethylor both isomers is not possible a t this time. enedioxy-(20R)-pregn-4-ene-3-one.It follows from Acknowledgments.-We are indebted to Dr. the interconversions represented by the products in M. W. Miller for assistance with certain of the reTable I that levorotatory BMD's in general must actions, to Mr. 117. H. RlcMullen for the hydrogen have the 20R-configuration and, by difference, that bonding spectrum, and to Drs. J. S.Buckley and G. D. Laubach for their advice and encouragement. the dextrorotatory isomers are 20s. It had been anticipated that displacement prodE~perirnental?~ ucts realized from the lGp-brom0-(20R)-BMD I would result from approach of the entering group 16p-Bromo-l7~~,2 l-dihydroxypregn-4-ene-3,20-dione .Solution Of 58.0 g. (dry basis) Of 16p-brom0-17a2l-dito the relatively unhindered Cu-faceand would have A hydroxypregn-4-ene-3,20-dione21-acetate9 in 552 ml. of the inverted (lGa) configuration.21 The chloroform and 2.00 1. of methanol was stirred a t room temshow, however, that the displacement proceeded perature with a solution of 145 i d . of concentrated (37%;) exclusively with net retention of configuration and hydrochloric acid and 224 ml. of water for 24 hours. The organic solvents were removed under reduced pressure a t an without rearrangement. This stereochemical internal temperature below 30" with concurrent portionwise participation by One of the ether oxygens, addition of 880 ml. of water. T h e resulting precipitate was presumably that attached a t C-17because it is the filtered and washed with water to yield (dry basis) 42.9 g. nearest even though it does not appear to be very (8176) of the bromoalcohol. Attempts to recrystallize this invariably resulted in decomposition and loss of favorably situated. ~h~ unreactive nature of 1 6 ~ - compound bromine. A carefully dried sample of the material was bromine in the 2OS-isomer 11 suggests that this prepared as (lescrihed; 1n.p. 128' tlec., [cY]"D (c 3 ) ; participation occurs as shielding on the a-side which CI:;x;1;2 2.81, 2.88, 5.72 p . determines the configuration of the products but Anal. Calcd. for C2,HZ9O4Br: C, 59.29; H, 6.87; Br, does not materially assist in the removal of bro- 18.179. Found: C, 59.33; H , G.74; Br, 18.99. mine.22 The greater reactivity of bromine in the 16P-Bromo-17~~,20;20,2 1-bismethylenedioxypregn-4-ene- 0 1 ~ .Levorotatory (20R)-Isomer (1).-.4 slurry of 10.0 g. 20R-isomer I may be the result of steric accelera- 3(dry weight) of moist bromohydrin from the previous step tion22 due to extraordinary crowding by the in 96 ml. of benzene, 96 ml. of 37% aqueous formaldehyde, nearby c-21methylene group. The inert charac- 96 ml. of concentrated hydrochloric acid and 27 ml. of ter of bromine in the 20S-form 11 must then be a water was stirred at 37" for 3 hours. Hexane (500 ml.) was added dropwise over the next 3 hours and the resulting consequence of exchanging this methylene group for mixture stirred 18 hours a t 37". The crude product, isolated oxygen. The now-proximal 20-ether oxygen, be- from the chilled mixture by filtration, was washed thorcause of smaller size and/or altered electronic oughly with water and recrystallized from aqueous dimethylcharacter, apparently cannot aid in halogen re- formainitle to give 5.43 g. (50%) of the 16p-bromo-(20R)BMD ( I ) , m.p. 205" dec. Recrystallization from methanol moval from this neopentyl-like position. The conditions for the BMD-forming reaction (23) Compare the complete hydrolysis of IVa after a 90 min. reflux are such that the ratio of isomers might be expected in co. 0.2 N methanolic sulfuric acid with the stability reported' fur the levorotatory isomers in ca. 1 N acid after 11 hours. to be the net result of relative rates of formation (24) Unless otherwise specified, rotations are in dioxane, ultraviolet
s
(20) Cf. L. P. Kuhn, J . A m . Chem. Soc., 74, 2492 (1952). (21) Ref. 2a, p. 14; cf. G. S. Hammond, M . F. Hawthorne, J. XI. Waters and n. M. Grnybill, J. A m . Chem. Soc., 83, 704 ( I O l j O ) , and rrlrrrnrcs therein for