1978 1529 Hydroxy-steroids. Part 202 Distinction between 1g-Norergosta-5,7,9-trien-3P-01 (Dihydroneoergosterol) and its 3a-Epimer By Gerald Felsky, Peter M. Fredericks, and G. Denis Meakins," Dyson Perrins Laboratory, Oxford University, South Parks Road, Oxford OX1 3QY Two methods have been used for epimerising the 3-hydroxy-group of 19-norergosta-5,7,9-trien-3~-ol. Although the 3p- and 3a-alcohols show a close resemblance in many of their properties it is possible to distinguish between them by examining the 19F n.m.r. spectra of their a-methoxy-a-trifluoromethylphenylacetates and the 'H n.m.r. spectra of their complexes with shift reagents. Purification of the alcohols is most reliably achieved by crystallising their 3,5-dinitrobenzoates. Measurement of optical rotation is the best method for the quantitative analysis of mixtures of the epimers.CONFORMATIONALmobility in ring A of steroids having an R2= CgHI7)have an equatorial hydroxy-group; this aromatic ring B was first demonstrated by i.r. measure- was attributed to their preferences for different half- ments, which showed that both neorgosterol (19-nor- chair conformations.2 The consideration that these ergosta-5,7,9,22-tetraen-3p-ol(1; R1 = H, R2 = CgH17) forms should be ' about equally free of destabilising and epineoergosterol (the 3-epimer) (2; R1 = H, interactions ' led Levine and Ghosh to question the D. H. R. Barton, R. C. Cookson, W. Klyne, and C. W. Part 19, A. D. Boul, R. Macrae, and G. D. Meakins, J. Chem. Shoppee, Chein.and Ind., 1954, 21. SOC.Perkin I, 1974, 1138. S. G. Levine and A. C. Ghosh, Tetrahedron Letters, 1969, 39. early report * that neoergosterol is converted completely into epineoergosterol by treatment with sodium in boiling pentyl alcohol. After developing a sequence (3P-OH --+3p-O-S02Me 3a-OCHO +3a-OH) for preparing epineoergosterol, these authors found that (i) the 3-epimers had almost identical i.r. and lH n.m.r. spectra, and (ii) mixtures of the alcohols did not exhibit m.p. depression and could not be separated chromato- graphically. Assessment of purity was based on rot- ations; from neoergosterol (a, -6.0") the new method J.C.S. Perkin I stereospecificity of the inversion procedure. Since several of the derivatives to be studied were expected to have crucial 1H n.m.r.signals in the region 7 4-5 it was decided to work with 22,23-dihydro-compounds based on dihydroneorgesterol (1; R1 = H, R2= C,H,,) so as to avoid interference from the neoergosterol olefinic signal at T 4.75. (This change of substrate does not appear to influence the problems under investigation.) In the first approach, summarised in the Scheme, di- hydroneoergosterol (la) was epimerised by the method SCHEME The epimeric 3-hydroxy- 19-norergosta-5,7,9-trienesand their derivatives OMeI R' represents (a)H, (b)Ac, (c ,(f) Ph-C-CO--, R2 =CgHlgin all compoundsI (+ 1 CF3 (+I H Rotations were measured using solutions in CHCl, (c 0.5-1.0). Tlie i.r. bands (CHCl, solutions, positions in cm-1) are the two strongest absorptions of each compound associated with C(3)-0 stretching. N.m.r.spectra were examined using solutions in CDCl,: the positions of the 3-H signals are T values, and those of the IsF signals are p.p.m. downfield from external CF,C02H M.p. Epimeric pairs M.p.("C) ah 1.r. 3-H '9F i, ii, iii ("c) La1D 1.r. 3-H 'OF 146-147 -2" 1050, 1029 5.89 (2a) b 165-166 +49" 1088, 1045 6.90 120-121 +2 1040, 1 032 4.88 120-122 +46 1208, 1015 4.90 121-122 -8 1120, 1029 4.57 124-126 $50 1117, 1026 4.55 220-222 -4 1175, 1041 4.51 (2d) 165-167 +46 1172, 1042 4.47 V 115-117 +17 1171, 1055 4.82 120-122 +48 1171, 1053 4.80 108-109 +19 1171, 1022 4.57 3.75 (If) 98-100 +67 1171, 1022 4.54 3.95 Reagents : i, Me.S0,C1-C,H5N : ii, NEt,+ HC0,--Me,CO ; iii, NaHCO, ; iv, Ac,O-C,H,N ; v, Appropriate acid chloride-C,H,N ; vi, EtO,C*N=N*CO,Et-PPh,-Ph*CO,H: vii, KOH-MeOH.* The product of this reaction had m.p. 126-128 "C, a=+49". Ref. 7: (la),m.p. 146-148 "C, a10 -3""; (lb), m.p. 120-121 "C, au+lo. * Ref. 4: (h),m.p. 167 "C; (2b), m.p. 83 "C. C G. A. D. Haslewood and E. Roe, J. Chem. Soc., 1935, 465: (Id),m.p. 216-218 "C. gave epineoergosterol with a, +50.7", and the various used with neoergosterol, and corresponding series of literature products (a,values no higher than +27") were derivatives were prepared from the original alcohol and concluded to be mixtures of epimers. the product. Compounds (la-f) and (2a-f) were Although the inversion sequence is expected to be obtained in crystalline form and their properties were stereospecific this feature is not rigorously established. examined ; repeated recrystallisation of the samples had It is conceivable that in the displacement of the 38- little effect on the properties, and at no stage did t.1.c.methylsulphonyloxy-group there is some SN1 contri-reveal evidence of inhomogeneity in any of the samples. bution or participation by the neighbouring aromatic The i.r. spectra of solutions of epimeric pairs are similar ring, and the presence in the final product of a certain but not identical; although there are distinct differences amount of the 3P-alcoho1, however formed, would not in the bands associated with the C(3)-0 stretching modes have been revealed by the previous st~dy.~ The main of the alcohols (la) and (2a) these differences are not objects of the present investigation were, therefore, to sufficient to exclude the possibility that the 3a-alcohol find other criteria for distinguishing between such similar epimeric alcohols, and to use these in assessing the A.Windaus and M. Deppe, Chem. Bey., 1937, 70, 76. (2a) contains a small amount (say, up to 5) of the start- ing material (la). Even greater similarity was found in the 1H n.m.r. spectra of epimeric pairs, including those (le) and (2e), (If) and (Zf) in which it had been hoped that esterification with a chiral acid would produce appreciable differences in the steroidal 3-H signals. The 19F n.m. r. signals of the met hoxytrifluoromet hylphenyl- acetates, (If) and (2f), provided the first clear spectro- metric distinction between the series.These signals are sufficiently well separated that contamination of the 3a-ester (2f) by ca. 2 or more of the 3p-epimer (If) would have been detected. With five of the pairs in the Scheme, mixtures of epimers had melting points between those of the compo- nents, and only the 3,5--dinitrobenzoates, (Id) and (2d) 1,531 at the head of the Table.) Here, the intention was to differentiate between the signals of specific protons (say the 3-H) in the diastereoisomeric alcohols (la)and (2a) by studying the spectra of their complexes. Since chiral reagents are known to induce a difference between the chemical shifts of enantiomeric protons and achiral reagents to increase the difference between those of diastereoisomeric protons, a reagent of each type was employed.The most complete analysis emerged with the system 3p-alcohol (la) plus Eu(facam), * where the resonance of each ring A proton was identified as a separate signal. Elsewhere difficulties of assignment arose from the over- lapping and crossing over of some signals caused by varying rates of shift (with respect to concentration). The effect of shift reagents on the 1Hn.m.r. signals of the epimeric alcohols (la) and (2a) S = steroid, I, = shift reagent, and p = L/SJ. For each system portions of S were added to a solution of L in CDC1, to give 10 solutions in the range p = 0.1-0.5; a selection of the results obtained by examining these solutions at 90 MHz is shown below.The descriptions low and high refer to the field strength at which the signals of a CH, group occur; the positions (in Hz downfield from SiMe,) of a signal in the presence and absence of L are denoted by Vah and vo respectively, and (Vah-Vo) = Av. Extrapolation to p = 0 of vah versus p plots (which were linear up to p = ca. 0.30) gave the v, values. Plots of S verszu l/Av were also linear; the gradients of these plots are L.vb,where Vh represents the bound shifts (in Hz) of particular protons 3p-Alcohol (la) in 0.13~-Eu(fod), * 1-euro;1 2-H 3-H 4-H P Low High Low High Low 584 417 673 474 777 524 0.23 426 364 462 839 568 0.26 vh8::: 453 990' Sashy; 384 517 590 Alcohol (2a) in O.lO 910 2130 ~-Eu(fod), 613 1475 562 394 623 437 0.17 vsh 679 477 0.24 788 555 0.27 832 586 Vt) 1870 1100 .Average v0 values (both alcohols) 247 t 240 238t 238: 364 274 High P 372 0.11 415 0.17 463 0.22 491 0.28 526 0.34 1320 3p-Alcohol (la) in O.lO~-Eu(facam), t l-H 2-H 3-H 4-H Low High Low High Low High '310 292 575 403 364 684 470 418 vat 333 409 377 773 530 463 382 350 451 410 836 573 500 410 362 483 451 881 605 544 vh '670: 580 860: 780: 2 120 1250 1190 3a-Alcohol (2a) in 0.13~-Eu(facam), 355 383 409 0.14::::vah{ 312 326 607 684 747 418 472 513 377 415 444 460 0.29 341 801 549 470 481 0.33 351 835 570 489 1050 vb 410 1770 1000 820 250 t Tris (trifluoroacetyl-d-camphorato)europiuni-* Tris(6,6,7,7,8,8,8-11eptafluoro-2,2-dimethyl-3,5-octanedionato)europium(111). (111).:Values less precise than the others in the same set. (mixed m.p. 154-161 "C), behaved normally. (The reluctance of 3,5-dinitrobenzoates to form mixed crystals has long been recognised in the traditional use of these derivatives for separating sterol mixtures.) Crystallis-ation was effective in purifying samples of the esters (Id) and (2d) to which 10 of the other isomer had been added; hydrolysis of the esters gave the parent alcohols (la) and (2a) with unchanged characteristics, thus establishing the purity of the 3a-alcohol (2a) produced in the epimerisation. The more direct method of inver- sion using diethyl azodicarboxylate gave the 3a-benzoate (2c) stereoselectively, and this is a more efficient route to compounds of the 3a-series.A different approach, involving n.m.r. shift reagents, is shown in the Table. (The standard techniques used in obtaining and processing the data are described briefly * See footnote to Table. A. K. Bose, B. Lal, W. A. Hoffman, and M-S. Manhas, Tetrahedron Letters, 1973, 18,1619. With both reagents the epimeric alcohols gave different values for the bound shifts (vb) of their 3-H signals. As expected from this result mixtures of the alcohols in a solution of either reagent give distinguishable 3-H signals ; however, these are so broad and closely overlapped that their use for estimating 3p : 3ct ratios leads to inaccurate values.The sharper 4-H signals of the Eu(facam), complexes are much more suitable for quantitative work, and with solutions of known composition the epimer ratios found by integration (at g values of ca. 0.26) were accurate to amp;5. The present work establishes the stereochemical homo- geneity of the 3a-alcohol (2a) having the constants reported in the Scheme and, almost certainly, of the epi- neoergosterol described by Levine and Ghosh., There is a close correspondence between the a, values of the G. A. Webb, ' Annual Reports on N.m.r. Spectroscopy,' Vol. GA, Academic Press, 1975; ' Nuclear Magnetic Resonance Shift Reagents,' ed. R. E. Sievers, Academic Press, London, 1973. 1532 J.C.S.Perkin I 3p-and 3a-alcohols in the two series (see earlier) and a large difference between those of each 3p/3a pair; hence, once reference data are available, optical rotation measurement is the most accurate and convenient method for analysing mixtures of such similar diastereo- isomeric alcohols. EXPERIMENTAL General directions were as described in J. Chem. Soc.(C), 1968, 2674 except that routine lH n.m.r, spectra were recorded at 100 MHz. Petrol refers to light petroleum, b.p. 60-80 "C. The constants and main spectrometric properties of compounds are shown in the Scheme and are repeated here only when required by the context. Epimerisation of 19-Norergosta-5,7,9-trien-3~-01(Dilzydro-neoergosterol) (la) .-(a) Three-stage sequence (see Scheme).Reactants, solvents, and apparatus required in the first stage were carefully dried, and atmospheric moisture was excluded during this stage. A solution of 19-norergosta- 5,7,9- trien- 3 @-yl methane-sulphonate (following paper; 6 g) in Me,CO (240 ml) was contained in a dry box under N,. NEt,+HCO,-(10 g) was weighed in the box and added to the solution, which was then removed from the box, boiled under reflux for 12 h, and evaporated. The residue was partitioned between H,O and C,H,, and the organic layer was washed twice with water, and evaporated. Saturated aqueous NaHCO, (22 ml) and MeOH (420 ml) were added, and the mixture was boiled under reflux for 1 h. After evaporation the residue was extracted with Et,O to give material which was ad- sorbed on to neutral Al,O, (200 g).Elution with petrol gave olefinic material (2.9 g). Et,O eluted material which 20 "C for 2 d. Work-up gave 19-norergosta-5,7,9-trien-Sp-yl benzoate (lc) (210 mg, from petrol) (Found: C, 83.6; H, 9.5. C34H4602 requires C, 83.9; H, 9.5). The 3a- alcohol (2a) (200 mg) similarly gave the 3a-benzoate (2c) (212 mg, from Et,O). Separate solutions of these esters (150 mg) in MeOH (10 nil) containing KOH (200 mg) were boiled under reflux for 1h. Work-up gave the alcohols (la) and (2a)in 90 yield. Compounds (Id) and (2d). A solution of the 313-alcohol (la) (500 mg) and 3,5-dinitrobenzoyl chloride (450 mg) in C,H,N (6 ml) was kept at 40 "C for 1d. Work-up gave the 3p-ester (Id) (678 mg, from EtOAc) (Found: C, 70.6; H, 7.7. Calc.for C,,H,,N,O,: C, 70.8; H, 7.7). The 3a- alcohol (2a) (500 mg) similarly gave 19-norergosta-5,7,9-trien-3a-yl 3,5-dinitrobenzoate (2d) (672 nig, from EtOAc) (Found: C, 70.5; H, 7.7). Separate solutions of these esters (200 mg) in MeOH (30 m1)-saturated NaHCO, aq. (1 ml) were boiled under reflux for 3 h. Work-up gave the alcohols (la) and (2a) in 85 yield. Two crystallisations from EtOAc of a mixture of the ester (Id) (180 mg) and the ester (2d) (20 mg) gave the ester (Id) (105 mg), m.p. 219-221 OC, a, -3 "C (c 0.6). A mixture of the ester (2d) (180 mg) and the ester (Id) (20 mg) similarly gave the ester (2d) (98 mg), m.p. 166-167 "C, aID +47" (c 0.7). Compounds (le) and (2e). A solution of the 3p-alcohol (la) (125 mg) and (+)-camphor-lO-sulphonylchloride (100 mg) in C,H,N (2 ml) was kept at 20 "C for 2 d.Work-up gave 19-norergosta-5,7,9-trien-3fi-yZ( +)-camphor-lO-sulphon-ate (le) (175 mg, from petrol) (Found: C, 75.4; H, 7.8. C,,H,,04S requires C, 75.7; H, 7.9). The 3a-alcohol (2a) ( 125 mg) similarly gave 19-norergosta-5,7,9-2rien-3a-yd 3a-01 (dihydroepineoergosterol) (2a) (1.62 g) (Found: C, 84.8; H, 11.2. Calc. for C2,H,,0: C, 84.75; H, 11.1). (b) Two-stage sequence (see Scheme). Diethyl diazodi-carboxylate (309 mg) was added to a solution of the 3p- alcohol' (la) (300 mg), PPh, (450 mg), and PhC0,H (195 mg) in dry tetrahydrofuran (6 ml). The solution was kept in a stoppered flask at 20 "C for 22 h, and then applied to a l-m SiO, p.1.c.plate which was developed with petrol- Me,CO. The material in the band of highest RF was collected and applied to a second plate which was developed with petrol. The material of lowest RF was collected and crystallised from MeOH to give 19-norergosta-5,7,9-trien-(+)-camphor-lO-szdphonate (2e) (171 mg, from petrol) (Found: C, 75.5; H, 7.7). Compounds (If) and (2f). Following the pub-lished method,g ( amp;)-a-methoxytrifluoromethylphenylacetic acid was treated with (+)-a-methylbenzylamine, a,, +37.5" (liquid), and the less soluble salt was crystallised three times from EtOH to give the a-methylbenzylammonium salt, m.p. 193-195 "C, all, +62" (c 0.8) (lit.,O m.p. 195-198 "C, ar, +59.1"). The acid was liberated, dried, and treated with SOC1, to give (+)-a-methoxy-a-trifluoromethyl-phenylacetyl chloride, b.p.56-58 "C/l mmHg (lit.,g 54- 56 OC/l mmHg), a, +94" (c 0.8),vmx. (liquid film) 1790 crystallised from Et,O to give 19-norergosta-5,7,9-trien-3~-c m-l . yl benzoate (2c) (181 mg) (Found: C, 83.9; H, 9.4. C34-H4,02 requires C, 83.9; H, 9.5). Hydrolysis of this material (140 mg) as described later gave the 3a-alcohol (2a) (94 mg), m.p. 164-165 OC, aID+48" (c 0.8). Esters of the Epimeric Alcohols (la) and (2a) .-Compounds (lb) and (2b). Treatment of the 3p-alcohol (la) (200 mg) with Ac,O (2 m1)-C5H5N (2 ml) at 90 "C for 1 h gave the 3p-acetate (lb) (195 mg, from EtOH). The 3a-alcohol (2a) (200 mg) similarly gave the 3a-acetate (2b) (191 mg, from Et,O). Compounds (lc) and (2c). A solution of the 313-alcohol (la) (200 mg) in BzCl (0.15 m1)-C,H,N (2 ml) was kept at 7 E. L. McGinnis, G. D. Meakins, and D. J. Morris, J. Chem. SOC.(C), 1967, 1238. * P. D. Bartlett and L. H. Knox, Org. Synth., 1965, 45, 14. A solution of the 3p-alcohol (la) (200 mg) and the fore- going chloride (170 mg) in CCl, (1 m1)-C5H,N (1 ml) was kept at 20 "C for 2 d. Work-up gave 19-norergosta-5,7,9- trien-3p-yl ( +)-a-methoxy-a-trifluoromethylphenylacetate(If) (245 mg, from petrol) (Found: C, 73.9; H, 8.3. C,,H4g-F@, requires C, 74.2; H, 8.2). The 3a-alcohol (2a) (200 mg) similarly gave 19-norergosta-5,8,9-trien-3a-yl (+)-a-methoxy-a-trz$uoromethylphe~yZucetate(2f) (251 mg, from petrol) (Found: C, 74.0; H, 8.2). We thank Glaxo Research Ltd. for a grant (to P. M. F.) and a gift of chemicals. 8/328 Received, 24th February, 19781 9 J. A. Dale, D. L. Dull, and H. S. Mosher, J.Org. Chern., 1969. 34, 2543.
展开▼
