Purification of organic compounds, especially those designated for pharmaceutical use, is of considerable importance for chemists synthesizing such compounds. Preparation of the compound usually requires many synthetic steps and, therefore, the final product can be contaminated not only with side-products derived from the last synthetic step of the procedure but also with compounds that were formed in previous steps. Even chromatographic purification, which is a very efficient but relatively time- consuming process, does not usually provide compounds which are sufficiently pure to be used as drugs.

Depending on the method used to synthesize la-hydroxyvitamin D compounds, different minor undesirable compounds can accompany the final product. Thus, for example, if direct C-1 hydroxylation of 5, 6-trans geometric isomer of vitamin D is performed, followed by Se02/NMO oxidation and photochemical irradiation [see Andrews et al., J. Org. Chem.

51, 1635 (1986); Calverley et al., Tetrahedron 43,4609 (1987); Choudry et al., J. Org. Chem. 58,1496 (1993)], the final la-hydroxyvitamin D product can be contaminated with l-hydroxy-as well as 5,6-trans isomers. If the method consists of C-1 allylic oxidation of the 4-phenyl-1, 2,4-triazoline-3,5- dione adduct of the previtamin D compound, followed by cycloreversion of

the modified adduct under basic conditions [Nevinckx et al., Tetrahedron 47,9419 (1991) ; Vanmaele et al., Tetrahedron 41,141 (1985) and 40,1179 (1991); Vanmaele et al., Tetrahedron Lett. 23,995 (1982)], one can expect that the desired la-hydroxyvitamin can be contaminated with the previtamin 5 (10), 6,8-triene and lés-hydroxy isomer. One of the most useful C-1 hydroxylation methods, of very broad scope and numerous applications, is the experimentally simple procedure elaborated by Paaren et al. [see J.

Org. Chem. 45,3253 (1980) and Proc. Natl. Acad. Sci. U. S. A. 75,2080 (1978)]. This method consists of allylic oxidation of 3, 5-cyclovitamin D derivatives, readily obtained from the buffered solvolysis of vitamin D tosylates, with SeO2/t-BuOOH and subsequent acid-catalyzed cycloreversion to the desired la-hydroxy compounds. Taking into account this synthetic path it is reasonable to assume that the final product can be contaminated with lp-hydroxy epimer, 5,6-trans isomer and the previtamin D form.

The vitamin D conjugated triene system is not only heat-and light- sensitive but it is also prone to oxidation, leading to the complex mixture of very polar compounds. Oxidation usually happens when a vitamin D compound has been stored for a prolonged time. Other types of processes that can lead to a partial decomposition of vitamin D compounds consist of the some water-elimination reactions; their driving force is the allylic (la-) and homoallylic (3p-) position of the hydroxy groups. The presence of such above-mentioned oxidation and elimination products can be easily detected by thin-layer chromatography. Thus, for example, using precoated aluminum silica sheets [with UV indicator; from EM Science (Cherry Hill, NJ)] and solvent system hexane-ethyl acetate (3: 7), the spot of la, 24 (S)- (OH) 2D2 (Rf 0.40) and its elimination products (Rf's ca. 0.8-0.9) are visible in ultraviolet light. Also, after spraying with sulfuric acid and heating, an additional spot can be visualized (Rf 0), derived from oxidation products.

Usually, all la-hydroxylation procedures require at least one chromatographic purification. However, even chromatographically purified la, 24 (S)-dihydroxyvitamin D2 although showing consistent spectroscopic data, suggesting its homogeneity, does not meet the purity criteria required for therapeutic agents that can be orally, parenterally or transdermally administered. Therefore, it was evident that a suitable method of purification of la, 24 (S)-dihydroxyvitamin D2 is required.

Since it is well known that the simplest procedure that can be used for compound purification is a crystallization process, it has been decided to investigate purification of la, 24 (S)- (OH) 2D2 by means of crystallization. The solvent plays a crucial role in the crystallization process, and is typically an individual liquid substance or a suitable mixture of different liquids. For crystallizing la, 24 (S)-dihydroxyvitamin D2, the most appropriate solvent and/or solvent system is characterized by the following factors: (1) low toxicity; (2) low boiling point; (3) significant dependence of solubility properties with regard to temperature (condition necessary for providing satisfactory crystallization yield); and (4) relatively low cost.

It is believed that highly apolar solvents (e. g. hydrocarbons) were not suitable due to their low solubility potency. Quite the reverse situation occurred in the highly polar media (e. g. alcohols), in which la, 24 (S)- (OH) 2D2, showed too high solubility. Therefore, it is concluded that for the successful crystallization of la, 24 (S)- (OH) 2D2, a solvent mixture is required, consisting of two (or more) solvents differing considerably in polarity. After numerous experiments, it was found that several binary solvent systems were useful for the crystallization of la, 24 (S)- (OH) 2D2, namely: acetone- hexane, 2-propanol-hexane and ethyl formate-petroleum ether. These solvents are all characterized by low toxicity, and they are very easy to

remove by evaporation or other well known methods. In all cases it is believed the crystallization process will occur easily and efficiently, and the precipitated crystals will be sufficiently large to assure their recovery by filtration.

BRIEF DESCRIPTION OF THE DRAWINGS Figures la-lh are graphs of 1H NMR spectrum (CDC13, 500 MHz) of the solid la, 24 (S)-dihydroxyvitamin D2 material before crystallization (Figs. la and lb) as well as the spectra of the crystals of la, 24 (S)- (OH) 2D2 which resulted after two crystallizations using the following solvent systems: acetone-hexane (Figs. lc and ld), 2-propanol-hexane (Figs. le and li) and HCOOEt-petroleum ether (Figs. lg and lh) ; Figures 2a-2d are HPLC (10 mm x 25 cm Zorbax-Sil column, 20% 2- propanol in hexane, 4 mL/min; UV detection at 260 nm) profiles of the solid la, 24 (S)-dihydroxyvitamin D2 material before crystallization (Fig. 2a) and the crystals resulted after two crystallizations using the following solvent systems: acetone-hexane (Fig. 2b), 2-propanol-hexane (Fig. 2c) and HCOOEt-petroleum ether (Fig. 2d). In the region indicated by asterisk (ca.

34 mL) sensitivity was decreased 20 times.

Figures 3a-3i are HPLC (10 mm x 25 cm Zorbax-Sil column, 20% 2- propanol in hexane, 4 mL/min; UV detection at 260 nm) profiles of the crystals of la, 24 (S)-dihydroxyvitamin D2 resulted after single crystallization using the following solvent systems: acetone-hexane (Fig. 3a), 2-propanol- hexane (Fig. 3d) and HCOOEt-petroleum ether (Fig. 3g); the HPLC profiles of mother liquors after single crystallization using the following solvent systems: acetone-hexane (Fig. 3b); 2-propanol-hexane (Fig. 3e) and HCOOEt-petroleum ether (Fig. 3h); and the HPLC profiles of mother liquors after two crystallizations using the following solvent systems: acetone- hexane (Fig. 3c); 2-propanol-hexane (Fig. 3f) and HCOOEt-petroleum ether

(Fig. 3i). Region with decreased sensitivity (ca. 34 mL) is indicated by asterisk.

Figures 4a-4d are HPLC (4.6 mm x 25 cm Zorbax-Eclipse XDB-C18 column, 20% water in methanol, 1 mL/min; UV detection at 260 nm) profiles of the solid la, 24 (S)-dihydroxyvitamin D2 material before crystallization (Fig. 4a) and the crystals resulted after two crystallizations using the following solvent systems: acetone-hexane (Fig. 4b), 2-propanol- hexane (Fig. 4c) and HCOOEt-petroleum ether (Fig. 4d). In the region indicated by asterisk (ca. 22 mL) sensitivity was decreased 20 times.

Figures 5a-5c are microscope-magnified images of the crystals of la, 24 (S)-dihydroxyvitamin D2 resulted after two crystallizations from: acetone-hexane (Fig. 5a), 2-propanol-hexane (Fig. 5b) and HCOOEt- petroleum ether (Fig. 5c).

Figure 6 is an illustration of the three dimensional structure of la, 24 (S)-dihydroxyvitamin D2 as defined by the atomic positional parameters discovered and set forth herein.

DETAILED DESCRIPTION OF THE INVENTION The present invention provides a valuable method of purification of la, 24 (S)-dihydroxyvitamin D2, a pharmacologically important compound, characterized by the formula shown below:

The purification technique involves obtaining the la, 24 (S)-dihydroxyvitamin D2 product in crystalline form by utilizing a crystallization procedure wherein the la, 24 (S)-dihydroxyvitamin D2 material to be purified is dissolved using as the solvent system one of the following: (1) acetone and hexane; (2) 2-propanol and hexane; or (3) ethyl formate and petroleum ether.

Thereafter, the solvent or solvent system can be removed by evaporation, with or without vacuum, or via other means as is well known. The technique can be used to purify a wide range of final products containing la, 24 (S)-dihydroxyvitamin D2 obtained from any known synthesis thereof, and in varying concentrations, i. e. from microgram amounts to kilogram amounts. As is well known to those skilled in this art, the amount of solvent utilized should be minimized and/or adjusted according to the amount of la, 24 (S)-dihydroxyvitamin Dz to be purified.

The usefulness and advantages of the present crystallization procedures is shown in the following specific Examples. Solid la, 24 (S)- dihydroxyvitamin D2 product which was purified by chromatography on silica and was used as a suitable starting material. This material showed reasonably good 500 MHz 1H NMR spectrum (Figures la, lb), but concomitant compounds were detected by straight-and reverse-phase HPLC (Figures 2a and 4a, respectively) and, moreover, the presence of some oxidation products was confirmed by TLC (presence of the spot of Rf 0).

After recrystallization from the solvents listed above, the precipitated material was observed under microscope to confirm its crystalline form (Figures 5a-5c). Additionally, in the case of crystals precipitated from 2- propanol-hexane, X-ray diffraction analysis was performed (Figure 6). The corresponding crops of crystals were then carefully analyzed and their significantly improved purity was confirmed by straight-phase HPLC

(Figures 2b, 2c, 2d), reverse-phase HPLC (Figures 4b, 4c, 4d), TLC and 500 MHz 1H NMR (Figures lc-lh). Yields of crystallizations were high and the obtained crystals showed a relatively sharp melting point.

The corresponding straight-and reverse-phase HPLC profiles of the recrystallized la, 24 (S)-dihydroxyvitamin D2, shown in Figures 2b-2d and 4b-4d, respectively, clearly indicate a considerable improvement in the compound purity. The important observation consists of the significantly diminished proportion of the concomitant la, 24 (R)-dihydroxyvitamin D2 (peak of retention time ca. 30 mL on Figures 2b-2d and ca. 25 mL on Figures 4b-4d) in the recrystallized compound; content of this R-isomer impurity has decreased more than 3 times (3.3-5.3) in respect to its value in the starting la, 24 (S)-dihydroxyvitamin D2 product and does not exceed 0.5%.

The described crystallization process of the synthetic la, 24 (S)- dihydroxyvitamin D2 product represents a valuable purification method, which can remove products derived from the synthetic path, including its concomitant 24-epimeric compound, namely, la, 24 (R)-dihydroxyvitamin D2.

Such impurity is a result of the nonstereospecific construction of the side chain (U. S. Patent No. 5,786,348 and 5,789,397). In such case the separation of both epimeric vitamins is necessary and it is usually performed during the last stage of the synthesis. However, column chromatography and straight-phase separation of the 24-epimers is practically impossible due to their similar chromatographical properties, and larger-scale separation is also difficult by reverse-phase HPLC.

CRYSTALLIZATION OF la, 24-DIHYDROXYVITAMIN D2 Example 1 Crystallization from acetone-hexane (a) la, 24 (S)-dihydroxyvitamin D2 product (25 mg, m. p. 136-142.5° C) was dissolved in boiling acetone (0.18 mL, Fisher Scientific) under argon

atmosphere and hexane (0.72 mL, Burdick & Jackson) was added. The solution was left at room temperature (68°F) for a few hours and then in a refrigerator (35-45°F) overnight. The precipitated crystals were filtered off, washed with a small volume of a cold hexane and dried. The yield of crystalline material was 15 mg (60%). HPLC profiles of crystals and mother liquor are shown in Figures 3a, 3b.

(b) These crystals of la, 24 (S)-dihydroxyvitamin D2 (12 mg) were recrystallized from acetone (0.07 mL) and hexane (0.3 mL) as described in Example l (a) and the precipitated crystals (9 mg, 75%), m. p. 146.5-151°C were observed under a microscope (Figure 5a) and analyzed by straight- phase HPLC (crystals: Figure 2b; mother liquors: Figure 3c), reverse-phase HPLC (Figure 4b) and 1H NMR (Figures lc, ld).

Example 2 Crystallization from 2-propanol-hexane (a) la, 24 (S)-dihydroxyvitamin D2 product (25 mg) was dissolved in a boiling 2-propanol-hexane mixture (1: 4,0.75 mL; Burdick & Jackson) under argon atmosphere, left at room temperature (68°F) for a few hours and then in a refrigerator (35-45°F) overnight. The precipitated crystals were filtered off, washed with a small volume of a cold hexane and dried.

The yield of crystalline material was 17 mg (68%) HPLC profiles of crystals and mother liquor are shown in Figures 3d, 3e.

(b) These crystals of la, 24 (S)-dihydroxyvitamin Dz product (15 mg) were recrystallized from 2-propanol-hexane mixture (1: 4,0.43 mL) as described in Example 2 (a) and the precipitated crystals (8 mg, 53%) m. p.

147-151.5°C, were observed under a microscope (Figure 5b) and analyzed by straight-phase HPLC (crystals: Figure 2c; mother liquors: Figure 3, reverse-phase HPLC (Figure 4c) and 1H NMR (Figures le, l.

Example 3 Crystallization from ethyl formate-petroleum ether (a) la, 24 (S)-dihydroxyvitamin D2 product (25 mg) was dissolved in boiling ethyl formate (0.5 mL, Aldrich) under argon atmosphere and petroleum ether (1 mL, b. p. 35-60°C ; Aldrich) was added. The solution was left at room temperature (68°F) for a few hours and then in a refrigerator (35- 45°F) overnight. The precipitated crystals were filtered off, washed with a small volume of a cold hexane and dried. The yield of crystalline material was 17 mg (68%). HPLC profiles of crystals and mother liquor are shown in Figures 3g, 3h.

(b) These crystals of la, 24 (S)-dihydroxyvitamin D2 (15 mg) were recrystallized from ethyl formate (0.25 mL) and petroleum ether (0.5 mL) as described in Example 3 (a) and the precipitated crystals (9 mg, 60%), m. p.

142-146.5°C, were observed under a microscope (Figure 5c) and analyzed by straight-phase HPLC (crystals: Figure 2d; mother liquors, Figure 3i), reverse- phase HPLC (Figure 4d) and 1H NMR (Figures lg, lh).

Experimental A colorless prism-shaped crystal of dimensions 0.42 x 0.17 x 0.08 mm was selected for structural analysis. Intensity data for this compound were collected using a Bruker SMART ccd area detector, (a) Data collection: SMART Software Reference Manual (1994). Bruker-AXS, 6300 Enterprise Dr., Madison, Wisconsin 53719-1173, U. S. A., (b) Data Reduction: SAINT Software Reference Manual (1995). Brunker-AXS, 6300 Enterprise Drive, Madison, Wisconsin 53719-1173, U. S. A., mounted on a Bruker Platform goniometer using graphite-monochromated Mo Ka radiation (k = 0.71073 A).

The sample was cooled to 133K. The intensity data, which nominally covered one and a half hemispheres of reciprocal space, were measured as a series of c3 oscillation frames each of 0.3 ° for 90 sec/frame. The detector was operated in 512 x 512 mode and was positioned 5.00 cm from the

sample. Coverage of unique data was 99.9 % complete to 25.00 degrees in 0. Cell parameters were determined from a non-linear least squares fit of 5435 peaks in the range 2.30 < 6 < 28.28°. The first 50 frames were repeated at the end of data collection and yielded 156 peaks showing a variation of 0.01 % during the data collection. A total of 9896 data were measured in the range 1.61 < A < 28.30°. The data were corrected for absorption by the empirical method G. M. Sheldrick (1996). SADABS.

Program for Empirical Absorption Correction of Area Detector Data.

University of Göttingen, Germany, giving minimum and maximum transmission factors of 0.769 and 0.970. The data were merged to form a set of 7019 independent data with R (int) = 0.0389.

The monoclinic space group P2 (1) was determined by systematic absences and statistical tests and verified by subsequent refinement. The structure was solved by direct methods and refined by full-matrix least- squares methods on F2, (a) G. M. Sheldrick (1994). SHELXTL Version 5 Reference Manual, Bruker-AXS, 6300 Enterprise Drive, Madison, Wisconsin 53719-1173, U. S. A. (b) International Tables for Crystallography, Vol. C, Tables 6.1.1.4,4.2.6.8, and 4.2.4.2, Kluwer : Boston (1995). Hydrogen atom positions were initially determined by geometry and refined by a riding model. Non-hydrogen atoms were refined with anisotropic displacement parameters. A total of 353 parameters were refined against 7 restraints and 7019 data to give wR (F2) = 0.1274 and S = 0.952 for weights ofw= l/ [cy2 (F2) + (0.0670 P) 2], where P = [Fo2 + 2Fc2j/3. The final R (F) was 0.0504 for the 5187 observed, [F > 4a (F) I, data. The largest shift/s. u. was 0.006 in the final refinement cycle. The final difference map had maxima and minima of 0.241 and-0.155 e/A3, respectively. The absolute structure was determined by refinement of the Flack parameter, H. D. Flack, Acta Cryst.

A39,876-881 (1983). The polar axis restraints were taken from Flack and Schwarzenbach, H. D. Flack and D. Schwarzenback, Acta Cryst. A44,499- 506 (1988).

The displacement ellipsoids are drawn at the 50% probability level in Figure 6. The solvent molecule of 2-propanol was disordered and modeled in two orientations with occupancies of 0.473 (10) and 0.527 (10) for the unprimed and primed atoms (not shown in Figure 6). Restraints on the positional parameters of the solvent were required for the refinement to achieve convergence.

The three dimensional structure of la, 24 (S)-dihydroxyvitamin Dz as defined by the following physical data and atomic positional parameters described and calculated herein is illustrated in Figure 6.

Table 1. Crystal data and structure refinement for 1 a, 24 (S)-dihydroxyvitamin D2.

Empirical formula (C28 H44 03) (C3 H8 O) Formula weight 488.73 Crystal system Monoclinic Space group P2 (1) Unit cell dimensions a = 12.8196 (11) A a-90° b = 7.6363 (6) A 99.270 (2) ° c = 15. 4915 (13) A γ=90° Volume 1496.7 (2) A3 Z 2 Density (calculated) 1.084 Mg/m3 Wavelength 0.71073 A Temperature 133 (2) K F (000) 540 Absorption coefficient 0.069 mm-1 Absorption correction Empirical Max. and min. transmission 0.970 and 0.769 Theta range for data collection 1.61 to 28.30°.

Reflections collected 9896 Independent reflections 7019 [R (int) = 0. 0389] <BR> <BR> <BR> Data/restraints/parameters 7019/7 (disorder)/353 wR (F2all data) wR2 = 0. 1274 R (F obsd data) Ru-0. 0504 Goodness-of-fit on F2 0. 952 Observed data [I > 2# (I)] 5187 Absolute structure parameter-1.3 (11) Largest and mean shift/s. u. 0.006and 0.000 Largest diff. peak and hole 0. 241 and-0.155 e/A3 Table 2. Atomic coordinates and equivalent isotropic displacement parameters for la, 24 (S)-dihydroxyvitamin D2. U (eq) is defined as one third of the trace of the orthogonalized Uij tensor. x y z U (eq) O (1) 0.74103 (14) 0.26699 (19) 0.64786 (9) 0.0504 (4) 0 (2) 0.71159 (11) 0.81818 (19) 0.65629 (8) 0.0398 (3) 0 (3) 0.71370 (12) 0.0537 (2)-0. 21338 (9) 0.0422 (4) C (1) 0.75494 (16) 0.4442 (3) 0.62449 (12) 0.0340 (4) C (2) 0.78823 (16) 0.5444 (3) 0.70961 (11) 0.0358 (4) C (3) 0.80863 (15) 0.7362 (3) 0.69351 (12) 0.0347 (4) C (4) 0.88914 (16) 0.7581 (3) 0.63196 (12) 0.0377 (5) C (5) 0.85897 (14) 0.6546 (3) 0.54858 (12) 0.0332 (4) C (6) 0.85205 (15) 0.7322 (3) 0.47018 (12) 0.0363 (4) C (7) 0.82031 (15) 0.6503 (3) 0.38543 (12) 0.0351 (4) C (8) 0.81345 (15) 0.7288 (3) 0.30735 (12) 0.0355 (4) C (9) 0.83815 (19) 0.9172 (3) 0.29164 (13) 0.0438 (5) C (10) 0.83535 (15) 0.4674 (3) 0.56360 (12) 0.0332 (4) C (11) 0.90937 (19) 0.9379 (3) 0.22164 (13) 0.0439 (5) C (12) 0.86813 (17) 0.8370 (3) 0.13713 (12) 0.0381 (5) C (13) 0.85111 (14) 0.6444 (3) 0.15568 (11) 0.0302 (4) C (14) 0.77427 (15) 0.6343 (3) 0.22337 (12) 0.0332 (4) C (15) 0.74419 (18) 0.4419 (3) 0.22396 (13) 0.0426 (5) C (16) 0.73673 (17) 0.3872 (3) 0.12732 (12) 0.0386 (5) C (17) 0.78577 (14) 0.5385 (3) 0.07984 (11) 0.0310 (4) C (18) 0.95630 (15) 0.5526 (3) 0.18950 (13) 0.0408 (5) C (19) 0.88104 (18) 0.3357 (3) 0.52951 (15) 0.0485 (6) C (2Q) 0.83816 (16) 0.4686 (3) 0.00427 (12) 0.0344 (4) C (21) 0.89386 (18) 0.6079 (3)-0.04330 (13) 0.0431 (5) C (22) 0.75293 (16) 0.3790 (3)-0.05931 (12) 0.0345 (4) C (23) 0.74996 (16) 0.2140 (3)-0.08077 (12) 0.0352 (4) C (24) 0.66492 (16) 0.1167 (3)-0.14141 (12) 0.0331 (4) C (28) 0.57047 (17) 0.2288 (3)-0.17729 (14) 0.0428 (5) C (25) 0.63467 (17)-0.0501 (3)-0.09445 (13) 0.0389 (5) C (26) 0.5682 (2)-0.1799 (3)-0.15392 (16) 0.0490 (6) C (27) 0.5808 (2)-0.0051 (3)-0.01633 (15) 0.0537 (6) O (1S) 0.6790 (4) 0.0433 (9) 0.5132 (5) 0.0525 (15) C (1S) 0.5785 (5) 0.0407 (9) 0.4589 (4) 0.0579 (19) C (2S) 0.568 (2) 0.193 (2) 0.3972 (15) 0.082 (6) C (3S) 0.5668 (13)-0.1295 (19) 0.4094 (16) 0.068 (4) O (1S') 0.6310 (6) 0.0515 (8) 0.5266 (3) 0.0531 (15) C (1S') 0.6298 (6) 0.0414 (9) 0.4348 (3) 0.0622 (19) C (2S') 0.589 (2) 0.216 (2) 0.4011 (14) 0.091 (5) C (3S') 0.5714 (16)-0.117 (2) 0.3957 (17) 0.104 (8) Table 3. Bond lengths [A] and angles [°] for la, 24 (S)-dihydroxyvitamin D2.

O (1)-C (1) 1.419 (2) C (13)-C (17) 1.556 (2) 0 (2)-C (3) 1.429 (2) C (14)-C (15) 1.519 (3) 0 (3)-C (24) 1.446 (2) C (15)-C (16) 1.542 (3) C (1)-C (10) 1.516 (3) C (16)-C (17) 1.556 (3) C (1)-C (2) 1.525 (3) C (17)-C (20) 1.536 (3) C (2)-C (3) 1.515 (3) C (20)-C (22) 1.512 (3) C (3)-C (4) 1.523 (3) C (20)-C (21) 1.534 (3) C (4)-C (5) 1.511 (3) C (22)-C (23) 1.303 (3) C (5)-C (6) 1.341 (3) C (23)-C (24) 1.514 (3) C (5)-C (10) 1.487 (3) C (24)-C (28) 1.514 (3) C (6)-C (7) 1.452 (3) C (24)-C (25) 1.547 (3) C (7)-C (8) 1.340 (3) C (25)-C (26) 1.518 (3) C (8)-C (9) 1.501 (3) C (25)-C (27) 1.526 (3) C (8)-C (14) 1.502 (3) O (1 S)-C (1 S) 1.421 (5) C (9)-C (11) 1.534 (3) C (1S)-C (2S) 1.500 (6) C (10)-C (19) 1.317 (3) C (1 S)-C (3 S) 1.504 (6) C (11)-C (12) 1.537 (3) O (1 S')-C (1 S') 1.422 (5) C (12)-C (13) 1.521 (3) C (lS')-C (2S') 1.498 (6) C (13)-C (18) 1.535 (3) C (1S')-C(3S') 1.501 (6) C (13)-C (14) 1.552 (3) O (1)-C (1)-C (10) 113.37 (16) C (8)-C (9)-C (11) 112.23 (18) O (1)-C (1)-C (2) 106.70 (15) C (19)-C (10)-C (5) 123.8 (2) C (10)-C (1)-C (2) 110.83 (16) C (19)-C (10)-C (1) 123.4 (2) C (3)-C (2)-C (1) 112.00 (15) C (5)-C (10)-C(1) 112.72 (16) 0 (2)-C (3)-C (2) 109.14 (16) C (9)-C (11)-C (12) 112.84 (17) 0 (2)-C (3)-C (4) 109.44 (16) C (13)-C (12)-C (11) 111.33 (16) C (2)-C (3)-C (4) 111.21 (17) C (12)-C (13)-C (18) 111.18 (17) C (5)-C (4)-C (3) 111.82 (16) C (12)-C (13)-C (14) 107.58 (16) C (6)-C (5)-C (10) 125.50 (18) C (18)-C (13)-C (14) 111.39 (15) C (6)-C (5)-C (4) 120.90 (19) C (12)-C (13)-C (17) 115.88 (15) C (10)-C (5)-C (4) 113.57 (17) C (18)-C (13)-C (17) 110.89 (16) C (5)-C (6)-C (7) 126.6 (2) C (14)-C (13)-C (17) 99.28 (14) C (8)-C (7)-C (6) 126.2 (2) C (8)-C (14)-C (15) 120.64 (17) C (7)-C (8)-C (9) 126.19 (18) C (8)-C (14)-C (13) 113.65 (15) C (7)-C (8)-C (14) 122.04 (19) C (15)-C (14)-C (13) 103.99 (16) C (9)-C (8)-C (14) 111.73 (17) C (14)-C (15)-C (16) 103.42 (17) C (15)-C (16)-C (17) 106.93 (16) 0 (3)-C (24)-C (25) 105.13 (15) C (20)-C (17)-C (13) 120.53 (15) C (28)-C (24)-C (25) 113.08 (18) C (20)-C (17)-C (16) 111.23 (17) C (23)-C (24)-C (25) 108.79 (16) C (13)-C (17)-C (16) 103.81 (14) C (26)-C (25)-C (27) 110.29 (18) C (22)-C (20)-C (21) 110.13 (16) C (26)-C (25)-C (24) 114.26 (17) C (22)-C (20)-C (17) 107.15 (15) C (27)-C (25)-C (24) 111.56 (18) C (21)-C (20)-C (17) 114.81 (17) O (1 S)-C (1 S)-C (2S) 110.1 (12) C (23)-C (22)-C (20) 126.32 (19) O (1S)-C (1S)-C (3S) 108.7 (8) C (22)-C (23)-C (24) 128.42 (19) C (2S)-C (1S)-C (3S) 110.9 (15) 0 (3)-C (24)-C (28) 108.95 (16) O (1S')-C (1S')-C (2S') 104.3 (9) 0 (3)-C (24)-C (23) 106.50 (15) O (1S')-C (1S')-C (3S') 111.8 (12) C (28)-C (24)-C (23) 113.83 (17) C (2S')-C (1S')-C(3S') 117.1 (14) Table 4. Anisotropic displacement parameters (Å2x 103) for 1 a, 24 (S)- dihydroxyvitamin D2. The anisotropic displacement factor exponent takes the form :-2 [h2a*2U1l +... +2hka*b*U12] Uii U22 U33 U23 U13 U12 O (1) 84 (1) 39 (1) 30 (1) 0 (1) 15 (1)-10 (1) 0 (2) 49 (1) 41 (1) 29 (1)-8 (1) 4 (1) 8 (1) 0 (3) 61 (1) 40 (1) 31 (1)-8 (1) 22 (1)-2 (1) C (1) 43 (1) 34 (1) 25 (1)-1 (1) 7 (1)-1 (1) C (2) 44 (1) 43 (1) 22 (1) 0 (1) 10 (1) 3 (1) C (3) 44 (1) 40 (1) 21 (1)-7 (1) 4 (1) 1 (1) C (4) 41 (1) 46 (1) 26 (1)-7 (1) 3 (1)-9 (1) C (5) 28 (1) 47 (1) 25 (1)-7 (1) 7 (1)-3 (1) C (6) 37 (1) 44 (1) 29 (1)-5 (1) 8 (1)-10 (1) C (7) 35 (1) 47 (1) 25 (1)-4 (1) 9 (1)-10 (1) C (8) 35 (1) 47 (1) 25 (1)-4 (1) 8 (1)-6 (1) C (9) 59 (1) 46 (1) 28 (1)-8 (1) 9 (1)-6 (1) C (10) 33 (1) 43 (1) 23 (1)-4 (1) 2 (1) 2 (1) C (11) 60 (1) 40 (1) 32 (1)-3 (1) 10 (1)-14 (1) C (12) 46 (1) 43 (1) 27 (1) 0 (1) 9 (1)-8 (1) C (13) 30 (1) 40 (1) 21 (1)-1 (1) 6 (1)-5 (1) C (14) 33 (1) 44 (1) 24 (1)-2 (1) 7 (1)-6 (1) C (15) 49 (1) 54 (1) 25 (1)-5 (1) 8 (1)-20 (1) C (16) 44 (1) 48 (1) 25 (1)-5 (1) 7 (1)-12 (1) C (17) 32 (1) 39 (1) 22 (1)-3 (1) 4 (1)-2 (1) C (18) 35 (1) 54 (1) 32 (1)-2 (1) 0 (1) 0 (1) C (19) 49 (1) 57 (1) 40 (1)-8 (1) 9 (1) 6 (1) C (20) 38 (1) 41 (1) 24 (1)-5 (1) 7 (1) 1 (1) C (21) 47 (1) 54 (1) 31 (1)-5 (1) 15 (1)-5 (1) C (22) 41 (1) 41 (1) 22 (1)-2 (1) 5 (1) 5 (1) C (23) 37 (1) 43 (1) 26 (1)-1 (1) 9 (1) 3 (1) C (24) 41 (1) 35 (1) 26 (1)-7 (1) 10 (1) 1 (1) C (28) 47 (1) 45 (1) 35 (1)-5 (1) 1 (1) 3 (1) C (25) 43 (1) 43 (1) 33 (1) 0 (1) 11 (1) 2 (1) C (26) 60 (1) 43 (1) 48 (1)-7 (1) 20 (1)-9 (1) C (27) 68 (2) 63 (2) 35 (1)-4 (1) 23 (1)-16 (1) O (1S) 71 (4) 47 (2) 35 (3) 0 (2)-4 (3)-6 (3) C (1S) 46 (4) 86 (4) 44 (4) 2 (3) 13 (3) 3 (3) C (2S) 52 (7) 116 (10) 76 (9) 42 (9) 7 (6) 29 (9) C (3S) 43 (6) 108 (10) 51 (6)-34 (5)-1 (4)-2 (6) O (1S') 78 (4) 50 (2) 30 (2) 5 (2) 4 (2) 3 (3) C (IS') 49 (4) 102 (5) 35 (3)-4 (3) 4 (2) 3 (3) C (2S') 73 (10) 139 (11) 56 (6) 33 (6)-3 (5)-10 (6) C (3S') 81 (9) 165 (15) 66 (9)-44 (8) 10 (6)-40 (9) Table 5. Hydrogen coordinates and isotropic displacement parameters for la, 24 (S)-dihydroxyvitamin D2. x y z U (eq) H (10) 0.7154 0. 2011 0. 5980 0.060 H (20) 0.6925 0. 9129 0. 6947 0.048 H (30) 0.7209 0.1479-0.2497 0.051 H (1) 0.6854 0. 4912 0. 5950 0.041 H (2A) 0.7320 0. 5346 0. 7462 0.043 H (2B) 0.8532 0. 4910 0. 7422 0.043 H (3) 0.8363 0. 7938 0. 7506 0.042 H (4A) 0.8946 0. 8837 0. 6174 0.045 H (4B) 0.9593 0. 7188 0. 6619 0.045 H (6) 0.8697 0. 8530 0. 4701 0.044 H (7) 0.8027 0. 5294 0. 3853 0.042 H (9A) 0.8737 0. 9699 0. 3471 0.053 H (9B) 0.7713 0. 9816 0. 2730 0.053 H (11A) 0.9812 0. 8955 0. 2455 0.053 H (11B) 0.9148 1. 0637 0. 2076 0.053 H (12A) 0.8005 0. 8892 0. 1088 0.046 H (12B) 0.9196 0. 8477 0. 0961 0.046 H (14) 0.7088 0. 6975 0. 1962 0.040 H (15A) 0.6756 0. 4260 0. 2445 0.051 H (15B) 0.7990 0. 3732 0. 2619 0.051 H (16A) 0.6620 0. 3686 0. 1007 0.046 H (16B) 0.7761 0. 2770 0. 1227 0.046 H (17) 0.7260 0. 6153 0. 0532 0.037 H (18A) 1.0078 0. 5807 0. 1512 0.061 H (18B) 0.9452 0. 4256 0. 1899 0.061 H (18C) 0.9831 0. 5928 0. 2490 0.061 H (19A) 0.9324 0. 3576 0. 4929 0.058 H (19B) 0.8627 0. 2186 0. 5415 0.058 H (20) 0.8916 0. 3785 0. 0284 0.041 H (21A) 0.9193 0.5545-0.0935 0.065 H (21B) 0.9539 0.6561-0.0031 0.065 H (21C) 0.8440 0.7021-0.0636 0.065 H (22) 0.6960 0.4500-0.0862 0.041 H (23) 0.8088 0.1457-0.0550 0.042 H (25A) 0.7023-0.1109-0.0706 0.047 H (26A) 0.6064-0.2152-0.2010 0.074 H (26B) 0.5011-0.1250-0.1791 0.074 H (26C) 0.5543-0.2832-0.1200 0.074 H (27A) 0.6233 0. 0815 0. 0205 0.081 H (27B) 0.5739-0.1112 0. 0179 0.081 H (27C) 0.5105 0.0434-0.0372 0.081 H (28A) 0.5940 0.3296-0.2083 0.064 H (28B) 0.5360 0.2701-0.1290 0.064 H (28C) 0.5202 0.1593-0.2178 0.064 H (1S) 0.6790-0.0276 0. 5547 0.063 H (1S1) 0.5221 0. 0487 0. 4964 0.069 H (2S1) 0.5939 0. 2994 0. 4296 0.123 H (2S2) 0.4940 0. 2087 0. 3714 0.123 H (2S3) 0.6104 0. 1727 0. 3507 0.123 H (3S1) 0.6057-0.1162 0. 3603 0.102 H (3S2) 0.4918-0.1495 0. 3870 0.102 H (3S3) 0.5952-0. 2295 0. 4454 0.102 H (1S') 0. 6602-0. 0383 0. 5507 0. 064 H (1S2) 0.7047 0. 0316 0. 4246 0.075 H (2S4) 0. 5199 0. 2383 0. 4191 0. 136 H (2S5) 0. 5804 0. 2165 0. 3371 0. 136 H (2S6) 0. 6385 0. 3081 0. 4247 0. 136 H (3S4) 0.6045-0.2183 0. 4284 0.156 H (3S5) 0.5751-0. 1324 0. 3334 0.156 H (3S6) 0. 4973-0. 1090 0. 4037 0. 156 Table 6. Torsion angles [°] for la, 24 (S)-dihydroxyvitamin D2.

C (12)-C (13)-C (14)-C (15) 168.80 (16) O (1)-C (1)-C (2)-C (3) 178.48 (16) C (18)-C (13)-C (14)-C (15)-69.1 (2) C (10)-C (1)-C (2)-C (3) 54.6 (2) C (17)-C (13)-C (14)-C (15) 47.76 (18) C (1)-C (2)-C (3)-0 (2) 66.0 (2) C (8)-C (14)-C (15)-C (16)-166. 13 (18) C (1)-C (2)-C (3)-C (4)-54.8 (2) C (13)-C (14)-C (15)-C (16)-37. 2 (2) 0 (2)-C (3)-C (4)-C (5)-68.2 (2) C (14)-C (15)-C (16)-C (17) 11.8 (2) C (2)-C (3)-C (4)-C (5) 52.5 (2) C (12)-C (13)-C (17)-C (20) 80.8 (2) C (3)-C (4)-C (5)-C (6) 126.3 (2) C (18)-C (13)-C (17)-C (20)-47.1 (2) C (3)-C (4)-C (5)-C (10)-51. 6 (2) C (14)-C (13)-C (17)-C (20) C (10)-C (5)-C (6)-C (7) 0.1 (3)-164.38 (17) C (4)-C (5)-C (6)-C (7)-177.57 (18) C (12)-C (13)-C (17)-C (16) C (5)-C (6)-C (7)-C (8) 179.9 (2)-153.88 (17) C (6)-C (7)-C (8)-C (9) 0.6 (4) C (18)-C (13)-C (17)-C (16) 78.17 ( 19) C (6)-C (7)-C (8)-C (14)-176.76 (18) C (14)-C (13)-C (17)-C (16)-39. 08 (18) C (7)-C (8)-C (9)-C (11) 132.1 (2) C (15)-C (16)-C (17)-C (20) 148.70 (17) C (14)-C (8)-C (9)-C (11)-50. 3 (2) C (15)-C (16)-C (17)-C (13) 17.7 (2) C (6)-C (5)-C (10)-C (19) 55.8 (3) C (13)-C (17)-C (20)-C (22) C (4)-C (5)-C (10)-C (19)-126. 3 (2)-177.80 (17) C (6)-C (5)-C (10)-C (1)-125. 7 (2) C (16)-C (17)-C (20)-C (22) 60.4 (2) C (4)-C (5)-C (10)-C (1) 52.1 (2) C (13)-C (17)-C (20)-C (21)-55.2 (2) O (1)-C (1)-C (10)-C (19) 5.6 (3) C (16)-C (17)-C (20)-C (21) C (2)-C (1)-C (10)-C (19) 125.6 (2)-176.91 (17) O (1)-C (1)-C (10)-C (5)-172.83 (15) C (21)-C (20)-C (22)-C (23) 116.8 (2) C (2)-C (1)-C (10)-C (5)-52.9 (2) C (17)-C (20)-C (22)-C (23)-117.7 (2) C (8)-C (9)-C (11)-C (12) 50.6 (3) C (20)-C (22)-C (23)-C (24) 177.65 (18) C (9)-C (11)-C (12)-C (13)-54. 9 (3) C (22)-C (23)-C (24)-0 (3) 117.2 (2) C (11)-C (12)-C (13)-C (18)-65. 8 (2) C (22)-C (23)-C (24)-C (28)-2.8 (3) C (11)-C (12)-C (13)-C (14) 56.4 (2) C (22)-C (23)-C (24)-C (25)-129.9 (2) C (11)-C (12)-C (13)-C (17) 166.37 (17) 0 (3)-C (24)-C (25)-C (26)-53.6 (2) C (7)-C (8)-C (14)-C (15)-2. 1 (3) C (28)-C (24)-C (25)-C (26) 65.2 (2) C (9)-C (8)-C (14)-C (15)-179. 81 (19) C (23)-C (24)-C (25)-C (26) C (7)-C (8)-C (14)-C (13)-126.6 (2)-167.32 (18) C (9)-C (8)-C (14)-C (13) 55.7 (2) 0 (3)-C (24)-C (25)-C (27)-179.52 (17) C (12)-C (13)-C (14)-C (8)-58.2 (2) C (28)-C (24)-C (25)-C (27)-60.8 (2) C (18)-C (13)-C (14)-C (8) 63.9 (2) C (23)-C (24)-C (25)-C (27) 66.7 (2) C (17)-C (13)-C (14)-C (8)-179.21 (17) Table 7. Hydrogen bonds for la, 24 (S)-dihydroxyvitamin D2 [A and °].

D-H... A d (D-H) d (H... A) d (D... A) < (DHA) O (1)-H (1O)...O(1S) 0.94 1.79 2.714 (7) 169.7 O (1)-H (1O)...O(1S') 0.94 1.82 2.715 (7) 159.1 0 (2)-H (20)... 0 (3) &num 1 0.99 1.77 2.7009 (18) 154.8 0 (3)-H (30)... 0 (1) #2 0.93 1.88 2.764 (2) 158.0 O(1S)-H(1S)... 0 (2) #3 0.84 1.96 2.783 (7) 167.7 O(1S')-H(1S')... 0 (2) #3 0.84 1.99 2.759 (6) 151.7 Symmetry transformations used to generate equivalent atoms: &num lx, y+l, z+1 &num 2x, y, z-1 &num 3x, y-l, z