Totally synthetic steroid heterocycles. Part 2. Stereochemistry of hydride reduction of 16-oxa- and 16-thia-8,14-didehydro-D-homoestrone 3-methyl ether and related compounds

摘要

1252 J.C.S. Perkin I Totally Synthetic Steroid Heterocycles. Part 2.l Stereochemistry of Hydride Reduction of 16-Oxa- and 16-Thia-8,14-didehydro-~-homoestrone 3-Methyl Ether and Related Compounds By Tadao Terasawa and Toshihiko Okada, Shionogi Research Laboratory, Shionogi and Co., Ltd., Fukushima- ku, Osaka 553, Japan The stereochemistry of carbonyl reduction of 16-oxa- and 16-thia-8,14-didehydro-~-homoestronederivatives (1)-(3) was studied using various hydrides. Some related compounds (4)-(8) were also examined for com- parison. The observed stereochemical results are tentatively interpreted in terms of control of steric approach to the ring carbonyls in different conformations by the ring heteroatoms involved. Ring deformation caused by intro- duction of a sulphur atom often reverses the steric course of carbonyl reduction in steroids.STERICmodifications of a carbocyclic framework by (la) is inconsistent with the generalization that insertion of heteroatoms may influence various reactions steroidal 17(or 17a)-ketones usually undergo nucleo- commonly used for carbocycles. In Part 1,l we des- philic attack preferentially from the less hindered or-side. cribed the synthesis of the 16-0~~~-16-thia-~-and homoestrogens (1)-(3). Lithium aluminium hydride (1) R=Me (2)R =Et a; X=S b; X=O (4)R=H,Y=S (5) R =CH3,Y =O Me (6) R = H (7) R =Et (8) (LAH) reduction of heterocyclic ketones of type (1) demonstrated that there were substantial changes in stereochemistry upon introduction of a ring heteroatom.We observed that hydride reduction of the thia-ketone (la) gave the l7aa-01 preferentially, while the oxa-ketone (lb) afforded the 17ap-01 exclusively. The steric course observed for the hydride reduction of the thia-compound In order to explore this observation, we decided to examine in more detail the carbonyl reduction of the heterocyclic ketones (1)-(3) and that of the related compounds (4)-(8). Various hydrides other than LAH were also used under standard conditions. The stereoselectivity for reduction seemed to be unaffected by solvents and temperatures. The results are summarized in Table 1. The preference for formation of the a-01 was observed for various reductions of the thia-compounds (3)-(7), closely similar in structure to compound (la).In contrast, the 18-homologue (2) and the oxa-compound (3b) were almost exclusively reduced to the p-01. The stereoselectivity of reduction of the thia-ketone (la) decreased when LAH was replaced by its alkoxy-derivatives or by sodium borohydride, but a distinct preference for the a-01 was still retained. However, use of sodium bis-(2-methoxyethoxy)aluminium hydride demonstrated a profound effect, eventually reversing the product ratio in favour of the p-01. Furthermore, this reagent improved the stereoselectivity favouring the p-01 in the case of the 18-homologue (2a). It seemed that the unusual behaviour of hydride reduction might be limited to unsaturated thiacyclic ketones of types (1)-(7) but we have found recently that compound (8) also gives the corresponding a-01 predominantly. A number of factors may govern the stereochemistry of these reductions. We first supposed that a metal chelate complex properly situated between the ring sulphur atom and the incipient hydroxy-function might play an important role in directing the steric course of carbonyl reduction of the thia-ketones, leading to pre- dominant formation of the a-oriented alcohols. How-ever, this seemed unlikely from the fact that the 18-homologue (2a) is not subject to similar steric control.Alternatively, we thought that there might be electronic interactions3 between the carbonyl group and the sul- Dhur atom. which could affect the stereochemistrv of 1978 1253 oxa-ketone (1b), and the corresponding carbocyclic In agreement with our results, the steric requirement c~mpound.~However, these comparative data gave in the oxa-ketones favours hydride attack from the a-no clear evidence for the existence of a polar interaction.side, leading to the P-oriented equatorial alcohol exclu- Although other possible polar factors should not be sively. On the other hand, the p-side becomes less TABLE1 Stereochemical outcome of hydride reduction H ydride Conditions a-OH :P-OH * Yield () * LAH THF (r.t.) 4.4: 1 92.6 NaBH, MeOH (0-6") 2.5: 1 96.3 LiAlH(ButO), THF (r.t.) 2.7: 1 97.6 LiAlH (MeO), THF (0-5") 2.0: 1 86.0 NaAlH,(OCH,CH,OCH,), PhH (r.t.) 1:2.9 87.9 LAH THF (r.t.) 1:6.1 92.1 NaBH, EtOH-THF (r.t.) 1:5.2 97.3 NaAlH,(OCH,CH,OCH,) , PhH (r.t.) 1 :19.5 89.8 LAH THF (r.t.) Ponly 70.5 LAH THF (r.t.) Ponly 89.7 LAH THF (r.t.) 1.4: 1 88.0 LAH Et,O-THF (r.t.) Ponly 70.5 NaBH, EtOH-THF (r.t.) 2.3: 1 85.9 NaBH, EtOH-THF (r.t.) 2.2: 1 70.5 LAH THF (r.t.) 1.4: 1 88.0 NaBH, MeOH (r.t.) 1.9: 1 98.3 LAH THF (r.t.) 1.9: 1 45.0 LAH THF (r.t.) 3.2: 1 87.4 NaAlH,(OCH,CH,OCH,) , PhH-THF (r.t.) 1.3: 1 96.7 r.t.=Room temperature. *Isolated. excluded, steric control of approach of the hydride hindered with increasing 4by 35" in the case of the thia- species is probably of major importance. Inspection of relative to the oxa-ketones, despite the presence of the Dreiding models shows that introduction of a sulphur TABLE2 Spectral data for alcohols and acetates from hydride reduction Alcohol Starting V(CC14)/ Acetate ketone M.p.("C) cm-ld M.p. ("C) a(CDC1,) * (J/Hz) (la) aa 140-142 3 553 136-137 4.98 ,I (t, J 3)B" 140-141 3632 162.5-164 5.06(a,J 5, 10.5)(lb) 8" 140.5-142 3633 161-164 5.00. .. (9,J:;:O)(2a) aa 112.5-114 3 554 Oil.. (t, J 3)Pa 100-101.5 3 631 103-105 5.12 (q, J 6, 10.5)5.07(2b) 8" 105.5-107 3 632 104-106. .. 66;')J5(9J Conformations and dihedral angles about the C( 13)-C(17a) (3a) ab 175-177 3 550 169.5-171 bond for compounds (1)-(7) J5 2i2)B* 216-218 3 630 210.5-213 (98 (a,J 4.5, 10.5) 5.06relative to an oxygen atom deforms the six-membered (3b) Pb 194-196 3632 183-185 (a, J 5, 10.5)4.95brring to a great extent, due to variation of bond lengths (4) ac 127-128 3552 116-118and angles.* The deformation alters the ring geometry (tsJ 3)5.03so as to bend the carbonyl oxygen down.Thus, a pc 94-97 3630 115-117 (a, J 5, 10.5)5.09brsignificant difference in spatial environment around the (5) ac 127-130 3 548 121-123carbonyl function is brought about between the oxa- (t, J 4)8" 154-157 3 630 187-189 5.13and thia-rings. This can be represented by the dihedral (q, J 4, 11) 4.78 (tpJ 3)4.89 angles (4) about the C(13)-C(17a) bond, which are (6) ac 109-110 3552 estimated from the Dreiding models as illustrated in the P" 129-130.5 3631 (7) ac Oil 3 548 Figure. PC Oil 3 631 *C-X bond lengths are often appreciably different from C-C (8) ac 137-139 3 528 159-161 distances: C-0 1.82 A.The variation in 1.43, C-C 1.54, C-S bond angles is less important for oxygen but the C-S-C angle of Be 142-144 3631 180-181.5 ca. 100" is substantially ,different from the near-tetrahedral (99angles found in other heterocyclic compounds. Ref. 1. This work. J 6, 9) T. Terasawa and T. Okada, to be T. Terasawa and M. Takasuka, unpublished data. published. OH band. CHOAc signal. 13-alkyl substituent. In this situation, hydride attack produces predominantly the a-oriented axial alcohol. An increase in the steric bulk of the reducing agent as well as the angular substituent may influence the course of reduction, shielding and restricting p-approach. This is exemplified in the reduction of the thia-ketones (la) and (2a) with various hydrides.The stereo-chemical results for reduction of compound (8) ($ 130") may be similarly interpreted. However, a detailed examination of the data presented in Table 1 suggests that accessibility of hydrides to the carbonyl group from either the a-or the p-side depends on a delicate steric balance. This fact reflects the limitations of the purely steric argument. The configurations of the epimeric alcohols obtained in this study were determined from i.r. (intramolecular hydrogen bonding) and n.m.r. data of the acetates (Table 2). EXPERIMENTAL General details are given in Part 1.l Reduction of 3-amp;Iethoxy-l6-thia-~-homoestra-1,3,5(lo),6,-8,14-hexaen-l7a-one (3a) .-To an ice-cold stirred solution of the thia-ketone (3a) (103.5 mg, 0.33 mmol) in dry tetra- hydrofuran (8 ml), lithium aluminium hydride (19 mg, 0.5 mmol) was added in small portions.Stirring was continued for 30 min at room temperature. The mixture was poured into ice-cold water and extracted with dichloromethane. The usual work-up left a crystalline residue which was purified by preparative t.1.c. (9 : 1 benzene-ethyl acetate with double development), giving the 17aa-01 (less polar) (52.4 mg, 50.8) and the 17aQ-01 (more polar) (38.3 mg, 37.1) ; product ratio 1.4 : 1. Analytical samples were obtained by two crystallizations from dichloromethane-ether or acetone. The 17aa-ol had m.p. 175-177", vlnax. (dilute CCl,) 3 550 cm-l (bonded OH); A,,,.(EtOH) 227, 249, 271.5sh, 280, 311.5, and 321sh nm (E 24300, 20900, 20 300, 22 500, 29 900, and 27 600); G(CDC1,) 1.04 (3 H, s, 13-Me), 3.89 (3H, s, OMe), 6.57br (1 H, s, 15-H), and 7.0- 7.9 (5 H, m, ArH); m/e 312 (amp;I+)(Found: C, 72.85; H, 6.45; S, 10.3. ClgH2002Srequires C, 73.05; H, 6.45; S, 10.25). The acetate was prepared from acetic anhydride- pyridine as a crystalline solid, m.p. 169.5-171" (dichloro-J.C.S. Perkin I methane-ether) ; vmx. (CHCl,) 1 739sh, 1 730 (OAc), 1 623, 1 604, 1582, and 1502 cm-l (aromatic); G(CDC1,) 1.13 (3 H, s, 13-Me), 2.09 (3 H, s, OAc), 3.90 (3 H, s, OMe), 5.08(1H,q,J2and4Hz, 17a-H),6.69br(1HIs, 15-H),and 7.0-7.9 (5 H, m, ArH). The 17ap-ol had m.p. 216-218", v,,,. (dilute CCl,) 3 630 cm-l (free OH) ; ?,m,.(EtOH) 228, 248, 272sh, 279.6, 309, and 317sh nm (E 24 400, 21 900, 22 400, 24 800, 26 600, and 25 000) ; m/e 312 (M+)(Found: C, 72.8; HI 6.4; S, 10.2). The acetate had m.p. 210.5- 213O, vmax. (CHCl,) 1742, 1727 (OAc), 1627, 1600, 1580, and 1 497 cm-l (aromatic) ; A,,, (EtOH) 227.5, 247.5, 270, 278.5, 307.5, and 318 nm (c 23 800, 22 100, 24 000, 26 800, 27 100, and 24 900); G(CDC1,) 1.14 (3 H, s, 13-Me), 2.14 (3H, s, OAc), 3.90 (3 HI s, OMe), 5.14 (1 H, q, J 4.5 and 10.5 Hz, 17a-H), 6.51br (1 H, s, 15-H), and 7.0-7.9 (5 H, m, ArH). Reduction of 3-Methoxy-l6-oxa-~-homoestra-l,3,5(10),6,-8,14-hexaen-17a-one (3b) .-To a cold stirred solution of the oxa-ketone (3b) (90 mg, 0.3 mmol) in dry ether (6 ml) and dry tetrahydrofuran (4 ml), lithium aluminium hydride (12 mg, 0.3 mmol) was added in small portions. The mixture was stirred at room temperature for 10 min, quenched with ice-water, and extracted with 3 : 1 ether-dichloromethane.The usual work-up left a foam which was triturated with dichloromethane-ether, giving the l7ap-01 (46.8 mg) as a crystalline solid, m.p. 193-197". The residue was purified by preparative t.1.c. (4 : 1 benzene-ethyl acetate with triple development) to yield a second crop of the 17ap-ol(15.9 mg), m.p. 195-197" (dichloromethane-ether) ; total yield 70.5. Recrystallization from dichloromethane-ether gave an analytical sample, m.p. 194-196", vmx. (dilute CCl,) 3 632 cm-l (free OH); A,,,. (EtOH) 220, 247sh, 256.5, 265, 282, 292, and 304 nm (E 19 100, 28 300, 39 400, 41 600, 13 800, 22 100, and 19 500); m/e 296 (M+) (Found: C, 76.85; H, 6.8. C,,H2,03 requires C, 77.0; HI6.8). The acetate was prepared from acetic anhydride-pyridine as a crystalline solid, m.p. 183-185" (dichloromethane-ether), v,,,. (CHC1,) 1736 (OAc), 1630, 1601, 1567, and 1507 cm-1 (aromatic) ; A,,,. (EtOH) 220, 248sh, 255.5, 264, 281.5, 291.5, and 303 nm (E 16 800, 32 000, 43 800, 46 700, 14 100, 19 200, and 19 000); G(CDC1,) 1.15 (3 H, s, 13-Me), 2.13 (3 H, s, OAc), 3.89 (3 H, s, OMe), 5.06 (1 H, q, J 5 and 10.5 Hz, 17a-H), 6.94 (1 H, s, 15-H), and 7.0-7.9 (5 H, m, ArH). 7/2039 Received, 21st November, 19771