STUDIESON CHOLESTERYL OXIDES
Oct., 1942 [CONTRIBUTION FROM
THE
231i
RESEARCE LABORATORIES, THE UPJOHNCOMPANY, KALAMAZOO, MICHIGAN J
Studies on Cholesteryl Oxides’ BY
PURNENDU
h.’. CHAKRAVORTY AND ROBERT H. LEVIN
In connection with work in progress, i t became necessary t o study the properties of the oxide ring in cholesteryl-~5,6-oxides. Previous investigators have used perbenzoic acid to prepare the various oxido derivatives of cholesterol, and from their results the nature of the substituent a t CSappeared to influence the ratio of stereoisomeric products formed. Thus from cholesterol Westphalen and Windaus2 obtained a-cholesterol oxide (I, R = H) in 50% yield, with some of the @-isomer(11, R = H) being isolable from the crystallization mother liquors. Ruzicka and Bosshard3 found that treatment of cholesteryl acetate with perbenzoic acid produced largely P-cholesteryl oxide acetate. Spring and Swain4 reported that cholesteryl benzoate, under the same conditions, formed almost equal amounts of a- and @-cholesteryl oxide benzoates in excellent yield. In this Laboratory, preparation of oxides has been simplified by the use’ of monoperphthalic acid.6 This reagent is readily prepared and is more stable than perbenzoic acid. Further, the phthalic acid which is formed from the perphthalic acid during the reaction can be separated easily from the sterol oxides because of its insolubility in chloroform. Such ease of separation is particularly useful in the preparation of bile acid oxides because it eliminates the need for washing with alkali. However, when using perphthalic acid several precautions are necessary. The chloroform should be freshly distilled over potassium carbonate, since the sterol oxides will react with even traces of hydrogen chloride. In some of our experiments, the perphthalic acid-cholesteryl compound reaction mixture was not worked up immediately but allowed to stand overnight or longer. In these instances the reaction product was quite sticky and uncrystallizable, and the yield was small. This was presumably due to a secondary reaction of the cholesteryl oxide with phthalic acid, giving, very probably, a half phthalic ester as was indicated by the solubility of the product in alkali. (1) Presented in part before the Organic Division of the American Chemical Society, Memphis, April 20-24, 1942. (2) Westphalen and Windaus, Ber., 48, 1064 (1915). (3) Riizicka and Bosshard, H c h . Chim. A d a , 20, 244 (1937) (4) Spring and Swain, J . Chcm. SOC.,1356 (1939). (5) For preparation of the reagent see B&hme,B e . , 70,379 (1987) ; “OrgPnic Syntheses,” 10,70 (1940).
When cholesteryl acetate was treated with perphthalic acid, P-cholesteryl oxide acetate (11, R = Ac) was isolated in 50% yield and the (Y
0 ’ 116
16
R
=
H, Ac, Rz
isomer (I, R = Ac) in 20% yield. A similar reaction using cholesteryl benzoate gave a-cholesteryl oxide benzoate (I, R = Bz) in 50 t o 70% yield, but no @-oxidebenzoate could be obtained. With cholesterol, the same reagent produced CYcholesterol oxide (I, R = H) in 60% yield, and a small amount of the @-isomer. Several partially successful attempts to rearrange cholesteryl oxides t o ketocholestane compounds have been reported. Chinaeva and Ushakov’ used a Grignard reagent. Spring and Swain4 employed heat and vigorous dehydrating reagents. In neither instance was the yield of 6ketocholestanol satisfactory. It was thought that any tendency of the oxido group t o isomerize into a ketone might be accelerated through a mass action effect by removing the product from the sphere of reaction as a ketone derivative. Accordingly, 8-cholesteryl oxide acetate was heated with semicarbazide hydrochloride in pyridine. The product of this reaction was a halogen containing substance, m. p. 188-190°, which could be recovered unchanged after refluxing with acetic anhydride, and was therefore identified as a 3-acetoxy.5-hydroxy-6-chlorocholestane(111, R = Ac) instead of 3-acetoxy-5-chloro-6-hydroxycholestane (IV, R = Ac, R’ = H). A rational explanation of this reaction became possible when it was found I
I11
IV
(6) Throughout this paper the steric configurations assigned e t C I and CS are purely arbitrary. (7) Chinaeva and Ushekov, J. Gcri. Ckcm. (E.S. S.R . ) , 11, 335 (1941); C. A., 81, 5903.
2318
PLTNENDU K.CHAKRAVOKTY AND ROBERT H. LEVIN
Vol. A4
that a solution of semicarbazide hydrochloride in oxide linkage followed by addition of fragments of pyridine has a pH of 6.2 and will turn blue litmus the benzoyl chloride molecule, then in our reacred. Since ethylene oxides are, in general, stable tion, as well as Spring and Swain’s reaction with to alkaline reagents, but susceptible to acidic sub- the a-oxide and its benzoate, the 5-benzoxy comstances, this transformation is not surprising. pound would be expected to form.I2 Since the 5It has been known for some time that ethylene hydroxy derivative was obtained, it seems that the oxide rings in aliphatic molecules are also opened benzoyl chloride acts like the semicarbazide hyby heavy metal and even alkaline earth metal drochloride, that is, merely as a donor of hydrogen halides.8 In an extension of this reaction to cho- chloride. Formation of 6-benzoxy compounds lesteryl oxides, ferric chloride, zinc chloride and can be explained on the assumption that the primagnesium bromidey were found to react with p- mary reaction is fission by hydrogen chloride, cholesteryl oxide acetate to give compounds con- followed by benzoylation of the newly formed titaining halogen. The ferric chloride reaction hydroxy group with excess of the reagent. product was identified as 3-acetoxy-5-hydroxy-& It was thought that pyridine might function as chlorocholestane (111, R = 4c). The zinc chlo a carrier of hydrogen chloride in these experiride and magnesium bromide products were not ments. Therefore, the reaction of “pyridine hycompletely characterized. drochloride” with /3-cholesteryl oxide acetate was As indicated above, cleavage of the cholesteryl undertaken. This reagent was obtained as a hyoxide bond can produce two position isomers groscopic white solid by passing dry hydrogen (111, IV). Recently some conflicting reports chloride into a solution of pyridine in ether. In have been published concerning the orientation petroleum ether there was no reaction, “pyridine of products obtained by the action of certain hydrochloride” being insoluble in this solvent. acidic reagents on various cholesteryl oxide de- In absolute ethanol a reaction readily took place, rivatives. Ruzicka and Bosshardj treated the 0- with the formation of the same 3-acetoxychlorooxide acetate with dry hydrogen chloride in pyri- hydrin (111) previously obtained with semicarbadine and obtained 3-acetoxy-3-hydroxy-ti-chloro-zide hydrochloride, ferric chloride and benzoyl cholestane (111, R = Ac). Hattori‘” reported chloride. The reaction of 0-cholesteryl oxide acetate with this same compound resulting from the action of benzoyl chloride was also carried out under anhydrogen chloride on a-cholesteryl oxide acetate, hydrous conditions using carbon tetrachloride as and claimed that very pure &oxide acetate“ gave a 3-chloro-6-hydroxycholestanederivative (IV, a solvent. The substitution of a completely K = Xc, R’ = H). Spring and Swain4 have chlorinated solvent for pyridine, which contains found that a-cholesterol oxide and its benzoate hydrogen, did not alter the course of the reaction, react with benzoyl chloride in pyridine or with hy- and the same 5-hydroxy-6-chlorocholestanededrogen chloride to give 5-hydroxy-6-chlorochol- rivative was isolated from the reaction mixture estane compounds (111), whereas 0-cholesterol although in a poorer yield. Apparently the eleoxide and its benzoate form 5-chloro-&hydroxy ments of hydrogen chloride are obtained from the derivatives (IV, R = Bz, K’ = Bz, H ) . We benzoyl chloride, but the mechanism of the transtreated 6-cholesteryl oxide acetate, In. p. 113’, formation remains quite obscure. The reaction with “pyridine hydrochloride” was with benzoyl chloride in pyridine and obtained 3acetoxy-5-hydroxy-6-chlorocholestane (111, R = extended to a-cholesterol oxide, its acetate, and Xc), which was characterized by mixed melting benzoate; and 0-cholesterol oxide. These subpoint, analysis, and its non-reactivity toward stances, to our surprise, uniformly gave 5-hyacetic anhydride. If the action of benzoyl chlo- droxy-6-chlorocholestanederivatives (111) in good ride involves simply a cleavage of the cholesteryl yields. “Pyridine hydrochloride’ ’ is a convenient reagent to use. It is easily prepared, may be kept (8) Bodforss, “Die Athylenoxyde,” Sammlung chcm chrm. frchn Vortroge, Ferdinand Enke, Stuttgart, 1920 indefinitely in a desiccator, reacts smoothly, and (9) Ushakov and Madae%a,J. Gcn Chcm. (U S S R ) 9, 1690 any excess is readily disposed of. In the case of (1939). treated cholesterol oxide with magnesium bromide, but did not isolate any compound containing halogen a- and @-cholesteroloxides, some trouble was ex(10) Hattori J . Pharm Soc Japan, 60, 334 (1940) perienced in purifying the reaction products bei l l ) &$tori’s vdue for the m p of 8-cholesterol oxldc, 136”,is at 71
variance with all other published values and our own findings wkeh give a m p of 106-107° for the pure substance However, hi5 oxide Acetate h a s the nrreptrd m p I I 2 11.3
(12) See for example Ehrenstein’s acetolysis of the stereoisomeric 5,fj-arides of dehydroandrosterone acetate, J Ore Chem , 6 , 62b ‘ 1 9 4 1 ) ~ also Hattor1,wf 1 0
Oct., 1942
STUDIES ON CHOLESTERYL OXIDES
2319
cause of dehalogenation occurring during recrystallization. The difficulty was obviated by benzoylating the crude dry reaction product and then crystallizing. Apparently the presence of a free hydroxyl group at CBlabilizes the halogen a t CS,forming possibly a 3,6-oxide. From the results of previous investigators it appeared that the configuration of the oxide and the nature of the substituent a t CS affected the orientation of the chlorohydrin produced by VI16 fission of the oxide ring. Our experiments, however, indicated that the reagent might become a which, on losing hydrogen chloride, gave an oxide dominant factor in determining the course of the of the opposite configuration (VII). Windausl' reaction. It became desirable t o check Spring started with an a-oxide and obtained a 5-chloro-6and Swain's experiments with benzoyl chloride in hydroxycholestane derivative which lost halogen pyridine. Using a-cholesterol oxide, its acetate acid and formed a 6-oxide. Furthermore, we have and benzoate, and /%cholesterol oxide, the prod- obtained the same chlorohydrin from oxides of ucts again were derivatives of 5-hydroxy-6-chloro- different configuration. The current theory states cholestane (111). I n the reaction of p-choles- that in the cleavage of an oxide linkage a transterol oxide, a new substance was obtained to- compound is formed.l6 However, if an a-oxide gether with 3-benzoxy-5-hydroxy-6-chlorocholes-(VIII) forms a trans-5-hydroxychlorohydrin (IX), and a 6-oxide (X) similarly produces a trans-5tane. The new product has a m. p. of 197-198' and gave a sharp depression with the 3-benzoxy hydroxy compound (XI), then the two will neceschlorohydrin obtained above. It was not, however, the expected dibenzoate of 3,6-dihydroxy-5chlorocholestane, since that compound has a melting point of 184'. Work is in progress t o elucidate its structure. Thus Spring and Swain obtained 5-hydroxy-6chloro compounds from the a-oxide series and 5chloro-6-hydroxy derivatives from the p-oxides, while we could isolate only one type, regardless of which oxide we used. Since the experimental procedures were basically similar, this discxepsarily be of opposite configuration a t Ca,C6. ancy, a t the moment, seems difficult to explain. Mild saponification of 3-acetoxy-5-hydroxy-6- Even assuming that the a-oxide may give a 5chlorocholestane (111, R = Ac) with sodium car- hydroxy-6-chloro derivative of a geometrical configuration different from IX, then the /3bonate gave a-cholesterol oxide, m. p. 141'. This is in accordance with the work of Spring and oxide, following the same route, would still give Swain, la who obtained a-cholesteryl oxide benzo- the opposite stereoisomer. Thus, it is necessary ate by heating 3-benzoxy-5-hydroxy-6-chloro-to postulate additional Walden inversions in one cholestane (111, R = Bz) with quinoline. Pre- case but not in the other or an entirely different viously Windaus14 had reported that the 5-chloro mechanism to explain how the same steric concompound, 3,6-dihydroxy-5-chlorocholestane,figuration is obtained by cleavage of the stereoforms p-cholesterol oxide when treated with alkali. isomeric oxides of cholesterol, or how the stereoThe stereochemical aspects of these reactions isomers are interconverted through an intermeshould be pointed out. In tEe above case, for diate chlorohydrin. Further studies on the oxides of cholesterol will example, an oxide with a p-configuration (V) be reported in a subsequent communication. combined with the elements of hydrogen chloride (16) Lucas, Schlatter and Jones, THISJOURNAL, 68, 22 (lo.&]), to form a 5-hydroxy-6-chloro compound (VI), (13) Spring and Swain, J . C h s n . Soc., 83 (1941).
(14) Windaus, 2. physiol. Chcm., 117, 154 (1921).
describe the opening of the oxide rings in the isomeric 2,3-epoxypentanes. See also Ehrenstein, J . Org. C h e m . , 4, 506 (1938); and Ehrenstein and Decker, f b i d . , 6, 544 (1940).
2320
1'1-KNENIII'
S . CHAKRAVOKTV .IND ROBERT H. LEVIN
1'01. 64
A portion (200 mg.) of the chlorohydrin, refluxed for fifteen minutes with acetic &&hydride,was recovered unMonoperphthalic acid was prepared according to the method of Bohme' and kept in the refrigerator for as long changed, hence it Was Characterized as 3-acetoxy-5hydroxy-6-chlorocholestane (111). as a month without appreciable decomposition. @Oxide Acetate and Ferric Chloride.-@-Oxide acetate p- and a-Cholesteryl Oxide Acetates.-Ten granis 10.40 9.) was treated with 1.0 g. of ferric chloride in alco(0.023 mole) of cholesteryl acetate, in. p. 112-174", dissolved in 50 cc. of ether, was mixed with 8.4 g. (0.046mole) hol. The solution was allowed t o reflux for three hours, of monoperphthalic acid irr 266 cc. of ether. The solution then concentrated t o half its volume and water added. was allowed to reflux for six hours anti the solvent removed Ether extraction and isolation in the usual manner gave a hy distillation. The residue was dried in OUGUO and chlorohydrin which was purified by repeated crystallization digested with 260 cc. of chloroform dried over potassium from dilute methanol and dilute acetone. The final prodcarbonate. Filtration gave a residue of 6.7 g. (8770 re- uct melted at 183-185', and was recovered unchanged covery) of phthalic acid and a clear colorless solution which after refluxing with acetic anhydride. There was no dewas taken to dryness in vacuo. The residue was crystal- pression of the m. p. on admixture with authentic 3lized from 30 cc. of methanol, giving 6.0 g. (58%) of @- acetoxy-5-hydroxy-6-chlorocholestane. @-Oxide Acetate and Benzoyl Chloride. 1. In Pyricholesteryl oxide acetate, which on recrystallization from a solution of 0.45 g. of P-cholesteryl oxide methanol gave the pure product, ni. p. 111-112', \ale51) dine.-To -21.8"; mixed In. p. with cholesteryl acetate, 97--101'. acetate in 5 cc. of dry pyridine was added 0.5 cc. of benzoyl Concentratioii of the filtrate gave 1.55 g. (13%) of a- chloride. The mixture was allowed to stand at room cholesteryl oxide acetate. The a-isomer was purified by temperature for one-half hour, then heated on the steamcrystallization from ethanol and found to have a ni. p. of bath for the same period. The solution was poured into 101-103", [ a ] " D -44.6". The m. p. is 8" higher thair ice water and the semi-crystalline precipitate separated by previously reported values.2*10However, it gave a de- decantation. Crystallization from dilute methanol gave pression (m. p. 88-104°) when mixed with pure @-oxide 0.42 g. of chlorohydrin, m. p. 150-161°. Two recrystalacetate. Saponification with methanolic potassium hy- lizations from dilute acetone gave 0.25 g. of 3-acetoxy-5droxide and crystallization from dilute alcohol gave a- hydroxy-6-chlorocholestane,m. p. 183-185 ', which showed no depression of the m. p. on admixture with the product cholesterol oxide, m. p. 141-143". [cY]'~D -44:.5".4 In several runs the reaction mixture was left over the of the semicarbazide hydrochloride reaction. And. Calcd. for CaHleOsCI: C, 72.4; H, 10.3; C1, week-end. The yield of oxides in these instances was very poor and a considerable amount of alkali soluble substance 7.38. Found: C, 72.3; H, 10.7, C1, 7.26. The compound was recovered unchanged after refluxing was formed. Since this product remained dissolved in alkali after saponification, it is assumed to be a &half- with acetic anhydride. 2. In Carbon Tetrachloride.-A solution of 0.60 g . of phthalate derivative of cholestanr. a-Cholesteryl oxide benzoate was sinrilarly prepared. @-oxideacetate in 30 cc. of carbon tetrachloride distilled Cholesteryl benzoate (16.7 g., 0.034 mole) reacting with over potassium carbonate was treated with 0.50 cc. of perphthalic acid (12.7 g., 0.07 mole) gave 8.7 g. (50%) of benzoyl chloride under anhydrous conditions. The solua-cholesteryl oxide benzoate, m. p. 164.-166", [ a l Z b ~tion was allowed t o reflux for three hours, then kept a t -B.0.4 No @-oxidebenzoate could be isolated. room temperature for thirty-six hours. After distilling &!holesterol Oxide.-In the same manner, 5 g. (0.013 the solvent, the excess benzoyl chloride was removed in mole) of cholesterol (Wilson) and 4.7 g. (0.026 mole) of vacuo. Crystallization from dilute methanol gave a prodA perphthalic acid gave 3.1 g. (61%) of a-cholesterol oxide, uct which sintered at 100' and melted at 135-155'. rn. p. 141-143'. A small amount of the B-isomer was Beilstein test for halogen was positive. Four crystallizations from dilute acetone gave colorless needles of 3obtained from the mother liquors. @-Cholesterol oxide, ni. p. 105-107°, ( a ) 2 6-12.7, ~ was acetoxy-5-hydroxy-6-chlorocholestane,m. p. 182-185' There was no depression of the m. p. on admixture with an also obtained by saponification of the &oxide acetate. Reaction of P-Cholesteryl Oxide Acetate with Semi- authentic sample. carbazide Hydrochloride.--A solution of 2.3 g. of 6 Anal. Calcd. for C;pHloO&l: C, 72.4; H, 10.3; C1, cholesteryl oxide acetate in 10 cc. of pyridine was treated 7.38. Found: C, 72.3; H, 10.2; C1, 7.17. with 2.0 g. of semicarbazide hydrochloride and warmed or1 @-Oxide Acetate and "Pyridine Hydrochloride." the steam-bath for six hours. The solution was concen- "Pyridine Hydrochloride9' was prepared by passing dry trated to half its volume and poured over crushed ice. hydrogen chloride into a solution of pyridine in ether and The precipitate was separated by filtration, digested with allowing the saturated solution t o stand for one-half hour. methanol, and filtered while hot to remove a small amount The white solid was separated by filtration and found to of disemicarbazide. Concentration and cooling gave color- be very hygroscopic. It is insoluble in dioxane and less needle-like crystals, m. p. 183-186", which gave a petroleum ether and soluble in alcohol. Crystallization strong Beilstein test. These were combined with crystals from absolute alcohol-ether gave colorless platelets, m. p. : recovered from the mother liquor and recrystallized twice s. 81; m. 141-146". It can be kept for months in a desicfrom dilute acetone, giving 0.8 g. of cholesteryl acetate cator over calcium chloride with no evidence of decomposichlorohydrin, m. p. 187.5-189.5". tion. Anal. Cakd. for CzoH,eO,Cl: C, 72.4; H, 10.3; C1, To a solution of 1.0 g. of 6-cholesteryl oxide acetate in 30 7.38. Found: C, 72.2: H. 10.1; C1, 7.31. cc. of absolute ethanol was added 1.0 g of pyridine hydrochloride The solution wa5 allowed to reflux for fifteen (16) Microanalyses by &I.Emerson and q'. Struck.
Experimental16
Oct., 1942
STUDIES ON CHOLESTERYL OXIDES
2321
minutes, then concentrated to half its volume and the Anal. Calcd. for Qs&d&C1: C, 72.4; H, 10.3. sterol precipitated by the addition of water. Crystallii- Found: C, 72.6; H , 10.1. tion from dilute acetone gave 0.85 g. (77%) of crude 3Starting with 0.50 g. of a-oxide benzoate4 and 0.50 cc. acetoxy-5-hydroxy-6-chlorocholestane, m. p. 148-168". A of benzoyl chloride, 0.45 g. (84%) of 3-benzoxy-5-hydroxyBeilstein test was positive. After several crystallizations 6-chlorocholestane, m. p. 191-195", was obtained. Refrom dilute acetone, colorless needles. m. p. 184-1S6° were crystallization from dilute acetone gave a pure product, m. obtained. The product was recovered unchanged after p. 197-198.5'. A mixed m. p. with authentic benzoate treatment with acetic anhydride, arid gave no depression chlorohydrin gave no depression. in m. p. on admixture with authentic 3-acetoxy-5-hydroxyAnal. Calcd. for C J ~ H J I O ~ CC,~ :75.2; H, 9.46; C1, 6-chlorocholestane. 6.53. Found: C, 75.0; H, 9.29; CL6.46. Reaction of Pyridine Hydrochloride with the Other Using 0.15 g. of wcholesterol oxide and 0.20 cc. of Oxides of Cholesterol.-a-Cholesteryl oxide benzoate is benzoyl chloride, the benzoate chlorohydrin was obtained sparingly soluble in alcohol, so it was necessary t o dissolve iu fair yield. After several recrystallizations its m. p. it in benzene. One gram of the oxide in benzene was and mixed m. p. with authentic 3-benzoxy-5-hydroxy-6mixed with 2.0 g. of pyridine hydrochloride in ethanol. chlorocholestane was 193-196 '. The solution was refluxed for fifteen minutes, the solvent The reaction of 8-cholesterol oxide with benzoyl evaporated, and the reaction product crystallized from chloride waa also carried out according t o Spring and methanol as a halogen containing compound, m. p. 193Swain.' The results were, however, different, so the de196'. Recrystallization from ethyl acetate-methanol, tails of the experiment are given. The fl+xide (0.30 g.) in and dilute acetone gave 0.60 g. (56%) of 3-benzoxy-5- 4 cc. of dry pyridine was warmed for an hour with 0.60 cc. hydroxy-6-chlorocholestane,m. p. 196-198'. It was re- of benzoyl chloride. The solution was poured into icecovered unchanged after rduxing with acetic anhydride. water and the precipitate separated by filtration. The Anal. Calcd. for C~Hp0&!1: C, 75.2; H, 9.46; C1, residue was crystallized from 30 cc. of acetone-methanol, 6.53. Found: C, 75.0; H, 9.40; C1,6.33. giving 0.15 g. of 3-benzoxy-5-hydroxy-6-chlorochol~e, a-Cholesteryl oxide acetate (0.80 g.) and 1.0 g. of m. p. 190-193 Recrystallizations from acetone-methpyridine hydrochloride in alcohol gave 0.65 g. (74%) of 3- anol and dilute acetone brought the m. p. up t o 195-197". acetoxy-5-hydroxy-6-chlorocholestane, m. p., and mixed Anal. Calcd. for CwHst08C1: C, 75.2; H, 9.46; C1, rn. p,, 185-157". It was recovered unchanged after treat- 6.63, Found: C, 74.9; H , 9.50; C1, 6.47. ment with acetic anhydride. Concentration of the first mother liquor gave 0.14 g. of a a- and @-Cholesterol oxides also reacted smoothly with crystalline compound, m. p. 176-182'. It was recrystalpyridine hydrochloride. However, the product lost halo- lized twice from acetone-methanol, in which it seemed t o gen easily, thereby hindering its purification. Thus 0.30 g. be more soluble than the previously obtained benzoate of the a-oxide gave 0.25 g. of crude chlorohydrin, m. p. chlorohydrin. It had a m. p. of 197-198", and its mixed 154-160 '; Beilstein test positive. Recrystallization from m. p. with authentic 3-benzoxy-5-hydroxy-6-chloromethanol-acetone gave feathery needles, m. p. 99-105". cholestane was 179-186'. This compound is not 3,6-diThis product no longer contained halogen. benzoxy-5-chlorocholestane. Benzoylation of the crude product gave an easily pur& Anal. Calcd. for 3,0-dibenzoxy-5-chlorocholestane, able substance. Using 0.45 g. of a-oxide and 1.0 g. of m. p. l a ' , CllHMOdCl: C1, 5.48. Found: C1, 7.01,7.11. pyridine hydrochloride, a chlorohydrin was obtained. I t *Oxide from 5-Hydroxy-6-chlorocholestane Derivawas dried and benzoylated with 0.50 cc. of benzoyl chloride tives.-A solution of 0.7 g. of 3-acetoxy-5-hydroxy-0in pyridine. The product was isolated in the usual way and crystallized from ethyl acetate-methanol, giving 0.35 chlorocholestane in 30 cc. of dilute alcohol was treated g. (58%) of 3-benzoxy-5-hydroxy-6-chlorocholestane, m. p. with 10 cc. of 2 N sodium carbonate and warmed on the 193-196 '. Recrystallization from dilute acetone, and steam-bath for thirty minutes. The addition of water again from ethyl acetate-methanol, gave colorless needles, gave a precipitate of 0.40 g. of a-cholesteryl oxide which m. p. 196-198", which showed no depression on admixture was recrystallized twice from ethyl acetate, m. p. 141143'. A mixed m. p. with an authentic sample of a-oxide with an authentic sample. Similarly, from 0.40 g. of the ,&oxide there was obtained showed no depression. 0.28 g. (52%) of 3-benzoxy-5-hydroxy-6-chlorocholestane, A similar reaction was carried out with 3-benzoxy-5m. p. 192-195". When mixed with a n authentic sample hydroxy-6-chlorocholestane and methanolic potassium hydroxide, again giving a-cholesterol oxide, m. p. 141-143'. the m. p. was not depressed. Reaction of Benzoyl Chloride with the Other Oxides Summary of Cholesterol.-Reactions with benzoyl chloride did not a- and @-cholesteroloxide, a- and j3-cholesteryl proceed as smoothly as the corresponding reactions with pyridine hydrochloride and the products were less pure. oxide acetate and a-cholesteryl oxide benzoate have been conveniently prepared through the acTreatment of 0.70 g. of a-oxide acetate with 0.60 cc. of benzoyl chloride gave 0.42 g. (55%) of a product, m. p. tion of perphthalic acid on cholesterol and its de160-170". After heating with acetic anhydride and re- rivatives. crystallizing twice from dilute acetone, 3-acetoxy-5Semicarbazide hydrochloride and ferric chlohydroxy-6-chlorocholestane was obtained as colorless ride react with P-cholesteryl oxide acetate to form needles, m. p. 178-183'. A mixed m. p. with authentio 3-~e~oxy-5-hydroxy.6-~~o€~~l~~tane. acetoxychkmhydrin, however, wm 179-184'.
'.
2322
ELLISMILLER,FRANK S . CROSSLEYAND MAURICE L. MOORE
All five of the cholesteryl oxides prepared react with "pyridine hydrochloride" and benzoyl chloride to give 5-hydroxy-6-chlorocholestanederivatives. 8-Cholesterol oxide, as its acetate, has been con-
--
Vol. 61
verted to a-cholesterol oxide through the 5-hydroxy-6-chloro compound. The stereochemical implications of these reactions have been pointed out. KALA~IAZOO, MICHIGAK RECEIVED JULY 17, 1942 .--.
I -
-- - - __
[CONTRlHUTION FROM THE MEDICAL-RESEARCH n I V I S I O N . SHARP ASD D O H M E , INC ]
Synthesis of p-Hydroxyphenyl Amyl Sulfide B Y ELLISM I L L E R .
FRANK
s. CROSSLEY
.4XD h f A U R I C E
L.
MOORE
A previous publication from these Laboratories' experimental conditions for the reaction has shown confirmed the report2 that the hydroxyphenyl that it can be used to prepare the compound in alkyl sulfides were more powerful in their bac- yields of 25-30%',. tericidal activity than the corresponding alkylThe diazotization was carried out a t a temperaphenols. These results prompted us to seek a prac- ture of 0-10' and the diazothio ether decomposed tical method for the preparation of p-hydroxy- by heating a t 60". The resulting product was disphenyl n-amyl sulfide, which had been shown to tilled and gave a yield of 5040% of a material possess maximum activity for the compounds in which solidified a t room temperature. However, this series. on crystallizing the material from solvent naphtha The hydroxydiphenyl sulfides have been pre- the yield of p-hydroxyphenyl n-amyl sulfide pared by Hilbert and Johnson3 by use of the Zie- dropped to 25-30%. Careful fractionation of the gler4 reaction between diazotized anisidine and reaction mixture disclosed that two definite prodthiophenol. The diazothio ether thus formed ucts were formed during the reaction and perbreaks down a t 70" to give a methoxydiphenyl mitted the isolation of the second product with sulfide which is converted into the desired product its purification. The product was identified by by dealkylation. They reported that an attempt preparation of derivatives and synthesis as di-nt o prepare p-hydroxydiphenyl sulfide by treating amyl disulfide. The reaction has been applied to diazotized p-aminophenol with sodium thiophen- n- and isoamyl mercaptan with comparable reate was unsuccessful. Suter and Hansen? treated sults. diazotized anisidine with potassium ethyl xanExperimental Part thate and decomposed the intermediate diazonium p-Hydroxyphenyl n-Amyl Sulfide.-fi-Aminophenol, 21.8 ethyl xanthate to obtain p-methoxythiophenol g. ( 0 2 mole), was dissolved in 110 cc. of 4 N hydrochloric which then reacted with amyl bromide and the acid ( 0 44 mole) and diazotized in the usual manner by the product was dealkylated t o give p-hydroxyphenyl slow addition of a solution of 15 g. (0.23 mole) of sodium n-amyl sulfide. They also reported an unsuccess- nitrite in 30 cc. of water. The mixture was stirred until ful attempt to combine the sodium salt of butane- the diazotization was complete (starch-iodide paper test) and the deep purple solution of P-hydroxyphenyldiazonium thiol- 1 with diazotized p-aminophenol. chloride rapidly filtered from a small amount of insoluble I n the previous publication, l a more satisfactory material. The solution was then added slowly to a cold procedure was developed which involved the (10") sludge of 24 g. (0.23 mole) of n-amyl mercaptan in 75 synthesis of thiohydroquinone and subsequent cc. of water containing 37.5 g. (0 94 mole) of sodium hyreaction of this with amyl bromide to give the de- droxide. The temperature was maintained a t 10" throughsired p-hydroxyphenyl n-amyl sulfide The suc- out the addition and frothing was controlled by the addition of small amounts of n-butanol. After the reaction cess of the thiohydroquinone synthesis suggested had subsided, the cooling bath was removed and stirring that diazotized p-aminophenol should react with continued at room temperature until the diazonium salt the sodium salt of n-amyl mercaptan, even though disappeared (R-salt test). The mixture was then heated previous workers2p3had been unsuccessful with to 60" to complete the decomposition of the diazothio similar reactions. An extensive study of possible ether and allowed to stand overnight a t room temperature, (1) Miller and Read, THISJOURNAL, 66, 1224 (1933). (2) Suter and Hansen, dbrd , 64, 4100 (1932) (3) Hilbert and Johnson,ibid., 61, 1526 (1929) ( 4 ) Ziegler, Re, , 23, 2469 (18W)
after which it was diluted with five volumes of water, acidified with concentrated hydrochloric acid and extracted with 500 cc. of toluene. The extract was washed three times with 500-cc. portions of water, dried over anhydcous
