STUDIES OF VITAMIN-D (CALCIFEROL) AND ITS ANALOGS .42. 3-DEOXY-3-THIA-1-ALPHA,25-DIHYDROXYVITAMIN-D3 AND ITS 1-BETA-EPIMER - SYNTHESIS AND BIOLOGICAL EVALUATION

The syntheses of 3-deoxy-3-thia-1-alpha,25-dihydroxyvitamin D3 (5a) and its 1-beta-epimer 5b have been achieved starting from CD-fragment 10 and the enantiomerically pure A-rings 11a and 11b prepared from thia 1,3-diketone 13. For the preparation of 11a and 11b, the thia diketone 13 was converted in two steps to the trimethylsilyl enynone 18. In the most effective route, the latter was asymmetrically reduced using (R)- or (S)-oxazaborolidine 24 and 25 and catecholborane to afford 19a (87% ee) and 19b (85% ee), respectively. Final enrichment to 11a and 11b was achieved via their crystalline carbamates 23a and 22b and then hydrolysis to the enantiomerically pure enynols 12a and 12b and silylated to the desired 11a and 11b. The absolute configuration at C-1 of the A-ring was examined in detail through four empirical correlation methods. The elution order of carbamates 22 and 23 by HPLC correlates with Pirkle's empirical rules. The specific rotations of enynols (-)-12a and (+)-12b correlate with Mills' rules for absolute configurations of chiral cyclohexenols. The enantiomeric sense of asymmetric reduction using the (R)- or (S)-oxazaborolidines and a complementary asymmetric reduction using LiAlH4 and N-methylephedrine was exactly that predicted. The C-1 absolute configuration of (S,S)-carbamate 22b was in fact fully confirmed by a single-crystal X-ray crystallographic study. Hence, the four empirically derived methods for establishing absolute configuration of the various enynols were mutually consistent. The analogues 5a and 5b were submitted for biological evaluation of their relative ability, in relation to the reference 1-alpha,25-dihydroxyvitamin D3 (3, 1-alpha,25-(OH)2-D3), to stimulate intestinal calcium absorption (ICA) and bone calcium mobilization (BCM) under in vivo conditions. Analogue 5a was 20% and <10% while analogue 5b was <20% and <10% as active as a 100 pmol dose of 1-alpha,25-(OH)2-D3 for ICA and BCM, respectively. In addition, 5a and 5b were 14.5 +/- 5.7% and 1.23 +/- 0.38%, respectively, as effective as 1-alpha,25-(OH)2-D3 in binding, under in vitro conditions, to a chick intestinal nuclear receptor.