Reaction of some ketosteroids with thioglycolic acid

(1) Using thioglycolic acid, five mercaptols of 3-ketosteroids have been synthesized sine catalyst. Two of these mercaptols afforded crystalline methyl esters. (2) The reaction of 20-ketoseroids with thioglycolic acid has been investigated. Formation of 20-mercaptols ensues only when zinc chloride and sodium sulfate are employed as catalysts. (3) The date indicated that 17-ketosteroid compounds condense with thioglycolic acid when zinc chloride and sodium sulfate are present in the reaction mixture. (4) The addition of thioglycolic acid to the 5,6-double bond of dehydroepiandrosterone (XLIII) has been described. This reaction proceeds in a mode contrary to the rule of Markownikoff. The position of the thioether linkage in the addition product has been established by oxidation studies. The same reaction was applied to two other 5,6-unsaturated steroids. (5) The synthesis of ?5-3-carbomethoxymethylmercaptocholestene has been described. (6) The use of steroid mercaptols derived from thioglycolic acid as derivatives for characterization and separation purposes has been discussed.

Approved: REACTION OF SOlVIE KETOSTEROIDS WITH THIOGLYCOLIC ACID 'by Sylvester J. Sanfilippo A thesis submitted to the faculty of the University of Utah in partial fulfillment of the requirements for the degree of Master of Science. Department of Biological Chemistry Chairman, Thesis Committee Dean.&zu~ University of Utah 1951 The author wishes to express his appreciation to Dr. Hans Reich for his generous assistance and guidance during the course of this work. ii TABLE OF CONTENTS Section INTRODUCTION • • • • • • • • • • • • • • • • • • • REVIE\1\f OF THE LITERATURE • • • • • • • • • • • • • Reaction of Steroid Aldehydes with Mercaptans Reaction of Steroid Ketones with Mercaptans • Reaction of Carbonyl Compounds with Thiog1y-colic Acid • • • • • • • • • • • • • • • • Reaction of Steroid Ketones with Thioglycolic Acid • • • • • • • • • • • • • • • • • • • STATEMJi:NT OF THE PROBLEM • • • • • • • • • • • • • DISCUSSION OF RESULTS • • • • • • • • • • • • • • Reaction of 3-Keto~teroids and Thioglycolic Ac id • • • • • • • • • • • • • • • • • • • Reaction of 20-Ketosteroids and Thioglycolic Acid • • • • • • • • • • • • • • • • • • • Reaction of 17~Ketosteroids and Thioglyco1ic Acid • • • • • • • • • • • • • • • • • • • Reaction of A5-Unsaturated Steroids and Thio­glycolic Ac1d • • • • • • • • • • • • • ~ SynthesiS of A~-j-CarbomethoxyIDethylmercapto­ch61estene • • • • • • • ~ ~ • ~ • • • • • Tables of Derivatives •••••••••••• EXPERIAffiNTAL DETAILS • • • • • • • • • • • • • • • Reaction of j-Ketosteroids and Thioglycolic Page 1 3 ~ 13 14 16 17 17 18 20 21 28 29 31 Acid • • • • • • • • • • • • • • • • • • • 31 Reaction of 20~Ketosteroids and Thioglyco1ic Ariid • • • • • • • • • • • • • • • • • • • 34 Reaction of 17~Ketosteroids and Thioglycollc Acid • • • • • • • • • • • • • • • • • • • 38 Reaction of A5-Unsaturated Steroids and Thio-glycolic Acid • • • • • • • • • • • • • • 40 Synthesis of A5-.3-Carbomethoxymethylmercapto-cholestene • • • • • • • • • • • • • • • • 46 SUM::MARY • • • • • • • • • • • • BIBLIOGRAPHY • • • • • • • • • • iii • • • • • • • • • • • • • • • • • • 48 49 INTRODUCTION Aldehydes and ketones are known to react with mer­captans with the formation of thioacetals which are usually called mercaptals when derived from aldehydes and mercaptols when derived from ketones. The reactions proceed according to the following equations: RCEO + 2R'SH + 2R'SH RCR::::: SR r SRI + + The above condensations are generally effected by the action of zinc chloride or fuming hydrochloric acid on a mixture of the reactants (1). During the past decade, the reaction of carbonyl compounds with mercaptans has received the attention of steroid chemists. The mercaptans most often reported in their inves-tigations were methyl, ethyl, phenyl and benzyl mercaptan and also ethanedithiol. The following section is a resume of the pertinent literature concerning steroid mercaptals, mercaptols and thioenolethers. The experiments per.forrned as part of this work deal with the preparation of mercaptols resulting from the reaction of steroid ketones with the mercaptan, thioglycolic acid, and are reported in a later section. Significant emphasis will also be centered on the formation of thioethers by the - 1 - - 2 - addition of thioglycolic acid to the 5, b-double bond of un­saturated ketosteroids. REVIEVJ OF THE LITEHATURE Reaction of Steroid Aldehydes with Mercaptans In the steroid series, the reaction of mercaptans with carbonyl compounds has been applied to aldehydes as well as to the ketones. Schindler et ale observed that ~5_pregnen- 3~-ol-20-one-2l-al (I) reacted with ethyl mercaptan and hydrogen chloride to give the diethylmercaptal II. These workers pointed out that the formation of the mercaptal II was an effective means of blocking the 21-aldehyde group, thus making possible the selective oxidation of the 3-hydroxy group by the Oppenauer method (2). CHO c8:0 HO d) -"------Jl-> I Koechlin and Reichstein described the conversion of strophanthidin (III), a cardiac aglycone, to its dimethyl-mercaptal IV in the presence of zinc chloride and sodium sulfate (3). Alternatively, by reacting strophanthldin (III) with ethanedithiol and hydrogen chloride catalyst, Speiser prepared the corresponding ethylenemercaptal V. The action of Raney nickel on the mercaptal V effected the reduction of the IO-mercaptal group to a IO-methyl group thereby yielding another cardiac aglycone, periplogenin (VI)(4). - 3 - - 4 - /~:o HO OH III OH IV HO OH VI From the half-aldehyde VII obtained from equilenin, Heer and Miescher synthesized the dibenzylmercaptal VIII using benzyl mercaptan. The catalytic agents for this reaction were zinc chloride and sodium sulfate. Raney nickel reduction of the mercaptal VIII resulted in the formation of a compound possessing a l4-ethyl side chain. CH VII VIII - 5 - Reaction of Steroid Ketones with Mercaptans In 1887 Mylius (6) treated dehydrocholic acid (IX) with phenyl mercaptan in the presence of hydrogen chloride and obtained a product which was the 3-dlphenylmercaptol X. The structure of this compound was proven by Hauptmann in IX ) ~s ¢S x There appears to have been no attempt to prepare steroid mercaptols until 60 years after the experiment of Mylius. A new investigation was initiated in 1946 with the work of Bernstein and Dorfman (8). Somewhat later the preparation of several mercaptols of steroid ketones was reported by Hauptmann (7), Jones et al. (9), Levin and Thompson (10), Norymberska et al. (11), Ralls et al. (12), Rosenkranz at al. (13, 14) and ~eanloz (15). The reaction ~f A4-cholesten-3-one (XI) with various mercaptans depends largely upon the amount of mercaptan used and upon the condensing agent present. In the presence of zinc chloride the sodium sulfate and using an excess of benzyl mercaptan or ethanedithiol, the mercaptol XII was obtained (7). Wllen only one mole of benzyl mercaptan was used under the same catalytic conditions, the thioenolether XIII was - 6 - formed (13). The identical thioenolether XIII was obtained by employing p-toluenesulfonic acid or pyridine hydrochloride as condensing agents (13, 14). s~ o~ C'H~ <p XIII XI XII Other A4-3-ketosteroids react essentially in the same manner. Thus, under the influence of pyridine hydro­chloride l testosterone (XIV) yielded the thioenoletl1er XV, and the sapogenin A4-22-isospirosten-3-one (XVI);:~ furnished the thioenolether XVII (14). ) XIV XVI XVII *For nomenclature of steroid sapogenins, see (16). - 7 - The reaction of the 20-keto group is determined by the choice of catalyst. Rosenkranz et ale (lIt) found that it does not react with benzyl mercaptan in the presence of pyridine hydrochloride. Under these conditions progesterone (XVIII), 17 a.-hydroxyprogesterone (XIX) and desoxycortico­sterone acetate (XX) gave o~ly the 3-monothioenolethers XXI, XXII and XXIII respectively. o~ ~ XVIII XVIV xx: R=H R=OH ~~OAc =-0 ) XXII XXIII R=H R=OH Vfhen catalyzed by zinc chloride and sodil~ sulfate, pregnen­olone acetate (Y~IV) reacted with 'ethyl mercaptan to give a mixture of the mercaptol XXV (35~&) and the thioenolether XXVI (65%) (15).- Ac.o xxv XXVI - 8 - The reactivity of AI6-20-ketosteroids with benzyl mercaptan was found to be different from that of saturated 20- ketosteroids. In the presence of pyridine hydrochloride, piperidine or p-toluenesulfonic acid, one mole of benzyl mercaptan added to the 16, 17-double bond which is known to be very reactive~'". It was assumed that the benzyl thio group is attached to the carbon atom in 16-position (14). For example, A5, 16_pregnadien-3~-ol-20-one acetate (XXVII) gave the 16-benzylthio derivative XXVIII in the presence of either pyridine hydrochloride or piperidine. However, two different products, XXX and XXXI, were obtained from A4, l6_pregnadiene- 3, 20-dione (XXIX) depending on the catalyst used (14). These reactions are depicted in the following formulae. Ac XXVII CH-S ~ <t>l. XXXI S CH~ <P 1:·3ee the reaction of the 16, 17-double bond with diazomethane (17) and sodium alkoxides (18). - 9 - In the presence of piperidine the n, ~ -unsaturated 3-keto group did not react with benzyl mercaptan, but reaction occurred when p-toluenesulfonic acid was used as a catalyst. From their experiments, Romo et al. have concluded that while thioenolether formation proceeds only in the presence of acid catalysts, the 1, 4-(or 3, 4)- addition to a sterically un­hindered a, P -unsaturated carbonyl system can occur with either acid or base. The l7-keto group of dehydroepiandrosterone (10), de­hydroepiandrosterone acetate (11) and estrone acetate (7) reacted with ethyl and benzyl mercaptan as well as with ethanedithiol with the formation of the expected mercaptol, when zinc chloride and sodium sulfate or hydrogen chloride were used as catalysts. No reaction was apparent in the presence of pyridine hydrochloride (14). Under the latter condition~ A4-androstene-3, l7-dione (L~II) yielded only the 3-monothioenolether XXXIII (14). The same derivative was obtained when one mole of benzyl mercaptan was employed in the presence of p-toluenesulfonic acid (13). "'1------. = 0 "f--~=O XXXII XXXIII - 10 - As previously mentioned, the 3-keto group in dehydro­cholic acid reacted readily with phenyl mercaptan in the presence of hydrogen chloride. The same is true for other 3- ketosteroids, since several were found to react with mercaptans in the presence of either hydrogen chloride or zinc chloride and sodium sulfate. Mercaptols were obtained from cholestan- 3-one (8, 12), 3-ketocholanic acid (8), and 3-keto-7a, l2a­dihydroxycholnnic acid (9). Keto groups in 7-and l2-positions do not react with ethyl mercaptan (8, 9), phenyl mercaptan (7, 9) and p-acetylaminophenyl mercaptan (9) under the same catalyzed conditions. With ethanedithiol, however, the keto groups in both the 7-and l2-positions were brought into reaction when hydrogen chloride was used as a catalyst. This reaction did not take place in the presence of zinc chloride and sodium sulfate (7). Ethyl dehydrocholate (XXXIV) gave a trimercaptol X~~ with ethanedithio~while ethyl 3a -hydroxy- 7, l2-diketocholanate produced the expected 7, l2-dimercaptol (7). XXXIV The reaction of the a5-7-keto group was investigated in only one case. Ralls (12) obtained the mercaptol (XXXVII) from 7-ketocholesteryl acetate (XXXVI) and ethanedithiol using hydrogen chloride as a catalyst. - 11 - AcO AcO XXXVI Usually mercaptols are split by mineral acids with formation of the parent ketone. If acetoxy groups are present in the molecule, they are also saponified under such conditions. To avoid the latter saponification, cadmium carbonate in com­bination with mercuric chloride can be used for the selective fission of the mercaptol group. Like corresponding acetals, the mercaptols are suitable derivatives for the protection of keto groups, since the latter can be easily regenerated by acid hydrolysis. Thioenolethers show essentially the same behavior toward acids as mercaptols (13, 14). On treatment with alkali, mercaptols and also thio­enolethers can be recovered unchanged. This stability is especially valuable when a free' keto group is to be reduced with lithium aluminum hydride in the presence of a mercaptol or thioenolether group. Thus, androstenedione (XXXII) was converted to testosterone (XIV) via the 3-monothioenolethers (XXXVIII and XXXIX) (13). .. 12 - ctr° -::.0 O~Ct) ,. S~ XL'CVIII XXXII C$:1- \ L", A 11-1 'I OH OH • S~ CH'),. XX.XIX ¢ XIV The most important reaction of mercapto,ls and thio­enolethers is the reduction with a special Raney nickel catalyst, which provides a mild and convenient method of transforming mercaptols to their corresponding hydrocarbons (8, 7, 11, 14). This reaction usually gives better yields than the reductions according to Clemmensen and Wolff-Kishner and is superior to these methods especially when applied to a, ~-unsaturated ketones. For example, A4-cholestene can be obtained in good yield from A4-cholesten-3-one benzylmercaptol (7), while the Wolf'f-Kishner reduction of A4-cholesten-3- one yields mainly A3-cholestene along with other reduction products (19). - 13 - Reaction of Carbonyl Compounds with Thioglycolic Acid The earliest reaction of aldehydes, l{etones and keto-acids with tl1ioglycolic acid was described by Bongartz in 1888 (20). Several aldehydes were found to react with thio-glycolic acid at room temperature without the intervention of a catalyst, while others required the addition of zinc chloride or hydrogen chloride. The tllree ketones investigated by Bongartz reacted with thioglycolic acid only in the presence of zinc chloride or when a stream of hydrogen chloride was introduced into a mixture of ketone and thioglycolic acid. Under the same conditions keto acids also formed mercaptols. Pyruvic acid (XL), however, reacted with thioglyco~ic acid without a catalyst to give an addition product XLI. In the presence of hydrogen chloride the normal mercaptol XLII was formed. CH3 CIl3 CR3 I/OH I I/" SCH2COOH ~ ...... SCH2COOH lIE C=O HCl ~ C I I ....... SCH2COOH COOH COOH COOH XLI XL XLII Cyclic ketones appear to react more readily with thioglycolic acid than aliphatic ketones. Evidence for this remark was provided by Holmberg (21) who succeeded in synthesiz+ng a mercaptol fram cyclohexanone without addition of a catalyst. - 14 - Reaction of Steroid Ketones with Thioglycolic Acid The reaction of a steroid ketone with thioglycolic acid was first reported by Levin and Thompson (10). They treated dehydroepiandrosterone (XLIII) with thioglycolic acid in the presence of zinc chloride and sodium sulfate and ob-tained the 17-mercaptol XLIV. W"O I-WOO XLIII The yield of crude mercaptol was 73%. "'1---.<.. SCH,.cOO~ SCI~f:00H XLIV By refluxing with aqueous alcoholic hydrochloric acid the parent ketone was regenerated from the mercaptol XLIV. Jones et ala (9) investigated the reaction of 3- keto-7a , 12a-dihydroxycholanic acid (XLV) with thioglyco1ic acid and found that reaction occurred when the acid was dissolved in thioglyco1ic acid and the solution kept at room temperature for 24 hours. The mercaptol thus formed was represented as XLV!. Hooce ~~S 1-400CCH:l.S XLVI vVhen dehydrocholic acid (IX) was treated with thio­glycolic acid without addition of a catalyst, only tpe keto group in the 3-position reacted (9). ~Vhether or not the keto - 15 - groups in 7- and 12-positions react with thioglycolic acid in the presence of zinc chloride or hydrogen chloride, is not known. Further experiments to determine if thioglycolic acid behaves like ethyl and phenyl mercaptan or like ethane­dithiol are definitely in order (see page 10). A4-cholesten-3-one (XI) was found to react readily with thioglycolic acid with formation of the normal mercaptol XLVII. The reaction proceeded at room temperature, when A4-cholesten-3-one was dissolved in thioglycolic acid (9) • XI .. l-i ooee Hl. S HOOCCHJ.S XLVII Upon heating with aqueous alcoholic hydrochloric acid, the mercaptol was split, and A4-cholesten-3-one recovered. The mercaptol, however, was stable to alkali. These few experiments indicate that thioglycolic acid reacts more readily with keto groups than do mercaptans (22). Furthermore, the mercaptols of thioglycolic acid are soluble in sodium hydroxide solution and for this reason can be separated from the unreacted starting materials, as will be described in the experimental part. STATEMENT OF' THE PROBLEM Methods for the separation of ketones from nOD­ketones are relatively few in number. The frequent and ex­tensive use of Girard's method is evidence that this procedure is the only adequate means of accomplishing ketone and non­ketone separation. The studies to be presented here are an attempt to determine the efficacy of thioglycolic acid as a reagent for the preparation of steroid ketone derivatives which, since they are acids, can be separated from neutral non-ketones. Moreover, this investigation attempts to interpret the practicability of these mercaptol derivatives as compounds suitable for the characterization of ketones. Since only four thioglycolic acid mercaptols of steroid ketones (keto groups in. the J-and l7-positions) have been previously prepared, the synthesis of several condensation products of steroid ketones and thioglycolic acid was under­taken. .. 16 - DISCUSSION OF RESULTS Reaction of 3-Ketosteroids and Thioglycolic Acid By repeating Jones' experiment, namely, the reaction of 44-cholesten-3-one (XI)~and thioglycolic acid, (9) the mercaptol XLVII was realized. The neutral product isolated from this reaction was identified as cholestenone and accounted for approximately 10% of the starting cholestenone. Since the mercaptol XLVII crystallized only with difficulty, it was converted to the corresponding methylester which could be purified by chromatography. This methylester showed no maximum absorption in the ultraviolet, but typical bands were noticed in the infrared. Upon sublimation in high vacuum, decomposition occurred. Testosterone (XIV) reacted with thioglycolic acid in the same manner as A4-choiesten-3-one (XI). The methylester, obtained with diazomethane, failed to crystallize, even after purification by chromatography. It was isolated as an oil which turned yellow after a short period of time and con­tained less sulfur than calculated. \Vhen dissolved in aqueous alcoholic hydrochloric acid and kept at room tem-perature overnight, free testosterone was recovered from ,the reaction mixture. Cholestan-3-one (XLVIII) was the only saturated 3- ketosteroid to be included in this investigation. It - 17 - - 18 - reacted with thioglycolic acid without a catalyst to give the expected mercaptol XLVIX. As mentioned above, the reaction of cholestanone with ethyl mercaptan occurred only in the presence of zinc chloride and sodium sulfate (8) or hydrogen chloride (12). The methylester L was purified by cnroma­tography and obtained in crystalline form. ROOCCH),S ROOCC 1-1,,5 XLVIII XLIX R=H L R=CH3 Reaction of 20-Ketosteroids and Thioglycolic Acid Allopregnan-3 ~-ol-20-one (LI) was found to react with thioglycolic acid only to a slight extent without catalyst. In one case 94% of starting material was recovered; in another case 87%. In the presence of zinc chloride and sodium sulfate, however, reaction occurred and the 20-mer-captol was obtained which was characterized by its crystalline methylester LII. HO LI ~3 ==0 LII - 19 .. Since the keto group in 20-position does not react with thioglycolic acid in the absence of a catalyst, it was ex­pected that progesterone (XVIII) which has keto groups in 3- and 20-positions would give only the 3-monomercaptol. Un­fortunately, the methylester of the mercaptol LIII did not crystallize and showed signs of decomposition when stored at room temperature. XVIII ) H3COOCCH~S H3COOCCH~S LITI The sulfur analysis indicated that only one keto group (in this case the 3-keto group) had reacted with thioglycolic acid. The infrared spectrum showed a band at 5.81 ~ which is characteristic for a 20-keto group. vVhen the same reac-tion was carried out in the presence of zinc chloride and sodium sulfate, the methylester obtained after treatment of the acid with diazomethane also failed to crystallize. After chromatography most of the fractions exhibited a band in the infrared at 5.85 ~ (20-keto group), however, two of the later fractions did not show this band. It was assumed, therefore, that the oily methylester represented a mixture of 3-mono- and 3, 20-dimercaptol. - 20 - Desoxycorticosterone (LIV) which is distinguished from progesterone (XVIII) by an additional hydroxyl group in 21-position reacted with thioglycolic acid essentially in the same way as progesterone. In the absence of a catalyst an acid was formed which was transformed to the methylester. This ester did not crystallize and appeared to be rather unstable. The sulfur analysis indicated that a monomercaptol was formed, and the infrared spectrum showed the presence of a free 20-keto group. The ester wa.s therefore formulated as the 3-moriomercaptol LV. LIV Reaction of 17-Ketosteroids and Thioglycolic Acid The condensation of dehydroepiandrosterone (XLIII) with thioglycolic acid in the presence of zinc chloride and sodium sulfate gave the mercaptol XLIV as described by Levin and Thompson (10). The methylester of this mercaptol was obtained in crystalline form. A similar experiment with androsterone (LVI) resulted in the isolation of 18.5% neutral products. The acids were methylated, but the methyl-exter was obtained as an oil. Although the value for sulfur - 21 - was somewhat higher than calculated, there is no doubt that this oil contained the methylester LVII probably in admixture with a little thioglycolic acid methylester. ) HO··· LVI HO· LVII SCHrOOCH., scH:£:OOCH3 Androsterone (LVI) failed to react with thioglycolic acid in the absence of a catalyst. Unchanged starting material was recovered in quantitative amounts. Reaction of AS-Unsaturated Steroids and Thioglycolic Acid It was rather surprising that dehydroepiandrosterone (XLIII) which, like androsterone, also contains a l7-keto group, reacted with thioglycolic acid without the aid of a catalyst. In addition to 12-15% of neutral products (probably starting material) an acid was obtained which was different from the mercaptol XLIV as prepared by Levin and Thompson (10). The analysis showed that it contained only one sulfur atom. The methylester of this acid crystallized with diffi­culty_ Consequently, the yield of crystalline material was very low.· It is believed that the quality of the thioglycolic acid had some influence on the yield and purity of the crys-talline methylester, since different preparations gave varying results. - 22 - The analysis of the methylester was in agreement with the empirical formula C22H3404S. This formula indicates that one mole of thioglyc 1c acid had added to one mole of dehydroepiandrosterone (LXIII). Since androsterone (LVI) did not react under the same condition, a formula like LVIII could be excluded. It was therefore assumed that the double bond in dehydroepiandrosterone (LXIII) had reacted with thio-glycolic acid with formation of a thioether which was ten-tatively formulated as LIX.~ ~o HO LVIII The addition of thioglycolic acid to a double bond was first reported by Holmberg (23) who obtained {3-phenyl­ethylthioglycolic acid (LXI) from styrene (LX). According to the findings of Kharasch et al. (24) this reaction is accelerated by light, oxygen and peroxides (e.g. ascaridole) and inhibited by typical antioxidants such as hydroquinone. A similar reaction was the reaction of isobutylene (LXII) with thioglycolic acid whi gave rise to the formation of isobutylthioglycolic acid (LXIII). C6H5CH=CH2 LX C6li5CH2CH2SCH2COOH LXI - 23 - CII3'-C=CH C1l3'" 2 ,. LXII The addition of thioglycolic acid to l-methylcyclohexene (LXIV) produced 2-methylcyclohexylthioglycolic acid (LXV) (25). In all of these cases, addition products "abnormal" with respect to Markownikoff's rule were formed. Cunneen (26) demonstrated, however, that in the presence of a strong mineral acid l-methylcyclohexene (LXIV) formed l-methyl­cyclohexylthioglycolic acid (LXVI), the "normalfl product with respect to Markownikoff's rule. LXV LXIV LXVI Since the reaction of del1ydroepiandrosterone (XLIII) with thioglycolic acid was carried out in the absence of a mineral acid, it was not unreasonable to speculate, that the thio­blycolic acid residue entered the 6-position of the steroid molecule as was illustrated in formula LIX. The 5a, 6P configuration in compound LIX is arbitrary. In order to prove the assumed structure of the thioether LIX, the methylester of LIX was acetylated and the crude acetate oxidized with chromium trioxide. The only product isolated in crystalline form melted at 2040 -2070 , - 24 - and the analysis gave C and H values which were in agreement wi th the empirical formula C2IH300L~. Al though no direct comparison was possible, it is likely that this substance is identical with androstan-3~-ol-6, 17-dione acetate (LXVII) (27). The infrared spectrtlln showed a band at 5.78 ~ which is characteristic for a keto group in the 6-position in the knovYn absence of such a group in position 3. Al though the yield of this oxidation product was very low, there is little doubt that compound LXVII resulted from the methylester of LIX and not from traces of dehydroepiandrosterone acetate, because the latter compound 'Yields A5-androsten-3f3-o1-7, 17- dione acetate, upon oxidation with chromium trioxide (28). '4---..=0 AcO II o LXVII In another experiment the methylester of LIX was oxidized with N-bromoacetamide and the reaction product refluxed with aqueous alcoholic hydrochloric acid. It was postulated that N-bromoacetamide would not attack the thio­ether linkage and that thioglycolic acid would be split off by treatment with hot hydrochloric acid. In this case, the end product would have been A4-androstene-3, 17-dione (XXXII). Actually, crystals were obtained which melted at 1940-196.5° and showed no selective absorption in the ultraviolet. The - 25 .. analysis gave values which were in agreement with the formula C19H2603. It is rather certain that this sUbstance is androstane-3, 6, 17-trione (LXVIII) (29). I o ""'+-----=0 LXVIII II o LXIX The infrared spectrum of this substance exhibited two bands in the keto region at 5.73 ~ and 5.82~. The latter band is characteristic for keto groups in 3- and 6-positions. The mother liquors of LXVIII contained a substance which dis­played a strong ~bsorption at 250~ Although this substance could not be isolated in pure form, it was assumed to be identical with A4-androsten-3, 6, 17-trione (LXIX) m.p. 216°.217° (uncorrected) (30). The formation of androstane-3, 6, 17-trione (LXVIII) from the methylester of LIX was evidence that bromoacetamide oxidized not only the hydroxyl group in the 3-position, but also the thioether linkage. This seems feasible, since 3.4 moles of bromoacetamide were consumed in this oxidation, whereas the oxidation of the hydroxyl group alone would have required only one mole of bromoacetamide. Since the yield of the trione LXVIII was very low, there exists the pos-sibility that this triketone was formed exclusively from - 26 - dehydroepiandrosterone (XLIII) which might have been present as an impurity in the starting material. It is probable too, that the unsaturated triketone LXIX was formed froM dehydro­epiandrosterone rather than from the methylester of LIX. However, nothing definite can be said about the course of these reactions, since the oxidation of dehydroepiandrosterone (XLIII) with N-bromoacetamide has thus far not been described. After it had been found that the 5, 6-double bond of dehydroepiandrosterone (XLIII) reacted with thioglycolic acid in the absence of a catalyst, the same reaction was applied to other 5, 6-unsaturated steroids- Treatment of ~5-pregnen~3~-ol-20-one (LXX) with thioglycolic acid produced an acid which was converted to its methylester LXXI by means of diazomethane. The methylester failed to crystallize, even after purification by chromatogTaphy- The analysis of the methylester revealed the presence of approximately one atom of sulfur. Furthermore, the infrared spectrum showed a band at 5.84 p. indicating that the keto group in 20-position had not reacted as was the case with allopregnan-3~-ol-20-one (LI). H H LXX LXXI - 27 - Ana]ogous to the acid LIX derived from dehydroepiandrosterone (XLIII) the methylester obtained from pregnenolone (LXX) was forr.1ulated as LXXI. Two additional preliminary experiments were carr:i.ed out with ~5-androstene-3(3, 17(3-diol and cholesterol. The former conpound gave a crystalline ac which 'Ivas not purified. The methylester of this acid crystallized, but showed signs of dec osition when recrystallization from benzene-hexane was attempted. Cholesterol was found to be very insoluble in thioClycolic acid, but the reaction could be accomplished when dioxane was used as a solvent. In addi tion to unc}).anged starting material an acid was obtained which melted at 1940- 1970. The methylester of this acid iled to crystallize, but the analysis showed that it contained one atom of sulf~w. Furthermore, the infrared spectrum exhibited a strong band at 5.75 V- indicating the presence of a carbomethoxy group. It was assumed the methylester had the formula LXXII analogous to the methylester obtained from dehydroepiandrosterone (methylester of LIX). HO LXXII - 28 - Synthesis of A5-3-Carbomethoxymethylmercapto-cholestene The methyl ester of A5-3- carboxymethylmercapto-chol­estene (LXXIII) was prepared for purposes of comparison with the thioethers mentioned above. The free acid was obtained by condensation of thiocholesterol with chloroacetic acid (9). H~OOCCH~ LXXIII The methylester L~~III had not been previously described and was prepared using diazomethane. It melted at 810 , showed infrared bands typical for the carbomethoxymethylmercapto group and proved to be rather stable to hydrochloric acid. - 29 - Tables of Derivatives Table I Melting Points of Mercaptol Methyl Esters Steroid Ketone ~4-Cholesten-3-one Testosterone Cholestan-3-one Allopregnan-3 ~-01-20-one Progesterone Progesterone Desoxycorticosterone Dehydroepiandrosterone Androsterone Table II Melting Point of Mercaptol Methylester oil oil (3-monomercaptol) oil (3,20-dimercaptol) oil (3-monomercaptol) 146°-148.5° oil Melting Points of Thioether Methyl Esters Steroid Dehydroepiandrosterone ~5-Pregnen-3 ~-01-20-one Cholesterol Melting Point of Thioether Methylester oil - 30 - As can be seen from the Tables I and II, seven of the twelve derivatives whioh had been prepared did not crystallize. The purification of the five orystalline derivatives was rather difficult, and inevitably resulted in a low yield of eaoh com­pound. On the basis of the few experiments which have been exeouted, it can be stated that the mercaptols of thioglycolic acid do not possess the desired properties prerequisite for characterization of steroid ketones. The value of thioglycolic acid mercaptols as a method for the separation of ketones and non-ketones is limited, since only saturated and A4-unsaturated 3-ketosteroids react with thioglycolic acid in the absence of a catalyst. Under these conditions, further complications are introduced due to reactivity of the 5, 6-double bond. lIowever, in the presence of zinc chloride and sodium sulfate as catalysts, the condensation of steroid ketones with thio­glycolic acid seems to be more promising, since in this chemi­cal environment keto groups in 17- and 20-positions also react with thioglycolic acid, while the 5, 6-double bond is not affected. EXPERIMENTAL DETAILS All melting points were taken on a Kofler micro hot stage and are corrected. The ultraviolet spectra were measured in absolute ethanol with a Beckman spectrophotometer model DU; the infrared spectra in carbon disulfide with a Beckman spectrophotometer model IR2. For all chromatographies, aluminwn oxide Merck, Tfsuitable for chromatographic adsorption", was used. The micro analyses were carried out by Drs. Weiler and Strauss, Oxford, England. Reaction of 3-Ketosteroids and Thioglycolic Acid Mercaptol from A4-cholesten-3-one and thioglycoliC}. acid (XLVII): A solution of 200 mg. of cholestenone (XI) in 0.8 cc. of thioglycolic acid was kept at room temperature overnight. After addition of water, the crystals were fil-tered, washed thoroughly with water and drIed in a vacuum desiccator. They were suspended in ether and the ether solu-tion extracted five times with small amounts of 2 N sodium carbonate solution and once with water. The ether solution was dried and evaporated and gave 18.1 mg. of neutral material which displayed a strong absorption at 240 m\-l. By recrystal-lization from dilute acetone, some unchanged cholestenone, m.p. 78°-80.50 , was recovered. The alkaline solutions were acidified to congo red and the precipitate filtered, washed with water and dried. A sample of the crude acid was re­crystallized from acetone-hexane and benzene and melted at - 31 - - 32 - 1330 -138.5°. Jones (9) has reported melting points of 122- 1240 and Ih2°. Infrared bands were observed at 5.80 \1 (carboxyl group), 7.66 \1, 7.73 ~ and 8.32 \1- Methylester: The main amount of the crude acid was suspended in ether and an ethereal solution of diazomethane was added. After standing for 10 minutes at room temperature the solu-tion was evaporated i!!. vacuo and the residue chromatographed on 6 gm. of aluminum oxide. The fractions, eluted with mix­tures of hexane and benzene containing from 20 to 50% benzene, were combined and recrystallized from acetone-water 5:1. Leaflets with m.p. 79-800 were obtained which gave a strong depression on admixture with 114-cholesten-3-one and showed no absorption at 240 rnlL. Infrared bands were observed at 5.70 tJ. (carbomethoxy group) and 7.88~. The latter band was also observed in the spectra of thioglycolic acid methy1- ester and apparently is characteristic for the methylesters of thioglycolic acid mercaptols. Analysis: Calculated for C33H54 04 S2' C, 68·46; H, 9·40; 5, 11.08%. Found, OJ 68.21; H, 9·15; s, 11.18%. Mercaptol from testosterone and thioglycolic acid:: A solution of 200 mg. of testosterone (XIV) in 0.8 cc. of thioglycolic acid was maintained at room temperature over-night. Upon addition of water to the clear solution, a heavy oil separated which crystallized after standing in the refrigerator for 24 hours. The crystals were filtered, - 33 - washed with water and dried. They melted at 1050 -1100 , showed a weak infrared band at 5.80 \1 (due to their slight solubility in carbon disulfide) and no absorption in the ultraviolet. Methylester: The methylester was prepared from the crystalline acid in the usual manner and chromatographed on 6.5 gm. of aluminum oxide. The principal amount of substance was eluted with benzene and mixtures of benzene and ether containing 10 and 20% ether. It consisted of a colorless oil which showed no absorption in the ultraviolet. Infrared bands appeared at 2.72 \1 (hydroxyl group), 5.71 \1 (earbomethoxy group) and 7.86 \1. The oily methylester turned yellow after standing for 1-2 days. Two individual fractions from the chromatography were analyzed for sulfur and gave values which were lower than calculated, indicating some decomposition. Analysis: Calculated for C25H3S05S2' S, 13.29%. Found, S, 10.13%; 9.94%. Fission with hyd~ochloric acid: A small amount of the oily methylester was dissolved in a solution of 0.2 cc. concentra­ted hydrochloric acid in 2 cc. of absolute ethanol and stored at room temperature overnight. The mixture was evaporated .i!l vacuo and the residue taken up in ether. The ether solution was washed with water, sodium carbonate solution and water, dried and evaporated. A single recrystallization from acetone­hexane gave somewhat impure testosterone, m.p. 11+3 0 -1520, -.34 - which showed strong absorption at 240 ~L. A further purifi-cation was not attempted. Mercaptol from cholestanone and thioglycolic acid (XLIX): A solution of 81.3 mg. of cholestan-3-one (XLVIII) in 0.8 cc. of thioglycolic acid was allowed to stand at room temperature overnight. After addition of water, the crystals were filtered, washed with water and dried. The crude acid melted at 1930-2010 and after recrystallization from ether softened at 2000 and melted at 206 0 -211°. Since this substance was very slightly soluble in carbon disulfide, the infrared band at 5.79 ~ (carboxyl group) was rather weak. Methylester: The methylester was prepared in the usual manner and chromatographed on 3 gm.. of aluminum oxide. It was eluted from the column with mixtures of hexane-benzene containing from 20 to 50% benzene. These fractions were com-bined and recrystallized from dilute acetone. Plates were obtained with m.p. 950 -97°. Infrared bands were noticed at 5.71 ~ (carbomethoxy group) and 7.85 ~. Analysis: Calculated for C33H5604S2' C, 68.22; H, 9.72; S, 11.04%. Found, C, 68.38; H, 9.70; S, 10.87%. Reaction of 20-Ketosteroids and Thioglycolic Acid Mercaptol from allopregnan-3~-ol-20-one and thio­glycolic acid: A. With zinc chloride and sodium sulfate. A mixture of 144 mg. of allopregnanolone (LI), 140 mg. of zinc - 35 - chloride, 250 mg. of sodium sulfate and 0.5 ce. of thioglycolic acid was allowed to stand at room temperature for 21+ hours. After addition of water, the precipitate was filtered, washed with water and dried in a vacuum desiccator. It was found to be very insoluble in etb~r, acetone, chloroform and benzene. A sample was recrystallized from a large volume of ethyl acetate and showed a melting point of 1750 -181°. Since this product was very slightly soluble in carbon disulfide, the infrared spectrum showed only a weak band at 5.75 ~ (carboxy group) • Methylester LII: The main amount of the crude acid was methylated with diazomethane and chromatographed on aluminum oxide. The methylester was eluted with mixtures of benzene and ether and recrystallized from acetone-hexane. These crystals melted at 1470 -149° and showed infrared bands at 2.66 ~ (hydroxyl group), 5.68 ~ (carbomethoxy group) and 7.82 ~. Analysis: Calculated for C27H4405S2' C, 63.24; H, 8;65; S, 12.51%. Found, C, 62.98; H, 8.48; S, 12.64%. B. Without catalyst. A solution of 50 mg. allo­pregnanolone (LI) in 0.2 cc. of thioglycolic acid was kept at room temperature overnight, diluted with ether and washed several times with 2 N NaOH. The ether solution contained l}7.2 mg. neutral material which consisted of unchanged starting material. - 36 - Mercaptols from progesterone and thioglycolic acid: A. Without catalyst. A solution of 200 mg. of pro sterone and 1.6 cc. of.thioglycolic acid was allowed to stand at room temperature for 24 hours. After addition of water, the precipitate was filtered, washed with water and dried in a vacuum desiccator. The crude reaction product melted at 1190 -1220 and showed infrared bands at 5.82 ~ (20-keto group and carboxy group), 7.73 ~, 8.34 ~ and 8.67 ~. Methy1ester LIII: The methy1ester prepared from the crude acid with diazomethane was chromatographed on 6 of alumi-num oxide and was eluted from the column with mixtures of hexane and benzene as a colorless oil which turned slightly yellow after short standing. It showed no absorption in the ultraviolet and bands were observed in the infrared at 5.69 ~ (carbomethoxy group), 5.81 ~ (20-keto ) and 7.84 ll. Two individual fractions were analyzed for sulfur and gave values somewhat lower than calculated, probably due to de-composition. Analysis: Calculated for C27H4005S2' S, 12.61%. Found, S, 11.52%; 11.16%. B. With zinc chloride and sodium sulfate. A mixture of 200 mg. of progesterone (XVIII), 140 mg. of zinc chloride, 250 • of sodium sulfate and 0.5 cc. of thiogly~olic acid was allowed to stand at room temperature for 24 hours. The reaction mixture solidified in 15 minutes. After addition - 37 - of water, the precipitate was filtered, washed with water and dried. The crude material melted at 118°-124° and showed only" one band in the infrared at 5.84 (.1. (carboxy group and 20-keto group [?1 ) • Methylester: The methylester, prepared the usual manner, was chromatographed on 1.5 gm. of alumin1lln oxide. The frac­tions eluted with mixtures of hexane and benzene and with pure benzene showed infrared bands at 5.73 II (carbomethoxy group) and 5.85 (.l. (20-keto group) and probably consisted of the methylester of the 3-monomercaptol (LIII). Fractions eluted with mixtures of benzene and ether showed only the band at 5.73 (.1.- It is assumed that they contained the 3,20-dimercaptol. Mercaptal from desoxycorticosterone and thioglycolic acid (LIV): A solution of 100 mg. of desoxycorticosterone in 0.8 cc. of thioglycolic acid was kept at room temperature for 24 hours. The solution was then poured into water and the resulting suspension was extracted with ether. The ether solution was washed several times with water, dried and evaporated. Methylester LV: The crude non-c.rystalline reaction product was methylated with diazomethane and chromatographed on 6 gm. of aluminum oxide. The fractions eluted with mixtures of hexane and benzene and with pure benzene showed only one band in the infrared spectrum at 5.71 (.1.. It is possible that they contained thioglycolic acid methylester. The fractions eluted - 38 - with mixtures of benzene and ether showed two bands, one at 5.71 (J, (carbomethoxy group) and the other at 5.85 tJ· (20- keto group), ,but failed to crystallize. They were obta as an oil which turned slightly yellow after a short time. Analysis: Calculated for C27H4006S2' c, 60.39; H, 7.68; s, 12.22%. Found, 0, 59.68; H, 7·94; s, 11·90%. Reaction of 11-Ketosteroids and Thioglzcolic Acid 1iIercaptol from dehydroepiandrosterone and thioglycolic acid (XLIV): This mercaptol was prepared according to the method of Levin and Thompson (10). After one recrystalliza­tion from ethyl acetate, the acid XLIV melted at 169°-l7ho. Since it was very slightly soluble in carbon disulfide, the infrared spectrum exhibited only a weak band at 5.77 ~, (carboxy group). Methylester: The methylester was prepared with diazomethane and recrystallized twice from ether-pentane, m.p. 11+6°-11~8.5°. The tetranitromethane reaction was positive. Infrared bands appeared at 2.70 ~ (hydroxy group), 5.69 ~ (carbomethoxy group) and 7.88 t-t- Analysis: Calculated for 025H380582' C, 62.21; H, 7.9It; S, 13.29%- Found, C, 62.07; H, 7.98; S, 13.37%. Mercaptol from androsterone and thioglycolic acid: A. With zinc chloride and sodium sulfate. A mixture of 49.5 mg. of androsterone (LVI), 35 mg. of zinc chloride, 63 mg. of sodium sulfate and 0.2 cc. of thioglycolic - 39 - acid was allowed to stand at room temperature overnight. The reaction product was taken up in water and ether and the ether solution was washed with four 5 cc. portions of 2 N sodium carbonate solution, with water, dried and evaporated. The neutral products weighed 9.2 mg. and were not further investigated. The alkaline solutions, upon acidification, gave a precipitate which was filtered, washed with water and dried. Methylester LVII: The crude acid was methylated with diazo-methane and chromatographed on 1.8 gm. of alluninum oxide. Two fractions eluted with hexane-benzene 1:4 and with benzene showed a single band at 5.73 ~ in the infrared (probably thioglycolic acid methylester). An oily fraction, eluted with benzene-ether 9:1, exhibited bands at 2.72 ~ (hydroxy group) and $.72 ~ (carbomethoxy group) and contained the methylester LVII, possibly in admixture with small amounts of thioglyco1ic acid methylester. This £raction could not be obtained in crystalline form. Analysis: Calculated for C25H4005S2' S, 13.23%. Found, S, 14·17%. B. Without ~atallst. A solution of 50 mg. of androsterone (LVI) in 0.2 cc. of thioglycolic acid was kept at room temperature overnight, diluted with ether and wash~d several times with 2 N sodium carbonate solution and water. The ether solution yielded 49.5 mg. of neutral material upon evaporation. - 40 - Reaction of A5-Unsaturated Steroids and Thioglycolic Acid Thioether from dehydroepiandrosterone and thioglycolic acid (LIX): A solution of 200 mg. of dehydroepiandrosterone (XLIII) in 0.8 cc. of thioglycolic acid was allowed to stand at room temperature overnight. After dilution with ether, the solution was washed five times with small portions of 2 N sodium carbonate solution, finally with water, dried and evaporated. The residue weighed 25.1 mg. and consisted principally of unchanged starting material. The alkaline washes were acidified to congo red. The precipitates thus obtained were filtered, washed with water and dried in a vacuum desiccator. After two recrystallizations from acetone, the acid LIX melted at 226.5°-228.50 • Infrared bands were apparent at 5.72 ~ (l7-keto group) and 5.79 ~ (carboxy group) (the substance was only slightly soluble in carbon disulfide). Analy~is: Calculated for C2lH3204S' c, 66.28; H, 8.48; s, 8.43%. Found, C, 66.21; H, 8.41; S, 8.35%. Methylester: The methylester was prepared with diazomethane and chromatographed on 5.2 gm. of aluminum oxide. It was eluted from the column with mixtures of benzene and ether containing from 20 to 50% ether and crystallized after long standing with pentane in the refrigerator. After two recrys-tallizations from ether-pentane, the methylester melted at 152-1550 and gave a strong depression in admixture with the methylester of the mercaptol XLIV (m.p_ 1460 -148.50 ). - 41 - Infrared bands were observed at 2.67 ~ (hydroxy group), 5.70 l-L (17-:1{et~tgroup and carbomethoxy group), 7.83 l-L, 8.00 l-L and 8.17. \1- Analysis: Calculated for C22H3404S, C, 66.96; H, 8.69; s, 8.13%. Found, 0, 66.93; H, 8.98; 5, 8.13%. Degradation with chromic acid: To a solution of 138.8 mg. of the methylester of LIX (probably containing small amounts of dehydroepiandrosterone (XLIIT)) in 1 cc. of pyridlne, 0.5 cc. of acetic anhydride was added. After standing overnig~t at room temperature, water was added dropwisa with cooling_ A heavy oil separated. The supernatant liquid was decanted and the oil taken up in ather. The ether solution was washed several times with 2 N HCl, water and sodium carbonate solu­tion, dried and evaporated ~ vacuo at 270 • The crude acetate weighed 150.7 mg. and failed to crystallize. The infrared spectrum showed strong bands at 5.72 l-L (acetate, 17-keto group and carbomethoxy group) and 8.02 ~ (acetate) in addition to weaker bands at 7.80 ~ and 7.84~. The acetate was dissolved in 1 cc. of glacial acetic acid and 1 cc. of a 2% chromium trioxide solution in glacial acetic acid was added. The addition of 1 cc. of chromium trioxide solution was repeated after 1/2, 2 and 6 hours. After the last addition, the solution was allowed to stand at room temperature overnight. The residue obtained following evaporation in vacuo at room temperature was distributed between water and a mixture of - 42 - ether and chloroform. The organic phase was washed with 1 N sulfuric acid, water, sodium carbonate solution and water, dried and evaporated in a stream of air. The residue weighed only 96.4 mg. and was chromatographed on 3 gm. of aluminum oxide. Two fractions eluted with benzene and benzene-ether 1:1 crystallized on standing with cold pentane. crystals were combined, sublimed in hi crystallized from acetone-hexane and dilute acetone. The colorless needles thus obtained melted at 2040-207° and showed infrared bands at 5.72 ~ (acetate and 17-keto group), 5.78 ~ (6-keto group) and 8.05 ~ (acetate). Analysis: Calculated for C21H3004' c, 72.80; H, 8.73%. Found, C, 72.46; H, 8.58%. This substance is probably identical to androstan- 3~-01-6,17-dione (LXVII) for which a melting point of 1970- 1980 (uncorrected) was reported (27).- Other fractions of the chromatography were eluted with mixtures of benzene and ether containing increasing amounts of ether and failed to crystal­lize. Altogether only 23 mg. of material were recovered from the chromatography- Degradation with N-bromoacetamide: To a solution of 253 mg. of the methy1ester of LIX (probably containing some dehydro­epiandrosterone (XLIII)) in 13.5 cc. of tertiary butanol, a solution of 177 • of N-bromoacetamide in 6 cc. of water was added. The mixture was allowed to stand at room temperature - 43 - for 24 hours. During the first four hours it was orange­brown in color, later colorless. Titration with sodium thio­sulfate solution, after addition of potassium iodide and acetic acid, indicated that all the bromoacetamide was utilized. The suspension was diluted with water, extracted with ether and the ether solutions washed with sodium thio­sulfate solution, sodium carbonate solution and water, dried and evaporated. The residue was again treated in the same manner with 177 mg. of bromoacetamide. After 17 hours at room temperature, the brown solution contained 52 mg. of bromoacetamide. Thus, 3.4 moles of bromoacetamide per mole methylester were 11tilized. The reaction product isolated as described above weighed only 133 mg. and showed bands at 5.69 lJ., 5.78 tJ. (3- and/or 6-keto group), 7. 73 tJ. and 8.20 lJ·. It was dissolved in 3.5 cc. of ethanol and 3.5 cc. of 5 N hydrochloric acid and refluxed for 4 hours (cfo 9). After cooling and addition of water, the suspension was extracted with ether. The ether solutions were washed with water, sodium carbonate solution and wat'sr, dried and evaporated. The residue weighed 98.6 mg. and was chromatographed on 3 gro. of a11nninum oxide. The fractions, eluted with mixtures of benzene and ether containing from 10 to 40% ether, gave crystals melting between 1800 and 1940 which showed a maximum absorption at 250 m~. They weighed 29.3 mg. By two recrys­tallizations from acetone-hexane, needles, m.p. 1940 -196.5°, .. 44 - were obtained which exhibited no selective absorption in the ultraviolet. Infrared bands were observed at 5.73 ~ (17- keto group) and 5.82 ~ (3-keto and 6-keto groups). Analysis: Calculated for C19H2603' C, 75.46; H, 8.67%. Found, C, 75.16; H, 8.85%. This substance is probably identical with androstane- 3,6,17-trione (LXVIII) for which a melting point of 1910_192° (uncorrected) was reported (29). Since the absorption maxi-mum at 250 ~ is typical for two keto groups in 1- and 4- positions having a mutual a,~-double bond, it is assQmed that the mother liquors of LXVIII contained some A4-androstene-3, 6,17-trione (LXIX) (30}. Thioether from A5-pregnen-3~-01-20-on~ and thiogly­colic acid: A solution of 50 mg. of A5~pregnen-3~-ol-20- one (LXX) in 0.2 cc. of thioglycolic acid was a110wed to stand overnight at room temperature. After addition of ether, the solution was washed with four 5 cc. portions of 2 N sodium carbonate solution and water. The ether solution, after drying and evaporation, gave 5.1 mg. of neutral material which consisted chiefly of unchanged pregnenolone. The al­kaline washes were acidified and the precipitate filtered, washed wi th V'tater and dried. The crude acid mel ted at 111-00- 1450 after softening at 1300• Methylester LXXI: The acid was methylated with diazomethane and chromatographed on 1.5 gm. of aluminum oxide. The - 45 - fractions, elITt-ed with mixtures of benzene and ether containing from 10 to 40% ether, failed to crystallize, but gave infrared bands at 2.72 ~ (hydroxy g~oup), 5.73 ~ (carbomethoxy group) and 5.84 ~ (20-keto group). Analysis: Calculated for C24H3804S (thioether), s, 7.42%. Calculated for C27H4205S2 (mercaptal), S, 12.56%. Found, S, 8.79%. Since the sulfur value was rather high, it is poss~ble that the methylester LXXI also contained some 20-mercaptol. Thioether from ~5-androsten-3~,17~-diol and thiogly­colic acid: A solution of 20 mg. of ~5-androsten-3~,17~- diol in 0.2 cc. of thioglycolic acid was stored overnight at room temperature. After addi tion of ether, the _,solution was washed three times with small amounts of 2'N sodium carbonate solution and water. The alkaline solutions were acidified and the precipitate filtered, washed with water and dried. The crude acid melted at 1320 -139°. Methylester: The methylester obtained with diazomethane was chromatographed on aluminum oxide and eluted from the colum.n with ether. Upon sta~ding with pentane in the cold, the methylester crystallized. The crystals melted between 700 and 900 and displayed bands in the infrared at 2.72 ~ (hydroxy group) and 5.74 ~ (carbomethoxy group). After two recrystallizations from benzene-hexane 5:1, the melting point was 1540-165°. This product contained only 5.88% sulfur instead of 8.09% as calculated for one atom of sulfur. - 46 - Thioether from cholesterol and thioglycolic acid: To a solution of 100 mg. of cholesterol in 1 ce. of dioxane, 0.6 ce. of thioglycolic acid was added. The mixture was kept at room temperature overnight. After addition of water, the p~ecipitate was filtered! washed with water and dried in a vaouum desiccator. The crude material was dissolved in ether and washed with four 5 cc. portions of 2 N sodium carbonate solution and with water. The ether solution, after drying and evaporation, gave 19.7 mg. neutral material. The alkaline solutions were acidified and the precipitate filtered, washed with water and dried. By recrystallization from acetone­hexane, crystals, m.p. 194°-197°, were obtained. Methylester LXXII: The m.ethylester, prepared from the mother liquor of the crystalline acid by means of diazomethane, was chromatographed on aluminum oxide and eluted from the column with mixtures of benzene and ether. It was obtained as an oil which showed infrared bands at 2.73 ~ (hydroxy group) and 5.75 ~ (carbomethoxy group). Analysis: Calculated for C)oH5203S, C, 73.11; H, 10.64; s, 6.51%. Found, C, 72.24; H, 10.57; s, 6.98%. SynthesiS of A5-3-CarbomethoxymethylmercaEto-cholestene 65-3-CarboxymethylmercaEto-cholestene: This acid was prepared from thiocholestero1 and chloroacetic acid according to the method of Jones (9). It showed infrared bands at 5.80 ~ (carboxy group) and 7.70 ~. .. 47 - Methylester LXXIII: The acid was methylated with diazomethane and the methy1ester chromatographed on aluminum oxide. The fractions eluted with hexane and hexane-benzene 1:1 were re-crystallized twice from ether-methanol. Needles of m.p. 80°- 81° were obtained which showed infrared bands at 5.70 ~ (carbomethoxy group), 7.82 ~ and 8.38 ~. Analysis: Calculated for C30H5002S, C, 75.89; H, 10.62; S, 6.76%. Found, C, 75.23; H, 10.58; s, 7.00%. SUMMARY (l) Using thioglycolic acid, five mercaptols of 3- ketosteroids have been synthesized ~ catalyst. Two o~ these mercaptols afforded crystalline methyl esters. (2) The reaction of 20-ketosteroids with thioglycolic acid has been investigated. Formation of 20-mercaptols ensues only when zinc chloride and sodium sulfate are em­ployed as catalysts. (3) The data indicate that l7-ketosteroid compounds condense with thioglycolic acid when zinc chloride and sodium sulfate are present in the reaction mixture. (4) The addition o~ thioglycolic acid to the 5,6- double bond of dehydroepiandrosterone (XLITI) has been described. This reaction proceeds in a mode contrary to the rule of Markownikoff. The position o~ the thioether linkage in the addition product has been established by oxidation studies. The same reaction was applied to two other 5,6- unsaturated steroids. (5) The synthesis of A5-3- carbomethoxymethylmercapto­cholestene has been described. (6) The use of steroid mercaptols derived from thioglycolic acid as derivatives for characterization and separation purposes has been discussed. - 48 - 1. 2. 4· 5. 6. 7· 8. 10. 11. 12. 13· BIBLIOGRAPHY E. Baumann, Ber., 18, 884 (1885); 12, 2803 (1886). w. Schindler, ·H~ Freland T. Reichstein, Helv. Chim. Acta, ~, 360 (1941). H. Koechlin and T. Reichstein, Helv. Chim. Acta, lQ., 1673 (1947). P. Speiser, Helv. Chim •. Acta,.2.S, 1368 (1949)~_ J. Heer and' K. Miescher, Helv. Chim. Acta, n, 405 (1948). F. Mylius,. Ber.,. 20, 1968 (1887). H. Hauptmann, J. Am. Chem. Soc., 22, 562 (1947). S. Bernstein and L. Dorfman, J. Am. Chern. Soc., 68, 1152 (1946). A~ S~ Jones, F. Smith and M. Webb, Nature, 162, 857 (1948), A. s. Jones, M. Webb and F. Smith, J. Chem.-SC>c. l.2lL2., 2764. - R. H. Levin ana J. L. Thompson, J. Am. Chem. Soc., 70, 3140 (1948). L. Norymberska, J. Norymberski and A. 01alde, J. Am. C~em. Soc., 70, 1256 (1948). J. w. Ralls, R. M. Dodson and B. Riegel, J. Am. Chern. Soc., 71, 3320 (1949). ~ '-~ ,. . " G. Roserikrariz,St. Kaufmann and J. Romo, J. Am. Chern. Soc., 71, 3689 (1949). ~ . -,. ---... '" J. Ronio, M~ Romero, C. Djerassi' and G. Rosenkranz, J. Am. Chem. ~oc., ll, 152~ (1951). R. w. Jeanloz, J. Am. Chern. Soc., 72", 2281 (1950). G. Rosenkran~ and c. Djerassi, Nat~re, 166, 104 (1950). A. Wettstein, Helv. Chim. Acta, ll, 1803 (1944). - 49 - - 50 - 18. D. K. Fukushima ana T. F. Gallagher, J. Arno Chemo Soc_, ll, 196 (1951). 19. G. Lardel11 and o. Jeger, Helv. Chim. Acta, ~, 1817 (1949) • 20. J. Bongartz, Der., ~'. 478 (1888). 21. B. Holmberg, J. prakt. Chern., 112, 57 (1932). 22. Vi. J. Hickinbottom, "Reac tions of Orga.nic' Compounds" , . Longmans, Green and Co., London, 1948, 2nd ed., p.132. 23.~ B. Holmberg, J. prakt. Chern., 141, 93 (1934). 24. M. s. Kharasch, A.' T. Read and F. R. Mayo, Chern. and Ind., 57, 752 (1948). 25. N. Hellstrom andS. Sandstrom, Svensk Kern. Tid., 2l:t., 149 ( 1942 ); C • A., l§., .3 257 ( 1944 ) ~ 26. J. I. Cunneen, J. Chern. Soc., l2kl, 36. 27. A. B~tenandt and L. A. Suranyi, Ber., 75, 591 (1942). 28. J. R. Billeter and K. Miescher, HelVe Chim. Acta, 11, 629 (1948). 29. M. I. 'Ushakov and A. Lyutenberg, J. Gen. Chern. (U .S. S.R.), ~, 69 (1939); C. A., ll, 63)4 (1939)-, 30. A. Butenandt and B. Riegel, Ber., 29., 1163 (1936).