NITROGEN musrAms RELAED To HEXESTROL ’ iii-1}? AND DIETHYLSTILBESTROL t j Them; for His .8 me 43‘ pl'l D ‘ ‘ ' MICHIGAN STATE UNIVERSITY ~ " ”’ " 9" John F Benner 1962 F ' THESIS (”-2, 1,; .5 R ARY ‘ . . _ «umqnn State University fi I'JIICH'GAN STATE UNIVERSITY FAST LANSING, MICHIGAN , . _DC,IIY, 'E UNIVE".‘ MICHIGAN 57m ABSTRACT NITROGEN MUSTARDS RELATED TO HEXESTROL AND DIETHYLSTILBESTROL by John F. Benner The observation that certain biological alkylating agents vary in their biological properties independently of their chemical reactivities (1) has led to the search for nitrogen mustards which show some specificity of action. The introduction of a physiological "carrier moiety" into the nitrogen mustard molecule may influence its arrival at the site of action without affecting its chemical reactivity (2)., The effective concentration (3) at the site of action may increase its effectiveness as an oncotoxin and lower its general toxicity. Estrogenic molecules may act as carrier moieties for biological agents of the nitrogen mustard- type. Accordingly, the syntheses of certain nitrogen mustard derivatives of meso-hexestrol, meso-hexestrol dimethyl ether, d, l-hexestrol, g-hexestrol dimethyl ether, and diethylstilbestrol dimethyl ether were undertaken. In the hope of retaining as much of the estrogenic activity as possible in the hexestrol carrier, it was decided to leave the aryl-portion of the molecule unchanged and to replace one of the ethyl groups with an N, N—bis (2—chloroethyl)aminomethyl group. Thus, the desired nitrogen mustards corresponding to meso- and d, 1-hexestrol are the threo- and erythro-modifications of N, N-bis(Z-chloroethyl)—2, 3—di(p-hydroxyphenyl)— pentylamine (I, R = H) and the corresponding dimethyl ethers (I, R : CH3). John F. Benner CZH5 I Ro—Q—w-wOOR SSH: N(CHZCH2C1)Z N, N-Bis(2-chloroethy1)|erythro-Z, 3-di(p-hydroxyphenyl)-pentyl]amine hydrochloride (I-HCl, R = H) and the corresponding dimethyl ether (I-HCl, R = CH3) were successfully prepared. The less stable threo- modification could not be obtained in a pure state" The configuration of the erythro-dimethyl ether (I, R = CH3) was correlated to the estrogenicly active meso-hexestrol. This was accomp— lished by a series of transformations which converted erythro-Z, 3—di— (p-methoxyphenyl)pentanoic acid, an intermediate used in the synthesis of the nitrogen mustard (1), to meso—hexestrol without affecting either of the asymmetric centers. The higher melting isomeric 2, 3—di(p=methoxy— phenyl)pentanenitrile was also shown to be in the erythro-series during the course of this work. This substantiates the assumption made by previous workers (4) in this field that the higher melting nitrile corres— ponds to the higher melting meso-hexestrol. In order to evaluate the effectiveness of diethylstilbestrol dimethyl ether as an estrogenic carrier moiety it was decided to prepare com- pounds in which one of the ethyl groups is replaced by a bis(2—chloro— ethyl)aminoalkyl group. Two such compounds were considered: N, N-bis(2-chloroethyl)-3, 4—di(p-methoxyphenyl)-3—hexenylamine (II, n = Z) and N, N-bis(Z—chloroethyl)—Z, 3-di(p-methoxyphenyl)-Z-pentenyl— amine (II, n = l). John F. Benner 3 Csz l CH30-O-C : C- -OCH3 I (CEHZ n N(CH2CHZC1)Z II N, N-Bis(Z-chloroethyl)- 3, 4-di(p-methoxyphenyl)-3—hexenylamine (II, n = 2) was successfully prepared. Five synthetic routes for the preparation of N, N-bis(Z-chloroethyl)—2., 3-di(p—methoxyphenyl)—-2— pentenylamine (II, n = 1) were devised and investigated. None of these procedures lead to the successful synthesis of this nitrogen mustard. The nitrogen mustards, as their hydrochloride salts, were sub- mitted to the Cancer Chemotherapy National Service Center* for screening. LITERA TURE CITED 1. W. C. J. Ross, Ann. N. Y. Acad. Sci., g, 669 (1958). 2. F. Bergel, Ann. N. Y. Acad. Sci., .éé’ 1238 (1958). 3. A. G. Ogstom, Trans. Faraday Soc., 1%, 45 (1948). 4. J. H. Burckhalter and J. Sam, J. Am. Chem. Soc., E, 187 (1952). “\Cancer Chemotherapy National Service Center, National Institutes of Health, Bethesda 14, Maryland. NITROGEN MUSTARDS RELATED TO HEXESTROL AND DIETHYLSTILBESTROL BY John F. Benne r A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Chemistry 1962 ACKNOWLEDGMENTS The author wishes to express his sincere appreciation to Dr. Gordon L. Goerner for his assistance and cooperation during this investigation. ‘ Grateful acknowledgment is also extended to my associates, Sharon K. Slack, Robert A. Martin,. Curtis R. Hare, Paul M. Dupree, James S. Skelcey and Richard L. Titus, for many en- lightening discussions of this problem. And finally, the author wishes to gratefully acknowledge the aid and encouragement extended to him by his parents, Mr. and Mrs. Hugh A. Benner, during the course of his education. ii TABLE OF CONTENTS Page HISTORICAL INTRODUCTION. . . . . . . . . . . . . . . . . . l A. Estrogens . . . . . . . . . . . . . . . . . l B. Biological Alkylating Agents. . . . . . . . . . . . . 12 PART I NITROGEN MUSTARD ANALOGS OF meso-HEXESTROL, racemiC-HEXESTROL, AND THEIR DIMETHYL ETHERS 26 Discussion........................ 26 Preparation of the nitrogen mustards . . . . . .. . 26 Stereochemistry of the hexestrol nitrogen mustards.................... 45 Experimental...................... 49 Anisyl alcohol. . . . . . . . . . 49 a. By the hydrogenation of anisaldehyde. . . . 49 b. By the crossed Cannizzaro reaction with anisaldehyde and formalin . . . . . . . . . 50 p—Methoxyphenylacetonitrile . . . . . . . . . . . . 51 a. With potassium cyanide in aqueous \ dioxane...................I. 51 b. With sodium cyanide in anhydrous acetone . '552 2, 3-Di(p-methoxyphenyl)acrylonitrile . . . . . . . 53 2, 3-Di(p-methoxyphenyl)pentanenitrile . . . . . . 5.4 erythro- 2, 3— Di(p— methoxyphenyl)pentylamine < hydrochloride . . . . . . 55 erythro- 2, 3- Di(p- methoxyphenyI)pentylamine . . . 56 l-Ierythro- 2, 3— Di(p- methoxyphenyl)pentyl]— 3— (1— naphthyl)urea. . . . . . . . . . . 56 threo— 2, 3- Di(p— methoxyphenyl)pentylamine hydrochloride . . . . . . . . . . 57 threo— 2, 3- Di(p- methoxyphenyl)pentylamine. . . . 58 l--|threo 2,3- Di(p— methoxyphenyl)pentyl]— 3- (l—naphthy1)urea . . . . . . . . . . . . . . . . 58 iii TABLE OF CONTENTS - Continued Page N, N- Bis(2-hydroxyethyl)| erythro—Z, 3—di-(pwmethoxy- phenyl)pentyl]amine -— by the ethoxylation of the primary amine with ethylene oxide . . . . . . 59 Chromatography of N, N— bis(2-= hydroxyethyl)- er thro— 2, 3 di( -methox hen l) ent lamine. 60 P' YP Y P Y Purification of thionyl chloride . . . . . 60 N, N Bis(2- c-hloroethyl)[erythro- 2, 3- di(p- -methoxy— phenyl)pentyl]amine hydrochloride . . . . . . . 62 2, 3- Di(p- methoxyphenyl)propylamine . . . 63 N, N— Bis(2— hydroxyethyl)=- 2, 3- di(p methoxyphenyl)- propylamine hydrochloride . . . . . 64 N, N- Bis(2— chloroethyl)— 2, 3- di(p= smethoxyphenyl propylamine hydrochloride . . . . . . . 65 erythro— 2, 3— Di(p- methoxyphenyl)- l— -pentanol. . . . 66 threo— 2, 3— Di(p- methoxyphenyl)- l- -pentanol. . . . . 66 erythro— l— Chloro-2, 3-di(p-methoxyphenyl)pentane . 67 Alkylation of diethanolamine with erythro-l-chloro— 2, 3- di(p- methoxyphenyl)pentane with ethylene glycol as solvent . . . . . 68 Alkylation of diethanolamine with erythro- l chloro— 2, 3— di(p— methoxyphenyl)pentane without a solvent . . . . . . . . . . . . . . . 70 2, 3- Di(p- methoxyphenyl)pentanoic acid. . . . . . 70 Methyl threo— 2, 3- di(p— methoxyphenyl)pentanoate using diazomethane . . . . . . 71 Methyl threo- 2, 3— di(p— methoxypahenyl)pentanoate from the acid chloride and methanol . . . . . . 73 Saponification of methyl threo-2, 3-di(p—methoxy- phenyl)pentanoate with methanolic potassium hydroxide. . . . . . . . . . 74 Attempted saponification of methyl threo- 2, 3- di(p- methoxyphenyl)pentanoate with one equivalent of sodium hydroxide in aqueous methanol. . . . 75 Attempted acid hydrolysis of methyl threo—Z, 3—di- (p- methoxyphenyl)pentanoate. . . . . . . . 76 Attempted saponification of methyl threo- 2, 3— di- (p- methoxyphenyl)pentanoate with one equiva- lent of potassium hydroxide—ethylene glycol in benzene..................... 76 Differential pH separation of the isomeric 2, 3-di— (p-methoxyphenyl)pentanoic acids . . . . . . . 77 iv TABLE OF CONTENTS — Continued Page Attempted chromatographic purification of crude threo-2, 3-di(p-methoxyphenyl)pentanoic acid. . 79 Attempted acid hydrolysis of threo- 2, 3-di(p~ methoxyphenyl)pentanenitrile. . . . . . . . . . 79 Attempted preparation of threo— 3, 4- di(p- -met°hoxy- phenyl)- 2- hexanone and the hypochlorite oxi— dation to threo- 2, 3-di(p-methoxyphenyl)— pentanoic acid. . . . . . . . . . . 80 Zone precipitation of threo~ 2, 3— di(p- m-ethoxy— phenyl)pentanoic acid . .. . . . . . . 81 2- Pyranyl erythro— 2, 3- di(p- methoxypheonyl): pentanoate. . . . . 81 Hydrolysis of 2— —pyranyl erythroa 2, 3— “on-di(p methoxy— pheny11Pentanoate. . . . . . . . . 82 2- Pyranyl threo- 2, 3- di(p— methoxyphenyl)— pentanoate.................... 82 a. With p toluenesulfonic acid catalyst. . . . . 82 b. With sulfuric acid catalyst. . . . 83 N, N- Bis(2- hydroxyethyl)— erythro— 2, 3— di(p- -methoxy- phenyl)pentanamide . . . 84 N, N— Bis(2— hydroxyethyl)[eryth°ro- 2, 3- di(p— —methoxy- phenyl)pentyl]amine hydrochloride by the reduction of the corresponding amide. . . . . . 85 N, N-Bis(Z-chloroethyl)[erythro—Z, 3—di(p-methoxy- phenyl)pentyl]amine hydrochloride with a di- methylformamide catalyst. . . . . . 86 N,N- Bis(2— hydroxyethy1)- threo— 2, 3- di(p- -methoxy- phenyl)pentaneamide. . . . . . . 86 N, N— Bis(2- chloroethyl)— threo— 2, 3 di(p- -methoxy- phenyl)pentyl amine hydrochloride . . . . . 87 erythro- 2, 3- Di(p- hydroxyphenyl)pentanoic acid. . . 89 erythro- 2, 3- Di(p- acetoxyphenyl)pentanoic acid. . . 90 N,N— Bis(2— hydroxyethyl)— erythro—Z, 3-di(p-hydroxy— phenyl)pentaneamide. . . 90 N, N- Bis(2- hydroxyethylHerythro— 2, 3- di(p- hydroxy- phenyl)pentyl]amine hydrochloride . . . . . 91 N, N— Bis(Z- chloroethyl)[erythro- 2, 3- di(p- hydroxy— phenyl)pentyl]amine hydrochloride . . . . . . . 92 threo-2, 3-Di(p—hydroxyphenyl)pentanoic acid . . . . 93 erythro-Z, 3-Di(p-methoxyphenyl)pentanoyl chloride 94 Methyl erythro-3, 4-di(p-methoxyphenyl)hexanoate . 95 TABLE OF CONTENTS - Continued Page erythro-3, 4—DiIp-methoxyphenyl)hexanoic acid. . . 96 erythro—3,4-Di(p-methoxyphenyl)hexanol . . . . . . 97 erythro-3, 4—Di(p-methoxypheny1)-l—hexyl-p- toluenesulfonate . . . . . . . . . . . . . . . . .. 97 meso- Hexestrol dimethyl ether . . . .. . 98 Attempted Schmidt reaction with erythro- 3, 4- di(p- methoxypheny1)hexanoic acid . . . . . . . . . 99 erythro- 2, 3— Di(p— methoxyphenyl)pentylamine using the Curtius reaction with erythro- 3, 4— di(p~ methoxyphenyl)hexanoic acid .. . . . . . . . . . 100 PART II NITROGEN MUSTARD ANALOGS OF DIETHYLSTILBESTROL DIMETHYL ETHER. . . . . . . . . . . . . . . . . . . . 102 Discussion........................ 102 Experimental........................ 115 Anisoin........................ 115 Deoxyanisoin. . . . . . . . . . . . . . . . . . . . .. 115 a. By the reduction of anisoin with tin and hydrochloric acid . . . . . . . . . . . . . . 115 b. By the reduction of anisoin with tinIII) chloride and hydrochloric acid . . . . . . . 116 CL-Ethyldeoxyanisoin by the alkylation of deoxy— anisoin with ethyl iodide . . . . . . . . . . . . 116 d- -Methy1enedeoxyanisoin . . . . . . 117 0— Ethyldeoxyanisoin by the addition of methyImag-- nesium iodide to a- methylenedeoxyanisoin. . . 118 Methyl bromoacetate . . . . . . . 119 Methyl 3- --hydroxy 3, 4- di(p- methoxyphenyl)— hexanoate. . . . 120 Dehydration of methyl 3- -hydroxy— 3, 4- din- -methoxy- phenyl)hexanoate. . . . . . . . . . . . . . . . 121 a. With iodine in refluxing toluene . . . . . . 121 b. With p—toluenesulfonic acid in refluxing benzene.................... 121 c. With phosphorous pentoxide in refluxing benzene.................... 122 vi TABLE OF CONTENTS - Continued Page d. With oxalic acid in refluxing benzene . 123 e. With concentrated sulfuric acid . 123 f. With thionyl chloride and pyridine . . 124 g. With acetyl chloride . . . . 125 Isomerization of 3, 4- din- methoxyphenyl)- 2— hexenoic acid to 3, 4- din— methoxypheny1)- 3- hexenoic acid. . . . . . . . 126 a. With potassium hydroxide in benzyI alcohol. 126 b. With potassium hydroxide in ethylene glycol. . . . . 126 N, N- BisI2~= hydroxyethy1)- 3, 4— odi(-p methoxyphenyl)- 3— hexenamide . . 127 N, N- Bis(2-hydroxyethy1): 3, 4-din methoxyphenyl)- 3- -hexeny1amine hydrochloride . . . . 128 N, N- BisI2- chloroethyl) 3, 4-din-methoxypheny1)— 3— —hexeny1amine hydrochloride . 129 p—Methoxypropiophenone . 129 Condensation of phenylacetonitrile with propio- phenone and of p—methoxyphenylacetonitrile with p—methoxypropiophenone . 131 a. Phenylacetonitrile and propiophenone with freshly prepared sodium amide. . . . . 131 b. p- Methoxyphenylacetonitrile and p- methoxy- prophiophenone with freshly prepared sodium amide. 132 c. Phenylacetonitrile and propiophenone with ammonium acetate in acetic acid. . . 133 d. p— Methoxyphenylacetonitrile and p- methoxy- propiophenone with ammonium acetate in acetic acid. 133 e. Attempted condensation with other Knoevenagel catalysts. 134 Addition of ethylmagnesium bromide to a—benzoyl- phenylacetonitrile. . . . . . . 134 Addition of ethyllithium to a- benzoylphenylacetoo- nitrile. . . . 135 p— Methoxyphenylacetic acid. . . 136 a. By the hydrolysis of p- methoxyphenyI- acetonitrile . . . . . . . . . . 136 b. From p=methoxyacetophenone by the Willgerodt reaction . 137 vii TABLE OF CONTENTS - Continued Page 3- -—Hydroxy 2, 3— —dipheny1pentanoic acid. . . . 138 Attempted dehydration of 3— —hydroxy- 2, 3- diphenyl- pentanoic acid. . . . . . . . . . 139 3- -Hydroxy- 2, 3— di(p- methoxyphenyl)pentanoic acid . 139 Attempted dehydration of 3— —-hydroxy 2, 3- di(p- methoxyphenyl)pentanoic acid. . . . . . . 140 Attempted preparation of 3- -acetoxy- 2, 3- di(p- methoxyphenyl)pentanoic ethanoic anhydride . . 141 Attempted preparation of 2-Ip—methoxypheny1)=-3— [N, N-bis(2—hydroxyethyl)amine]-p-methoxy— propiophenone. . . . . . . . . . . . . . . . . . . 142 a. By the Mannich reaction of deoxyanisoin and diethanolamine . . . . . . . . . . . . . . 142 b. By the addition of diethanolamine to 0- -methylenedeoxyanisoin. . . . 143 Attempted preparation of 2, 3n-di(p=-methoxypheny1)I= 2- -pentenylamine hydrochloride . . . . . . . . . 143 LITERATURE CITED. . . . . . . . . . . . . . . . . . . . . . . 145 viii LIST OF TA BLES TABLE 1. Distance Between Hydrogen-bonding Groups vs. Estrogenic Activity. II. Effective Competition Factors, Ff, or Reactive Groups. III. Relation of Biological Activity to Chemical Reactivity IV. Variation in Biological Activity of Compounds of Com- parable Chemical Reactivity. V. Kinetic Data .for the Alkylation of Diethanolamine with erflhro- l--Chloro—2, 3-di(p-methoxyphenyl)pentane . VI. Titration Data for the Neutralization of Mixture of Isomeric 2, 3—Din—methoxyphenyl)pentanoic Acids . ix Page 18 20 21 68 77 FIGURE III. IV. VI. VII. VIII. IX. XI. XII. LIST OF FIGURES PART I . Dimensions of Estrone and Diethylstilbestrol. . Probable Configuration of Dienestrol and Diethylstil— bestrol. Conformation of meso—Hexestrol . Synthesis of the Isomeric 2, 3=Di(p~=methoxypheny1)— pentanenitriles. . Preparation of the Isomeric N, N—Bis(2—chloroethy1)— 2, 3-di(p-methoxypheny1)pentylamines by Scheme A Preparation of the Isomeric N, N-BisIZ—chloroethyl)- 2, 3-di(p-methoxyphenyl)pentylamines by Scheme B Preparation of the Isomeric N, N—Bis(2—chloroethy1)- 2, 3-di(p-methoxypheny1)penty1amines by Scheme C Preparation of N, N—BisI2-chloroethy1)-2, 3-din- hydroxyphenyl)pentylamine hydrochloride. Schematic for the Conversion of erythro—Z, 3-Din- methoxyphenyl)pentanoic acid to meso—Hexestrol di- methyl ether . . Apparatus Used for Hydroxyethylation . Weight of Residue vs. Eluent in the Chromatography of N, N- Bis(2-hydroxyethyl)[erythro-2, 3-din- methoxyphenyl)pentyl]amine. Second Order Plot for the Alkylation of Diethanol- amine. Page 11 11 28 31 34 36 43 46 59 61 69 'LIST OF FIGURES - Continued FIGURE Page XIII. Diazomethane Apparatus. . . . . . . . . . . . . . . . 72 XIV. Volume of Acid vs. pH for the Neutralization of the Isomeric 2, 3-DiIp-methoxyphenyl)pentanoic Acids . . 78 PART II I. The Preparation of N, N-BisI2-ch10roethyl)-3, 4—di— Ip—methoxyphenyl)-3-hexenylamine hydrochloride . . 103 xi HISTORICAL INTRODUC TION A. Estrogens Apart from knowing that the sexual functions of the female vertebra depend on the ovaries and are not dependent on the nervous connectors for control, our knowledge of the interrelations of hormones and sexual functions, as applied to ovarian activity, dates from the observation of Marshall and Jolly (1) in 1905 that estrus can be induced in spayed dogs by the injection of ovarian extracts removed from other dogs during estrus, or by the implantation of estral ovaries into the peritoneum. The interspecial commonness of ovarian extracts was first reported by Allen and Doisy (2, 3) in 1924. It was observed that extracts of sow's ovaries caused estrus-like changes in the rat vagina. The first successful isolation of a pure, crystalline estrogen was accomplished independently by Doisy, Valer, and Thayer (4) in 1929 and by Buterond (5) in the same year. Prior to this time biologists were confined to the use of crude extracts and their observations lacked any quantitative significance. The first compound to be isolated, now known as estrone (I), was also the first of the estrogens to have its structure elucidated. The isolation of pure estrogens stimulated the study of their physiological role. In the vertebra the estrogens are concerned mainly with the female sexual functions and with the formation of the subsidiary traits of appearance and behavior which characterize the female. The estrogens are known to effect the reproductive organs in four ways. 1) They provide an automatic check on 0 their own production by controlling the production of gonadotropin by the pituitary . HO 2) They stimulate the growth and function of organs concerned with the female sex, such as the ovaries. 3) They alter some of the actions of androgen and progestin. 4) They are responsible for ovagenesis. They also exert psychological effects in prompting the behavior characteristic of the female. As restricted by present day usage and testing techniques (6), the term "estrogen" denotes any substance that will produce cornification in the vagina of an adult test animal resembling that of natural estrus. The estrogens can be considered as belonging to one of two general classes: 1) the natural or b) the artificial (7). Even classes as general as these lead to confusion. The stipulation must be placed on the first class so that it must include all the natural estrogens produced outside the animal in which they act, but that it must not include natural products, plant or animal, not produced as estrogens by animals. It is almost impossible to limit the class of natural estrogens to compounds produced by animals as estrogens, due to our lack of understanding of the com- plete physiological role of these compounds. The ”artificial" estrogens include all the compounds produced in the laboratory and all those com— pounds produced in the plant kingdom which Show estrogenic activity, but which differ in structure from the naturally occurring estrogens. Thus, estrone isolated from palm kernel extracts and the estrone produced synthetically is still classed as a "natural" estrogen since it is, in all respects, similar to the compound produced by animals as an estrogen. Substances such as genislein, though isolated from natural sources, are classed as artificial estrogens because they are not produced by the animal as estrogenic hormones. In both classes the distinction must be made between the "estrogens" and the "pro-estrogens. “ By the use of spayed mice in which the vagina was divided into two pouches, a marked difference between these two types of compounds was clearly shown by Emmens (8). A pro-estrogen when introduced into one vaginal pouch caused a response in both pouches. An estrogen when applied, in a small amount, to one of the pouches caused it to cornify, but had no effect on the other because its concentration in the general circulation was below the minimum level necessary for estrogenic activity. Since the isolation of estrone there have been four other natural estrogens isolated from various sources. Estrone was first isolated from pregnancy urine, but has since been isolated from various other sources such as: mare and stallion urine, ovaries, testes, and palm kernel extracts. Pregnancy urine also contains the closely related estriol (II). Mare urine contains estradiol (III), the most potent of the natural estrogens, as well as the weakly estrogenic equilenin (IV) and equilin (V). The natural estrogens are structurally characterized by the absence of a methyl group at C—10 and the aromatic nature of ring-A. The structural features possessed by the natural estrogens that render them active have not been completely elucidated. Of the five natural estrogens (I-V) thus far isolated, it can be said that an increase in the degree of unsaturation decreases the activity. OH OH OH HO HO ‘ III 11 O HO / HO—UV” IV With regard to other structural variations, the most can be said of equilenin (IV), since a large number of closely related derivatives have been synthesized. The hydroxyl group in the 3-position seems to be necessary for high activity (9). When it is shifted to the 2- or 6— position the activity is considerably less (10, 11). The 3-deoxy com- pound is only weakly estrogenic in large doses (12). There have been a few 3-deoxy compounds related to equilenin that show about one- fifteenth to one—twentieth the activity of dl-equilenin (13). Thus, the hydroxyl group is not essential for activity, but is necessary for high activity. When the 17-keto group of equilenin is shifted to the 16- position the activity is decreased to about one-third (14). Data on 17—deoxyequilenin is not available, but 17-deoxyestrone is about one- thirteenth as active as estrone (15). Variations in the group at C—13 can either increase or decrease the activity. Replacement of the angular methyl group by an ethyl group decreases the activity,but replacing it with a n-propyl group (either cis or trans) increases the activity (10). 18-Norequilenin is inactive (16), but this compound belongs to the cis- series and should be compared to isoequilenin, itself only very weakly estrogenic. The absence of molecular specificity in the series of natural estrogens led Dodds and his co-workers to examine how the molecule could be changed and still retain its activity. This work was rewarded with the discovery of diethylstilbestrol (VI)(17), hexestrol (VII)(18), and dienestrol (VIII)(19). These compounds remain among. the most useful and potent artificial estrogens. They possess all of the quali— tative properties of the natural estrogens, are inexpensive to produce, and have the advantage of being active orally. Stimulated by the success of these compounds a great number of closely related compounds have been synthesized and tested. Though only a few of these compounds have proven to be of any use as therapeutic agents, the work has given valuable data on chemical constitution and estrogenic activity. Among the more important open-chain estrogens are the tri- phenylethylenes of Robson and co-workers (20, 21, 22), Csz C2H5 HOHQC SQ... WC}..- CH-©OH C2H5 C2H5 VI VII CH3 Csz HO— ”Q C- HC-QOH HO- JG CH— CIH— CH— CH—Q— OH C2H5 CH3 VIII IX 1, l-di(p-methoxyphenyl)—2—bromophenylethylene and benzestrol (IX) prepared by Blanchard, Stuart and Tallman (23). In 1944 Miescher and his co—workers (24) discovered a new type of estrogen more closely related to the natural estrogens than those of the stilbene type. Doisynolic acid (X) and bisdehydrodoisynolic acid (XI) are examples of this type of estrogen. The doisynolic acids can be considered as open D-ring analogs of the natural estrogens. Jacques and Horeau (25) achieved further structural simplification in the allenolic acids in which the C-ring of the diosynolic acids is replaced with an aliphatic chain. Horeau acid (XII) is an active member of this series. The estrogenic activity of various isoflavones, such as genislein (XIII) has been noted (26). By structural considerations these compounds can be considered as heterocyclic closed ring analogs of the stilbene type of artificial estrogen. .._- 6 CH3 CH3 COZH . COzH / 02H, CHZ—CH3 HO HO X XI .5... Csz—C—COZH o I \/ CH-CZHS HO / HO/ / 0 x11 x111 By the use of x—ray crystallography Giacomello and Bianchi (27) found that the molecules of diethylstilbestrol and estrone were 8. 55A long and 3. 88A wide as shown in Figure I. Schueler (28), using this data, postulated that a given substance may be estrogenic if it consists Dimensions of Estrone and Diethylstilbestrol. Figure I. of a rather rigid and inert molecule with two active, or potential, hydrogen-bond forming groups separated by an optimum distance of 8. 55A. On the basis of this hypothesis Schueler divided the estrogens into four classes: 1) Substances in which the distance between the hydrogen-bond forming groups is nearly optimum, i. e. , 8.55A. These sub- stances will show the highest activity. 2) Substances in which this distance is appreciably larger than optimum. These substances will show a decrease in activity as the distance increases. 3) Substances in which the distance is less than optimum. These substances will show a decrease in activity as the deviation becomes greater. 4) Substances in which there is no hydroxyl or keto group, but which still possess estrogenic activity. The evidence cited by Schueler (28) in support of this hypothesis is given in Table I. The high activity of triphenylchloroethylene indicates that the critical distance of 8. 55A may be the distance between two hydrogen— bond forming groups, such as hydroxyl, keto, or simply hydrogen atoms that have been activated by the inductive effect. The inductive effect has been enhanced by the insertion of the chlorine atom in this case. This explains the activity of the chloro-compound and the lack of activity of the corresponding hydrogen-compound. The following observations may be cited as supporting Schueler's postulate: 1) hydroxy derivatives are more active than the corresponding keto analogs; 2) estrogen ethers decrease in activity in the order of the difficulty of hydrolysis, ethers which are extremely difficult to hydrolyze being inactive; 3) triphenyl- chloroethylene is more active than triphenylethylene, indicating that the ability of the para—hydrogens to form hydrogen-bonds is the essential feature of the molecule; and 4) trans-p, p'-dihydroxyazobenzene meets Table 1. Distance Between Hydrogen—bonding Groups vs. Estrogenic Activity. Group Substance Distanc ea Activity 1) Diethylstilbestrol (trans) 8. 55 0 . 3 trans-l,2-Di[-1-(4-hydroxy— 8.56 10.0 naphthyl)]-ethylene 1-Methyl-2-(4-hydroxyphenyl)- 8. 56 0. 5 3, 4—dihydro—6-hydroxynaphthalene 2) 1, 3-Di- (4-hydroxyphenyl)- 9 . 8 5 mg. 1, 2-diethylpropane 1, 4-Di- (4-hydroxyphenyl) 12.0 inactive l, 3-diethylbutane 3) trans-3, 3'-Dihydroxy—o., o.'- 7. 7 c diethylstilbene trans-2, 2'-Dihydroxy-o.,a‘- 5.9 d diethylstilbene p, p' -Dihydroxydiphenyl ether 8. 0 80 mg p, p' ~Dihydroxyldiphenyl 7. 1 100 mg. 4) Triphenylchloroethylene 8. 56 65 a . . The d1stances g1ven are calculated from bond lengths and bond angles for the compounds where direct measurements have not been made. b c Le s s than diethylstilbe strol . Less than trans-3, 3'—dihydroxy-u, (1' —diethylstilbene. The activities given are in gammas (lO'é'g.) unless otherwise noted. the spacial requirements for an active estrogen and, in spite of the dif- ferences in its structure as compared to the natural estrogens, shows considerable activity. It is a matter of conjecture as to whether the hydrogen—bonding is directly responsible for or whether a position of low electron density is necessary for an in vivo nucleophilic attack to produce an active com- pound. In view of the prolonged activity of the estrogens and their slow elimination, the latter seems quite reasonable. Another compound which does not possess two active hydrogen-bond forming groups but which does show estrogenic activity is racemic phenanthrol methyl ether (XIV, R : -CH3)(29). It is difficult to explain the lack of activity of the ethyl homolog (XIV, R: —C2H5). R One possibility is that the methyl . CH3 compound is a pro-estrogen and is ac ox1dized to some other form £1. v1vo CH3O producing a true estrogen. XIV Macovske and Georgescu (30) have suggested that the optimum separation of hydrogen-bonding groups is 9. 6A. The concept of an optimum separation of hydrogen—bond forming groups is quite incomplete without consideration of the other factors involved. In any case the distance of 8. 55A, as shown in Figure I, is not the distance separating the hydrogen-bond forming groups. For diethylstilbestrol this separa— tion is 15. 5A. Keasling and Schueler (31) modified the earlier postulate, claiming that for weaker hydrogen-bonding groups a separation of 9-10A was associated with estrogenic activity. In agreement with the hydrogen- bonding concept the value of functional groups to an estrogenic molecule is in the order of alcoholic hydroxyl < phenolic hydroxyl < carboxyl group. Oki and his collaborators (32) considered that besides the proper spacing of the hydrogen—bonding groups the proper orientation of the groups is necessary. Although the aliphatic portion of the artificial 10 estrogens does not resemble that of the natural estrogens, it does appear that there may be stereochemical requirements which may not be common to both the natural and the artificial estrogens. The stereo- chemical configuration is probably the most important single factor affecting estrogenic activity. All of the stereoisomers of equilenin have been prepared and tested. d-Equilenin, the naturally occurring isomer, is the most active, whereas i—equilenin is only one-thirteenth as active. Both d- and l—isoequilenin are inactive (9). The work of Serini and Logemann (33) with 8-isoestrone, which is about one-half as active as estrone, suggests that a cis B:C ring fusion may not be too detrimental to activity. The comparison of diethylstilbestrol with estradiol, as shown in Figure I, may be in error. X—ray crystallographic studies (34,35) on dienestrol indicate it to have the configuration shown in Figure 11, with the aromatic rings rotated about 500 out of the hexadiene plane. This is partially supported by the observation that in (1, (1'—dialkylsti1benes the aromatic rings are unable to attain the coplanarity and the full resonance energy of the unsubstituted trans-stilbene system due to steric interactions. This indicates that the formal similarity between the artificial estrogens and the natural estrogens may have been over OH O E /\ HO I HO XV XVI emphasized. The high activity of (XV)(36) and the low activity of (XVI) (37) further argue against this formal similarity. Though there are many such examples which tend to discourage the concept of formal 11 / OH I O OH HO O HO\/ I Figure 11. Probable Configuration of Dienestrol and Diethyl- stilbestrol. similarity, there are several factors which cannot be overlooked. Of the two isomers of diethylstilbestrol, the more active trans-isomer more closely resembles estradiol in size and shape than the less active cis-isomer. The more active meso-isomer of hexestrol, when repre- sented in the extended form, as in Figure 111, corresponds at the bridge carbons to the trans-B:C ring conformation of estradiol. Figure 111. Conformation of meso—Hexestrol. Structural features, other than the size and configuration of a molecule, which might effect its activity are the hydrophilic-lipophilic distribution, the rate of inactivation and excretion, and the ease of adsorption on the receptor organ. Grundland (38, 39) has suggested that estrogens and lipids form van der Waal's adhesion complexes in the body, but this postulate does not account for the influence of the number and spacing of hydroxyl groups on the molecule. In connection with adsorption it was demonstrated by Rideal and Schulman (40, 41) that the peak of estrogenic activity in a series of c, u'—dialkyl-p, p'-dihydroxysti1benes coincides with the peak of adsorption into a unimolecular layer. 12 B. Biological Alkylating Agents The first reference citing the unique pharmacodynamic function of biological alkylating agents was made by Paul Ehrlich (42) in an address dealing with the relationships between chemical constitution, distribution, and pharmacological action. In this report Ehrlich stated: I have made extensive experiments with many hundreds of different compounds, and in all of these I have discovered only one substance to which I am inclined to ascribe such a substituting action on protoplasm. This substance, vinylamine, discovered by Gabriel and described by him in a masterly manner, is formed by abstracting bromine from bromoethylamine by means of potassium. . . . Since then, however, Marchwald has shown con— vincingly that this substance cannot, as was at first supposed, contain a double bond. . . . It can therefore only possess the constitution of a dimethyleneimine. In view of this a complete analogy exists between the ethyleneimine and ethylene oxide. These two substances, ethyleneimine and ethylene oxide, are highly toxic compounds as has been shown by the researches of Levaditi and myself. The pathological changes excited by dimethylene— imine are especially interesting. Administered to a great variety of animals. . . . in doses which cause death after 1%- to 2 days or more, this substance causes total necrosis of the kidney papillae. . marked changes extending from the pelvis of the ureter to the urethra, and consisting of necrosis of the lining epithelium, haemorrages, and oedema. . . . Everyone who has learned to recognize these changes-—changes absolutely unique in pathology-- will be forced to the assumption that this localization is dependent on a direct attack of the vinylamine on the effected epithelia, an ethylamine group entering the protoplasmic molecule. Although these observations had been made by Ehrlich prior to 1898, it was not until 1942 that an agent of the "alkylating" variety was suggested for the treatment of certain neoplasms. The classified studies of Goodman and Gilman on the pharmocology of certain nitrogen mustards had shown the effect of these compounds on lymphoid tissue and rapidly dividing cells. After studies on experimental neoplasms and subsequent clinical trials it was found that methyl—bis(2-chloroethyl)amine, HN2, \ was effective in the chemotherapy of Hodgkin‘ 5 disease and caused limited 13 regression in certain cases of lymphosarcoma. Because of the military aspect of the nitrogen mustards these observations were not made generally known until 1946, when Gilman and Philips (43) summarized the wartime work on the biological actions and chemotherapy of the 2-chloroethyl amines and sulfides. At this time only two alkylating agents, both of nitrogen mustard type, had been clinically investigated, triss (2-chloroethyl)amine and methyl—bis(2=-chloroethyl)amine. Since the publication of this report hundreds of alkylating agents of various types have been synthesized and their basic biological action studied (44). Well over fifty of these compounds have warranted clinical trials. Today there can be recognized four general types of biological alkylating agents which have achieved some clinical success in the treat- ment of cancer: 1) ethyleneimines, 2) epoxides, 3) esters of sulfonic acids, and 4) nitrogen mustards. Bis(2-chloroethyl)su1fide offers little opportunity for modification and its high general toxicity limits its use as a selective oncotoxin, thus it will not be considered. 1) Ethyleneimine s The first ethyleneimine to show oncotoxic activity was 2, 4, 6—tris— (1—aziridinyl)-5—triazine (I) (45, 46). The related 2—(1-aziridinyl-4, 6- dimethoxy—S-triazine (II) is of interest, being one of the few known active monofunctional alkylating agents. Several phosphoramides and thio— phosphoramides have been investigated and found to possess some activity. N—(3-Oxapentamethylene)-N',N"-diethylenephosphoramide (47) (III) and its thio-analog are compounds of this type and have been found to cause complete regression of the Flexner-Jcbling carcinoma. l4 0 CH CH N CH CH CH z\N_ \ -—N/ 2 CH 0 \ -N/ | I 2\N _ 1TD - N/| 2 3 If \ / I \C CH2/ \CH2 N /N CH2 CI 2 /N HZ N /N Y }/ OCH3 CH2 \CH2 / \ | l CHZ CH,2 CH,2 /CH2 \0 I II 111 2) Epoxide 3 Although no epoxide has met major success in clinical investigation, they have given information which correlates with the structure—activity relationships of the other types of alkylating agents. Ross (48, 49) has demonstrated that a certain degree of chemical reactivity is necessary for biological activity of the epoxides, as has been noted for the other alkylating agents. In the series of diepoxides (IV) it has been shown (50, 51) that the oncotoxicity decreases from n = 1 to n = 6. Butadiene dioxide (IV, n = 0) has found some use clinically. /O\ /O\ CHz-CH-(CHZ)n-CH-CHZ IV 3) Sulfonic Acid Esters Probably the most studied and indeed the most clinically success- ful sulfonic acid ester is 1, 4-dimethanesulfonoxybutane, Myleran (V, n : 4). Myleran has promise in the treatment of chronic myeloid leukemia. Variation of the number of methylene groups (n = 3 to 10) CH3SOZO(CHz)nOSOZCI-13 V 15 shows that peak biological activity occurs at n = 4 and 5, although there is little or no change in the chemical activity. This observation led Timmis (52) to propose his hypothesis of cycloalkylation. The suggestion was made that the order of biological activity could be explained by the formation of a ring structure, involving the carbon chain of the alkylat— ing agent, by the dialkylation of a single group. This concept could be applied to the reaction with secondary phosphate groups occurring in, for example, RNA or ATP molecules. The ring formed by Myleran and the secondary phosphate would be a seven membered ring. Khorana and co-workers (53) have shown that the seven membered ring is the most stable ring size in the case of the cyclic phosphates. The fact that the acetylene analog (V1) is active against the Walker tumor, although it cannot undergo this type of cyclization because of its rigidity, has raised some doubts concerning this hypothesis as a general mechanism for biological alkylation . CH3SOzO—CHz—C E C—CHz-OSOZCH3 VI 4) Nitrogen Mustards The term nitrogen mustard is applied to those substances containing a bisI2-haloalkyl)amine group (VII) because of the structural similarity between these compounds and the well—known mustard gas of World War I, bisIZ-chloroethyl)sulfide (VIII). The vesicant properties of the nitrogen mustards were first observed simultaneously in 1934 by McCombie and /CHz-CH2-Cl /CHz—CHz—Cl /CHz—CH2—Cl R-N s C1-CHz-CH2—N \CHz-CHz-Cl \CHZ-CHz-Cl \CHz-CHZ-Cl VII VIII IX l6 Purdie (54) in Great Britain and Ward (55) in this country when they prepared trisI2—chloroethyl)amine (IX). Ward observed that on contact with the skin both this amine and its hydrochloride produced deep blisters which require months to heal and which left dark discoloration, reactions which are identical with those of mustard gas. He also noted that bisI2-chloroethyl)amine (VII, R = H) was not a vesicant. This, however, was not the first time such compounds had been prepared. . As early as 1897 Gabriel and Eschenbach (56) had prepared bisI2—bromo= ethyl)amine. Prelog and Stephan (57) prepared several N—alkyl nitrogen mustards as intermediates for the preparation of N, N'—disubstituted pyrazines. Included in this series of compounds was the now important methyl-bisIZ-chloroethyl)amine, but no mention of any physiological properties or difficulties in manipulations were made. The aliphatic mustards (VII, R is alkyl) react under physiological conditions via the cyclic ethylene immonium ion (X). When R is a simple alkyl group, the cyclic species is a true intermediate and has been iso- lated as the picryl sulfate salt (58). Bartlett, Ross and Swain (59), by a detailed kinetic study, have shown that the cyclic immonium ion displays a pronounced selectivity toward competitive substrates in aqueous media. The weak electrophilic character of the immonium ion, formed by the cyclization of the nitrogen mustards, is very important to their biological CHz—CHz-Cl ®/CHZ\ + C16 / R-N R-N —-—— CHz \ _— CHz—CHZ—Cl \C HZ—CHz-Cl VII X activity. It is due to this factor that the nitrogen mustards have a lower general toxicity than the sulfur mustards. The sulfonium ion inter- mediate (XI) of the sulfur mustard (VIII), being less stable, is highly 17 reactive and is an indiscriminate alkylating agent. Hz—CHz—Cl ED CH2 (3 \ s s -/—-——CH2 + C1 CHZ—CHz-Cl . \CHz—CHz—Cl VIII XI The main types of nucleophilic centers found in biological material, capable of acting as substrates for these agents, are organic and inorganic anions, amine groups and sulfide groups. Under physiological conditions there will not be any reaction with undissociated acid groups or ammonium cations, since these are not nucleophilic at this pH. There is a wide variation in the rate at which the reactive forms of these various groups will combine with alkylating agents. This nucleophilic capacity has been described by Ogstom (60) as the so—called competition factor. The value of the competition factor, F, relates to the rate at which a given center is alkylated (in vitro) under standard conditions. High values for F indicate a high affinity between the group and the alkylating agent. The effective competition factor of a center at a specified pH is given by the product of the competition factor, F, of the reactive form and of "f", which is a factor denoting the proportion of the total groups in the reactive form at the specified pH. Some values of effective competition factors at pH 7. 5 are given in Table II. The concentration, c, of the particular group must also be considered. This is of great importance in tumor inhibition where a limited amount of alkylating agent is available for reaction with a large excess of reacting groups. The expression F-f- c is applicable to homogeneous solutions containing simple substrate molecules, but is seldom more than qualitative in biological systems. Absorption of the reagent onto the macromolecular substrate will increase the effective concentration. Steric factors may decrease the effective 18 concentration as in the case of proteins where the active groups are folded into the helix structure. Table 11. Effective Competition Factors, Ff, of Reactive Groups Group F f Ff Terminal Cysteinyl 105 10°"1 104 Aromatic amino 102 l 102 Phosphoryl 50-100 1 50-100 Carboxyl 10—80 1 10—80 Imidazole (=NH) 70 0. 8~1 56—70 Nonterminal cysteinyl 105 10‘“4 10 Aliphatic amino 4 x 102 10' -10' 0.4-4 On the basis that the dividing cell is more susceptible to the action of an alkylating agent than is a resting cell, Ross (48) has hypothesized that the reaction with nucleic acids is responsible for the biological effects of these reagents. There is an accumulation of evidence to sup- port this hypothesis. Herriott (61) has shown that the pneumococcus transforming principle, composed largely of DNA, exhibits the greatest sensitivity toward mustard gases of all the materials examined. Also, Revell (62) has shown that when chromosome aberrations are produced in dividing cells of Vicia root tips by the action of alkylating agents, the effected positions are not random, as in the case of radiation, but are localized in regions of high nucleic acid content. There is also the possibility that these agents produce their characteristic effects by enzyme inhibition. It has been shown that phos- phokinase, certain peptidases, choline oxidase, and acetylcholine esterase are extremely sensitive to various alkylating agents (63). \$ 19 The objection to the enzyme inhibition hypothesis is that the concen— tration of agent generally required for inactivation does not appear to be achieved invivo when cytotoxic effects can be demonstrated (64). One of the first modifications of the nitrogen mustard structure was the formation of the amine oxide (X). I The oxide is a weaker base than the free amine and therefore less reactive. Although the oxide (X) is devoid of any vesicant action and displays less general toxicity, it o T/CHz—CHZ-Cl CH3-N \CHz-CHZ-Cl X is reported to have greater antitumor efficiency than the free amine (65) Varying the R group in VII, as would be expected, alters both the chemical and biological reactivity of the mustard. Ross (50) has shown that a certain degree of chemical reactivity, as shown by ease of hydrolysis in 50% acetone, is necessary for biological activity. From the data in Table III it is apparent that electron-donating groups enhance the chemical reactivity and render the mustards biologically active, whereas an electron-withdrawing group decreases the chemical reactivity to the point where they will not alkylate under physiological conditions. This trend of chemical reactivity depends on the basic strength of the amine or, more specifically, on the availability of the free electron pair on the nitrogen for participation. It should be pointed out, however, that the formation of the cyclic ethyleneimmonium intermediate has not been proven in the case of the aryl mustards. Nevertheless, the increase in reactivity, both chemical and biological, over the simple alkyl halides must be due to anchimeric assistance from the fi—nitrogen, even if the cyclic structure, as such, is not formed. 20 Table III (48). Relation of Biological Activity to Chemical Reactivity __ /CH2-CHz-C1 R \ / -N\ CHz-CHz-Cl Chemical a Biological R— Reactivity Reactivityb NHZ- 100 + CH3O- 58 + NHAc— 41 + CH3— 38 + H— 20 + Cl- 9 _ COZEt- <1 _ CHO— <1 _ HZNOC- <1 _ aPer cent hydrolysis in 50% acetone at 660 in one-half hour. Inhibition of Walker rat carcinoma 256. There have been several series of closely related nitrogen mustards studied in which the trend of chemical reactivity and biological activity do correspond. Table IV shows such a series of compounds. Ross (48) attributed these differences to the so—called transport characteristics of these compounds. The fact that mustards do vary in their biological properties inde- pendently of the chemical reactivity led to a search for alkylating agents which show some specificity of action. The introduction of a certain 21 Table IV. Variation in Biological Activity of Compounds of Comparable Chemical Reactivity /CH2-CHz-Cl R _ N\CHz-CHz-Cl Chemical a Biologicalb R Reactivity Reactivity —COOH 15 i —CHZCOOI-l 39 +1, -CHZCHZCOOH 41 + —OCH2COOH 48 + -CHZCHZCH2COOH 42 +++ -OCH2CH2COOH 52 +++ —CHZCHZCHZCHZCOOH 39 i -OCHZCH2CH2COOH 50 + -OCH2CHZCH2CHZCOOH 44 ++ 8"Per cent hydrolysis in one-half hour in 50% acetone at 660C. Inhibition of Walker rat carcinoma 253. "carrier" group in the alkylating agent may influence its arrival at the site of action without affecting its chemical reactivity. An example (66) of this type of specificity is exhibited by Melphan, L-3-(p-[bis-(2-chloro- ethyl)amino]phenyl)alanine (XI) and its D-isomer. The L-isomer of NH,2 /CHz—CHz-C1 l CH -—CHZ- -N l \ COZI-I CHZ-CHz-Cl XI 22 this phenylalanine mustard shows intense inhibition of the Walker rat carcinoma 256; the D—isomer shows only slight inhibition under similar conditions, and the DL—form shows an intermediate degree of activity (67, 68). The reason for the lack of activity of the D—isomer has not, as yet, been adequately explained. Efforts to design nitrogen mustards that are activated in vivo have been made. Ross and Warwick (69) have synthesized a series of substi= tuted p-{bis(2—chloroethyl)amino]azobenzenes (XII), which are very poor alkylating agents, as such, but are reduced i_n vivo to N, N—bis(2-chloro- ethyl)—p-phenylenediamine (XIII), which is a good alkylating agent. By varying the ortho substituent (X in XII), it was shown that the /CHz-CHz—Cl /CHz—CHz—C1 —> HzN -N N—\CHZ CH2 C1 \CHz-CHz-Cl XII XIII biological activity of these compounds is directly proportional to the ease of reduction by the xanthine-xanthine oxidase system (70, 71). Another recently investigated class of mustards are the cyclic phosphamides (XIV), first synthesized by Arnold and Bourseaux (72) in 1957. \“Cyclophosphamide” (XIV, n = 2) has been given a brief clinical trial (73). NH ./\ O \P / n\ / /CHz—CHz-Cl \CHz—CHZ-Cl XIV It possesses a degree of antitumor activity comparable to nitrogen mustard (VII, R = CH3), but appears to cause less bone marrow 23 depression, one of the more serious side effects of nitrogen mustards. This phosphamide is inactive iii vitro, its biological activity being due to the breakdown of the cyclic structure in the host cells. The postulate is that activation of the chloroethyl groups occurs through the influence of the enzymes phosphamidase and phosphatase, which are present in higher concentrations in neoplastic cells than in non-neoplastic cells. Thus', this type agent holds promise of displaying a specificity of action. Extensive investigation of the steric requirements and the necessity for bifunctionality for oncotoxic activity has been investigated by a number of workers. In compounds of the general structure (XV) the activity is limited (74) to the two compounds where m = n = 2 and m = 3, n : 2. This work supplements that of Varghn e_t a1. (75), whose “mannitol mustard” (XVI) possesses a high degree of activity. The closely related dulcitol mustard (XVII) and the 1, 6—hexamethylenediamine mustard (XVIII) are inactive. Ar-1\lI-(CH2)m_1|\I_AR (CI-12),, (CHZ)n C1 C1 xv Clle-NH-CHz-CHz-Cl (IJHz—NH—CHz-CHz-Cl (IZHZ—NH—CHz-CHz-Cl HO— p-11 H-(IJ-OH cle, Ho- (|:-H HO-Cll-H CHZ I H- (ll-OH Ho-<|:-H CH, H- C—OH H-Cll-OH (III-12 I CHz-NH—CHZ-CHz-Cl CHz-NH-CHz-CHz-Cl CHz-NH-CHz-CHz-Cl XVI XVII XVIII 24 Other variations of nitrogen mustards that have shown oncotoxicity are those with a physiologically active carrier. The success of many chemotherapeutic agents can be attributed to the incorporation of a specific carrier moiety into the molecule. Ing (76) had distinguished between the pharmocodynamically active portion and the carrier group in several series of physiologically active compounds. The concept that biological alkylating agents can be considered as being composed of an active alkylating group and a carrier moiety was first introduced by Bergel (77). Typical of the nitrogen mustards, incorporating a physio- logically active carrier are the pyrimidine-containing nitrogen mustard, 5-[bisI2—chloroethyl)amine]-6-methyl-2,4-pyrimidinediol (XIX) (78), the pyridoxine derivative, 5-[bis(2-chloroethyl)aminomethy]]—4-(methoxy— methyl)~2-methyl-3-pyridinol (XX) (79), the antimalarial analogs (XXXI-XXIII) (80, 81), and the benzimidazole mustard (XXIV) (82). OH _ /CHZ-CHz-Cl HO —- CHZocH}CHZ_CHz—C1 HO N CH3 \ CHz-N \N / \CHz-CHZ-Cl N / \CHZ-CHz-Cl CH3 XIX XX CH, CHz-CHz-Cl "OH NH_CIH(CH2)3N/ CH -CH -C1 \ / Z 2 CH ~CH -C1 \ N\ \ z z CHz-CHZ-Cl / / C1 N C1 N XXI XXII 25 CH3 NH—CI—IICHZ)3N \ \ CHz—CHZ—Cl N\ /CHZ_CH2_C1 N/ CHZ'N\CHZ-CH2_C1 I? H /CHZ-CHz-Cl XXIII XXIV A recently prepared nitrogen mustard, which has been shown to be effective against a wide variety of transplantable rodent tumors and several human neoplasms, is 5-bis(2—chloroethy1)amino» uracil (XXV) (83, 84). Based on recent clinical evaluation this mustard has been sug- gested for chemotherapy in the "lymphoma—leukemia”category (85, 86) . Uracil mustard also has the advantage of being equally effective when administered by oral and parenteral routes. 0 /CHz—CH2-Cl H-N N 01\ I \CHz-CHz-Cl N | H XXV The present work involves the synthesis of nitrogen mustards incorporating an "estrogenic carrier. " In order to evaluate the effectiveness of an estrogen as a carrier moiety, the preparation of the nitrogen mustard analogs of meso—hexestrol, d, l—hexestrol, and diethylstilbestrol was undertaken. The dimethyl ethers of these com- pounds were also desired in order to compare the effectiveness of a ”pro—estrogenic carrier. ” The nitrogen mustards, as their hydro- chloride salts, were submitted to the Cancer Chemotherapy National Service Center for screening. II Cancer Chemotherapy National Service Center, National Institute of Health, Bethesda 14, Maryland. PART I NITROGEN MUSTARD ANALOGS OF meso-HEXESTROL, d, l-HEXESTROL AND THEIR DIMETHYL ETHERS DISCUSSION Preparation of the Nitrogen Mustards In the hope of retaining as much of the estrogenic activity as possible in the hexestrol carrier it was decided to leave the aryl—portion of the molecule unsubstituted and replace one of the ethyl groups with a N, N—bis(2—chloroethyl)aminomethy1 group. Thus, the compounds desired are the threo— and erythro-modifications of N, N—bis(2-chloroethyl)-2, 3—di- (p:hydroxyphenyl)pentylamine hydrochloride (1e and It,. R = H)1 and their dimethyl ethers (1e and It, R = CH3). The erythro-modification CHZN: H CHz—CHz-Cl HSCZ ‘. H H5Cz H. CHz-CHz-Cl 1e 11 1Throughout this thesis compounds are numbered 1, II, III, etc., when no specific reference to a stereoisomeric modification is being made, or to a single stereoisomeric form when the other forms are designated by a separate number. The designation “1e" refers to the erythro form of compound "I" and "It" to the corresponding threo form. 26 27 corresponds to the estrogenically active meso-hexestrol (II) and the threo—isomer to the less active d,1-hexestrol (III). OR QR S H\ /—CH2 CH2- C1 :1“ 2H CH2 CH2 C1 H CHZ—CH3 OR It 111 Three synthetic routes for the preparation of these compounds were devised and investigated. For convenience these routes will be discussed separately as schemes A, B, and C. The synthesis of 2, 3-di- (p-methoxyphenyl)pentanenitrile, the intermediate common to all three synthetic routes, is represented schematically in Figure IV. The anisyl alcohol was prepared by both the catalytic hydrogenation of anisaldehyde (87, 88) and by the crossed Cannizzaro reaction with anisaldehyde and formaldehyde (89, 90). The catalytic reduction, using Raney nickel, gave a product inferior to that obtained by the crossed Cannizzaro re- action, making the latter the method of choice despite the lower yields and the inherent experimental difficulties. The wide distillation range of the alcohol obtained from the hydrogenation is undoubtedly due to closely related products arising through partial hydrogenation of the ring. Commercial anisyl alcohol (Eastman white label) was found to be of the same general quality as that obtained by the hydrogenation of anisaldehyde. Z8 .mofinficodds‘nmaAacmfiminxoguogamrflIm .N 3.5803 93 mo mfimofifigm .>Hopdwfim as end 83 20 _ m I I I I n moo ©mo mw O 0 mo 5 Emma a; £5 E; H; a m N .N £00. .0” mo- ommo LE ZONEU- -onmo i; E .20 g > Z N l m. T N I m jlll I M. 5 mo ©o mo Gm no mo ©o mo mom ovum omo ©o mo 29 Rorig, (91) reported that the yields of Z, 3-di(p-methoxyphenyl)- acrylonitrile (VIII) obtained from the condensation of p—methoxyphenyl- acetonitrile and anisaldehyde varied greatly with the method used for the preparation of the p—methoxyphenylacetonitrile. He claimed that the procedure of Livshits (92), in which aqueous dioxane is used as the reaction medium for the p-anisyl chloride and potassium cyanide, gives a product superior to that obtained by the procedure of Lee (93) in which the potassium cyanide is suspended in anhydrous acetone. The procedure using anhydrous acetone was later improved and is described by Rorig in Organic Syntheses (94). The quality of the nitrile obtained by both of these procedures was good. It was found more convenient to prepare p—anisyl chloride from anhydrous hydrogen chloride, as described by Livshits, rather than to use concentrated hydrochloric acid as in the Organic Syntheses procedure. The use of concentrated hydrochloric acid often led to an emulsion. This made the isolation and drying of the p-anisyl chloride virtually impossible. 2, 3-Di(p-methoxyphenyl)acrylonitrile (VIII) was prepared by the condensation of anisaldehyde and p-methoxyphenylacetonitrile following Niederl's (95) modification of Kohler's (96) original procedure. The preparation 01:32, 3—di(p-methoxyphenyl)pentanenitrile (IX) by the conju- gate addition of ethyl magnesium bromide to Z, 3-di(p-methoxyphenyl)- acrylonitrile was a modification (97) of Kohler's (96) procedure. The most common difficulty usually encountered in following Kohler's pro- cedure has been the addition of the ether insoluble unsaturated nitrile to the reaction mixture. Kohler, using a-phenylcinnamonitrile, added sufficient solid nitrile to the refluxing Grignard reagent to obtain a permanent red coloration in the reaction mixture. The p-methoxy- substituted compound used in this work does not give a colored complex, necessitating the addition of a calculated amount of nitrile. Several recent investigators (98, 99) have extracted the nitrile into the reaction 30 mixture by means of a Soxhlet extractor. Neither of these methods of addition is satisfactory for large scale preparations. In this work it was found that a solution of the nitrile in benzene or tetrahydrofuran gave excellent results and was far more convenient. Scheme A The reaction sequence called "scheme A" is shown in Figure V.. It offered the distinct advantage of the facile separation of the two diastereomeric Z, 3-di(p-methoxyphenyl)pentanenitriles (IXe and IXt, R = CH3). It has been established by previous workers (96, 97) that one of these isomeric nitriles is an ether insoluble solid, whereas the other is an ether miscible oil. Since the subsequent steps do not affect either of the asymmetric centers, the stereochemistry of the resulting nitrogen mustards would be determined by the correlation of either of the inter— mediate disastereomeric nitriles (IX) or amines (X) with either meso- or d, l-hexestrol. This correlation has been accomplished and will be discussed in the second part of this section. The two isomeric pentanenitriles were conveniently separated by filtration to give pure solid erythro-Z, 3-di(p-methoxyphenyl)pentane- nitrile (IXe). The threo—Z, 3-di(p-methoxyphenyl)pentanenitrile, being a viscous oil, was not purified other than by simple distillation, leaving it contaminated with some dissolved erythro-isomer. Reduction of erythro-Z, 3—di(p-methoxyphenyl)pentanenitrile (IXe) with an equal molar mixture of lithium aluminum hydride and aluminum chloride (100) followed by acid hydrolysis with hydrochloric acid gave erythro—Z, 3-di(p-methoxyphenyl)pentylamine hydrochloride (Xe- HCl) in near quantitative yield. The amine hydrochloride, which is insoluble in the ether and in the concentrated aqueous salt solution, was isolated by simple filtration. The isolation of threo—2, 3-di(p-methoxypheny1)pentyl- amine proved to be more difficult. On acid hydrolysis of the lithium .< meflom >9. moswamfficom IS>G0£m>xo£meImvfipIm iNIATAJuooHoEoINVwfimIH/H .Z UHHoEOmH 05. mo coflsmhmmohna .> ohdwfim Mao-” m .H 1 8m on 3 kaokmo/ z NEONEUNmon Ho-~mo-~mo\_ wa Nmo _ _ m I I m n I I I I m m m - m Til m mo m i 00 ©mo w o o £00m moo o _ o o ”metro x £5 £6 / \ o .x on N 23w z_o n H + £00.\ / moio- .038 %.l mmoo.\ /-mo-mo- .035 II _ $35 I: _ mENO mENO 32 aluminum hydride complex this isomer formed a liquid hydrochloride which became a sirupy layer at the interface of the salt solution and the ether layer. Hydrolysis of the complex with base did not give clean alumina, which is formed in most cases, but instead gave a tan paste which occluded much of the amine. There was always a small amount of solid erythro-amine isolated from the reduction product of the threo-nitrile. This was probably due to impure nitrile rather than to partial equilibration taking place during the reduction, since no equi- libration was observed during the reduction of the erz. thro-nitrile. McKay (101) reported the successful preparation of N, N—bis- (Z—hydroxyethy1)-2, 3-di(p-methoxyphenyl)propylamine (XIII, R = R‘ = -OCH3) in good yield by the addition of the corresponding amine (XII, R = R' = -OCH;,) to two moles of ethylene oxide. Despite the apparent similarity between the amines used by McKay and those dealt with in this work, his procedures did not give satisfactory results with either of the isomeric 2, 3-di(p-methoxyphenyl)pentylamines (Xe or Xt). Addition of a slight excess of ethylene oxide to a cold methanol solution of either the threo— or the efl.’ thro-amine resulted in recovery of the Q .. CH;—/—->©H ~ HHHQ H—.H WC) NHz N(CH2CHZOH)Z XII XIII starting material. The addition of a slight excess of ethylene oxide to a methanol solution of erythro-Z, 3-di(p-methoxyphenyl)pentylamine at 500 gave an impure product from which only a small amount of the desired diethanolamine could be isolated by chromatography on alumina. This latter procedure for the hydroxyethylation of the erythro-amine, 33 accompanied by the chromatographic separation, resulted in a 43% overall yield. Considering the time required for the chromatography, this yield is too low to be of general preparative value. The threo- amine was recovered unchanged regardless of the conditions employed for the addition of the ethylene oxide. The identical procedure, which gave unsatisfactory results with the isomeric 2, 3-di(p—methoxyphenyl)pentylamines used in this work, gave results comparable to those reported by McKay when applied to the hydroxyethylation of 2, 3-di(p-methoxyphenyl)propylamine (XII, R = R! -OCH3) or 2, 3-diphenylpropylamine (XII, R = R' = -H). Scheme B One of the more obvious routes to the preparation of the isomeric N, N-bis(2-hydroxyethyl)-Z, 3—di(p—methoxyphenyl)pentylamines is by the alkylation of diethanolamine with the isomeric l—chloro-Z, 3—di(p—methoxy— phenyl)pentanes (XVI). This is the basis of Scheme B, which is repre— sented in Figure VI. The hydrolysis of 2, 3—di(p-methoxyphenyl)pentanenitrile to a mix- ture of the isomeric 2, 3-di(p—methoxyphenyl)pentanoic acids is discussed under ”Scheme C" in this section. The lithium aluminum hydride re- duction of either the free erythro—Z, 3-di(p-methoxyphenyl)pentanoic acid (XIVe) in tetrahydrofuran or its methyl ester in ether gave erythro-Z, 3- di(p-methoxyphenyl)—l—pentanol (XVe). This alcohol is a well defined solid which was isolated from the reaction mixture in almost an analytic- ally pure state. On reduction methyl threo—2., 3-di(p-methoxyphenyl)— pentanoate gave the threo-alcohol (XVt) as a colorless viscous oil, which gave good analytical data, but which still may have been contaminated with the erythro-alcohol. The less hindered erythro—Z, 3-di(p-methoxyphenyl)—l-pentanol was converted to the chloride (XVIe) with phosphorus pentachloride. 34 .m mammom >0. m.omflgmatwudogfitncomgwxomuo£7337.MrmnSkhsuoOHoEoINvmfimJ/H .2 3.5503 0%. mo mofidemQOHQ .H> ohswfih nmo. n m .momm mm Amononovz. mom :mONmonov n m m m moo -©-m_o mo- mo- .60 mo All-£00m. moo -Qmw mo- mo- _--0© mo A||I|zl:mo~mo~mov H>x m NH.U mo mm~o_ H>X >X INmO INHIHU n I m .Allli m I 6 Ali moo -mo- mo fl-Oo mo 30m moo -mo- mo _--0© mo £23 Ex mHIHNU mmN_ Ex VS m~oo 20 n On .0 “Mr M I I _ I I I n moo- -mo- mo- mo Alaa moo mo mo 0 mo mmmo “_o mmN 35 This chloride failed to crystallize. It was used for alkylation without purification. The infra-red spectrum of the chloride, however, did not show the characteristic hydroxyl bands at 3660 and 1050 cm'1 which were well defined in the spectrum of the alcohol. The alkylation of diethanolamine was carried out in refluxing ethylene glycol. The progress of the reaction was followed by titrating the remaining free amine with standard hydrochloric acid. As shown in the equation below, one mole of amine is neutralized for each mole of diethanolamine alkylated. The details of this alkylation procedure are given in the experimental section. The alkylation was found to be too slow to be of any synthetic value. R-CHz-Cl + HN(CH2CH20H)z—-9 [R-CHz-NH(CH2CH20H)Z@+ (JICa Changing the solvent to formamide, which has a relative quaternization rate approximately three times that of ethylene glycol (102), gave no better results. Treating the chloride with a ten-fold excess of di- ethanolamine at 2000 for 72 hours gave no isolatable tertiary amine. The failure of hindered halides to alkylate diethanolamine has been reported by several investigators (103, 104). Scheme C The successful preparation of N, N-bis(2-hydroxyethyl)-Z, 3-di- (p—methoxyphenyl)pentylamine (XI) was realized through the reduction of N, N—bis (Z-hydroxyethyl)-Z, 3-di(p-methoxyphenyl)pentanamide (XVIII), as shown in "Scheme C, " Figure VII. Hydrolysis of the pentane- nitrile (IX) with sodium hydroxide in ethylene glycol gave a mixture of the two diastereomeric Z, 3-di(p-methoxyphenyl)pentanoic acids (XIVe and XIVt). Since either isomeric nitrile is completely equilibrated to a mixture of the isomeric acids by this alkaline treatment, usually the rm 36 .O mammom >3 momfigmfifidomfitnmommtnxomwoSIB$7m .NIAthmuoohoafloINVmflmIH/H .Z oflmoaofl 9.3 «0 mofludpmmonnm .HH> oudwfim nmo- I m .H on monono / . moamono/ z mom \2. mom monono\_ m0~mo~mo~m\m Nmo__ nmoo. -mo m_o- -0mmo I'll mmoo. -mo mo- onmo IIIHmIthmE/x N50m £25: mmwnw mHIHNU E>x m>x mo~mo~mo/ /z N N m0 mo moo m o 0 o m m n I n m m j m m i > oo -0©o 0 mo _--0© mo Amo :monmoz moo ©-_ mo- o _©0 0 N60m mm mm~_ mE-L fix on m...oo n m n I n 300 I©I_ ID. ED _IIO© OTD Alli-$002 $00 IQ. EDI OEOI© 0 EU mmuw mmNU 37 crude nitrile from the Grignard reaction was used without any attempt to separate the isomers. The pure erythro—acid could be obtained by recrystallization of the mixture of crude isomeric acids from methanol. The residue from the filtrate, containing the crude threo—acid, failed to solidify. Hunter and Korman (99) esterified this crude acid with diazo— methane, purified the resulting methyl ester, and upon saponification with methanolic potassium hydroxide obtained the pure threo—acid. Their procedure for obtaining the free threo—acid failed in this work. Other investigators (105) have also reported difficulties in attempting to repeat this preparation. Esterification of the crude threo-acid did give the methyl ester as reported. This ester could be purified, but saponifi— cation, by the procedure of Hunter and Korman gave impure acid which could not be recrystallized to purity. The methyl ester could also be made by converting the crude threo-acid to the acid chloride with thionyl chloride, followed by treatment with anhydrous methanol. This latter method was more convenient for large scale preparations, but it gave slightly poorer yields. The quality of the ester seemed to be the same in both cases. Treatment of the ester with dimethyl sulfate in order to methylate any free hydroxyl groups, prior to hydrolysis, did not improve the quality of the acid obtained. The impure acid probably arises through the partial isomerization of the threo-isomer to the erythro—isomer during the saponification. This seems analogous to the observation of Hauser and co-workers (106) that threo-Z, 3—diphenylbutanenitrile is completely isomerized to the more sterically favored erythro-form in liquid ammonia with a catalytic amount of lithium amide. Attempts to saponify the threo— ester with one equivalent of potassium hydroxide—ethylene glycol in benzene (107) were unsuccessful. Only unreacted ester was recovered. Assuming that the difficulty experienced in the isolation of pure threo-2, 3-di(p-methoxyphenyl)pentanoic acid from the saponification of its methyl ester is due to partial isomerization of the threo-isomer to the 38 more stable erythro—isomer, then the use of an acid catalyzed hydrolytic procedure should afford pure threo-acid. Acid hydrolysis of the methyl ester could not be effected in aqueous hydrochloric acid. Severe reaction conditions had to be avoided in order not to cleave the methoxy groups. Z-Pyranyl esters are known to be hydrolyzed readily under mild con_ ditions in dilute acid (108). If the Z—pyranyl ester of threoHZ, 3-=di(pH= methoxyphenyl)pentanoic acid (XIXt) could be prepared and purified, it appeared that it might be hydrolyzed to give the pure threo-acid. The Z-pyranyl ester (XIXe) of the more readily available erythro—acid was prepared in order to explore the experimental details of the preparation (109). It was obtained in an 89% yield employing p-toluenesulfonic acid 0 XIV (impure) W CH3O-©-CC CH— CH- =OCH3 XIXt (3sz CH CO H, H 0 _ CO;_ > XIXt —;——Z———Z—-> CH3OO -CH- CH 0CH3 \ XIVt as the catalyst. The threo-ester could not be prepared with the procedure developed with the erythro-acid. By using sulfuric acid as the catalyst Z-pyranyl threo-Z, 3-di(p-methoxyphenyl)pentanoate could be prepared in low yield. This ester did not crystallize readily and was extremely difficult to purify. The threo-acid obtained by hydrolyzing this Z—pyranyl 39 ester was of no better quality than that used for the preparation of the ester. There is sufficient difference in the pKa values of the isomeric pentanoic acids (XIVe and XIVt) to permit the selective precipitation of the threo—acid. This procedure afforded an initial precipitate of impure threo—acid, and upon further acidification pure erythro—acid separated from solution. The crude threo—acid obtained by this pr0= cedure could not be purified by recrystallization from any of the solvents tried. Attempted purification of the crude threo-acid by chromatographic adsorption on alumina was also unsuccessful. Only unchanged nitrile was isolated from the reaction mixture resulting from an attempted hydrolysis of the threo—nitrile (IXt) with methanolic hydrogen chloride . An attempt was made to prepare threo—3, 4—di(p—methoxyphenyl)- Z-hexanone (XXt). It was anticipated that the hypochlorite oxidation of this ketone would afford the threo-acid under conditions mild enough to prevent the partial isomerization to the erythro-acid. Rorig (110) has reported that the ketone did not form from the reaction of an equi- molar amount of methyl magnesium bromide and threo-Z, 3—di(p—methoxy- phenyl)pentanenitrile. This is undoubtedly due to the destruction of the Grignard reagent by the abstraction of the acidic alpha-hydrogen. Burckhalter and Sam (105) attempted to prepare the threo-ketone (XXe) by using an excess of methyl magnesium bromide with the threo—nitrile (IXt) but they could not isolate a product which met analytical require— ments. However, they did isolate a small amount of the threo-ketone as a side product from the reaction between the erythro—nitrile (IXe) and methyl magnesium bromide. Wawzonek (111) succeeded in isolating a mixture of both isomeric ketones by the complete equilibration of either isomeric nitrile during the reaction with an excess of methyl magnesium 40 Csz Csz 30 zCHs XXt :iNHaOCI CH30 ©_CH_ CH HQ OCH3 30 COZH XIVt bromide. In this work the product formed from the reaction of a five- fold excess of methyl magnesium iodide and the threo-nitrile was similar to that obtained by Burckhalter and Sam. No solid was obtained on work— ing up the product. The only acidic material isolated from the hypochlorite oxidation of this crude product was a very small amount of erythro—Z, 3-di- (p—methoxyphenyl)pentanoic acid. Zone precipitation of the crude acid, obtained by the methanolic potassium hydroxide saponification of methyl threo—Z, 3—di(p-=methoxy- phenyl)pentanoate, with an anisole solvent afforded a small amount of pure threo—2, 3-di(p-methoxyphenyl)pentanoic acid. All of the threo—acid used was first purified by this method. The preparation of erythro-Z, 3—di(p-methoxyphenyl)pentanoyl chloride (XVIIe) was straight forward when redistilled Matheson, Coleman and Bell thionyl chloride was used. Eastman white label thionyl chloride, even after an elaborate purification (112, 113, 114), gave an impure product. Other than the careful removal of the excess thionyl chloride, the acid chloride was not purified before being used to acylate diethanolamine. The acylation of diethanolamine according to the procedure of Genster and Sherman (115), using benzene as the solvent, gave poor yields. Of the several solvents tried for this acylation, dioxane was found 41 to give the best results. The N, N—bis(2-hydroxyethyl)—erythro-Z, 3- di(p-methoxyphenyl)pentanamide (XVIIIe) prepared by this procedure could not be recrystallized to analytical purity. A melting point range of less than 200 was seldom noted, even after repeated recrystallization. The product obtained from the acylation is probably a mixture of N— and O—acylated diethanolamines (XVIIIa). Separation of this mixture was not deemed necessary since the hydride reduction will reduce the ester-= linkages to the desired hydroxyl groups. The ether soluble alcohol (XV), the only major impurity, was easily removed by forming the ether insoluble amine hydrochloride (XIH HCl) in an ether solution. The lithium aluminum hydride reduction of the erythro-diethanol— amide (XVIIIe) proceeded very well in refluxing tetrahydrofuran when sufficient reagent was used to combine with the free hydroxyl groups (or to reduce the esters XVIIIa) as well as to reduce the amide. The hydrolysis of the reaction complex was accomplished by complexing the aluminum salts with an ammonium tartrate solution. The addition of 2H5 ‘f R = CH3O-©-CH—(|3H—©-OCH3 /CHZCHzO-(COR) R-CO-Cl + HN(CH2CHZOH)Z ——> XVIII + R-CO—N \CHZCHzO-(COR) XVII XVIIIa CH CH 0H - / Z Z XVIIIa 141—“£44. R-CI—Iz-N + RCHZOH \CHZCHZOH XI XV anhydrous hydrogen chloride to an etheral solution of the crude diethanol- amine (Xle) precipitated relatively pure amine hydrochloride. 42 The erythro-nitrogen mustard hydrochloride(Ie-HC1, R = -CI-I3) was prepared from bis(2=hydroxyethyl)[erythro—Z, 3—di(p=methoxyphenyl)- pentyl]amine hydrochloride in a chloroform solution using freshly dis- tilled Matheson, Coleman and Bell thionyl chloride with a trace of dimethyl- formamide as a catalyst (116, 117). Attempts to prepare the nitrogen mustard hydrochloride without the dimethylformamide or with purified (112, 113, 113) Eastman white label thionyl chloride gave tarry products. It was necessary to remove the unreacted N—alkyldiethanolamine from the crude product with calcium chloride (118) and then carry out a chromato- graphic purification on alumina in order to obtain pure product. The identical procedure which had been used for the successful preparation of the erythro-nitrogen mustard hydrochloride (Ie.HCl, R 2 -CH3), was employed in an attempt to prepare the threo-nitrogen mustard hydrochloride (Ital-1C1, R = -CH3). threo-2, 3—Di(p—methoxyphenyl)pentanoic acid, which had been purified by zone precipitation, was used for this preparation. The acid chloride (XVIIt), the diethanolamide (XVIIIt), the diethanolamine hydrochloride (XIt) and the resulting threo—nitrogen mustard hydrochloride (It. HCl, R 2 -CH3) were all found to be viscous oils which failed to solidify. The final product possessed a very dark brown coloration which was not lightened by chromatography on alumina followed by precipitation from ether with anhydrous hydrogen chloride. erythro-Z, 3-Di(p—hydroxyphenyl)pentanoic acid (XXIe) was prepared by treating the dimethoxy acid with pyridine hydrochloride. The free hydroxy acid was too insoluble in most solvents to be conveniently recrystal- lized to maximum purity or to be used in subsequent reactions. By con- verting it to the diacetyl derivative (XXIIe) the acid was easily purified ‘ and was soluble enough to be conveniently converted to the acid chloride (XXIIIe). The amide (XXIVe) produced by the acylation of diethanolamine with the acid Chloride (XXIIIe) was, in part, the free phenolic compound, produced by aminolysis of the acetyl group. The loss of the acetyl groups 43 . opfimozoonpknn omfigdatnpmomAH>Comm>x0Hp>m Imvflou m .N IA m- I m .mom N50m BOA :mowmoflmovz _ Nmo _ THO Q THU HAM—v OOH-H AwmiH‘oJu-w . D mHIHN HHHXN HOOD mmNo HunmuoOHoEoINmemIZ .2 mo Gofiudndaonm HHH> ohdmwrm .H Exx NAmONmonovz _ .._. I I I T! «0 Omo mn_o Oom NAmONmonovzm Exx ..mNo HHXX m~oo _ _ m N I I I IN m m N I I M m m 00 m m m i m m mo Ali x o 0 o ”_0 oo o :08 mooo o ”.0 oo oroommov H x 0 USA mmo o _ EO I©IEOIEOI _ mm...o mHIHN 2x .3. \ / -0m 3 mmoono-mo-Oommo om z m o ... ~_ 1 m o 44 made the isolation of this compound different from the isolation of the dimethoxy-analog (XVIII). The amide with the free phenolic groups was insoluble in all the water-immiscible solvents which were tried for extraction. It was finally allowed to solidify from the water solution and was isolated by filtration. This amide was reduced with lithium aluminum hydride in tetra- hydrofuran without an attempted purification. Here the amide can exist not only as a mixture of O— and N-acylated products, as was true in the case of the dimethoxy—analog, but also as a mixture of the p-acetoxy- (XXIV, R = -OZCCH3) and p-hydroxy- (XXIV, R = H) compounds. Separation of this mixture was not considered necessary since the product of the hydride reduction would be the desired N, N-bis(Z—hydroxyethyl)— Ierythro-Z, 3—di(p-methoxyphenyl)pentyl]amine no matter which components were present. The amine hydrochloride was formed in solution by passing anhydrous hydrogen chloride through a methanol solution of the crude reaction residue. The addition of ether to this methanol solution pre— cipitated solid amine hydrochloride. The preparation of the nitrogen mustard hydrochloride (1» HCl, R = —H) was carried out as in the case of the dimethox y-ana10g except that a large excess of thionyl chloride was used in order to obtain a homogenous reaction mixture. Attempts to prepare threo-Z, 3-di(p-hydroxyphenyl)pentanoic acid by the same procedure used for the preparation of the erythro-acid led to a mixture of the two isomeric acids. There was not enough of threo- Z, 3-di(p-methoxyphenyl)pentanoic acid available to prepare a sufficient amount of the threo-dihydroxy acid for further work in this series. 45 Stereochemistry of the Nitrogen Mustard Analogs In order to assess the effectiveness of the hexestrol carrier moiety it was necessary to determine which of the isomeric N, N-bis(Z-=chloro-= ethyl)-Z, 3-di(p—methoxyphenyl)pentylamines (I) corresponds to the estro—H genically active meso—hexestrol. This correlation has been accomplished by determining the configuration of the intermediate 2, 3—di(p-methoxyphenyl)— pentanoic acid (XIV). Since the preparation of the nitrogen mustard (1, R 2 -CH3) from this acid by scheme C does not affect either of the asymmetric centers, the configuration of the two asymmetric carbon atoms in the resulting nitrogen mustard should be the same as they were in the inter- mediate acid. At the beginning of this work the assumption was made, as has been made by the many other investigators who have worked in this area, that the higher melting acid was the more symmetrical erythro= isomer and would correspond to meso—hexestrol. This assumption was proven correct by the sequence of conversions represented schematically in Figure IX. The erythro—Z, 3—di(p—methoxyphenyl)pentanoic acid (XIVe) which was used was the best portion obtained through a systematic fractional recrystallization (119). The acid chloride (XVIIe), prepared from the acid and freshly distilled Mathe son, Coleman and Bell thionyl chloride, was treated with Norite and was then recrystallized from benzene before use in the preparation of the diazoketone (XXVIe). Alcohol-free diazo- methane, prepared as described by Vogel (120), was permitted to react with the purified acid chloride in approximately a three-molar excess. The diazoketone (XXVIe) was not isolated prior to its conversion by the Wolff rearrangement to methyl erythro-3, 4-di(p—methoxyphenyl)— hexanoate (XXVIIe). The Wolff rearrangement of the ab0ve diazoketone gave a good yield of the methyl ester (XXVIIe) with silver oxide as the catalyst (121). However, the modification of Klein and Bergmann (122), 46 .Homuo T300530 HoxamoonIOmoE on. Cflod ofiocdumomSxmoflmxxomuoaImvHQIm .NIo-Hmfnno mo comm-H0280 can no“ omamEomom .XH ohdmfim Hxxx mxxx mmo. I m .m N N I IIIIIIVI N N I 11W: m N I m0 mo mo _m 58H 8.5 mo mo _m amid-m mo mo ..m «mg-3 Hm>xx m>xx H>xx m m~oo~mo-_m E mmoNOommoé monmoou < ...zmooo-.m mox : \ mmoo N 80m Nszo Q Hxxx m>x Ex , m mooommonm mooo-pm 11.11130 0 m m~oo._m - m ' mmNo 60mm 3 szz .o .m A. I ..m . modz . xxx x xH «moxw :03. m N I N N I I 200 mo _m [IllY mz mo _m AlliImmE-H 20 m 47 using silver benzoate in triethylamine, gave poor yields of inferior product. Methyl erythro-3, 4-di(p-methoxyphenyl)hexanoate was saponified to erythro-3, 4-di(p-methoxyphenyl)hexanoic acid, m. p. 183- 184. 50. A sample of this acid was re—esterified with diazomethane back to the original erythro-ester. This proved that no isomerization had taken place during the saponification. 3, 4-Di(p-methoxyphenyl)hexanoic acid has been previously prepared by Hofstetter and Smith (123) by the oxidation of a mixture of 3, 4-di(p—methoxyphenyl)hexanols. The melting point of this acid reported by these workers is 181. 5-182. 50. Lithium aluminum hydride reduction of the methyl ester (XXVIIe) afforded erythro-3, 4-di(p-methoxyphenyl)-l-hexanol (XXXIe) in a near quantitative yield. This alcohol was zone refined before being converted to erythro-3, 4—di(p-methoxyphenyl)hexyl p—toluenesulfonate (XXXIIe). The tosylate, which darkened on standing, was immediately reduced with lithium aluminum hydride to meso-hexestrol (II, R 2 -CH3). The hexestrol, obtained by the reduction of the tosylate (XXXIIE), proved to be identical to an authentic sample in melting point, mixed melting point, and U. V. spectrum. The foregoing steps thus established that the high melting isomer of 2, 3-di(p-methoxyphenyl) pentanoic acid is indeed the erythro- modification corresponding to meso—hexestrol. It could also be reasonably assumed that the solid 2, 3-di(p-methoxy- phenyl)pentanenitrile (IX) is in the erythro—series. This has been definitely established by the conversion of erythro-3, 4-di(p-methoxyphenyl)- hexanoic acid (XXVIIIe) to erythro—Z, 3-di(p—methoxyphenyl)pentylamine (Xe). The desired amine (Xe) could not be isolated from the reaction mixture obtained by an attempted Schmidt reaction (124) with erythro— 3, 4—di(p-methoxyphenyl)hexanoic acid. This was probably due to sulfonation of the activated aromatic rings. The amine (Xe) was success- fully prepared by the Curtius reaction. Addition of an aqueous solution of sodium azide to an acetone solution of erythro-3, 4-di(p—methoxyphenyl)— hexanoyl chloride (XXIXe) gave the azide (XXXe), which was rearranged ..3 48 to the amine without prior purification. This amine proved to be identical to the erythro-Z, 3-di(p-methoxyphenyl)pentaylamine (Xe) obtained by the reduction of the solid nitrile (IXe) with lithium aluminum hydride. This substantiated the assumption of Burckhalter and Sam (105) that the solid isomeric 2., 3-di(p-methoxyphenyl)pentanenitrile corresponds to the higher melting meso-hexestrol. 49 EXPERIMENTAL" Z Anisyl Alcohol a) By the Hydrogenation of Anisaldehyde. A solution of 68. l g. (0. 5 mole) of freshly distilled anisaldehyde in 125 ml. of absolute ethanol was placed in a 500 ml. pressure bottle with approximately 5 g. of Raney nickel catalyst (Mozingo type) (125). The bottle was placed in a Parr low pressure hydrogenation apparatus, and after evacuation and flushing with hydrogen, it was shaken with hydrogen at an initial pressure of 52 psi. at a temperature of 500. After two hours the pressure had fallen to 25 psi. , whereupon the system was recharged to 50 psi. The pressure dropped to 41 psi. within an hour and remained at this pressure for an additional hour. The total hydrogen up—take was 35 pounds (0. 44 mole-~based on an uptake of eight pounds of hydrogen per 0. 1 mole of hydrogen acceptor). The catalyst was removed by filtration and the solvent distilled through a short Vigreux column, leaving a colorless residue which weighed 69. 8 g. and which colored fuchsin aldehyde re— agent, indicating the presence of some unreduced anisaldehyde. This crude product was fractionated through a Fenske column into the follow-— ing fractions. Fraction B. p. , OC. (4 mm.) Grams Hg 1 110-115 2.6 1.5190 2 115—117 2.4 1.5453 3 117—118 58.2 1.5436 4 117-120 3.1 1.5440 residue -— 2.4 -— 1Microanalyses by Micro—Tech Laboratories, Skokie, Illinois. 2Melting points were taken in open capillaries and are uncorrected, unless otherwise noted. Boiling points are all uncorrected. 50 The third fraction was redistilled through a Fenske column to give the following . Fraction B.p. , OC. (4 mm.) Grams n3 1 115-116 2.8 1.5451 2 116-117 5.2 1.5440 3 117-117 46.8 1.5435 4 117—118 3.8 1.5445 Fractions 2. and 4, of this second distillation, were taken as pure anisyl alcohol, giving a total yield of 55. 8 g. (81%). b) By the Crossed Cannizzaro Reaction with Anisaldehyde and Formalin. To a solution of 200 g. (1.46 moles) of anisaldehyde and 375 ml. (4. 8 moles) of 36% formalin in 450 ml. of 95% ethanol contained in a 2-1. three necked flask fitted with a reflux condenser, an addition funnel and a stirrer, was added with vigorous stirring 525 ml. of 55% potassium hydroxide solution at such a rate that the temperature remained below 650 without external cooling. This addition required approximately one hour. The flask was transferred to an oil bath which had been preheated to 600 and was heated at this temperature for one hour. The bath temperature was then raised to 900 and was maintained at this temperature for ten minutes. The reflux condenser was replaced with a Claisen head and the reaction mixture was concentrated to one— half of its volume by distillation at 15 mm. pressure. Stirring was con- tinued throughout the distillation to achieve even boiling. Five hundred milliliters of water was added to dissolve the solid which had separated. The dark brown organic layer was separated and the water layer was extracted twice with 200 ml. portions of ether. The combined organic material was extracted with 250 ml. of 10% sodium bisulfite in several portions, was washed twice with water and was dried over anhydrous sodium sulfate. The ether was removed at atmospheric pressure and 51 the very dark brown residue was distilled without fractionation to give 156 g. of light yellow product, b.p. 142-1440 (16 mm.). Redistillation through a Fenske column gave 153 g. (76%) of pure anisyl alcohol, b.p. 143—144° (16 mm.), 531.5422. Dankova e_t a_1. (90), using a similar procedure, reported a 75% yield of anisyl alcohol, b.p. 140—1410 (12 mm.). p—Methoxyphenylacetonitrile a) Potassium cyanide in aqueous dioxane. A solution of 835 g. (6 moles) of anisyl alcohol in 3—1. of benzene was placed in a 5—1. flask equipped with a 75 cm. Allihn condenser and a sintered glass gas addition tube. Anhydrous hydrogen chloride was passed directly from the cylinder into the solution for two hours without external cooling. After the water layer had been separated, the benzene was removed by flash distillation at 15 mm. without the application of external heat and the colorless residue was taken up in 2-1. of dry dioxane. This dioxane solution was added over a period of one and one-half hours to a well—stirred solution of 1171 g. (18 moles) of potassium cyanide in one liter of boiling water. After the resulting mixture had been stirred at the reflux temperature for three hours, the reflux condenser was replaced with a Claisen head and 2-1. of the solvent was removed at atmospheric pressure. The residue was diluted with 1500 m1. of water and 500 ml. of benzene. The water layer was separated and was extracted with 1200 ml. of benzene in four portions. The benzene solutions were combined, washed with two 200 m1. portions of hot water (750), and dried for ten minutes over anhydrous calcium chloride. After removal of the benzene by flash distillation (900 at 15 mm.) the light brown residue was distilled through a short Vigreux column giving the following. 52 Fraction B. p. , OC. (0. 5 mm.) Grams n3 1 114-116 272 1.5320 2 116-117 136 1.5322 3 117-121 380 1.5331 The second and third fractions were combined and were redistilled through a Fenske column. Fraction B. p. , OC. (0. 5 mm.) Grams nZD5 1 114-116 12 1.5320 2 116-117 354 1.5322 3 117-122 142 1.5338 The second fraction from this redistillation was taken as the pure p—methoxyphenylacetonitrile and was used for the condensation with anisaldehyde. The remaining fractions from both distillations were combined and were used for the preparation of p-methoxyphenylacetic acid. The second fraction of 354 g. represents a 39% yield of pure p-methoxyphenylacetonitrile. The three fractions obtained from the first distillation amounted to 788 g. and represented a 90% yield of crude nitrile. This procedure is essentially that of Livshits (92), as described by Rorig (92). The latter obtained a 72% yield of p-methoxyphenyl— acetonitrile, b.p. 105-1080 (0.5 mm.), r12]; 1.5324. b) Sodium cyanide in anhydrous acetone. A mixture of 138 g. (1 mole) of anisyl alcohol and 248 m1. of concentrated hydrochloric acid was placed in a 1-1. round bottomed flask and vigorously stirred for 15 minutes. This mixture was transferred to a separatory funnel and the layers were separated. It usually required from one to four hours for the layers to separate completely. The anisyl chloride was dried for 30 minutes over anhydrous calcium chloride. It was then filtered and the calcium chloride rinsed with 500 ml. of anhydrous acetone. The anisyl 53 chloride and the filtered acetone rinse were added to a suspension of 73.6 g. (l. 25 moles) of sodium cyanide and 10 g. of sodium iodide in 500 ml. of anhydrous acetone contained in a 2-1. flask, equipped with a stirrer and a reflux condenser topped with a drying tube. After the reaction mixture had refluxed on a steam bath for 20 hhrs. with constant sitrring, it was allowed to cool to room temperature. The solid was filtered out and was washed with two 100-ml. portions of dry acetone. Distillation of the acetone through a short Vigreux column at 30 mm. left 144 g. of light brown liquid. This crude product was taken up in 300 ml. of benzene and was washed with four portions of hot water (750). The benzene solution was then dried for 15 min. over anhydrous calcium chloride, during which time it acquired a dark brown coloration. Removal of the benzene irivacuo left 142 g. of brown oil, which was distilled without fractionation to give 138 g. of crude p—methoxyphenylacetonitrile, b.p. 91—101o(0.3 mm.), n3 1.5297. Fractionation of the crude nitrile through a Fenske column gave the following. Fraction B. p. , OC. (0. 3 mm.) Grams n3 1 90—93 2.2 1.5274 2 93-94 10.0 1.5289 3 94—94 120.0 1.5303 4 94—103 3.4 1.5333 Fractions 2 and 3, taken as pure p-methoxyphenylacetonitrile, amounted to 130 g. (88%). According to Organic Syntheses (94) this procedure should give a 74-81% yield of product, b.p. 94-970 (0.3 mm.), oz]; l.5285-1.5291. 2, 3~Di(p—methoxyphenyl)acrylonitrile. A solution of sodium ethoxide (from 3. 5 g. of sodium and 75 ml. of absolute ethanol) was slowly added to a mechanically stirred solution of 220 g. (l. 5 moles) of anisaldehyde and 204 g. (1. 5 moles) of p-methoxyphenylacetonitrile 54 in 2-1. of 95% ethanol contained in a 4—1. beaker. The reaction mix- ture became deep yellow during the addition, but there was no notice- able heat evolved. The stirring was continued for an additional hour, during which time a flocculent solid separated. The reaction mixture was cooled to 00 in an ice—salt bath and filtered. The light yellow solid was washed once with water. It was then slurried in the funnel three times with 200-m1. portions of ether. There was obtained, after air drying, 334 g. of light yellow product, m.p. 107. 5—1090. The filtrates from the reaction mixture and the ether washings were combined and concentrated to approximately 200 ml. The concentrate was cooled to 0° and filtered to give 28 g. of dark yellow solid, m.p. 101—1080. Recrystallization from ethanol gave 25. 5 g. of light yellow solid, m.p. 106. 5—1090. The total yield of 2, 3-di(p-methoxyphenyl)acrylonitrile was 359.5 g. (90.3%). Niederl and Ziering (95) used this procedure to prepare 2, 3-di- (p-methoxyphenyl)acrylonitrile (in an unspecified yield) with a melting point of 1080. 2, 3-Di(p-methoxyphenyl)pentanenitrile. A solution of 40 g. (0. 15 mole) of 2, 3—di(p-methoxyphenyl)acrylonitrile in 500 ml. of dry benzene was added over a period of 30 min. to a solution of ethylmagnesium bromide (prepared from 4. 9 g. (0. 2 mole) of magnesium and 21. 8 g. (0. 2 mole) of ethyl bromide) in 1—1. of ether. After being refluxed for 3 hrs. the reaction mixture was allowed to cool to room temperature and was then poured onto 1-1. of ice containing 100 ml. of concentrated hydrochloric acid. An additional 500 ml. of benzene was added to dis— solve the product. The aqueous layer was separated and was extracted with two 50 ml. portions of benzene. The combined organic solution was washed successively with two 100 ml. portions of water, 50 ml. of 10% sodium carbonate, and 50 m1. of water. After drying over anhydrous 55 potassium carbonate, the solvent was removed in a current of air and the semi-solid residue was slurried with 50 ml. of ether, cooled, and filtered. The solid was recrystallized from ethanol to give 19.2 g. of erythro-Z,3-di(p—methoxyphenyl)pentanenitrile, m.p. 130.1320 (43.2%). The filtrate was concentrated and distilled (bath temperature, 1200; pressure, 0.02 mm.) through an ether bridge to give 18.4 g. (41. 5%) of threo-2, 3-di(p—methoxyphenyl)pentanenitrile as a light yellow viscous sirup. Burckhalter and Sam (105), using a similar procedure, obtained a 42% yield of the erythro-isomer, m.p. 130-1310, and a 49% yield of the threo—isomer with a boiling point of 205C) (1 mm. ). erythro—Z, 3—Di(p-methoxyphenyl)penty1amine hydrochloride. Anhydrous aluminum chloride (31.1 g., 0. 25 mole) was added in small portions to 500 ml. of anhydrous ether contained in a 1-1. Erlenmeyer flask. The flask and contents were cooled in an ice-salt bath and the solution stirred with a magnetic stirrer. Throughout the addition of the aluminum chloride a slow stream of dry nitrogen was passed into the flask to exclude moisture. This solution was filtered into an addition funnel and was added over a period of 20 min. to a well-stirred slurry of 10.0 g. (0. 316 mole) of lithium aluminum hydride in 500 ml. of dry ether. The resulting suspension was stirred at room temperature for 30 min. To this suspension a solution of 36. 3 g. (0. 25 mole) erythro—Z, 3- di(p-methoxyphenyl)pentanenitrile in 750 ml. of dry benzene was added over a period of 1. 5 hrs. The reaction mixture was refluxed for 20 hrs. and was cooled to 00 in an ice-salt bath. It was then hydrolyzed by the dropwise addition of 20 m1. of water. This was followed by the addition of 350 m1. of 9N hydrochloric acid. The resulting gray slurry gave a suspension of colorless amine hydrochloride after being stirred for 20 hrs. at room temperature. The mixture was filtered and the solid, 56 after being washed twice with 100 m1. portions of cold water, was dis— solved in 900 m1. of boiling water. The resulting solution was filtered through a bed of Celite 0. 25 in. thick in a steam jacketed Buchner funnel. The filtrate was allowed to come to room temperature slowly and was then kept at 80 for 18 hrs. The solid was removed by filtration and was dried over phosphorus pentoxide for two days to give 80. 3 g. (93%) of pure erxthro-Z, 3-di(p-methoxyphenyl)pentylamine hydrochloride, m. p. 248-2490. I Analysis. Calculated for C19H26NOZC1: C, 67. 94; H, 7.80; Cl, 10. 56. Found: C, 68.05; H, 7.92; Cl, 10.74. The above procedure is based on the work of Nystrom (100), who reported yields in the range of 75-91% for the reduction of a variety of nitriles. erythro-Z, 3—Di(p-methoxyphenyl)pentylamine. The crude hydro— chloride obtained by the reduction of 0. 25 mole of erythro-Z, 3—di— (p—methoxyphenyl)pentanenitrile was stirred mechanically for several hours in a 1-1. beaker with a mixture of 300 m1. of benzene and 150 m1. of 15% potassium hydroxide. The mixture was filtered and the colorless solid was washed several times with water. The benzene solution was separated and the benzene was evaporated in a Rinco evaporator, leaving a colorless solid. The combined solid after recrystallization from methanol, weighed 70. 2 g. (90. 0% based on nitrile). The pure erythr - 2, 3—di(p-methoxyphenyl)pentylamine melted at 140. 50. Derivative 1- | erythro-Z, 3-Di(p-methoxyphenyl)pentyl l- 3- (1 -naphthyl)urea. This derivative was prepared by the procedure described for the threo- isomer below. Three grams of amine gave 4. 5 g. of derivative, m.p. 175-1770. 57 Analysis. Calculated for C30H3ZNZO3: C, 76.89; H, 6.89; N, 5.98. Found: C, 76.77; H, 6.43; N, 6.26. threo-2, 3-Di(p-methoxyphenyl)pentaylamine hydrochloride. A solu- tion of 14. 2 g. (0. 106 mole) of anhydrous aluminum chloride in 300 m1. of dry ether was slowly added to a vigorously stirred slurry of 4.42 g. (0.106 mole + 10% excess) of lithium aluminum hydride in 300 ml. of dry ether. After the aluminum chloride —lithium aluminum chloride mixture had been stirred for an additional hour, a solution of 31. 3 g. (0. 106 mole) of m2, 3-di(p-methoxyphenyl)pentanenitrile in 250 ml. of dry ether was added over a period of 2 hrs. After being refluxed for 24 hrs., the reaction mixture was cooled to 100 and was hydrolyzed by the dropwise addition of 15 m1. of water. This was followed by the addition of 100 ml. of 9N hydrochloric acid. This hydrolysis procedure resulted in the formation of an upper ether layer containing unreduced nitrile, a middle layer of amine hydrochloride and a lower aqueous layer. These layers were separated. The aqueous layer was extracted with two 50 ml. portions of chloroform, which were then used to dissolve the amine hydrochloride. This resulting chloroform solution was filtered to free it of suspended gray inorganic impurities and was then washed with two 50 ml. portions of saturated sodium chloride. The chloroform was removed with a Rinco evaporator and the residue after solution in 500 m1. of boiling water, was treated with Norite. After the solution was cooled, a light tan oil and a small amount of solid separated. The mix- ture was filtered and the solid recrystallized from water. There was obtained 3. 2 g. of erythro-Z, 3-di(p-methoxyphenyl)pentylamine hydro— chloride, m.p. 248-2490. The oil from the filtrate, after separation and drying in vacuo over phosphorus pentoxide, amounted to 20. 2 g. (56. 5%) in threo-2, 3-di(p-methoxyphenylpentylamine hydrochloride. This light tan oil failed to solidify. Evaporation of the ether solution from the hydrolysis mixture led to the recovery of 4.. 6 g. of unreduced 2, 3 —di(p-methoxyphenyl)pentanenitrile . _d '1 58 threo-2, 3-Di(p-methoxyphenyl)pentylamine. A 5.42 g. portion of W the above amine hydrochloride was added to a solution of aqueous base and the resulting free amine extracted with benzene. The benzene solu- tion was dried over barium oxide and the benzene removed with a Rinco evaporator. The threo-2, 3—di(p—methoxyphenyl)pentylamine, a light brown oil, weighed 4. 8 g. (99% based on amine hydrochloride). A small quantity of the amine was added to cold dilute sulfuric acid. The sulfate salt was soluble in the warm solution, but it separated as a dark gum from the cold solution and could not be purified to a well- defined solid. The perchlorate salt was soluble in water, but could not be made to solidify on concentration of the solution i_n vacuo. All attempts to isolate a solid picrate salt were unsuccessful. The addition of an etheral solution of anhydrous oxalic acid to a solution of the amine in dry ether produced a solid oxalate. This salt sintered at approximately 1400 and decomposed with the evolution of gas at approximately 1800. Derivative 1- threo-2, 3-Di(p-methox henyl) ent 1 -3-(l-na hthyl)urea. A solution of 3 g. (0. 01 mole) of threo-2, 3—di(p-methoxyphenyl)penty1- amine in 25 m1. of dry benzene was dried for 24 hrs. over 4A molecular sieveszl: and was pipetted into a dried flask containing 1. 7 g. (0. 01 mole) of CI-naphthyl isocyanate. The flask was immediately fitted with a short reflux condenser topped with a drying tube. The solution was warmed in a hot water bath at 500 for 3 hrs. The colorless solid which separated on cooling the solution was removed by filtration and was washed with cold benzene. It was recrystallized several times from benzene to give :fiProduct of Linde Air Products Company. 59 4. 2 g. of 1—[_thr_eo_—2, 3—di(p—methoxypheny1)pentyl]-3-(1-naphthy1)urea, m.p. 168-1700. ,Analysis. Calculated for C30H32NZO3: C, 76.89; H, 6.89; N, 5. 98. Found: C, 76.69; H, 7.07; N, 6.04. The above procedure, for the use of a-naphthyl isocyanate for the preparation of derivatives of amines, is that of French and Wirtel (126). N, N— Bis (2 - hydroxyethyl) [erythro— 2 , 3 -di(p—methoxyphenyl)pentyl] amine. The apparatus used, see Figure X, consisted of a 500 ml. three necked flask (F), a dry-ice condenser (C) with a U-tube containing a 0. 5 cm. column of mercury and a sintered glass gas addition tube (T) connected to a cold trap (CT). The reaction flask was kept in a 500 thermostated water bath during the entire reaction. Eleven grams (9.8 ml., 0. 25 mole) of ethylene oxide was distilled through a cooled (100) spiral condenser and collected in the cold trap, which was cooled in dry ice. The cold trap was connected to the gas inlet tube and the ethylene oxide was distilled during a period of Figure X. Apparatus used for 3 hrs. into a solution of 30. 94 g. hydroxyethylation (0.10 mole) of erythro-Z, 3—di(p- methoxyphenyl)pentylamine in 250 ml. of methanol. Once the air was expelled from the system no gas escaped through the "U"-tube during the addition of the ethylene oxide. When the addition had been completed the reaction mixture was allowed to remain at 500 for an additional 8 hrs. and was then gently refluxed on a steam bath for 1 hr. The solvent was evaporated in a Rinco evaporator, leaving 38. 2 g. (98. 5% —crude yield) of light brown sirup which could not be made to solidify. 60 The product did not give a precipitate with 5—nitrosalicyladehydeg-nickel chloride reagent, (127) indicating the absence of a primary amine. All attempts to prepare a solid hydrochloride from this crude product were unsuccessful. Chromatography of N, N-bis(2—hydroxyethyl)[erythro-2, 3-di(p- methoxyphenyl)pentylIamine. A 150 x 4 cm. column was packed with Fischer chromatographic grade alumina (600 gm.). A solution of 29 g. of crude N, N-bis(2-hydroxyethyl)|erythro-Z, 3—di(p~methoxyphenyl)= pentyl]amine in 30 ml. of hexane was placed on the column and was developed with hexane. Fractions of 500 ml. were collected as the eluent was changed from hexane, hexaneebenzene, benzene, benzene- ethanol, to ethanol. The solvent was removed from each fraction through a short Vigreux column at atmospheric pressure and the residue was transferred to a small beaker with ether. The ether was evaporated and the residue was dried in a vacuum oven at 500 overnight before being weighed. Figure XII shows both the weight of the residue of the individual fractions and the eluent used to remove each from the column. Fractions 1 to 28 consisted of 1.6 g. of light tan sirup which could not be crystallized. Fractions 29—40 amounted to 17. 51 g. of an off—white solid which, after repeated recrystallization from methanol, gave 14. 2 g. of pure N,N—bis(2-hydroxyethyl)|erythro-Z, 3-di(p-methoxy— phenyl)pentyl]amine, m.p. 85-860. Analysis. Calculated for C23H33NO4: C, 71. 28; H, 8. 58; N, 3.62. Found: C, 71.21; H, 8.37; N, 3.66. Purification of Thionyl Chloride. Eastman "white label" thionyl chloride was distilled through a Fenske column and the fraction distilling between 74 and 760 was kept for further purification. Except for the removal from a non—volatile residue, this first distillation effected no Weight in grams 13.2 1.8 0.55 .50 .45 .40 .35 .30 O .25 .20 .15 .10 000000000 .05 20 40 60 80 100 61 W I -5 10 15 20 25 3O 35 40 45 Fraction % Benzene in Hexane % Ethanol in Benzene Figure XI. Weight of residue vs. eluent in the chromatography of N, N-Bis(2-hydroxyethyl)| erythro-Z, 3-di(p- methoxyphenyl)pentyl]amine. r—f - 62 separation. The sulfur chlorides were next removed by refluxing the distillate with 4% of its weight of flowers of sulfur and an anhydrous ferric chloride catalyst for 6 hrs. (112). The sulfuryl chloride was removed by refluxing the thionyl chloride, from which the sulfur chlorides had been removed, for 4 hrs. with 2% of its weight of naphthalene and an aluminum chloride catalyst (113). A typical run is as follows: A newly opened 500 g. vial of Eastman ”white label" thionyl chloride was distilled through a Fenske column to give 480 g. of material, b.p. 74-760 (750 mm.). Less than one gram of black residue remained. This crude thionyl chloride was refluxed for 6 hrs. with 20 g. of flowers of sulfur and 0.1 g. of anhydrous ferric chloride. The deep yellow thionyl chloride was distilled from the sulfur without fractionation. It was redistilled through a Fenske column to give a deep yellow fore—run and a nearly colorless main fraction of 440 g. The latter was refluxed for 4 hrs. with 9 g. of naphthalene and 0. 2 g. of anhydrous aluminum chloride. The thionyl Chloride was distilled from this mixture without fractionation and was redistilled through a Fenske column to give 380 g. of thionyl chloride, b.p. 75—75.5O (750 mm.).. N, N-Bis(2-chloroethyl) erythro-2', 3—di(p-=methoxyphenyl)penty1]= amine hydrochloride. A solution of 6. 3 g. (0. 053 mole) of freshly distilled Matheson, Coleman and Bell thionyl chloride in 150 m1. of alcohol-free chloroform was placed in a 500 m1. flask equipped with a thermometer, a reflux condenser topped with a drying tube, an addition funnel with a pressure equalizing tube, and a nitrogen inlet tube. The flask was placed in an ice bath and a stream of dry nitrogen was passed through the system. When the contents of the flask had been cooled to o 1 , a solution of 8. 7 g. (0.025 mole) of N, N-bis(2—hydrosyethyl)[erythro- 2, 3—di(p-methoxyphenyl)pentyl]amine in 100 m1. of chloroform was added 63 at such a rate that the temperature remained below 40. The reaction mixture was vigorously stirred with a magnetic stirrer during this addition, which required approximately 2 hrs. The homogeneous reaction mixture was allowed to stand at room temperature for 18 hrs. before the excess thionyl chloride and the benzene were removed with a Rinco evaporator. A brown semi—solid residue remained. This residue was slurried with 50 ml. of dry benzene and the benzene was evaporated in order to remove the last traces of thionyl chloride. All attempts to recrystallize this product did not lead to a well-defined solid. The residue was dissolved in 100 ml. of chloroform and was allowed to stand over 50 g. of anhydrous calcium chloride, with occasional shaking, for 24 hrs. This chloroform solution was filtered from the calcium chloride and was then passed through a 2 x 15 cm. column of alumina. The column was washed with an additional 200 ml. of chloroform, which had previously been used to rinse the calcium chloride. The resulting light yellow eluate was concentrated with a Rinco evaporator and gave 6. 2 g. of tan sirup. This was dissolved in 200 ml. of dry ether, cooled to 00, and treated with anhydrous hydrogen chloride to precipitate a light tan semi-=crystalline mass. This residue was repeatedly recrystal= lized from dry acetone-chloroform to give 4.6 g. of pure N, N=bis- (2-chloroethyl)[erythro-Z, 3—di(p-methoxyphenyl)pentyl]amine hydrochloride, m.p. 133. 5-1340. Analysis: Calculated for C23H33NOZC13: C, 59.94; H, 6.99; N, 3.04; C1, 23.08. Found: C, 59.91; H, 7.05; N, 2.93; Cl, 23.20. 2, 3-Di(p-methoxyphenyl)propylamine. A solution of 40 g. (0. 15 mole) of 2, 3-di(p-methoxyphenyl)acrylonitrile in 200 ml. of dry tetra- hydrofuran was added over a period of 3 hrs. to a well—stirred slurry of 11.4 g. (0. 3 mole) of lithium aluminum hydride in 500 ml. of tetra- hydrofuran at the reflux temperature. After being refluxed for 24 hrs. 64 under a nitrogen atmosphere, the reaction mixture was cooled and was hydrolyzed by the dropwise addition of 50 ml. of water. This was followed by the addition of 600 ml. of saturated ammonium tartrate- ammonium sulfate solution. Two hundred milliliters of methylene chloride was added and the water layer was separated. The latter was extracted with two 100 m1. portions of methylene chloride. The come bined organic solution was washed successively with 100 ml. of saturated ammonium tartrate=ammonium sulfate, with four 100 ml. portions of water and with 100 ml. of saturated sodium chloride solution. After the solution was dried over barium oxide, the solvent was removed with a Rinco evaporator leaving 39. 6 g. of nearly colorless sirup. An iso» propyl ether solution of this residue deposited 38. 2 g. (94%) of a color-a less solid on cooling. A second recrystallization from isopropyl ether did not raise the melting point of 74—750. McKay and Brownell (101) prepared this amine in a 34% yield from 2, 3—di(p—methoxyphenyl)acrylo- nitrile by hydrogenation over Adam's catalyst. They reported a melting point of 73—740. N, N-Bis(2-hydroxyethyl)=2, 3-di(p—methoxyphenyl)propylamine hydrochloride. Using exactly the same procedure and apparatus as described for the preparation of N, N-bis(2-hydroxyethyl)[erythro-Z, 3-di- (p-methoxyphenyl)pentyl]amine, 12. 7 g. (0. 3 mole) of ethylene oxide was distilled into a solution of 35 g. (0.13 mole) of 2, 3—di(p—methoxy- phenyl)propy1amine in 200 ml. of methanol. After aging for 8 hrs. and after refluxing for 1 hr., the solution was evaporated with a Rinco evaporator to give 46. 2 g. of a nearly colorless sirup. Anhydrous hydrogen chloride was passed into a cold ethereal solution (100) of this sirup and precipitated a colorless non-crystalline hydrochloride. After decanting the ether and removing the last traces of solvent i_I_l vacuo there remained 40. 2 g. (81%) of N, N—bis(2—hydroxyethyl)—2, 3-di— (p—methoxyphenyl)propylamine hydrochloride as a very viscous sirup. 65 McKay and Brownell (101) obtained this hydrochloride in an un-= specified yield by the treatment of the free amine with concentrated hydrochloric acid. N, N- Bis(2-chloroethyl)-2, 3-di(p=—methothenylmropylamine hydrochloride. A solution of 8. 8 g. (0. 074 mole) of freshly distilled Matheson, Coleman and Bell thionyl chloride in 20 m1. of alcohol free chloroform was added with stirring over a period of 1 hr. to a solution of 5.4 g. (0. 014 mole) of N, N-bis(2-hydroxyethyl)-2, 3==di(p=-methoxy— phenyl)propylamine hydrochloride in 50 m1. of chloroform. The temperature was not allowed to rise above 500 during the course of the addition. After the addition the reaction mixture was refluxed for 1 hr. and was then allowed to stand overnight at room temperature. The ex- cess thionyl chloride and chloroform were removed with a Rinco evapor- ator and the semi-solid residue slurried with 50 ml. of dry benzene. The latter was evaporated to remove the last traces of thionyl chloride. After recrystallization from anisole, the product amounted to 4. 62 g. (78. 5%) of colorless N, N—bis(2—chloroethyl)-2, 3-di(p-methoxyphenyl)— propylamine hydrochloride, m.p. 166—1680. McKay and Brownell (101) obtained this nitrogen mustard, using this procedure, in an 80% yield. They reported a melting point of 165-1670. erythro-Z, 3-Di(p-methoxyphenyl)-l-pentanol. A suspension of 12. 5 g. (0.04 mole) of erythro—Z, 3-di(p-methoxyphenyl)pentanoic acid in 100 ml. of dry tetrahydrofuran was added over a period of 20 min. to a well stirred slurry of 1. 9 g. (0. 05 mole) of lithium aluminum hydride in 150 ml. of tetrahydrofuran. There was only a slight rise in the temperature of the reaction mixture during this addition. After the resulting slurry had been stirred at room temperature for 10 hrs. an 66 additional 0. 2 g. of lithium aluminum hydride was added and the mix- ture was refluxed for 14 hrs. The reaction mixture was then cooled to room temperature and the excess reagent destroyed by the dropwise addition of 75 ml. of a 10% solution of ethyl acetate in ether. Three hundred milliliters of methylene chloride and 200 ml. of saturated sodium chloride solution were added, followed by sufficient 10% sulfuric acid to dissolve the basic salts. The clear water layer was separated and extracted with three 50 ml. portions of methylene chloride. The combined organic solution was washed with two 100 ml. portions of saturated sodium chloride solution, was extracted with 100 ml. of 10% sodium carbonate solution and was again washed with 100 ml. of saturated sodium chloride solution. Acidification of the combined basic extract and the final wash solution did not afford the recovery of any unreacted acid. The organic solution was dried with anhydrous sodium sulfate and was concentrated to approximately 30 ml. by distilling off the solvent through a short Vigreux column at atmospheric pressure. The remainder of the solvent was removed i_n vacuo leaving 11. 7 g. of colorless crude product. One recrystallization from methanol afforded 11.4 g. (95%) of erythro-Z, 3-di(p—methoxyphenyl)-1-pentanol, m. p. 122=1250. Repeated recrystallization from methanol did not raise the melting point of this alcohol. One gram of this product was zone refined in a 2 mm. tube, giving an analytical sample with a melting point of 124. 5—1260. Analysis. Calculated for C19H3403: C, 75.97; H, 8. 15. Found: C, 75.76; H, 8.33. threo-2, 3~Di(p-methoxyphenyl)-1—pentanol. A solution of 7. 3 g. (0. 022 mole) of methyl threo-2, 3-di(p—methoxyphenyl)pentanoate in 150 m1. of ether was added over a period of 2 hrs. to a well-stirred slurry of 0. 64 g. (0. 015 mole) of lithium aluminum hydride in 100 ml. of ether. After the reaction mixture had been stirred for 6 hrs. at room temperature, 67 the excess reagent was destroyed by the dropwise addition of 30 ml. of 10% ethyl acetate in ether. Eighty milliliters of eight percent sulfuric acid was added in order to dissolve the basic salts. The water layer was separated and extracted several times with ether. The combined ether solution was washed with saturated sodium chloride solution and was dried over anhydrous sodium sulfate. The solution was concen= trated to approximately 30 ml. by the distillation of the ether through a short Vigreux column at atmospheric pressure. Removal of the remain- ing solvent with a Rinco evaporator afforded 5. 7 g. of a colorless oil which did not solidify on standing. This oil was transferred to a Hickman still and was distilled at 0° 01 mm. at a bath temperature of 1200. The distillate amounted to 5. 6 g. (85%) of threom2y 3—di(p-methoxyphenyl)= l —pentanol . Analysis. Calculated for C19HZ4O3: C, 75.97; H, 8.15. Found: C, 75.71; H, 8.02. erythro- l-Chloro-2, 3-=di(p—methoxyphenyl)pentane. A solution of 3. 7 g. (0. 012 mole) of erythro—Z, 3-di(p—methoxyphenyl)—l-pentanol in 50 ml. of freshly distilled chloroform was placed in a 100 ml. flask equipped with a reflux condenser topped with a drying tube, a thermometer and a 15 ml. dry addition pistol. Finely ground dried calcium carbonate (4. 7 g. , 0. 047 mole) was added to the solution and the resulting slurry was cooled to 50 in an ice bath. The addition pistol was charged with 4. 9 g. (0. 024 mole) of phosphorous pentachloride, which was added to the well—stirred slurry of the alcohol Over a period of 45 min. The reaction mixture was then poured directly into a separatory funnel con— taining 75 ml. of saturated sodium bicarbonate solution. The chloroform solution was separated and the water solution was extracted with three 50 ml. portions of chloroform. The insoluble calcium carbonate re— mained in the water layer making filtration unnecessary. The chloro— form solutions were combined, washed with water and dried over 68 anhydrous sodium sulfate. Removal of the solvent with a Rinco evapor— ator left 3. 73 g. (99% of crude erythro-l«chloro=2, 3=di(p~methoxyphenyl)— pentane as a light yellow sirup. Alkylation of diethanolamine with erythro-l-—chloro=2, 3‘=-=di(p=-= methoxyphenyl)pentane with ethylene glycol as solvent. A solution of 3. 73 g. (0. 012 mole) of erythro-l—chloro—Z, 3—di(p-methoxyphenyl)= pentane and 2. 52 g. (0. 024 mole) of diethanolamine in 75 ml. of ethylene glycol was placed in a 100 ml. flask equipped with a thermometer and a short reflux condenser. This reaction mixture was heated in a thermo— stated oil bath at 1501:_2O for a period of 31 hrs. At various times during the course of the reaction l-ml. samples were withdrawn and analyzed for free amine. The analytical procedure consisted of withdrawing a sample with a l—ml. pipette, transferring the sample to a 125 ml. Erlenmeyer flask, diluting it with 25 ml. of 50% aqueous methanol and titrating to the bromcresol green end point with standard hydrochloric acid. Table V. Kinetic data for the alkylation of diethanolamine with erythro—l- chloro-Z, 3-di(p-methoxyphenyl)pentane. Time ml. HCl [x] [a] [b] log b a - x] a[b - x] (hours) (0. 0626N) (product) (chloride) (amine) 0 4.81 0 .151 .301 0 3 4.76 .003 .149 .298 .005 7 4.69 .007 .147 .294 .010 9 4.51 .019 .141 .282 .012 21 4.16 .041 .130 .260 .093 27 4.02 .049 .126 .252 .120 31 3 .91 .056 .122 .245 .154 69 .16 .14 .12 .10 .08 .06 .04 00000000 .02 5 10 15 20 25 30 Time (hrs. ) Figure XII. Second order plot for the alkylation of diethanolamine. In the calculations shown, the free amine content of the reaction mixture was determined when the temperature reached 1500. This concentration was taken as the concentration of diethanolamine, [b], at t = 0. Since the chloride concentration was exactly one-half the concen= tration of the amine, the concentration of the chloride, [a], at t = 0 must be exactly one-half that of the amine. The second order rate constant calculated from this data is approximately 2. 5 x 10'5 1. -mole'1 sec. which is too low to be of any preparative value. Since no product could be isolated from this reaction mixture there is no assurance that the alkylation of diethanolamine is responsible for the decrease in the free amine content of the reaction mixture. Dehydrohalogenation or ring closure could also account for the results observed. In any case this does not alter the fact that the alkylation of diethanolamine with erythro—l-chloro-Z, 3-di(p-methoxyphenyl)pentane is too slow to be of any use as a. synthetic method. 70 Alkylation of diethanolamine with ergthro—l—chloro—Z, 3-di-= (p—methoxyphenyl)pentane without a solvent. A solution of 2. 0 g. (0.0063 mole) of erythro—l=chloro—2, 3-di(p-methoxyphenyl)pentane and 15 g. (0. 14 mole) of diethanolamine was heated in a thermostated oil bath at 2001150 for 72 hrs. The resulting dark brown reaction mixture was diluted with 100 ml. of water and adjusted to a pH of approximately 12 by the addition of a few drops of sodium hydroxide solution. This basic solution was extracted four times with 50 ml. portions of ether. The combined ether extract was washed twice with saturated sodium hydroxide and dried over anhydrous sodium sulfate. Removal of the solvent with a Rinco evaporator left 1. 9 g. of dark brown sirup. This residue gave a positive halogen test, but a negative nitrogen test (sodium fusion). 2, 3-Di(p-methoxyphenyl)pentanoic acid. A mixture of 48.4 g. (0. 16 mole) of the isomeric 2, 3=di(p—methoxyphenyl)pentanenitriles, 16 g. (0.4 mole) of sodium hydroxide, 32 ml. of water and 300 ml. of distilled ethylene glycol was refluxed with constant stirring for 38 hrs. in a 1—1. flask equipped with a 75 cm. Allihn condenser and a copper Hershberg stirrer. The insoluble amide separated from the solution during the course of the reaction and caused excessive bumping, which could be controlled only by stirring the reaction mixture. The reaction mixture was cooled to 900 and was diluted with an equal volume of water. The solid was filtered off and after recrystallization from ethyl acetate weighed 2.1 g. This material was shown to be erythro—Z, 3—di- (p-methoxyphenyl)pentanamide, m. p. 223—2250, by mixed melting point, 223-2250, with an authentic sample prepared from the acid chloride and ammonia. The cooled filtrate was made acid to Congo red with 6N hydro- chloric acid and the solution was decanted from the semi-solid product which clung to the sides of the beaker. This residue was dissolved in 71 400 ml. of boiling methanol and the resulting solution was treated with Norite. The solution, after cooling to 00 with constant stirring and filtering, yielded 14. 5 g. of erythro—Z, 3-di(p-methoxyphenyl)pentanoic acid, m.p. 177—1790. After removal of the solvent from the filtrate with a Rinco evaporator, the liquid residue was dissolved in benzene. After several extractions with water to remove the ethylene glycol, the benzene solution was dried over anhydrous sodium sulfate. Upon removal of the benzene with a Rinco evaporator, 31. 7 g. of light amber sirup remained. This residue was dissolved in 300 ml. of methanol-petroleum ether (1:5) and the solution was cooled to -50 for 18 hrs. Upon filtra» tion an additional 3.1 g. of erythro-acid, m.p. 168—1740, was obtained. The total yield of erythro-Z, 3-di(p-methoxyphenyl)pentanoic acid was 20.6 g. (41%). The crude threo—acid obtained by concentrating the filtrate failed to crystallize. This procedure is essentially that of Snyder and McIntosh (107) as modified by Hunter and Korman (99). The latter workers obtained a 36% yield of erythro—Z, 3-di(p-methoxyphenyl)pentanoic acid, m. p. 117. 5-1790, which they described only as the ”high melting" isomeric acid. Methyl threo-2, 3-di(p—methoxyphenyl)pentanoate using diazomethane. The apparatus used for the preparation of diazomethane is shown in FigureXIIl Flask "A” is a 250 ml. Wurtz flask with a small water con- denser attached to an extended side arm. The receiving flasks, B and C, are 500 m1. and 125 ml. Erlenmeyer flasks respectively. The side arm of flask "A" extended to within 1 cm. of the bottom of flask "B" and the glass bridge connecting flasks "B" and "C" extended wo within 1 cm. of the bottom of flask "C". All glass ends were fire polished and all the flasks were free of abberations. A solution of 14 g. of potassium hydroxide in 21 ml. of water and 70 ml. of methanol was placed in flask "A", 125 ml. of ether was placed in flask "B" and 50 ml. of ether was placed in flask "C”. 72 Figure XIII. Diazomethane Apparatus Flask ”B" was cooled with an ice bath and flask ”C" with a dry icemethyl cellosolve bath. The reaction flask, A, was placed in a water bath maintained at 60—650 during the entire course of the reaction. The re- action mixture in flask "A" was agitated by a slow stream of nitrogen passed through tube "T". The lower tip of this tube was approximately 0. 5 cm. above the bottom of flask "A". Passing nitrogen through the reaction mixture not only kept it homogeneous, but also maintained a positive pressure in flask ”A" to prevent the distillate in the receiving flasks from backing up through the side arm back to the generator. A solution of 60 g. (0. 28 mole) of N-nitrosomethyl p-toluenesulfonamide in 350 ml. of ether was added from the addition funnel "S" to the potassium hydroxide solution at such a rate that the volume in the generating flask remained constant. This process required approximately 45 min. When the nitrosoamide solution had been added, 40 ml. of ether was added in one portion and the mixture distilled until the distillate became colorless. 73 The diazomethane solution was transferred to a 1—1. Erlenmeyer flask which was cooled in an ice bath. A solution of 35. 7 g. (0.11 mole) of crude threo-2, 3—di(p—methoxyphenyl)pentanoic acid in 200 ml. of methanol was dropped from a separato/ry funnel into the diazomethane solution which was constantly stirred with a Teflon covered magnetic stirring bar. When the addition of the acid was complete, the flask was covered with a watch glass and the flask was allowed to stand in the ice bath as the ice melted. After standing at room temperature for 18 hrs. , the solution was concentrated to 100 ml. i_n vacuo. The resulting mixture was warmed to dissolve the solid which had separated, was treated with Norite, and was cooled to 00 overnight and filtered. The solid was fractionally recrystallized (119) from methanol to give the following fractions. Fraction m. p. , OC. Grams Color A 95-96 14.4 colorless B 88. 5—96 13.0 colorless C 83. 5-90 3. 7 faint yellow D 79. 5-112 2.6 yellow Fractions A and B were combined and recrystallized from methanol to give 22. 8 g. (61%) of pure methyl threo -2, 3—di(p—methoxyphenyl)pentanoate, m.p. 95-96. 5°. This procedure for the esterification of the crude acid is that of Hunter and Korman (99). These workers obtained a 78% yield of ester, m.p. 93—94. 50, which they described only as the methyl ester of the "low melting" 2, 3-di(p-methoxyphenyl)pentanoic acid. Methyl threo-2, 3-di(p—methoxyphenyl)pentanoate from the crude acid chloride and methanol. A solution of 29. 2 g. (0. 094 mole) of crude threo—2, 3—di(p-methoxyphenyl)pentanoic acid and 0. 5 ml. of pyridine in 75 ml. of dry benzene was cooled to 100. Thionyl chloride (24 g., 0. 2 mole) was distilled directly into this cooled solution as it was vigorously 74 stirred with a magnetic stirrer. After the resulting solution had been allowed to stand at room temperature for 2 hrs. it was heated to reflux on the steam bath for 30 min. Approximately 50 ml. of the excess thionyl chloride and benzene was removed by distillation at atmospheric pressure. The remaining solvent was then removed from the cooled solution with a Rinco evaporator. One hundred milliliters of anhydrous methanol was added to the light brown residue and the resulting solution refluxed for 2 hrs. . Removal of the excess methanol from the cooled solution with a Rinco evaporator left 26. 3 g. of tan solid. Two recrystal- lizations of this crude product from methanol afforded 22. 8 g. (74%) of impure methyl threo-2, 3—di(p—methoxyphenyl)pentanoate, m.p. 89=94O. Further recrystallization of this material did not seem to improve its purity. Approximately 20 g. of this impure product was slurried in 200 ml. of cold methanol. An ethereal solution of diazomethane was added to this slurry until the color of the diazomethane persisted. After standing at room temperature for 12 hrs. the solution was warmed on a steam bath to remove most of the ether and the remaining solvent was removed with a Rinco evaporator. After one recrystallization from methanol there was obtained pure methyl threo-2, 3—di(p-methoxyphenyl)pentanoate, m.p. 94.0-95.00. Saponification of methyl threo—2, 3-di(p-methoxyphenyl)pentanoate with methanolic potassium hydroxide. A slurry of 7. 0 g. (0. 021 mole) of finely ground methyl threo-2, 3—di(p—methoxyphenyl)pentanoate and 150 ml. of 15% methanolic potassium hydroxide was stirred at room temperature for 4 hrs. The suspension was then heated on a steam bath to effect solution and the resulting solution was refluxed for 30 min. The hot solution was transferred to a beaker, diluted with an equal volume of water and the methanol was removed in a current of air. 75 Sufficient water was added to redissolve the solid which had separated and the resulting water solution was treated with Norite. Acidification of the cooled solution with 6N hydrochloric acid resulted in the separa- tion of a light yellow semi—solid, which was easily removed from the solution on the end of a stirring rod. The acidic solution was extracted several times with methylene chloride. The solid obtained by the evapor- ation of the methylene chloride and the original semi—solid were combined and recrystallized from methanol. The first crop amounted to 3. 1 g. of colorless solid, m.p. 163—168. 50. After concentrating the filtrate there was obtained a second crop of 3.6 g. of light yellow solid, m.p. 125. 5- 1380. No additional solid was obtained from this second filtrate. The second crop of solid was recrystallized two more times from methanol to give 3.4 g. of colorless solid, m.p. 128-1380. The melting point of neither crop of impure threo-2, 3wdi(p—methoxyphenyl)pentanoic acid was improved by further recrystallization from methanol, aqueous methanol, benzene or ethyl acetate. The neutralization equivalent of the product from the first crop was determined to be 314. 9 and that from the second crop was 315. 2. The calculated neutralization equivalent for 2, 3—di(p- methoxyphenyl)pentanoic acid is 314.4. This is the procedure of Hunter and Korman (99) who obtained a 60% yield of the "low melting" acid, m.p. 163-164. 50. Attempted saponification of methyl threo-2, 3-di(p-methoxypheny1)— pentanoate with one equivalent of sodium hydroxide in aqueous methanol. A solution of 0. 68 g. (0. 002 mole) of methyl threo-2, 3-di(p—methoxyphenyl)- pentanoate and 2 ml. of 1N sodium hydroxide in 50 ml. of 80% aqueous methanol was allowed to stand at room temperature for 48 hrs. in a stoppered flask. The stopper was then replaced with a reflux condenser and the solution was heated to reflux for 2 hrs. The addition of 50 ml. of water caused a solid to separate from the solution. This solid was re— moved by vacuum filtration and was washed several times with warm water, 76 which was added to the filtrate. The methanol was removed from the filtrate in a current of air. The solid which separated as the methanol was removed was filtered off and was washed with warm water, which was added to the filtrate. Acidification of the cooled filtrate resulted in a slightly opalescent solution from which no free acid could be iso» lated. Recrystallization of the two solid fractions recovered during the work-up gave 0. 65 g. (96%) of recovered methyl threo-2, 3~di(p—methoxy«= phenyl)pentanoate, m.p. 92—940. This procedure was modified and repeated several times, employing reflux periods up to 72 hrs. No threo-2, 3-di(p-methoxyphenyl)pentanoic acid was isolated from any of these attempted saponifications. Attempted acid hydrolysis of methyl threo—2, 3-di(p—methoxyphenyl)~ pentanoate. A solution of 3. 3 g. (0.01 mole) of methyl threo-2, 3-di(p— methoxyphenyl)pentanoate in 50 ml. of concentrated hydrochloric acid and 20 m1. of acetic acid was refluxed for 72 hrs. The cooled reaction mixture was diluted with 50 ml. of water and the resulting slurry was extracted with three 50 ml. portions of ether. The combined ether extract was washed once with water and was extracted with 10% sodium carbonate solution. Acidification of the basic extract gave no insoluble product. The ether was removed with a Rinco evaporator to give 3. 3 g. (100%) of recovered methyl threo-2, 3-di(p-methoxyphenyl)pentanoate, m.p. 92-940. Attempted saponification of methyl film—2, 3-di(p—methoxyphenyl)- pentanoate with one equivalent of potassium hydroxide-ethylene glycol in benzene. One milliliter of a 1N solution of potassium hydroxide in ethylene glycol was added to a solution of 0. 16 g. (0.0005 mole) of methyl threo-2, 3—di(p—methoxyphenyl)pentanoate in 50 m1. of benzene. This solu- tion was refluxed under an atmosphere of nitrogen for 144 hrs. The cooled solution was extracted with 25 m1. of water and this water solution was 77 acidified to Congo red with 6N hydrochloric acid. No free acid could be isolated from this acidified solution. The benzene solution was dried over anhydrous sodium sulfate and the solvent was removed with a Rinco evaporator. After one recrystallization from methanol the residue amounted to 0.12 g. of impure methyl threo-2, 3—di(p=— methoxyphenyl)pentanoate, m.p. 86-980. This procedure resulted in the partial isomerization of the threo-ester to the more stable erythro- ester, but no appreciable saponification took place. Differential pH separation of the isomeric 2, 3-di(p-methoxyphenyl)— pentanoic acids. A mixture of 12.1 g. (0.041 mole) of erythro-Z, 3—di(p- methoxyphenyl)pentanenitrile and a solution of 4 g. of sodium hydroxide in 8 ml. of water and 75 m1. of ethylene glycol was refluxed for 24 hrs. The reaction mixture, diluted with water to 200 ml. , was treated with Norite. Five milliliters of this solution was neutralized with 0. lO6N hydrochloric acid. The solution was kept at room temperature (320) during neutralization. The pH of the solution was measured with a Beckman pH meter, Model H-Z, after each addition of acid. The titration data given in Table VI is represented graphically in Figure XIV. Table VI. Titration data for the neutralization of mixture of isomeric 2, 3-di(p-methoxyphenyl)pentanoic acids. nu pH nfl. pH nu. pH 0 13.0 16.0 10.1 23.0 5.8 2 12.6 16.5 9.4 23.5 5.7 3 12.6 17.0 8.2 24.0 5.6 5 12.4 17.2 7.2 24.5 5.5 7 12.3 17.4 7.1 25.0 5.3 8 12.1 18.0 7.3 25.5 5.0 9 12.0 18.2 7.4 26.0 4.5 10 12.0 20.0 7.4 26.5 3.3 12 11.9 20.5 7.3 27.0 2.8 13 11.6 21.0 7.1 27.5 2.5 14 11.3 21.5 6.7 28 2.1 15 10.9 22.5 5.9 78 2 4 6 pH 8 10 12/ 14 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 ml. Figure XIV. Volume of acid vs. pH for the neutralization of the isomeric 2, 3-di(p-methoxyphenyl)pentanoic acids. As shown in Figure IXIV one isomeric acid separated at approximately pH. 7, whereas the second isomer separated at pH 5. 3. In this trial run the isomeric acids were not separated. Since the solution containing the sodium salt contained approximately 20% ethylene glycol these results cannot be taken as the true pKa values of these isomeric acids. A 90 ml. aliquot of the ethylene glycol water solution was diluted with 110 ml. of water and neutralized to pH 7 with dilute hydrochloric acid. The solid was extracted from the solution with benzene. The benzene solution was washed with water and dried with anhydrous sodium sulfate. Evaporation of the benzene left a residue, which when recrystal— lized from methanol amounted to 2.4 g. of impure threo-2, 3-di(p—methoxy- phenyl)pentanoic acid, m.p. 142—1650. This acid could not be further purified by repeated recrystallization from any solvent tried. Acidification of the aqueous solution to Congo red precipitated the re- maining acid. This second portion of the acid, after recrystallization from methanol, amounted to 2. 9 g. of pure erythro-Z, 3-di(p-methoxy- phenyl)pentanoic acid, m.p. 177-1790. 79 Attempted chromatographic purification of crude threo—2, 3—di- _________________.__._.__———— (p-methoxyphenyl)pentanoic acid. A solution of 4. l g. of crude threo— 2, 3-di(p-methoxyphenyl)pentanoic acid, m.p. 128—1580, in 200 ml. of ether was passed through a 1. 5 x 75 cm. column packed with 150 g. of acid-washed alumina. The column was eluted with 1-1. of ether=hexane (1:4), 1—1. of ether, 1-1. of ether-methanol (3:2) and 3-1. of methanol. The fractions were collected in 250 ml. portions and evaporated. The first appearance of solid was from the first pure methanol fraction. The total solid obtained from the 3-1. of methanol amounted to 0. 12 g. and melted at 130—1480. The alumina was removed from the column and extracted with methanol in a Soxhlet extractor for 72 hrs. Evaporation of the methanol gave 3. 6 g. of impure methyl threo-2, 3—di- (p—methoxyphenyl)pentanoate, m.p. 91—96. 50. Extraction of a benzene solution of this ester with dilute sodium carbonate solution and recrystal— lization of the residue from methanol gave 3. l g. of pure methyl threo— 2,3-di(p—methoxyphenyl)pentanoate, m.p. 94-960. Attempted acid hydrolysis of glam-2, 3-di(p-methoxypheny_l)— pentanenitrile. A solution of 5. 9 g. (0.02 mole) of threowZ, 3—di(p- methoxyphenyl)pentanenitrile in 150 m1. of dry methanol was saturated with dry hydrogen chloride at 250. After the solution had been refluxed for 9 hrs. , most of the solvent was distilled off at atmospheric pressure. The remainder of the solvent was removed from the cool solution with a Rinco evaporator and left 6. l g. of light yellow sirup, which contained nitrogen (sodium fusion). This residue was refluxed with 100 ml. of 15% methanolic potassium hydroxide for 2 hrs. , and was diluted with an equal volume of water. The methanol was removed in a current of air. _A heavy oil separated from the solution as the methanol was removed. After the insoluble oil was extracted from the aqueous solution with benzene, the solution was acidified to Congo red with 6N hydrochloric 8O acid. No acidic material could be isolated from this solution. Evaporation of the benzene extract gave 5. 8 g. of recovered nitrile. Attempted preparation of threo-3, 4—di(p-methoxyphenyl)-2-hexanone and the hypochlorite oxidation to threo-2, 3-di(p-methoxyphenyl)pentanoic acid. A solution of 3.0 g. (0.01 mole) of threo-2, 3—di(p==methoxyphenyl)— pentanenitrile in 100 ml. of dry benzene was slowly added to a well— stirred solution of 0. 05 mole of methylmagnesium iodide in 100 ml. of ether. The ether was distilled off and the reaction mixture was refluxed for 48 hrs. The cool reaction mixture was hydrolyzed by adding sufficient 6N hydrochloric acid to give a clear water layer. The mixture was then heated to 600 for several hours to insure complete hydrolysis. The water layer was separated and extracted with two 50 ml. portions of benzene. The combined benzene extract was washed successively with water, with 10% sodium carbonate, and with saturated sodium chloride solution and was dried over anhydrous sodium sulfate. Removal of the solvent with a Rinco evaporator left 2. 8 g. of a light yellow sirup, which would not solidify. This product did not contain nitrogen (sodium fusion). This oil was added to 150 ml. of 5. 25% sodium hypochlorite and sufficient dioxane was added to effect solution. After standing for 24 hrs. , the mixture was warmed in a water bath to 600 for several hours. The cool reaction mixture, acidified to Congo red with 6N hydro— chloric acid, was extracted with three 100 ml. portions of methylene chloride. After drying over anhydrous sodium sulfate and after removal of the solvent with a Rinco evaporator, the semi—solid residue was recrystallized from methanol to give 0. 9 g. of erythro-Z, 3—di(p-methoxy-— phenyl)pentanoic acid, m.p. 177-1790. The residue obtained by the evaporation of the filtrate would not solidify. This light tan oil had a neutralization equivalent of 311. 1 (calculated for C19H22042314.4) and was assumed to be impure threo—2, 3—di(p-methoxyphenyl)pentanoic acid. 81 Zone precipitation of threo-2, 3-di(p—methoxyphenyl)pentanoic acid. A dry paste of 2. 0 g. of impure threo-2, 3=di(p-methoxyphenyl)- pentanoic acid and 0. 5 ml. of anisole was placed,as a liquid melt, in a VPyrex zone refining tube 2 mm. in diameter. The heaters of 3. Fisher zone refiner were set to give a temperature of 12.00 at the center of the circular coil. With the refining speed set at the slowest position, resolidification was achieved using the compressed air cooling rings. The sample was melted eight times, in the descend— ing direction, using double heaters and cooling rings. The glass tube containing the sample was then cut into 1 cm. lengths and the content of each section was dissolved in 5 ml. of methanol. After removing the glass, these solutions were evaporated to a volume of approximately 1 ml. , were cooled in an ice bath for several hours and were filtered. The melting points of the fractions were as follows. 0 Fraction m.p., C. 1-6 175-179 7-9 174-170 10-12 167-171 13-17 164-170 18-22 166-171 Fractions 10 through 22 were combined and were recrystallized from methanol to give 1 . l g. of threo-2, 3-di(p-methoxyphenyl)pentanoic acid, m.p. 168-170.50. The zone refining of this acid by the conventional method without using anisole as a solvent was unsatisfactory. The tubes would invariably burst, presumably because of the decarboxylation of the acid. 2—Pyranyl erythro—Z, 3—di(p-methoxyphenyl)pentanoate. To a solu- tion of 2. 53 g. (0.03 mole) of 3, 4-dihydro—2H—pyran and of 20 mg. of p-toluenesulfonic acid—monohydrate in 70 ml. of dry benzene contained in a 125 ml. iodine flask was added 6.49 g. (0.02 mole) of finely powdered 82 erythro-Z, 3-di(p-methoxyphenyl)pentanoic acid. This solution was stirred in the glass stoppered flask at room temperature for 3. 5 hrs. with a magnetic stirrer. One drop of pyridine was then added and the resulting solution was extracted successively with dilute sodium carbonate solution, with water and with saturated sodium chloride solu- tion. After the solution was dried over anhydrous sodium sulfate, removal of the solvent with a Rinco evaporator left 7.42 g. of colorless solid residue. Recrystallization from ethanol gave 7. 1 g. (89. 0%) of Z—pyranyl erythro-Z, 3-=di(p—methoxyphenyl)pentanoate, m.p. 12211240, This procedure is a modification of that described by Johnson (128) for the preparation of 2-pyranyl esters of steroidal acids. Hydrolysis of 2-pyranyl erythro—Z, 3-di(p—methoxyphenyl)pentanoate. One gram of 2-pyranyl erythro-Z, 3-di(p-=methoxyphenyl)pentanoate was refluxed in 10 m1. of acetic acid with 5 mg. of p-toluenesulfonic acid monohydrate for 30 min. and the hot solution was poured onto ice. The resulting solid was taken up in ether and the water solution was ex- tracted several times with ether. The combined ether solution was washed once with water and was then dried over anhydrous sodium sulfate. There was obtained, upon removal of the solvent and recrystallization of the residue from methanol, 0.76 g. (96%») of erythro-Z, 3—di(p-methoxy- phenyl)pentanoic acid, m.p. 177-1790. 2 - Pyranyl threo- 2, 3 - di(p-methoxyphenyl)pentanoate . a) With p-toluenesulfonic acid. This procedure was identical with that employed for the erythro—pyranyl ester. Treatment of 5. 5 g. (0. 017 mole) of crude threo—2, 3-di(p—methoxyphenyl)pentanoic acid, m.p. 130— 165°, with 2.14 g. (0.025 mole) of 3,4-dihydro—2H-pyran and 20 mg. of p-toluenesulfonic acid monohydrate yielded less than 0. 5 g. of oil as the only neutral product. Hydrolysis of this oil in acetic acid with p—toluene— sulfonic acid, as described for the erythro—pyranyl ester, failed to yield an isolatable amount of acidic material. 83 b) With concentrated sulfuric acid. A solution of 4. 92g. (0. 015 mole) of crude threo-2, 3-—di(p-methoxyphenyl)pentanoic acid, m.p. 130—1650, and of 1.90 g. (0.023 mole) of 3,4—dihydro=2H—pyran in 75 ml. of dry benzene containing 2 drops of concentrated sulfuric acid was stirred at room temperature for 48 hrs. with a magnetic stirrer. After one drop of pyridine and 40 ml. of ether were added to the reaction mixture, the solid was removed by filtration. This recovered starting acid weighed 0. 32 g. and melted at 140—1690. The filtrate was washed several times with 5% sodium carbonate solution and several times with saturated sodium chloride solution. It was dried over anhydrous sodium sulfate. After removal of the solvent with a Rinco evaporator, the residue was recrystallized from methanol, giving 4. 82 g. of a colorless solid, m. p. 105—1150 with previous softening. Concentration of the filtrate gave a yellow oil which failed to solidify. Repeated recrystal- lization of the solid from methanol afforded 2. 2 g. of 2=pyranyl threo— 2, 3-di(p—methoxyphenyl)pentanoate, m. p. 116-1180. This procedure is essentially that of Bowman and Fordham (129), who used a sulfuric acid catalyst for the preparation of pyranyl malonates. One gram of the above ester and 5 mg. of p—toluenesulfonic acid monohydrate were refluxed in 10 ml. of acetic acid for 30 min. The hot solution was poured over ice and the resulting mixture was extracted with several portions of ether. The combined ether solution was washed once with water and was dried over anhydrous sodium sulfate. Removal of the solvent with a Rinco evaporator left a semi-solid residue. Recrystallization of the latter from benzene-petroleum ether (60-900) gave 0. 62 g. (77%) of impure threo-2, 3—di(p-methoxyphenyl)pentanoic acid, m.p. 141-1680. 84 N, N—Bis(2-hydroxyethyl)-erythro-2, 3—di(p-methoxyphenyl)— pentanamide. A solution of 30 ml. (0.42 mole) of freshly distilled Matheson, Coleman and Bell thionyl chloride in 30 ml. of dry benzene was added over a period of 30 min. to a stirred slurry of 28.4 g. (0. 088 mole) of erythro—Z, 3-di(p-methoxyphenyl)pentanoic acid in 50 ml. of dry benzene. After the addition had been completed, the homo» geneous solution was allowed to stand at room temperature for 1 hr. and was then refluxed on the steam bath for 1. 5 hr. The benzene and excess thionyl chloride were stripped off with a Rinco evaporator and the residue was slurried with 100 ml. of dry benzene and the process repeated several times in order to remove the last traces of thionyl chloride. The residual acid chloride, dissolved in 100 ml. of purified dioxane (130) was added over a period of 3 hrs. to a well-stirred mixture of 37 g. (0. 352 mole) of diethanolamine in 100 ml. of dioxane. The temperature rose to 550 during this addition. After the reaction mixture had been heated at 800 with constant stirring for several hours, the dioxane was distilled directly from the reaction flask at 20 mm. pressure. The resulting liquid residue was dissolved in 200 ml. of benzene and 100 ml. of water. The water layer was separated and was extracted with two 50 m1. portions of benzene. The combined organic extracts were washed successively with dilute hydrochloric acid, water, 10% sodium carbonate solution and water. The sodium carbonate wash and the last water wash were combined and acidified to give 0. 29 g. of recovered erythro—2, 3—di(p-methoxyphenyl)pentanoic acid, m.p. 173-1770. The residue remaining after the benzene was removed with a Rinco evaporator was dried at 560 at 0. 05 mm. pressure for 6 hrs. There remained 28.8 g. (80%) of crude N, N-bis(2—hydroxyethyl)-erythro-Z, 3-di(p-methoxyphenyl)— pentanamide, m.p. 75.5—910. After repeated recrystallizations from methanol, the melting point range of this material was not greatly improved. Attempted zone 85 purification was unsuccessful due to the reluctance of the material to solidify from the melt in the tube. Analysis showed this material to be impure. _Analysis. Calculated for C23H31NOS: C, 68.80; H, 7. 78; N, 3.49. Found: C, 66.77; H, 7.99; N, 2.86. N, N— Bis (2-hydroxyethyl)| erythro— 2, 3 —di(p—methoxyphenyl)pentyl]- amine hydrochloride by the reduction of the corresponding amide. A solution of 21. 2 g. (0. 05 mole) of crude erythro—N, N-bis(2-hydroxyethyl)— 2, 3—di(p—methoxyphenyl)pentanamide in 200 ml. of freshly distilled tetra- hydrofuran was added over a period of 3 hrs. to a well-stirred slurry of 2. 1 g. (0. 06 mole) of lithium aluminum hydride in 800 ml. of tetra— hydrofuran. After the reaction mixture had been refluxed under a nitrogen atmosphere for 24 hrs. , an additional 0. 2 g. of solid lithium aluminum hydride was added and the refluxing was continued for 12 hrs. The cool reaction mixture was hydrolyzed by the dropwise addition of 10 ml. of water and the basic aluminum salts complexed by the addition of 400 ml. of saturated ammonium tartrate—ammonium sulfate solution. The inorganic salts were not completely complexed, but the dark gray solids remained suspended in the water layer and filtration was un- necessary. After the addition of 200 m1. of methylene chloride the organic layer was separated and the water suspension was extracted several times with methylene chloride. The combined organic solution was washed with water and dried over anhydrous sodium sulfate. Removal of the solvents with a Rinco evaporator left 19.15 g. (93. 5%) of nearly colorless oil. This oil was dissolved in 100 ml. of methanol, anhydrous hydrogen chloride was added to the cool solution until it was acidic to moist litmus and the whole permitted to stand for an hour at room temperature. The solvent was then removed, leaving 21. 25 g. (100%--based on free amine) of crude N, N—bis(2-hydroxyethy1)[erythro- 2, 3—di(p-methoxyphenyl)pentyl]amine hydrochloride as a very thick sirup. 86 This product was not further purified before its use in the preparation of the nitrogen mustard. N, N-Bis(2-chloroethy1)|erythrg-Z, 3—di(p-methoxmhenyl)pentyl l— amine hydrochloride with dimethylformamide catalyst. A solution of 19.6 g. (0. 045 mole) of N, N—bis(2-hydroxyethyl)|erflhro-aZ, 3—di(p~methoxy— phenyl)pentyl]amine hydrochloride and of 5 ml. of dimethylformamide in 250 m1. of freshly distilled chloroform was prepared in a 500 ml. flask equipped with a thermometer, a reflux condenser topped with a drying tube, a nitrogen inlet tube and an addition funnel. The flask was cooled in an ice bath to 100 as a stream of dry nitrogen was passed through the system. As this solution was stirred with a magnetic stirrer, a solution of 80 ml. of freshly distilled thionyl chloride in 50 ml. of chloroform was added over a period of 1 hr. After the addition of the thionyl chloride the reaction mixture was allowed to stand at room temperature for 1 hr. The reaction mixture was then slowly heated to 400, at which temperature there commenced a rapid evolution of gas. When this initial reaction had subsided, the reaction mixture was refluxed for 1. 5 hr. Concentra- tion of the latter with a Rinco evaporator gave a residual semi—solid. This was slurried with 100 ml. of dry benzene and the solvent was evapor— ated to strip off the last traces of thionyl chloride. The solid product was rinsed with dry ether and recrystallized twice from dry anisole. Pure N, N-bis(2-chloroethyl)[ernhro-2, 3—di(p—methoxyphenyl)pentyl]~ amine hydrochloride was obtained in a yield of 20. 2 g. (95%), m.p. 134—1350. This compound proved to be identical with the nitrogen mustard hydrochloride prepared in a much lower yield and in an impure state without the addition of the dimethylformamide catalyst (116, 117). N, N—Bis(Z-hydroxyethyl)-threo-2, 3-di(p—methoxyphenyl)pentanamide. A slurry of 4. l g. (0. 013 mole) of threo—2, 3-di(p—methoxyphenyl)pentanoic acid in 10 m1. of dry benzene containing two drops of pyridine was cooled 87 to 100 under a nitrogen atmosphere. To this slurry there was added 3 ml. of freshly distilled Matheson, Coleman and Bell thionyl chloride in one portion. The resulting clear solution was maintained at 100 for 1 hr. , was allowed to stand at room temperature for 3 hrs. and was finally refluxed for 30 min. The excess thionyl chloride and the benzene were removed from the reaction mixture with a Rinco evaporator, leaving a light tan oil which was dissolved in 50 ml. of dry dioxane. This dioxane solution was added over a period of 1. 5 hr. to a well=stirred solution of 5. 5 g. (0. 052 mole) of diethanolamine in 75 ml. of dioxane which had been heated in an oil bath to 550. The temperature of the reaction mixture was maintained at 55—600 during the entire course of this addition. The reaction mixture was then heated at 750 with constant stirring for 12 hrs. , after which the solvent was removed by distillation at 15 mm. pressure directly from the reaction flask. The yellow sirup was taken up in 100 ml. of benzene and 100 ml. of water. The water layer was separated and extracted with three 50 ml. portions of benzene. The combined benzene solution was extracted successively with dilute hydrochloric acid, water, dilute sodium carbonate, and saturated sodium chloride solution. Acidification of the basic extract led to the recovery of approximately 0. 1 g. of impure starting acid, m. p. 163-1700. The benzene solution was dried over anhydrous sodium sulfate and the benzene was then removed with a Rinco evaporator. The last traces of solvent were finally removed by heating the sirup at 400 at a pressure of O. 02 mm. for 72 hrs. The crude N, N—bis(2-hydroxyethyl)-th_r_eg_:2, 3-di(p- methoxyphenyl)pentanamide, a light brown, viscous sirup, weighed 5.1 g. (97%). N, N-Bis(2—chloroethyl) m-Z, 3-di(p-methoxyphenyl)penty1 - amine hydrochloride. A solution of 5. 2 g. (0. 013 mole) of N, N-bis- (2-hydroxyethyl)—threo-—2, 3—di(p—methoxyphenyl)pentanamide in 100 ml. of anhydrous tetrahydrofuran was added over a period of 1 hr. to a 88 well-stirred slurry of 1. 0 g. (0.026 mole) of lithium aluminum hydride in 75 ml. of tetrahydrofuran. . After being refluxed for 9 hrs. , the reaction mixture was cooled to 100 and the excess lithium aluminum hydride was destroyed by the dropwise addition of 9 g. (0. 1 mole) of ethyl acetate. The aluminum salts were complexed by the addition of 150 ml. of saturated ammonium tartrate—ammonium sulfate solution. A slight turbidity remained in the water layer, but the layers separated cleanly making filtration unnecessary. After the addition of 200 ml. of methylene chloride, the layers were separated and the water layer was extracted with two 50 ml. portions of methylene chloride. The combined organic solution was washed several times with saturated ammonium tartrate—ammonium sulfate solution and was dried over anhydrous sodium sulfate. Removal of the solvent with a Rinco evaporator left 5. 3 g. of light tan sirup. This was dissolved in 100 m1. of dry ether and after cooling to 100, the solution was saturated with anhydrous hydrogen chloride. The amine hydrochloride separated as a dark yellow gum, which was dissolved in hot acetone and was reprecipitated by the addition of ether. Decantation of the solvent and drying i_n vacuo left 5. 3 g. (96%) of N, N-bis(2—hydroxyethy1)-£1_r_e_o_-2, 3—di(p-methoxyphenyl)pentylamine hydrochloride as a light tan glass. To a cold solution of the above amine hydrochloride in 50 ml. of freshly distilled chloroform containing 0. 5 ml. of dimethylformamide there was added 6. 2 g. (0. 052 mole) of freshly distilled Matheson, Coleman and Bell thionyl chloride in 20 ml. of chloroform over a period of 30 min. The temperature of the reaction mixture was maintained below 100 during this addition. After standing at room temperature for 2 hrs., the reaction mixture was heated to 600 for 1 hr. The solvent and excess thionyl chloride were removed with a Rinco evaporator and the residue was slurried with dry benzene. Evaporation of the benzene removed the last traces of thionyl chloride. Drying i_p vacuo at 400 at 89 a pressure of 0. 2 mm. afforded 5.4 g. (90%) of impure N, N-bis(2-chloro— ethyl)|threo-2, 3—di(p-methoxyphenyl)pentyl]amine hydrochloride as a brown sirup. This product could not be obtained as a crystalline solid from any of the solvents tried. Treatment of a chloroform solution of this nitrogen mustard hydrochloride with anhydrous calcium chloride followed by chromatography on alumina and reprecipitation of the amine hydrochloride from an ether solution with anhydrous hydrogen chloride did not seem to improve its purity or color. erghro—Z, 3—Di(p—hydroxyphenyl)pentanoic acid. A mixture of 6. l g. (0. 019 mole) of erythro-Z, 3-di(p—methoxyphenyl)pentanoic acid and 22 g. (0. 14 mole) of pyridine hydrochloride was heated for 2. 5 hrs. at 2100. The resulting light tan, homogeneous solution was allowed to cool slightly and was poured into 150 ml. of cold water. This mixture was extracted with four 50 m1. portions of methylene chloride. After suc- cessively extracting the combined methylene chloride solution with dilute hydrochloric acid and with saturated sodium chloride solution and after drying over anhydrous sodium sulfate, the methylene chloride was removed with a Rinco evaporator. The resulting tan solid was dissolved in 100 m1. of 40% aqueous methanol, treated with Norite and the light yellow solu- tion cooled at —50 for 72 hrs. The colorless solid was removed by filtration and was air dried. The impure erythro-Z, 3-di:(p-hydroxyphenyl)- pentanoic acid, m.p. 219-2350 (with dec.) weighed 5.4 g. Two more recrystallizations from 40% aqueous methanol followed by vacuum drying over calcium chloride gave 4.9 g. (88%) of pure acid, m.p. 245—2480. Foss,. Freund and Stove (131) isolated erythro—Z, 3-di(p-hydroxyphenyl)- pentanoic acid, described as the "high—melting isomer, " from a mixture of the isomeric acids obtained by treating erythro—Z, 3-di(p-methoxy- phenyl)pentanoic acid with refluxing concentrated hydrochloric acid. 90 These workers obtained the erythro-acid in a 6. 5% yield and reported . a melting point of 234-2440. By preparing and purifying the diacetate of the acid these investigators obtained on hydrolysis an acid with a melting point of 243-2450. erythro—Z, 3-Di(p—acetoxyphenyl)pentanoic acid. A solution of 11. 5 g. (0.04 mole) of erythro—Z, 3—di(p—hydroxyphenyl)pentanoic acid in 30 m1. of acetic anhydride and 60 ml. of pyridine was refluxed under an atmosphere of nitrogen for a period of 10 min. The hot reaction mix— ture was poured into 300 ml. of ice water and the resulting mixture extracted several times with ether. The combined ether extract was washed with saturated sodium chloride solution and was dried over anhydrous sodium sulfate before removing the solvent with a Rinco evaporator. The light tan residue was recrystallized twice from cyclo— hexane —benzene (1:1) to give 9. 2 g. (62%) of pure erythro~2, 3-di(p— acetoxyphenyl)pentanoic acid, m.p. 215. 5—2170. A 1.0 g. portion of this product gave 0. 72 g. (93%) of pure erythro—Z, 3-di(p—hydroxyphenyl)pentanoic acid, m.p. 245-2470, on mild hydrolysis with 5% methanolic potassium hydroxide. The procedure used for this acetylation is that of Dodds and co- workers (132) . N, N-Bis(2ahydroxyethyl)-erythro—2, 3—di(p—hydroxyphenyl)- pentanamide. A solution of 12. 0 g. (0.034 mole) of erythro—Z, 3-di- (p—acetoxyphenyl)pentanoic acid and 2. 7 g. (0. 034 mole) of pyridine in 60 ml. of dry benzene was cooled to 100 under a nitrogen atmosphere. To this solution was added a solution of 17 g. (0. 14 mole) of freshly distilled Matheson, Coleman and Bell thionyl chloride in 10 ml. of dry benzene over a period of 30 min. with constant stirring. After being kept at 100 for 1 hr. and at room temperature for 4 hrs. , the reaction mixture was refluxed for 30 min. After removing the excess thionyl 91 chloride and the benzene from the reaction mixture with a Rinco evaporator, the residue was slurried with 20 ml. of dry benzene. The benzene was again evaporated to remove the last traces of thionyl chloride. The resi— due was warmed with 75 ml. of dry dioxane and the warm solution was filtered into an addition funnel. The solid pyridine hydrochloride was rinsed with an additional 20 ml. of dioxane. The dioxane solution was added to a well-stirred solution of 28. 5 g. (0. 25 mole) of diethanolamine in 100 ml. of dioxane over a period of 2 hrs. The temperature of this reaction mixture was maintained at 500 during the course of this addition. After the heterogeneous reaction mixture had been stirred for an addi— tional 12 hrs. at 500, the solvent was distilled directly from the reaction flask i_n vacuo. The reaction mixture was poured into 200 ml. of water and the solution was made slightly acidic with dilute hydrochloric acid. Neither chloroform, ether, nor benzene extracted the insoluble oil from the water layer. However, the oil solidified after stirring with a glass covered magnetic stirring bar overnight in an ice bath. One recrystalliza- tion from methanol gave 13.2 g. of colorless solid, m.p. 122-1680. The theoretical yield of N, N—bis(2—hydroxyethyl)-erythro-2, 3—di(p-hydroxy- phenyl)pentanamide is 12. 2 g. and that of N, N—bis(2—hydroxyethyl)-erythro-' 2, 3—di(p—acetoxyphenyl)pentanamide is 15. 0 g. The product obtained here is undoubtedly a mixture of both the p-hydroxy— and the p-acetoxy- compounds, and may well contain some of the O-acylated product. No attempt was made to purify this mixture further before its reduction to the corresponding amine. N, N—Bis(2—hydroxyethyl) I erythro-Z, 3—di(p—hydroxyphenyl)pentyl]— amine hydrochloride. A solution of 12.0 g. of the crude N, N-bis(2— hydroxyethyl)-erythro—Z, 3-di(p—hydroxyphenyl)pentanamide in 100 ml. of dry tetrahydrofuran was added to a mechanically stirred slurry of 6 g. of lithium aluminum hydride in 250 m1. of tetrahydrofuran over a period of 30 min. This reaction mixture was maintained at the reflux temperature 92 during the course of this addition. After an additional 24 hrs. of reflux- ing, the resulting slurry was cooled to 50 in an ice bath. The excess lithium aluminum hydride was destroyed by the dropwise addition of moist ethyl acetate until no further reaction was noted (about 5 ml. was required). The aluminum salts were complexed by the addition of 300 ml. of saturated ammonium tartrate-ammonium sulfate solution. _After the addition of 200 ml. of methylene chloride, the water layer was separated and was extracted with two 50 m1. portions of methylene chloride. The combined organic solution was washed with 100 ml. of saturated ammonium tartrate-ammonium sulfate solution and was dried over anhydrous sodium sulfate. The solvent was removed with a Rinco evaporator. The resulting light tan sirup was dissolved in 200 ml. of dry ether, the solution was cooled to 100 and was treated with anhydrous hydrogen chloride until it was acid to moist Congo red paper. This re- action mixture stood overnight at room temperature and the ether was then decanted from the semi—solid residue. The latter, recrystallized twice from 75% aqueous methanol, gave 8. 6 g. of N, N-bis(2—hydroxyethyl)— Ierythro—Z, 3-di(p—hydroxyphenyl)pentyl]amine hydrochloride, m. p. 213-2160. An analytical sample of this amine hydrochloride was prepared by the repeated recrystallization of 1. 0 g. of the product from methanol- acetone (1:2), giving 0.4 g. of pure compound, m.p. 217-2180. Analysis: Calculated for C21H30C1NO4: C, 63.70; H, 7.64; Cl, 8.96. Found: C, 63.51; H, 7.64; Cl, 9.14. N,N-Bis(2-chloroeth 1) er thro-Z, 3—di( -h drox hen l) ent l - amine hydrochloride. A slurry of 5.0 g. (0.013 mole) of N,N—bis(2- hydroxyethyl)| erythro—Z, 3—di(p-hydroxyphenyl)pentyl]amine hydrochloride in 100 m1. of freshly distilled chloroform containing two drops of di- methylformamide was cooled to 50 under a nitrogen atmosphere. To this there was added a solution of 25 ml. of freshly distilled Matheson, 93 Coleman and Bell thionyl chloride in 50 ml. of chloroform at such a rate that there was no rise in temperature. The large excess of thionyl chloride was used in order to effect a homogeneous reaction mixture. After the solution had been allowed to stand at room temperature for 30 min. , the temperature was slowly increased by warming the reaction flask in a water bath. When the reaction mixture reached 450 a slow reaction was evidenced by the evolution of a gas. After maintaining the temperature of the reaction mixture at 500 for 30 min. , it was re= fluxed for 15 min. on a steam bath. Removal of the excess thionyl chloride and of the chloroform with a Rinco evaporator left a tan solid residue. This was slurried with chloroform several times and the latter evaporated in order to remove the last traces of thionyl chloride. Several recrystallizations of the solid from anisole gave 4. 6 g. (84%) of N, N-bis(2—chloroethyl)[ erythro-2, 3-di(p—hydroxyphenyl)pentyl]amine hydrochloride, m.p. 210—2200 (with decomp.). Repeated recrystalliza= tions of this material failed to raise the melting point further. Analysis. Calculated for C21H38C13NOZ: C, 58.27; H, 6.52; CI, 24. 58 Found: C, 56.71; H, 6.45; Cl, 24.59; ash, 0.75. The appearance of a residue in this analysis is unreconcilable, considering that the diethanolamine hydrochloride used gave a satis— factory analysis showing no ash. Also, no solid reagents were used in this preparation. threo-2, 3-Di(p—hydroxyphenyl)pentanoic acid. A mixture of 4.0 g. (0. 01 mole) of threo-2, 3-di(p-methoxyphenyl)pentanoic acid and 15. 1 g. (0.13 mole) of pyridine hydrochloride was heated for 2. 5 hrs. at 2100. The resulting light brown solution was poured onto approximately 200 g. of crushed ice. The tarry residue was extracted from the water with several portions of methylene chloride. After the methylene chloride solution was washed several times with saturated sodium chloride 94 solution and after it was dried over anhydrous sodium sulfate, the solvent was removed with a Rinco evaporator. Recrystallization of the light brown, semi-solid residue from 40% aqueous methanol afforded a first crop of 1. 3 g. of crude erythro-Z, 3-di(p-hydroxyphenyl)pentanoic acid, m.p. 230-2420, and a second crop of 0. 9 g. of impure threo—2, 3-di(p- hydroxyphenyl)pentanoic acid, m.p. 191-1960. . Concentrating the filtrate yielded 0. 8 g. of non—crystalline residue from which no further solid acid could be isolated. Two further recrystallizations of the erythro-acid from 40% aqueous methanol gave 1.1 g. of pure compound, m.p. 245-2470. After recrystallizing the threo-acid three times from 40% methanol there was obtained 0.6 g. (17%) of pure compound, m.p. 195-1980, threo—2, 3-Di(p-hydroxyphenyl)pentanoic acid has been prepared previously by Foss, Freund and Stove (131). These workers obtained the acid in a 1. 5% yield by refluxing erythro—Z, 3-di(p—methoxyphenyl)pentanoic acid in concentrated hydrochloric acid. They reported this acid only as the “low—melting" acid with a melting point of 188—1970. erythro—Z, 3a-Di(p-methoxyphenyl)pentanoyl chloride. A suspension of 10. 0 g. (0. 032 mole) of erythro-2, 3-di(p-methoxyphenyl)pentanoic acid in 20 ml. of dry benzene was prepared in a 100 m1. flask, equipped with a reflux condenser, thermometer and a nitrogen inlet tube. This suspension was cooled to 100 and the flask purged with dry nitrogen. Matheson, Coleman and Bell thionyl chloride (20 ml., 0. 22 mole) was distilled directly into the reaction flask as the slurry was vigorously stirred with a magnetic stirrer. After the thionyl chloride had been added, the homogeneous reaction misture was allowed to come to room temperature. It was then heated on the steam bath for 30 min. and finally allowed to stand at room temperature overnight. The excess thionyl chloride and the benzene were stripped off with a Rinco evaporator. Any residual thionyl chloride was removed by repeatedly slurrying the solid residue 95 with 10 ml. portions of dry benzene and by evaporating the solvent on the evaporator. The light tan residue, m.p. 126-1300, was recrystal- lized from benzene. This gave 8.8 g. (83%) of colorless erythro—Z, 3-di— (p-methoxyphenyl)pentanoyl chloride, m.p. 136-138.50. Concentration of the filtrate produced no further solid. The solvent was completely removed from the filtrate and the residue was treated with boiling water. Recrystallization of the solid led to the recovery of l. 3 g. of starting acid, m.p. 174-1780. The yield of the acid chloride was 95% when based on unrecovered starting material. Burckhalter and Sam (105) obtained a 60% yield of erythro-Z, 3— di(p—methoxyphenyl)pentanoyl chloride, m.p. 136=l37o, by essentially the same procedure as that described here. They described their product only as the acid chloride of the "high-melting" acid. Methyl erythro-3, 4—di(p-methoxyphenyl)hexanoate. An ethereal solution of alcohol-free diazomethane was prepared by adding a solution of 25 g. N-nitrosomethyl p—toluenesulfonamide in 175 ml. ether to a solu— tion of 7. 2 g. potassium hydroxide in 12 ml. of water and 45 ml. carbitol. The procedure used for the preparation of this diazomethane solution is the same as that described under the preparation of methyl threo—2, 3-di- (p-methoxyphenyl)pentanoate with the exception that the methanol was replaced by the less—volatile carbitol. This solution of diazomethane was transferred to a 1—1. Erlenmeyer flask and was cooled to 00 in an ice bath. To this solution there was added over a period of 30 min. a suspension of 10. 8 g. (0. 033 mole) of erythro-Z, 3—di(p-methoxyphenyl)— pentanoyl chloride in 200 ml. of dry ether. The solution was stirred with a Teflon covered magnetic stirring bar during this addition. The mixture was stirred at 00 for an additional hour, during which time the solid diazoketone had separated as light yellow plates, and at room temperature for 4 hrs. the solution was concentrated to one-half its 96 volume on the steam bath (a fine stream of nitrogen from a small capillary was used to achieve even boiling), and 50 ml. of anhydrous methanol and 0. 8 g. of freshly prepared silver oxide were added. The solution was boiled gently for 2 hrs. with an additional 0. 8 g. of silver oxide being added in three portions at 30 min intervals and with fresh methanol being added occasionally to maintain a volume of approxi— mately 100 ml. The solution was treated with 2 g. of Norite, was filtered, and was passed through a 0. 5 x 3 cm. column of alumina to remove the collodial silver. After the solvent was removed with a Rinco evaporator, 8.4 g. of an off—white solid, m.p. 79:960 was obtained. Several recrystallizations from methanol gave 5. 0 g. (46%) of pure methyl erflhro—3, 4-di(p-methoxyphenyl)hexanoate, m. p. 102-1040. Analysis. Calculated for C21H2604: C, 73.66; H, 7. 63. Found: C, 73.62; H, 7.65. erythro-3, 4—Di(p-methoxyphenyl)hexanoic acid. Five ..grams (0. 014 mole) of methyl erythro-3, 4-di(p—methoxyphenyl)hexanoate was added to a solution of 22. 5 g. potassium hydroxide in 130 ml. of absolute methanol. The mixture was gently warmed and the resulting solution was stirred at room temperature for 3 hrs. , during which time a color- less solid separated and the solution became pale yellow. The reaction mixture was then refluxed for 2 hrs. , and was diluted with an equal volume of water. The methanol was removed in a current of air. The water solution contained a solid which did not dissolve on further dilu- tion. The latter was extracted with methylene chloride. Acidification of the cold aqueous solution to Congo red with cold 6N hydrochloric acid caused the separation of a pasty solid, which was extracted with methylene chloride. From the latter solution, the methylene chloride was removed with a Rinco evaporator and the semi-solid residue was recrystallized from ethanol, giving 3.0 g. of the free acid, m.p. 183-184. 50. Concentration of the filtrate to approximately one-half of its original 97 volume and cooling for several days led to the isolation of an additional 0. 3 g. of acid, m.p. 180-1830. Recrystallization of the combined solid gave 3. 3 g. (69%) of pure erythro-3, 4-di(p-methoxyphenyl)hexanoic acid, m.p. 183484.50. Hofstetter and Smith (123) obtained erythro-3, 4-di(p-methoxyphenyl)— hexanoic acid, m.p. 181. 5-182. 50, by the oxidation of a mixture of 3, 4-di(p-methoxyphenyl)hexanols. The stereochemistry of this acid was not discussed by these workers. erghro-3, 4—Di(p—methoxyphenyl)hexanol. A solution of 3. 0 g. (0. 009 mole) of methyl erythro—3, 4-di(p-methoxyphenyl)hexanoate in 75 ml. of anhydrous ether was added to a well-stirred slurry of l. 7 g. (0.05 mole) of lithium aluminum hydride in 100 ml. of ether at such a rate that the reaction mixture did not attain its reflux temperature. This addition required approximately 30 min. After being stirred at the reflux temperature for 24 hrs. , the reaction mixture was cooled to 50 and was hydrolyzed by the successive dropwise addition of 5 ml. of water and 55 ml. of 6N hydrochloric acid. The water layer was separated and was extracted with two 50 ml. portions of ether which were added to the original ether solution. Removal of the ether with a Rinco evaporator afforded 2. 7 g. of crude product which was recrystallized twice from methanol to give 2. 6 g. (94%) of pure erythro-3, 4-di(p-methoxyphenyl)- 1-hexanol, m.p. 139-140.50. Analysis. Calculated for Gaol—126033 C, 76.40; H, 8. 34. Found: C, 76.48; H, 8.41. erythro—3, 4-Di(p—methoxyphenyl)- l-hexyl p-toluenesulfonate. A solution of 0. 5 g. (0.015 mole) of erythro—3, 4-di(p-methoxyphenyl)- 1-hexanol and 0. 3 g. (0.015 mole) of p-toluenesulfonyl chloride in 10 ml. of dry pyridine was kept at 100 for 24 hrs. in a tightly stoppered flask. The solution was then poured into 20 ml. of ice water and the entire solution 98 was made acid to litmus with cold 3N sulfuric acid. A colorless semi— solid separated upon acidification. The solution was extracted four times with 20 ml. portions of ether-pentane (2:1). The combined ether solution was washed once with saturated sodium chloride and was dried briefly over sodium sulfate. Removal of the solvent with a Rinco evaporator left 0. 73 g. of a pasty residue which on recrystallization from methanol afforded 0. 29 g. (41%) of erythro-3, 4—di(p-methoxy— phenyl)-1—hexy1 p-toluenesulfonate, m.p. 108—1100. No further solid could be obtained from the filtrate. This compound darkened on stand— ing. It was reduced immediately to meso—hexestrol. meso—Hexestrol dimethyl ether. A solution of 0. 20 g. (4. 2 x 10"4 mole) of erythro-3, 4-di(p—methoxyphenyl)—l—hexyl p—toluenesulfonate in 50 ml. of anhydrous ether was added in one portion to a slurry of 0. 13 g. (4. 2 x 10“3 mole) of lithium aluminum hydride in 100 ml. of ether. This reaction mixture was stirred at room temperature for 18 hrs. under an atmosphere of nitrogen. After cooling to 50 the reaction mixture was hydrolyzed by the dropwise addition of 3 ml. of water followed by 25 ml. of 6N hydrochloric acid. The water layer was separated and was extracted with two 25 ml. portions of ether which were added to the original ether solution. This ether solution was washed successively with dilute sodium hydroxide and with saturated sodium chloride solution and was dried over anhydrous sodium sulfate. Distillation of the solvent through a short column at atmospheric pressure left 0.12 g. of colorless, solid residue, m.p. 137-1430. Three recrystallizations of this crude product from methylcyclohexane gave 0. 11 g. (85%) of pure meso— hexestrol dimethyl ether, m.p. 143. 5-1440. This product gave a mixed melting point of 1440 with a sample of o authentic meso-hexestrol dimethyl ether, m.p. 144 . 99 Attempted Schmidt reaction with erythro—3, 4-di(p-methoxyphenyl)- hexanoic acid. A solution of 1.69 g. (0. 005 mole) of erfihroa3, 4—di- (p-methoxyphenyl)hexanoic acid in 40 ml. of chloroform was added to 5 m1. of concentrated sulfuric acid in a 100 ml. three-necked flask, equipped with a reflux condenser and a thermometer. The reaction mixture was heated to 500 in an oil bath and 0. 39 g. (0. 006 mole) of solid sodium azide was added in small portions over a period of one hour as the solution was being vigorously stirred with a magnetic stirrer. Following each addition of sodium azide there was a slight increase in temperature and a rapid evolution of gas. The evolution of nitrogen was allowed to subside before addition of the next portion of sodium azide. After an additional hour of stirring at 500 the evolution of gas had ceased. The sulfuric acid layer was a dark red color. The cool reaction mixture was poured into 200 ml. of ice water, the chloroform layer was separated, and the water layer was extracted with benzene. The combined organic solutions were concentrated in a Rinco evaporator leaving about 0. 2 g. of dark brown tar. Solid potassium hydroxide was added to the water solu— tion to adjust the pH to about 12. The addition of base did not lighten the red color of the water solution and no insoluble material separated. The basic solution was continuously extracted with benzene for several days. The benzene did not remove any of the red colored material and, on evaporation of the benzene, no residue remained. erythro-Z, 3-Di(p-methoxyphenyl)pentylamine using the Curtius reaction with erythro-3, 4-di(p-methoxyphenyl)hexanoic acid. A mixture of 2. 87 g. (0.0085 mole) of erythro—3,4—di(p—methoxyphenyl)hexanoic acid and 6 ml. of freshly distilled Mathe son, Coleman and Bell thionyl chloride was refluxed on the steam bath for 25 min. The acid went into solution almost immediately on heating and the evolution of gas was complete within 15 min. The excess thionyl chloride was removed from 100 the solution with a Rinco evaporator leaving a tan solid product. The last traces of thionyl chloride were removed by the addition of 10 ml. portions of dry benzene followed by evaporation with a Rinco evaporator. A. solution of this acid chloride in 40 ml. of dry acetone was placed in a 100 ml. round bottom flask and cooled to 00 in an ice—salt bath. As the acetone solution was magnetically stirred a solution of 2. 75 g. (0. 043 mole) of sodium azide in 18 ml. of water was added in one portion. The resulting mixture was stirred for 25 min. while being cooled in the ice—salt bath. A colorless solid separated during this reaction period. The ice bath was removed, 70 ml. water was added and the reaction mixture was stirred for 10 min. at room temperature. The solid azide was filtered off and was washed twice with 10 ml. portions of water. The wash water was used to rinse the flask, but the azide which clung to the sides of the reaction flask was not removed. The moist azide was transferred back to the reaction flask and was dissolved in 50 ml. of benzene. This solution was refluxed for 14 hrs. on the steam bath with a Dean and Stark water separator, which had been filled previously with benzene in order to maintain 50 m1. of benzene in the reaction flask. Twenty milliliters of concentrated hydrochloric acid was added to the cool solution and the mixture was heated for 20 min. During this heating period the solid amine hydrochloride separated as a pasty mass. The contents of the flask was added to 300 ml. of water in an open beaker and boiled for 30 min. to remove the benzene. The resulting water solu- tion was treated with Norite, was allowed to cool to room temperature very slowly, and was cooled at 00 for several hours. The resulting colorless solid was filtered off, was air dried overnight and was then dried over phosphorus pentoxide for 24 hrs. The erythro-2, 3-di(p- 101 methoxyphenyl)pentylamine hydrochloride, thus obtained, weighed 1. 92 g. (65%, based on free acid) and melted at 247. 5—2490. The mixed melting point of this amine hydrochloride and the erythro—Z, 3=di(p— Inethoxyphenyl)pentylamine hydrochloride, m.p. 248—2490, obtained by the reduction of the solid nitrile with lithium aluminum hydride was 248—249.5°. PART II NITROGEN MUSTARD ANALOGS OF DIETHYLSTILBESTROL DIMETHYLETHER DISCUSSION In order to evaluate the effectiveness of diethylstilbestrol di— methylether as an estrogenic carrier moiety for nitrogen mustards it was decided to prepare compounds in which one of the ethyl groups is replaced by a bis(2-chloroethyl)aminoalkyl residue. Two such com- pounds of this type were considered: N, N-bis(2-chloroethyl)—3-4—di(p— methoxyphenyl)—3-hexenylamine hydrochloride (1' HCl) and N, N-bis(2- chloroethyl)- 2, 3-di(p—methoxyphenyl) — Z—pentenylamine hydrochloride (IIo HCl) . ('3sz <|32H5 CH30©-C=(|;©OCH3 CH3O©—C:T-©—OCH3 lCH2 CHZ-N(CHZCHZOH)Z CHZ-N(CHZCHZOH)2 HCl HCl 1- HCl 11- HCl N, N— Bis(2-chloroethyl)— 3, 4—di(p-methoxypheny1)- 3-hexenylamine hydrochloride (I- HCl) was conveniently prepared via the corresponding diethanolamide as shown in Figure I. This procedure is essentially that used for the preparation of the hexestrol nitrogen mustards. 102 103 H HHH> o ..m. _ 01m _ mmo~oo~mo GN H; 0 mo mmuw _ _ I m A- III m I I I I I n AI llllllll 0 ©0 mo Goonmo moo O ”_o mo © 0 mo m$0.00.ngme H> nmo...oo...mo 9 > mEND NED _ __ M I I I I I M M I I IOI IOMHIHU O _ _ O H: E nmoo. -okmo- ommo AlbIml foo- -o-mo-_\ /-o...mo 1| o 030 opfiuoHAoOHpafl ocfiadfincoxoflumIAH>Qo£Q>x0£u®EIQVAU -e .méFtohofiuécflmz .2 so 52358an 2:. H oudwfim 11 2 25.2.0 00.x omo©o£o 104 HOE .H Gm ”sofiofovz Nmo ”mm -.Q .-. -.Q mm~_ HHX NEONEUNEUVZ- oo Nmo £00 .0“ o ommo 11m :moEonovz AMEN— X mmno £0006" o-©- ommo mNOUNmo H03 .3va H03 NEONmonovz -Nmo EEO JUOm 1.59 £3 mom Nmo Gm S M I n 11 moo -Q oI Io_- .\ l/Yo mo £23 : ..mNo Um Hoooumo XH ..mNo «TEQO- _H_oIIImn_o- mmouoomo Q mm~_ m m Tll moo- O._ on o _--o© omo N50m onmo All mom EX X HHH> 105 The anisoin (III) was prepared by the benzoin condensation of anisaldehyde. A modification of the procedure of Nazarov and Kotlyarevskii (133) was employed. The essential feature of this pro-= cedure is the use of a proper concentration of aqueous ethanol as the reaction solvent in order to insure that during the course of the re— action the anisoin separates from solution as it is formed. This removes the anisoin from the reaction medium allowing the reaction to proceed to near completion (134,135). Reduction of anisoin to deoxyanisoin (IV) was effected both by the use of metallic tin and hydrochloric acid (136) and by the use of tin(II) chloride and hydrochloric acid (133). The latter procedure gave a product of far superior quality in comparison to that obtained from the conventional metal-acid reduction. The preparation of pure u-alkyldeoxyanisoin has always been a serious problem in the synthesis of stilbestrol derivatives. Alkylation of deoxyanisoin with ethyl iodide using sodium ethoxide as described by Dodds and co-workers (132) gives a virtually inseparable mixture of deoxyanisoin and a-ethyldeoxyanisoin. o—Ethyldeoxyanisoin, being an oil, can be purified only by distillation. This does not afford a separa- tion from the dissolved unalkylated deoxyanisoin. Preparing u-methylene— deoxyanisoin (V) by taking advantage of the facile elimination of piperidine from the Mannich base affords a solid intermediate which can be readily purified. The preparation of the Mannich base from deoxyanisoin, formaldehyde and piperidine and the elimination of the piperidine to form a—methylenedeoxyanisoin can be carried out simultaneously. The quantitative 1, 4-addition of methylmagnesium iodide to this unsaturated ketone gives pure u-ethyldeoxyanisoin (140). Methyl 3-hydroxy—3, 4—di(p-methoxypheny1)-hexanoate (VII) was prepared in good yield from u-ethyldeoxyanisoin and methyl bromoacetate by the Reformatsky reaction. Confirming the observations of Martin and 106 Stoffyn- Thomas (137), the use of ethyl bromoacetate in the present work was found to give much lower yields of the corresponding ethyl ester. Several methods for the dehydration of this B—hydroxy ester were explored, 'none of which proved to be altogether successful. Only unchanged hydroxy ester was isolated after refluxing a toluene solution of the hydroxy ester with iodine. Attempted dehydration with acidic reagents such as p—toluenesulfonic acid, phosphorous pentoxide, finely powdered fused potassium bisulfate or oxalic acid in refluxing benzene did not lead to the isolation of an appreciable amount of acidic product, directly or after subsequent saponification. It was presumed that the dehydration was accompanied by hydrolysis of the ester and subsequent decarboxylation in the acid reaction medium (138). This was verified by the isolation of 2, 3—di(p~methoxyphenyl)-Z-pentene from the attempted dehydration with oxalic acid. Silverman and Bogert (139) isolated 3, 4—di(p—methoxyphenyl)-3—hexenoic acid, in an unreported yield, by dehydrating ethyl 3-hydroxy-3, 4-di(p-methoxyphenyl)-3‘= hexanoate with cold concentrated sulfuric acid. This method was examined and was found to give yields of approximately 10% oflthe free acid after saponification of the dehydrated ester. Treating a pyridine solution of the B-hydroxyester (VII) with thionyl chloride in the cold gave the B—chloroester (VIII). This B-chloroester was dehydrohalo- genated and saponified to give 3,4-di(p—methoxyphenyl)-3—hexenoic acid (IX) in moderate yields. Refluxing the B—hydroxyester in acetyl chloride followed by dehydrohalogenation and saponification gave sufficiently good yields of the 0., B-unsaturated acid to warrant this as the method of choice, despite the fact that it necessitated an additional isomerization step. The 0., fi-unsaturated acid (IX) was isomerized to the more stable 3, 4—di(p-methoxyphenyl)—3—hexenoic acid (X) by refluxing it in either a benzyl alcohol or an ethylene glycol solution of potassium hydroxide. A somewhat higher yield was obtained when the isomerization was carried 107 out in a benzyl alcohol solution, but the isolation of the acid was extremely tedious. The ethylene glycol solution of the sodium salt of the isomerized acid could be diluted with several volumes of water and acidified to precipitate the free acid. Benzyl alcohol, being immiscible with water, had to be removed by steam distillation before isolating the acid. 3, 4—Di(p-methoxyphenyl)-3-hexenoyl chloride (XI) was prepared by treating a chloroform solution of the acid with freshly distilled Matheson, Coleman and Bell thionyl chloride and an equivalent of pyridine. This acid chloride was used, without purification, for the acylation of diethanolamine° The N, N—bis(2—hydroxyethyl)-3, 4-di(p— methoxyphenyl)-3-hexenamide (VII) thus formed was reduced with lithium aluminum hydride to N, N-bis(Z-hydroxyethy1)—3, 4-di(p-methoxy— phenyl)=-3—hexenylamine (XIII) without purification. Unlike the diethanol— amine in the hexestrol series, this unsaturated amine formed a well defined crystalline hydrochloride and was easily purified before con- version to the nitrogen mustard. N, N-Bis(2—chloroethyl) — 3, 4-di(p-methoxyphenyl) — 3-hexenylamine hydrochloride (LI-1C1) was prepared by treating a chloroform solution of the diethanolamine hydrochloride (XII' HCl) with a sufficient excess of thionyl chloride to give a homogeneous reaction mixture, using a dimethylformamide catalyst. Several synthetic routes for the preparation of N, N—bis(Z-chloro- ethyl)2, 3-di(p—methoxyphenyl)-Z-pentenylamine hydrochloride (IL HCl) were devised and investigated. Although none of these procedures was successful, they will be briefly described under separate headings. The complete reaction sequence for the proposed synthetic route will be given and the sequence will then be discussed up to the point of failure. 108 a.) O ([3sz II Ar—C-CZHs + Ar-CHz-CN fl Ar-C=C}I-Ar CN XIV ('3sz ('3sz XIV 1% Ar-C=C‘;-Ar Ell—fl.) Ar-C=(|3—Ar CH2 CH2 I I NH2 N(CHZCHZOH)Z XV XVI XVI fl II- HCl The condensation of aromatic aldehydes with arylacetonitriles is well-known (145) and proceeds, in most cases, to give good yields of a, B-diarylacrylonitriles. The condensation of arylacetonitriles with alkaryl ketones is usually complicated by the formation of chalcones, arising from the self—condensation of two molecules of the ketone. The yields of the desired a-aryl—B-alkylcinnamonitriles are generally low (147). Rorig (146) however succeeded in condensing phenylacetonitrile with propiophenone and p-methoxyphenylacetonitrile with p-methoxy- propiophenone employing freshly prepared sodium amide in refluxing xylene. The yields obtained from these condensations, especially in the case of the desired p-methoxy—compounds, were very low. Attempts to improve the yields of product from the condensation of phenylacetonitrile with propiophenone by employing a wide variety of catalysts were unsuccessful. Details of these attempted condensations are given in the experimental section. 109 I? 9sz Ar—C-CH-Ar C2H5M Bf Ar-C CH-Ar | or C2H5L1 l I CN OH CN XVII XVIII (IZZHS (17sz L' _ XVIII flfi> Ar-Cll—5—(iZI-I—Ar 322+ Ar-C= (II—Ar OH ('le2 (III-12 NH; NII2 XIX XV c XV J—IIE—> XVI soc12 ILHCI The exploratory work on this reaction was done with the more readily available a-benzoylphenylacetonitrile (XVII, Ar: C6H5‘). The selective addition either of one mole of ethylmagnesium bromide or of one mole of ethyl lithium to the carbonyl group of the ketonitrile (XVII) would result in the formation of the hydroxy nitrile (XVIII). This nitrile could be reduced to the alkanolamine (XIX) and then de- hydrated to the unsaturated amine (XV). All attempted additions of ethylmagnesium bromide or of ethyl lithium to c-benzoylphenylacetonitrile (XVII, Ar = C6H5-) gave non-nitrogenous products. Both direct and inverse additions of both ethylmagnesium bromide and ethyl lithium were investigated. 110 C) OMgCl / Ar-CHZCOZH (CH3)ZCHM Cl Ar-CH-COZMgC1 4—Ar-C1-1:C,\ OM c MgCl g l XX 1) ArCOC H ' -H o ' XX ——_,_—-2——5-> Ar-C CI-l-Ar —2—> Ar-C= C-Ar 2) H20 l I | OH COZH COZH XXI XXII (I32H5 (.3sz L'A P XXII —1-—1-Ii‘——> Ar—C= (II-Ar 233—» Ar-CzclI—Ar II- HCl 113 The substitution of an N, N—bis(2—hydroxyethy1)aminomethy1 group for an (II-hydrogen in deoxyanisoin by means of the Mannich reaction would give a—[N, N-bi5([3-hydroxyethyl)aminomethyl]deoxyanisoin (XXIX). Treatment of this diethanolaminoketone with an excess of ethylmagnesium bromide should give N, N-bis(Z—hydroxyethyl)-3-hydroxy—Z, 3—di(p— methoxyphenyl)pentylamine (XXX), which could then be dehydrated to the unsaturated amine (XVI). Attempted Mannich reactions with deoxybenzoin and with deoxy- anisoin employing formaldehyde or trioxyrnethylene and diethanolamine did not give the expected Mannich base. The ketone was recovered un— changed in all attempts to carry out this reaction. Other workers (153) have reported similar difficulties when diethanolamine was used as the base in the Mannich reaction. The addition of diethanolamine to s O O ..-c-é£-.. W A.-c...£-.. 'c'... (I... N(CHZCHZOH)Z V XXIX a-methylenedeoxyanisoin (V) was also unsuccessful. A variety of Mannich bases have been prepared in this fashion by Fiesselmann and Ribka (140). 6) (€sz $21—15 Ar-C: C-Ar $50—$29 Ar-C= c _ Ar | 2) NaN3 ’ CH2 CH2 I I COZH CON3 X XXXI (€sz XXXI Ma; Ar—C: C-Ar % XVI 2) HCl | CH2 I NHz-HC1 XV XVI s_o%_> II-HCI 114 3, 4-Di(p-methoxyphenyl)-3-hexenoic acid (X), available from the preparation of the 3-hexenylamine nitrogen mustard (I), was converted to the corresponding azide (XXXI). This azide was prepared from the acid chloride by the method described for the preparation of erythro—3, 4- di(p—methoxyphenyl)hexanoyl azide. Heating this azide in refluxing benzene, toluene, or xylene did not cause a rearrangement to take place. The azide was recovered unchanged even after 72 hrs. in refluxing xylene. EXPERIMENTAL Anisoin. A mixture of 80 g. (0. 59 mole) of anisaldehyde, 100 m1. of 95% ethanol, 40 m1. of water and 16 g. of potassium cyanide was refluxed for 2 hrs. after the addition of a second portion of 16 g. of potassium cyanide, the mixture was refluxed for an additional 3 hrs. The hot reaction mixture was filtered and the solid potassium cyanide was washed with 50 ml. of hot 95% ethanol. After the addition of 20 m1. of water, the filtrate was rapidly cooled to -50 in an ice-salt bath with constant stirring, was allowed to remain at this temperature for 1 hr. with occasional stirring, and was filtered. The solid was washed once with 50 m1. of cold ether and air dried to give 52. 9 g. (66%) of pale yellow anisoin, m. p. 11 1—1120, This procedure is based on the work of both Sumrell, Stevens and Goheen (135) and Bosher (134). The reported yields are 44-50% and the reported melting point is 110-1120. The major modification of these procedures which was made in the present work was to use less alcohol and more water in order to promote the formation of a two phase system. Nazarov and Kotlyarevskii (133) obtained a yield of 68%, using . . . o _ a Similar ratio of reactants and solvents but heated them to 105 in an autoclave. Deoxyanisoin a) By the reduction of anisoin with tin and hydrochloric acid. A solution containing 27.2 g. (0.1 mole) of anisoin, 25 m1. of concen- trated hydrochloric acid and 30 m1. of 95% ethanol was refluxed vigorously for 24 hrs. with 20 g. (0. 16 g. atom) of powdered tin. The tin settled to the bottom of the flask if the reaction mixture was not vigorously refluxed. The hot solution, decanted from the remaining tin, was filtered and then 115 116 was cooled to 90 with constant stirring. The yellow solid was collected by suction filtration and the hot filtrate was used to wash the tin remain— ing in the reaction flask. After filtering and concentrating this solution to half its original volume, it was cooled to 00 with constant stirring. , A second crop of solid was collected and was combined with the first crop. Recrystallization from 95% ethanol gave 22. l g. of impure deoxyanisoin, m.p. 105—1090. After three further recrystallizations from 95% ethanol, there remained 18.4 g. (72%) of pure deoxyanisoin, m.p. 110-1110. The above procedure is that of Carter e_t a_1_l. who reported in Organic Syntheses (136) that they obtained an 86a92% yield of deoxy- anisoin with a melting point of 108-1110. b) By the reduction of anisoin with tin(II) chloride and hydrochloric acid. A solution containing 166 g. (0.61 mole) of anisoin and 415 g. (l. 83 moles) tin(II) chloride dihydrate in 400 ml. of concentrated hydro- chloric acid and 425 ml. 95% ethanol was refluxed for 3 hr. on a steam bath. This solution was transferred to a 2—1. beaker and was cooled in an ice bath with constant stirring. As the solution cooled 1-1. of water was added in small portions at such a rate that the liquid phase remained homogeneous. The solid obtained after filtration was recrystal- lized from ligroin (90-1200) and gave 142 g. (91%) of pure deoxyanisoin, m.p. 109-1110. Nazarov and Kotlyarevskii (133) reported that a 100% yield of deoxyanisoin, m.p. 109-1100, was obtained by the above procedure. u—Ethyldeoxyanisoin by the alkylation of deoxyanisoin with ethyl iodide. A slurry of 107 g. (0.42 mole) of deoxyanisoin and 65.8 g. (0.42 mole) of ethyl iodide was added in one portion to a boiling solution of sodium ethoxide (prepared from 9. 7 g., 0.42 mole, of sodium and 180 ml. of absolute ethanol). After the flask was swirled several times, 117 a vigorous reaction took place and the flask was cooled in ice water until the refluxing had nearly ceased. The reaction mixture was refluxed until neutral to moist litmus (approximately 20 min. ). A solution of sodium ethoxide (from 3.6 g. , 0.16 mole, sodium and 60 m1. of absolute ethanol) and 24. 5 g. (0. 16 mole) of ethyl iodide was added and the reaction mixture was again refluxed until neutral (approximately 4 hrs. ). The solution was then diluted with 250 ml. of water, the reflux condenser was replaced with a Claisen head and most of the ethanol was removed by distillation at reduced pressure. After the solution was cooled, the product layer was separated and the water layer was extracted with two 100 ml. portions of ether. These were combined with the product layer. This ether solution was washed with two 25 ml. portions of 10% sodium thiosulfate and once with saturated sodium chloride. The dried ether solution (anhydrous sodium sulfate) was concentrated to one-half its volume and was cooled to —50 for 3 hrs. .Filtration gave 14.6 g. of recovered deoxyanisoin, m.p. 108—1100. The remaining ether was removed from the filtrate and the product was distilled with- out tractionation to give 93. 6 g. (82%) of u-ethyldeoxyanisoin, b.p. 162-164O(0.1mm.‘), as a colorless oil. Using a similar procedure, Dodds (132) obtained a 91% yield of a-ethyldeoxyanisoin with a boiling range of 192-1950 (0. 65 mm.). a-Methylenedeoxyanisoin. A slurry of 25. 6 g. (0.1 mole) of deoxyanisoin in a solution of 110 m1. of methanol, 14 m1. of 35% formalin, 0. 5 m1. of pipridine and 0. 5 m1. of glacial acetic acid was refluxed for 5 hrs. on a steam bath. The reaction mixture became homogenous after being refluxed for 2. 5 hrs. The hot reaction mixture was transferred to a beaker and cooled in an ice—salt bath with constant stirring until crystallization began. As the solid separated, water was added in small portions until 100 ml. had been added. After the mixture 0 . had been maintained at -5 for several hours the solid was removed by 118 suctionfiltration and was washed once with cold 50% ethanol. The nearly colorless solid was recrystallized twice from methanol, giving 24. 8 g. (93%) of pure c-methylenedeoxyanisoin, m.p. 61-62. 50. Fiesselmann and Ribka (140) reported that a 67% yield of u-methylene- deoxyanisoin, m.p. 610, was obtained with a similar procedure. The 2, 4-dinitrophenylhydrazone prepared from the above product had a melt- ing point of 170-170. 50 as compared to 1700 reported by Fiesselmann and Ribka. c-Ethyldeoxyanisoin by the addition of methylmagnesium iodide to a-methylenedeoxyanisoin. A solution of 26.8 g. (0. 1 mole) of a-methylene— deoxyanisoin in 200 m1. of anhydrous tetrahydrofuran was added dropwise to a solution of 0. 1 mole of methylmagnesium iodide (prepared from 2. 5 g. of magnesium and 14.2 g. of methyl iodide) in 150 ml. of ether. The temperature of the reaction mixture was maintained between 00 and 50 during this addition. The solution was stirred at room temperature for 2. 5 hrs. and then at reflux temperature for 30 min. After being cooled to 100, the reaction mixture was poured onto cracked ice and was neutralized with cold 6N hydrochloric acid. The water layer was separated and extracted with four 50 m1. portions of ether. The combined ether solution was washed successively with cold 6N hydrochloric acid, with water, with 6N ammonium hydroxide, and with saturated sodium chloride solution and was dried over anhydrous sodium sulfate. Removal of the ether with a Rinco evaporator left a lightly tan oil amounting to 28. 2 g. Distillation of this oil gave only a single fraction of 28. 0 g. of colorless a-ethyldeoxyanisoin, b. p. 170-1730 (0. 2 mm.). Dodds (132) reported a boiling point of 192-1950 (0.65 mm.) for a-ethyldeoxyanisoin which had been prepared by the ethylation of deoxyanisoin with ethyl iodide. 119 Methyl bromoacetate. A mixture of 1—1. of glacial acetic acid, 200 ml. acetic anhydride and 1 ml. of pyridine was placed in a 5=1. flask equipped with an all glass "Turbore" paddle stirrer, with an addition funnel with a stem extending below the surface of the acetic acid and with a parallel adapter holding two reflux condensers topped with standard taper calcium sulfate drying tubes. The solution was heated with a free flame to the reflux temperature, approximately 2 m1. of bromine was added and the solution was refluxed until the bromine color had disappeared. The remainder of 360 ml. (1119 g. , 7. 0 moles) of bromine was added to the stirred solution, at its reflux temperature, at such a rate that no free bromine vapors entered the reflux condensers. This addition required approximately 3 hrs. The evolution of hydrogen bromide was sufficiently rapid to carry acetic acid through the reflux condensers, making it necessary to replace the drying tubes three times during the course of the addition. When the addition had been com— pleted, the solution was refluxed until the bromine color had dissipated and the evolution of hydrogen bromide had ceased. The cool solution was treated with 75 m1. of water to destroy the acetic anhydride and the reflux condensers were replaced with a Claisen head. The acetic acid was then removed by distillation at 35 mm. pressure. Throughout this distillation the solution was stirred to achieve even boiling. After the addition of 960 ml. (24 moles) of dry methanol, 25 ml. of concentrated sulfuric acid, and 2400 ml. of ethylene dichloride, the mixture was refluxed for 8 hrs. The reaction mixture was cooled and was washed successively with two 1—1. portions of water, three 500 ml. portions of 1% sodium bicarbonate solution, once with 500 m1. of saturated sodium chloride solution and was finally dried over anhydrous sodium sulfate. After distillation of the solvent at atmospheric pressure, the product was distilled through a short Vigreux column. The following fractions wer e obtained . 120 B.p. , oC Fraction (30 mm.) Grams I 30-650 16 2 65-65 646 residue - approx. 20 The second fraction constituted a yield of 60% of pure methyl bromo- acetate based on the 7 moles of bromine used in the bromination of acetic acid. The procedure for the bromination of acetic acid is a modification of that described by Natelson and Gottfried in Organic Syntheses (141) and the procedure for the esterification, employing a water immiscible sol- vent, is the general procedure of Clinton and Lasowski (142). Methyl 3-hydroxy—3, 4=di(p—methoxyphenyl)hexanoate. A solution of 94. 5 g. (0. 33 mole) of u—ethyldeoxyanisoin in 240 ml. of anhydrous ether and 360 m1. of dry benzene was placed in a 2—1. flask equipped with a mechanical stirrer, a thermometer, a reflux condenser topped with a calcium chloride drying tube and a nitrogen inlet tube. To this solu— tion was added 80 g. of 30 mesh zinc, which had been previously cleaned and dried (143), 56 m1. of methyl bromoacetate and 1 g. of iodine. After being refluxed for 1. 5 hrs. the reaction mixture became cloudy and began to reflux vigorously. The reaction was not vigorous enough to necessitate the removal of external heat. This initial reaction sub- sided after about 2 hrs. and a second portion of 80 g. of zinc and 56 ml. of methyl bromoacetate was added. Again the reaction mixture was refluxed for an hour with vigorous stirring. A third portion of 5 g. of zinc (making a total 165 g.) and 28 m1. of methyl bromoacetate was added and the refluxing was continued for a further 30 min. After the addition of a fourth portion of 10 m1. of methyl bromoacetate (making a total of 140 m1. ), the reaction mixture was refluxed for an additional 1. 5 hr. The cool reaction mixture was poured into a mixture of 400 m1. of ethanol and 400 ml. of acetic acid and the clear solution was decanted 121 from the unreacted zinc. The zinc was washed with amixture of 100 ml. of ethanol and 100 ml. of acetic acid, divided into three portions, and the washings were added to the first solution. The resulting solution was diluted with 1-1. of water. The aqueous layer was separated and was extracted with four 100 ml. portions of ether. The combined organic solution was washed successively with four 100 ml. portions of 3% ammonium hydroxide and once with 250 m1. of saturated sodium chloride. After the solution had been dried over anhydrous sodium sul- fate, the solvent was removed with a Rinco evaporator and the residue was recrystallized from 80% ethanol. The colorless product amounted to 89. 3 g. (81%) of methyl 3—hydroxy-3, 4-di(p—methoxyphenyl)hexanoate, m.p. 97—1000. This procedure for the Reformatsky reaction is a modification of that reported by Bachrnann and Wilds (144). Methyl 3-hydroxy—3, 4—di- (p—methoxyphenyl)hexanoate has been prepared previously by Martin and Stoffyn-Thomas (137) in a 67% yield by a similar procedure. The melting point is reported to be 96-980. Dehydration of methyl 3-hydroxy-3, 4—di(p—methoxmhenyl)— hexanoate . a) With iodine in refluxing toluene. A solution of 3. 3 g. of methyl 3-hydroxy—3, 4-di(p—methoxypheny1)hexanoate and 0. 5 g. of iodine in 250 m1. of dry toluene was refluxed for 48 hrs. The cool solution was extracted several times with 50 ml. portions of 10% sodium bisulfite, once with water and was dried over anhydrous sodium sulfate. The toluene was removed with a Rinco evaporator and left a colorless solid residue, which was recrystallized from ethanol. Three and one-tenth grams of . 0 starting ester, m.p. 96—100 , was recovered. b) With p-toluenesulfonic acid in refluxing benzene. A solution of 0. 7 g. of methyl 3-hydroxy-3, 4—di(p-methoxyphenyl)hexanoate and 0. 06 122 g. of p-toluenesulfonic acid in 20 m1. of dry benzene was refluxed for 12 hr. The cool solution was extracted with 20 m1. of water and was dried over anhydrous sodium sulfate. . Evaporation of the benzene with a Rinco evaporator left a light tan oil which was refluxed for 2. 5 hrs. with 20 ml. of 15% methanolic potassium hydroxide. This basic solu— tion was diluted with an equal volume of water and the methanol removed in a current of air. The oil, which had separated only partially, dis- solved on the addition of water. After extracting this insoluble .oil with methylene chloride, the cold water solution was acidified to Congo red with 6N hydrochloric acid. The colorless gum which separated on acidification of this solution was extracted with ether. After drying the ether solution over anhydrous sodium sulfate and after removing the ether with a Rinco evaporator there remained O. 32 g. of tanoil. This residue became partly solid after several days, but had turned bright green. The neutralization equivalent of this oil was found to be 271. The calculated value for ConzzO4 is 326. This residue gave a negative phenol test with ferric chloride. c) With phosphorous pentoxide in refluxing benzene. Five grams of methyl 3—hydroxy-3, 4—di(p-methoxyphenyl)hexanoate was dissolved in 50 ml. of dry benzene which contained 2. 5 g. of suspended phosphorous pentoxide. This mixture was refluxed for 12 hrs. with constant stirring. After the mixture was cooled to 100, 25 m1. of water was added at such a rate that the temperature did not rise above 120. The water layer was separated and was washed twice with 10 m1. portions of benzene. The combined benzene solution was washed twice with water and was dried over anhydrous sodium sulfate. Removal of the benzene with a Rinco evaporator left 4. 6 g. of yellow oil, which was refluxed for 2. 5 hrs. . with 80 m1. of 15% methanolic potassium hydroxide. The resulting clear solution was diluted with an equal volume of water and the methanol was removed in a current of air. Acidification of this water solution caused 123 the separation of a light ten semi—solid, which after three recrystalli- zations from 80% ethanol gave 1.4 g... (30%) of 3, 4-di(p-=methoxyphenyl)— 3-hexenoic acid, m.p. 142-1440. No further solid could be obtained from the filtrate. The residual oil obtained by evaporating the filtrate was completely soluble in sodium bicarbonate solution. This residue was found to have a neutralization equivalent of 254. The calculated value for Gaol—12204 is 326. This residue also gave a negative phenol test with ferric chloride . d) With oxalic acid in refluxing benzene. A solution of 6. 6 g. (0. 02 mole) of methyl 3—hydroxy—3, 4-di(p-methoxyphenyl)hexanoate in 50 m1. of benzene containing 0. 5 g. of oxalic acid dihydrate was refluxed for 24 hrs. The cool solution was washed with water and was dried over anhydrous sodium sulfate. After the solvent had been removed with a Rinco evaporator, the sirupy residue was saponified by refluxing it in 50 m1. of 15% methanolic potassium hydroxide for 2. 5 hr. When this solution was diluted with an equal volume of water a colorless solid separated. This solid was removed by filtration and was recrystallized from ethanol, giving 3.8 g. (66%) of 2, 3-di(p-methoxyphenyl)-2-—pentene, m.p. 88n89O (137). Acidification of the basic filtrate gave less than 1 g. of light tan oil, which dissolved completely in sodium bicarbonate solu— tion, which had a neutralization equivalent of 240 and which gave a faint phenol test with ferric chloride. e) With concentrated sulfuric acid.’ Two grams of methyl 3-hydroxy- 3, 4-di(p—methoxyphenyl)hexanoate was added in small portions to 20 ml. of ice-cold sulfuric acid. An immediate deep red coloration resulted. Stirring was continued for 15 min. at 00 and then the solution was allowed to stand for 15 min. at room temperature. The solution was poured onto 40 g. of crushed ice and the reddish—brown gum which separated was extracted with ether. This ether solution was washed with water, was treated with Norite and was dried over anhydrous sodium sulfate. 124 The ether was removed, leaving a light tan paste which was refluxed for 2. 5 hrs. with 40 ml. of 15% methanolic potassium hydroxide. The solution was diluted with 100 ml. of water and the methanol was re- moved in a current of air. Acidification of the resulting water solution to Congo red with 6N hydrochloric acid produced a semissolid which was extracted with ether. After washing the ether solution with water and after drying it over anhydrous sodium sulfate, the ether was removed with a Rinco evaporator. Two recrystallizations from ethanol gave 0.2 g. (11%) of 3,4=di(p-methoxyphenyl)~3=hexenoic acid, m.p. 144-1460. Silverman and Bogart (139), following a similar procedure with the corresponding ethyl ester, obtained 3, 4—di(p=methoxyphenyl)-3- hexenoic acid, m.p. 145—1460, in an unreported yield. f) With thionyl chloride and pyridine. A solution of 10 g. (0.03 mole) of methyl 3-hydroxy-3, 4~di(p—methoxyphenyl)hexanoate in 90 g. of pyridine was cooled to 00. To this solution there was added 100 ml. of cold thionyl chloride at such a rate that the temperature of the reaction mixture did not rise above 30. This addition required approximately 2 hrs. The reaction flask was tightly stoppered and was kept in an icewsalt bath, contained in a Dewar flask, for 12 hrs. The dark brown reaction mixture was slowly poured with vigorous stirring over 500 g. of crushed ice covered with 200 ml. of'ether. After the ice had melted, the water layer was separated and was extracted with three 50 ml. portions of ether. The combined ether extracts were washed with four 50 m1. portions of 5% sodium bicarbonate solution, with two 50 m1. portions of 3N hydrochloric acid and finally with 100 ml. of saturated sodium chloride solution. After the solution had been dried over anhydrous sodium sulfate, the ether was removed with a Rinco evaporator leaving a brown viscous sirup. This residue was refluxed for 4 hrs. in a solution of 5 m1. of water and 75 ml. of methanol containing 8.4 g. (0. 15 mole) of potassium hydroxide. The warm basic solution was 125 diluted with 100 ml. of water and was concentrated to approximately 75 ml. in a current of air. After the resulting brown solution had been treated with Norite, it was cooled to 100 and acidified to Congo red with 6N hydrochloric acid. Three recrystallizations of the semi- solid residue from 80% methanol gave 2. 8 g. (29%) of 3, 4—di(p—methoxy- pheny1)-3—hexenoic acid, m.p. l44==145. 50. i g) With acetyl chloride. A solution of 53.0 g. (0.15 mole) of methyl 3-hydroxyl-3, 4-di(p-methoxyphenyl)hexanoic acid in 116 g. (1. 5 mole) of acetyl chloride was refluxed for 3. 5 hrs. There was an immediate evolution of hydrogen chloride which continued throughout the first 2 hrs. of refluxing. The mixture was then poured onto 300 g. of ice and was allowed to stand with occasional stirring for 30 min. The pasty solid was extracted into ether and the ether solution was washed once with water. After the ether solution had been dried over anhydrous sodium sulfate, the ether was removed with a Rinco evaporator, leaving 53.0 g. of yellow oil. A solution of 42 g. (0. 75 mole) of potassium hydroxide in 300 ml. of 95% ethanol containing 20 ml. of water was added to this oil and the mixture was refluxed for 4 hrs. This solution was then diluted with 300 ml. of water and the total volume reduced to 250 ml. by evaporation in a current of air. This basic solution was treated with Norite and was then acidified to Congo red with 6N hydro- chloric acid. One recrystallization of the solid residue from 80% ethanol gave 32. 5 g. (69%) of 3, 4—di(p-methoxypheny1)-2-hexanoic acid, m.p. 113—1150. This acid was found to have a neutralization equivalent of 309. The calculated value for Gaol-12204 is 326. Martin and Stoffyn— Thomas (137) have reported that this acid has a melting point of 109-1100. 12.6 Isomerization of .3, 4-di(p-methoxyphenyl)—2—hexenoic acid to 3, 4-di(p-me.thoxyphenyl)- 3—hexenoic acid. a) With potassium hydroxide in benzyl alcohol. A solution of 10 g. (0.03 mole) of 3, 4—di(p~methoxyphenyl)-2—hexenoic acid and 25 g. of potassium hydroxide in 200 ml. of benzyl alcohol was refluxed for 20 hrs. This reaction mixture became bright red during the first 12 hrs. of refluxing. The color faded to a light tan by the end of the 20 hrs. reflux period. The hot solution was filtered and the benzyl alcohol was removed by steam distillation. It required over 20 l. of distillate to completely remove the benzyl alcohol from the basic solution. The water solution remaining in the distillation flask was cooled and was acidified to Congo red with 6N hydrochloric acid. The semiusolid residue was recrystallized from 80% methanol twice, giving 8. 9 g. (89%) of 3, 4-di(p=methoxyphenyl)=-3—hexenoic acid, m.p. 143-1450. b) With potassium hydroxide in ethylene glycol. A solution of 32. 6 g. (0. 10 mole) of 3, 4—di(p=-methoxyphenyl)—2—hexenoic acid and 56 g. (1. 0 mole) of potassium hydroxide in 1-1. of redistilled ethylene glycol was gently refluxed for 20 hrs. The solution was permitted to cool and was diluted with l-kg. of ice. The pH of the solution was adjusted to approximately 8 (pHydrion paper) with 6N hydrochloric acid. To this solution there was added three 1-ml. portions of dimethyl sul- fate at 30 min. intervals. The solution was kept in an ice bath during this treatment. The solution was stirred slowly overnight and was then warmed on the steam bath to destroy any excess dimethyl sulfate. As the solution was warmed dilute base was added to maintain a basic solution. The solution was cooled and acidified to Congo red with 6N hydrochloric acid. The pasty solid which separated on acidification was removed with a stirring rod and the aqueous solution extracted with eight lOO—ml. portions of methylene chloride. The solid portion of the product was dissolved in the combined methylene chloride solution which ”1 127 was then washed with six 100-m1. portions of water to remove any dissolved ethylene glycol. After the solvent had been removed with a Rinco evaporator, the light tan residue was recrystallized three times from 80% methanol, giving 26.5 g. (73%) of 3, 4-di(p—methoxyphenyl)- 3—hexenoic acid, m.p. 143.5—1450. N, N-Bis(2-hydroxyethyl)-3, 4-di(p—methoxyphenyl) —3-hexenarnide. A slurry of 9. 5 g. (0.03 mole) of 3,4-di(p-methoxyphenyl)—3—hexanoic acid in 20 ml. of dry benzene and 2. 5 m1. of pyridine was cooled to 100 in an ice bath. As this slurry was vigorously stirred with a magnetic stirrer, a solution of 6 ml. of freshly distilled Matheson, Coleman and Bell thionyl chloride in 10 ml. of dry benzene was added over a period of 30 min. After the homogeneous reaction mixture had been allowed to stand at room temperature for 2 hrs. , it was refluxed for 30 min. Removal of the benzene and the excess thionyl chloride from the reaction mixture with a Rinco evaporator left a brown sirup. Solution of this residue inqdry benzene and evaporation of the benzene was repeated several times in order to remove the last traces of thionyl chloride. The residue was dissolved in 100 m1. of purified dioxane and the solution was filtered in order to remove the pyridine hydrochloride. This dioxane solution was added to a well stirred solution of 7. 0 g. (0. 15 mole) of diethanolamine in 100 m1. of dioxane over a period of 2 hrs. The reaction mixture was slowly heated to 600 in an oil bath and stirred at this temperature for 12 hrs. After the dioxane had been dis- tilled directly from the reaction flask at a pressure of 15 mm., the residue was dissolved in 50 m1. of benzene and 50 ml. of water. The water layer was separated and extracted twice with benzene. The benzene solutions were combined and the whole was successively washed with one 50 ml. portion of water, two 50 m1. portions of 2% potassium hydroxide, one 50 ml. portion of water, two 50 ml. portions of 3N hydrochloric acid and one 100 ml. portion of saturated sodium chloride. 128 Acidification of the basic extract led to the recovery of 1. 2 g. of starting acid, m.p. 145-1510. The benzene solution was dried over anhydrous sodium sulfate and was treated with Norite. The solvent was removed with a Rinco evaporator leaving 9. 8 g. of impure N, N-bis- (2-hydroxyethyl)-3, 4-di(p—methoxyphenyl)-3-hexenamide. This product was not purified before reduction to the correspond-- ing amine. . N, N-Bis(2-hydroxyethyl)-3, 4-di(p-methoxyphenyl)-3=hexenylamine hydrochloride. A solution of 6.4 g. (0. 015 mole) of crude N, N-bis- (2-hydroxyethyl)-3, 4-di(p—methoxyphenyl)-3-hexenamide in 30 m1. of dry tetrahydrofuran was added over a period of 1 hr. to a well stirred slurry of 1. 2 g. (0. 032 mole) lithium aluminum hydride in 75 ml. of dry tetrahydrofuran at room temperature. The mixture was then refluxed under a nitrogen atmosphere for 18 hrs. Two 0. 2 g. portions of solid lithium aluminum were added after 6 and 12 hrs. of refluxing. After the reaction mixture was cooled to 100, there was added 5 m1. of water, followed by 100 ml. of saturated ammonium sulfatecammonium tartrate solution. Fifty milliliters of chloroform was added and the water layer was separated and was extracted with four 25 ml. portions of chloroform. The combined organic solution was successively washed with two 50 m1. portions of saturated ammonium sulfate-ammonium tartrate solution and one 100 m1. portion of saturated sodium chloride. The solution was dried over anhydrous sodium sulfate and the solvent was removed with a Rinco evaporator leaving 6. 2 g. of light yellow sirup. This sirup was dissolved in 100 m1. of absolute ether and anhydrous hydrogen chloride was passed into the solution until it was acid to moist Congo red paper. The ether was removed with a Rinco evaporator and the semi- solid residue recrystallized from benzene-methanol (10:1) giving 4.4 g. (65%) of impure N, N—bis(2-hydroxyethy1)-3, 4-di(p—methoxyphenyl)- hexenylamine hydrochloride, m.p. 119-1250. Several further recrystal- lizations from methanol left 3. 7 g. (55%) of reasonably pure product, 129 In.p. 120-1230. An analytical sample was repeatedly recrystallized from acetone to give a low yield of pure product, m.p. 124—1250. . Analysis: Calculated for CMH34NO4C1: C, 66. 11; H, 7. 86; N, 3.21; CI, 8.13. Found: C, 66.09; H, 7.92; N, 3.43; Cl, 8.07. N, N- Bis(2- chloroethyl) - 3, 4-di(p-methoxyphenyl) =3=hexenylamine hydrochloride. A well stirred slurry of 4. 5 g. (0.01 mole) of N, N—bis- (2=hydroxyethy1)-3, 4-di(p—methoxyphenyl)-3—hexenylamine hydrochloride in a solution of 5 drops dimethylformamide in 30 ml. of dry carbon tetrachloride was cooled to 50. To this slurry there was added a solution of 13. 5 ml. (22.4 g. , 0. 2 mole) of freshly distilled Matheson, Coleman and Bell thionyl chloride in 25 m1. of carbon tetrachloride at such a rate that the temperature did not rise above 50. This addition required approximately 20 min. After standing at room temperature for 4 hrs. the homogeneous reaction mixture was refluxed for 30 min. and was permitted to stand overnight at room temperature. The solvent was removed from the light brown reaction mixture with a Rinco evaporator leaving a brown sirup. The last traces of thionyl chloride were removed from this by repeated solution in 10 ml. portions of benzene and removal of the latter with a Rinco evaporator. Three recrystallizations of the semi—solid residue from anisole gave 3. 9 g. (80%) of N, N-bis(2-chloro- ethyl)-3, 4-di(p-methoxypheny1)—3—hexenylamine hydrochloride, m. p. 139-140°. Analysis. Calculated for CMH3ZOZNC13: C, 60.95; H, 6.82; Cl, 22.49. Found: C, 61.06; H, 6.96; Cl, 22.49. p-Methoxypropiophenone. A solution of 108 g. (1.0 mole) of anisol in 400 ml. of freshly distilled carbon disulfide, contained in a 2-1. flask equipped with a stirrer, an addition funnel, a reflux condenser topped with a calcium chloride drying tube and a thermometer, was cooled to 130 100 in an ice bath. To this solution there was added 293 g. (2. 2 1110165) of anhydrous aluminum chloride in small portions at such a rate that the temperature of the solution did not exceed 150. After the aluminum chloride had been added, the ice bath was removed and the mixture was stirred at room temperature for 30 min. To this resulting clear solu-'- tion there was added 130 g. (1.0 mole) of propanoic anhydride over a period of 30 min. The reaction mixture was refluxed for 15 min. in a water bath, was cooled to 50 and was poured over 1-kg. of cracked ice and 100 m1. of concentrated hydrochloric acid. This hydrolysis mixture was then allowed to stand uncovered in the hood overnight. The water layer was separated and was extracted with three 100 ml. portions of ether. The ether solutions were added to the product layer and the whole was washed successively with two 50 ml. portions of water, with 100 m1. of 5% sodium hydroxide and with saturated sodium chloride solution. After the solution had been dried over anhydrous sodium sulfate the ether was removed at atmospheric pressure and the residue was distilled through a short Vigreux column. B.p., OC. Fraction (16 mm.) Grams 1 54-140 12 2 140-148 144 residue -- 10 The second fraction representing an 88%‘yield of crude p-methoxy— propiophenone, was redistilled through a Fenske column. 0 B.P. , C. Fraction (16 mm.) Grams 1 138-140 2. Z 2 140—141 3.6 3 141-142 68.0 4 142—143 58. Z 5 143-144 6. 3 131 The third and fourth fractions constitute a yield 126.. 2 g. (77%) of pure p-methoxypropiophenone. This procedure is a modification of that of Noller and Adams (154), who reported an 87% yield of p-methoxypropiophenone, b.p. 1250 (4 mm.). Condensation of phenylacetonitrile with propiophenone and of p—methoxyphenylacetonitrile with p-methoxypropiophenone. a) Phenylacetonitrile and propiophenone with freshly prepared sodium amide. A 1-1. flask equipped with a stirrer, a reflux condenser topped with a soda-lime drying tube and an inlet tube was cooled in a dry—ice acetone bath. One hundred milliliters of liquid ammonia was introduced into the flask and a small piece of sodium was added followed by approximately 1 mg. of anhydrous ferric chloride. 7 When the blue color of the solution had dissipated, the remainder of 11. 5 g. of sodium was added in small portions over a period of 5 minutes. After the solu- tion had lost its blue coloration the flask was swirled by hand to dissolve the small pieces of sodium from the sides of the flask. The flask was allowed to stand at room temperature overnight to allow the ammonia to evaporate. After the flask was evacuated for 1 hr. with a water aspirator (the vacuum was broken with dry nitrogen), 100 ml. of sodium dried xylene was added to the dry sodium amide. This slurry was cooled to 100 and a cold solution of 58. 6 g. (0. 5 mole) of phenylacetonitrile in 50 ml. of dry xylene was added over a period of 30 min. To the resulting bright red solution there was added over a period of 1 hr. a solution of 67. 1 g. (0. 5 mole) of propiophenone in 50 ml. of dry xylene. After being refluxed for 1 hr., the reaction mixture was cooled to 100 and was decomposed by the dropwise addition of 10 ml. of water. After the addition of 200 ml. of 20% acetic acid, the water layer was separated and was extracted with two 50 m1. portions of ether. The ether was added to the product layer and the solution was washed with saturated 132 sodium chloride and was dried over anhydrous sodium sulfate. Removal of the solvent by distillation at 15 mm. left a dark brown residue which was distilled without a fractionating column. B.p. , C Fraction (1.0 mm.) Grams 1 40—134 33.1 2 134-136 20.0 3 136—138 22.2 4 138-146 5.5 residue approx. 20.0 Fractions 2, ,3 and 4 were combined and were dissolved in 100 m1. 0 of petroleum ether and the solution was maintained at -5 for several days. The solid obtained amounted to 16.4 g. (14.2%) bf trans—2, 3=di~ phenyl-Z—pentenenitrile, m. p. 109—1110. No additional solid could be obtained from the filtrate which consisted of impure cis-=2, 3-diphenyl—2— pentenenitrile. The procedure for the preparation of sodium amide is essentially that described by Jacobs (155). The method for the condensation of phenylacetonitrile and propiophenone is similar to that given by Rorig (156) who obtained a total yield of 40. 9% of mixed isomers. The trans- isomer is reported by Rorig to be a solid, m.p. 109—1100, and the cis- isomer to be a liquid, b.p. 117—1190(0.1mm.). b) p—Methoxyphenylacetonitrile and p-methoxypropiophenone with freshly prepared sodium amide. This condensation was carried out exactly as described above, except that 73. 6 g. (0. 5 mole) of p-methoxy- phenylacetonitrile and 82. 1 g. (0. 5 mole) of p~methoxypropiophenone were used. The product amounted to 1. 2 g. of material, b.p. 171-1790 (0.1 mm. ). .No solid could be isolated from this product. Rorig (156) reported a 0.17% yield of mixed isomers. The trans- isomer is a solid with a melting point of 131.5-132. 50 and the cis— isomer is a liquid with a boiling range of 160—1660 (0. 05 mm.). 133 This condensation was repeated several times using longer reflux periods and an excess of sodium amide. In no case were the yields better than those described here. c) Phenylacetonitrile and propiophenone with ammonium acetate and acetic acid. A mixture of 29. 3 g. (0. 25 mole) of phenylacetonitrile, 33.8 g. (0. 25 mole) of propiophenone, 3. 9 g- (0.05 mole) of ammonium acetate and 12 g. (0. 2 mole) of acetic acid were dissolved in 50 m1. of benzene. This solution was placed in a 250‘m1. flask equipped with a Dean and Stark water separator pre—filled with benzene and was refluxed for 24 hrs. At the end of this reflux period no water had been collected. The reaction mixture was cooled and was poured into 100 m1. of water. The water layer was separated and the organic layer washed with dilute sodium bicarbonate solution followed by water. After the solution was dried over anhydrous sodium sulfate, the solvent was removed by distil- lation at atmospheric pressure leaving a light brown residue amounting to 61. 5 g. This residue was distilled at reduced pressure to give 60 grams of a mixture of phenylacetonitrile and propiophenone, b. p. 82—950 (5 mm.), n25 1. 5218. A 1:1 mixture of phenylacetonitrile and propio- D phenone was prepared and distilled, b.p. 81-940 (5 mm.), n3 1. 5215. The above procedure is that described by Cope (157) for the con- densation of ethyl cyanoacetate with aromatic and hindered aliphatic ketones. d) p—Methoxyphenylacetonitrile and p—methoxypropiophenone with ammonium acetate and acetic acid. The procedure used was exactly as described above except that 36. 8 g. (0. 25 mole) of p—methoxyphenyl- acetonitrile and 41. 3 g. (0. 25 mole) of p—methoxypropiophenone were used. The only material rec0vered proved to be a mixture of the un— reacted starting materials . 134 e) Attempted condensation with other Knoevenagel catalysts. Solutions of 0. 1 mole of p—methoxyphenylacetonitrile and 0. 1 mole of p—methoxypropiophenone in 50 m1. of dry benzene were refluxed for 72 hrs. with the given amounts of the following catalysts. 1. 0.1g. B—alanine 2. 0.5 g. fi-alanine 3. 1.0 g. B-alanine 4. 0.5 g. p-aminophenol 5. 0.5 g. a-aminophenylacetate 6. 0. 5 g. u-aminocaproic acid 7. 1.0 g. fused zinc chloride and 1.0 g. aniline 8. 2.0 g. fused zinc chloride and 1.0 g. aniline 9. 2.0 g. anhydrous sodium sulfate and 1.0 g. piperidine 10. 5.0 g. anhydrous sodium sulfate and 1.0 g. piperidine After the reflux period in each case, the reaction mixture was poured into water and the water layer was extracted with benzene. The combined benzene solution was washed successively with dilute sodium bicarbonate solution, with 3N hydrochloric acid, and with saturated sodium chloride solution. It was dried over anhydrous sodium sulfate and the solvent was removed. Distillation gave a mixture of the reactants and in no case was there any appreciable residue. The catalysts used in runs 1 through 6 are those described by Prout (158) for the condensation of ethyl cyanoacetate with acetone. Scheiber and Meisel (159) reported that the fused zinc chloride-aniline system, which was used in runs 7 and 8, was a general catalyst for the condensation of ethyl cyanoacetate with ketones. Cowan and Vogel (160) used anhydrous sodium sulfate and piperidine and considered them to constitute an effective catalyst system for the condensation of ethyl cyanoacetate with aldehydes and ketones. Addition of ethylmagnesium bromide to a-benzoylphenylacetonitrile. A solution of ethylmagnesium bromide (prepared from 2. 6 g. , 0.12 mole of magnesium and 12.0 g., 0.11 mole of ethyl bromide) in 200 m1. of ether was added to a well stirred solution of 22. 1 g. (0.1 mole) of 135 o.—benzoylphenylacetonitrile in 150 ml. of ether over a period of 30 min. By the time this addition had been completed the insoluble complex had formed a hard ball in the flask. The addition of 100 ml. of dry benzene and the accompanying removal of 150 ml. of ether from the refluxing solution increased the temperature sufficiently to dissolve the complex. After the solution had been refluxed for 2 hrs. , the reaction mixture was cooled and was poured over a mixture of 800 g. of ice and 100 ml. of concentrated hydrochloric acid. After the ice had melted the water layer was separated and was extracted twice with 100 ml. portions of ether. The combined organic phase was washed successively with two 100 m1. portions of 6N hydrochloric acid, with one 100 ml. portion of dilute sodium carbonate and once with water. The solution was dried over anhydrous magnesium sulfate, the solvent was removed at 15 mm. pressure and the brown viscous residue distilled. B.p., OC. Fraction (5 mm.) Grams N(sodium fusion) 1 78—84 10.8 (--) 2 85-110 4.5 H 3 110-160 2. 3 (-) residue --— 3 2 (-) Since this product did not contain nitrogen and hence could not be the desired hydroxynitrile, the identity of the product was not investi— gated further. Addition of ethyllithium to o.-benzoylphenylacetonitrile. Forty milliliters of sodium-dried pentane was placed in a 250 ml. flask equipped with a stirrer, a reflux condenser topped with a calcium sulfate drying tube and an addition funnel. To this pentane there was added 2. 0 g. (0. 29 mole) of lithium wire (approx. 0. 5 x 2.0 mm. ), which had been brightened by dipping into pure ethyl bromide. As the lithium—pentane mixture was vigorously stirred 13.0 g. (0.12 mole) of ethyl bromide in 40 ml. of pentane was added over a period of 4 hrs. The flask was heated 136 to the reflux temperature during this addition. . After 2 additional hrs. of heating, the contents of the flask was cooled and the pentane solution was filtered in a closed system into an addition funnel. Most of the lithium metal had reacted. This solution of ethyllithium was added to a well stirred solution of 22.1 g. (0. 1 mole) of a-benzoylphenylacetonitrile in 50 ml. of reflux- ing ether over a period of 2 hrs. After an additional hour of refluxing, the reaction mixture was cooled and was poured over 400 g. of ice con— taining 50 ml. of 6N hydrochloric acid. The water layer was separated and was extracted twice with 50 ml. portions of ether. The combined organic solution was washed once with water and was dried over sodium sulfate. Removal of the solvent at reduced pressure left a brown viscous sirup. The latter was dissolved in 40 ml. of ethanol and allowed to stand at -50 for one week. Filtration led to the recovery of approxi— mately 5 g. of a—benzoylphenylacetonitrile. The filtrate was concen- trated to a dark brown viscous sirup. This residue gave a very faint nitrogen test (sodium fusion) which could have been due to starting material. .Attempted distillation at 0. 001 mm. in a Hickman still was unsuccessful. The preparation for ethyllithium which was used in this work is that described by Bryce-Smith and Turner (161). p—Methoxyphenylac etic acid. a) By the hydrolysis of p-methoxyphenylacetonitrile. A mixture of 100 ml. of 50% sulfuric acid, 50 ml. of acetic acid, and 50 ml. (52. 5 g. , 0. 354 mole) of p-methoxyphenylacetonitrile was placed in a 300 ml. flask equipped with a stirrer, a thermometer, and a reflux condenser. This mixture was gently refluxed with constant sitrring for 1 hr. in an oil bath heated to 1300. The oil bath was then replaced with an ice bath . . . . O and the stirring was continued until the temperature had fallen to 10 . 137 The crude product was collected in a sintered glass funnel, was washed twice with cold water' and was air dried, giving 62.4 g. of damp, light purple solid. This solid was dissolved in 150 ml. of 15% sodium hydroxide, the solution was cooled to 100 and 0. 5 ml. of dimethylsulfate was added in one portion. The solution was now stirred with a magnetic stirrer for one hour at 100. The cold—water bath was then removed and an additional 0. 5 ml. of dimethylsulfate was added and the stirring was continued for 2 hrs. at room temperature. The reaction mixture was heated on the steam bath for one hour to destroy the excess dimethyl— sulfate, after which it was cooled to 100 and was acidified to Congo red with 6N hydrochloric acid. The solid was filtered off and was washed once with cold water to give a colorless solid, which weighed 54. 8 g. (93. 5%) after recrystallization from ethanol. This p—methoxyphenyl— acetic acid melted at 86-870. This procedure is based on that described by Adams and Thal for the preparation of phenyl acetic acid in Organic Syntheses (162). b) From p—methoxyacetophenone by the Willgerodt reaction. A mixture of 220 g. (1.45 moles) of p—methoxyacetophenone, 192 g. (2. 2 moles) of morpholine, and 70. 2 g. (2. 2 g. atoms) of sulfur was refluxed with constant stirring for 14 hrs. The hot reaction mixture was then poured into 500 ml. of warm ethanol. After cooling, this solu— tion deposited light yellow 4-(p-methoxythiobenzoyl)morpholine. The latter was removed by filtration, was washed once with 100 ml. of cold ethanol, and was air dried giving 267 g. of product. The filtrate was concentrated to approximately 150 m1. and was cooled. Filtration gave an additional 34 g. of the morpholide. The combined morpholide was then refluxed for 3 hrs. with 3-1. of a 10% aqueous ethanol (1:1) solution of potassium hydroxide. After cooling, the reaction mixture was acidified to Congo red with cold 6N hydrochloric acid and 142 g. of light brown product was filtered off. Two recrystallizations of the latter from 50% 138 ethanol gave 126 g. (52%) of pure p-methoxyphenylacetic acid, m.p. 86-87.5°. This procedure is a modification of that given by Shirley (163), who obtained a 48% yield of p-methoxyphenylacetic acid, m. p. 86—870. 3—Hydroxy-2, 3-diphenylpentanoic acid. A 3-1. flask was equipped with a stirrer, a parallel adaptor holding two reflux condensers topped with calcium chloride drying tubes, a nitrogen inlet tube and an addition funnel. After the flask was purged with dry nitrogen, 54. 7 g. (2. 2 moles) of magnesium turnings and 250 ml. of dry ether were placed in the flask. To this there was added 172.6 g. (2.1 moles) of isopropyl chloride in 750 m1. of dry ether over a period of 4 hrs. After being refluxed for an additional hour, the Grignard solution was cooled to room temperature and a solution of 136. 1 g. (1.0 mole) of phenylacetic acid in 800 m1. of dry ether was added over a period of 5 hrs. The resulting reaction mixture was then refluxed for an additional 10 hrs. under an atmosphere of dry nitrogen. To the resulting Ivanov reagent a solution of 134. 2 g. (l. 0 mole) of propiophenone in 400 ml. of dry ether was added over a period of 4 hrs. The reaction mixture was refluxed for an additional hour. It was cooled to 50 in an ice bath and was poured onto a mixture of 2 kg. of ice and 500 ml. of concen— trated hydrochloric acid. The product was insoluble in the ether and was filtered from the hydrolysis mixture. The water layer was separated from the ether layer and the latter was washed three times with water. Removal of the ether in a current of air left a residue which was added to the solid collected by filtration. The solid was recrystallized twice from ethanol—ethyl acetate (121). After air drying for two days and after a final drying in a vacuum oven (600 at 5 mm.) for 12 hrs. there was obtained 220.8 g. (82%) of 3—hydroxy-2, 3-dipheny1pentanoic acid, m.p. 179. 5—182O (with softening at 1600). The neutralization equivalent of this acid was found to be 268. 2. The calculated value for C17H1803 is 270. 3. 139 The procedure used for the preparation of this acid is essentially that of Zimmerman and Traxler (151). ,. Attempted dehydration of 3=hydroxy-2, 3-diphenylpentanoic acid. A solution of 10. 0 g. of 3-hydroxy-2, 3-diphenylpentanoic acid and l. 0 g. of sodium acetate in 25 m1. of acetic anhydride was heated on the steam bath for 5 hrs. The reaction mixture was cooled and was poured over 20 g. of ice. Stirring was continued until all of the acetic anhydride had reacted. The solid which separated was dissolved in ether and the ether layer was removed. The water layer was extracted with three 50 ml. portions of ether and the combined ether solution was washed three times with water. Extraction of this ether solution with 3N sodium hydroxide did not remove any acid material. Evaporation of the ether left 6. 2 g. of a nearly colorless oil. a—Ethylstilbene is reported to be a liquid boiling at 300—305° (164). Although the identity of this neutral product was not investigated further, it seems reasonable that u—ethylstilbene could be formed as a non-acidic product by an accompanying decarboxylation of the unsaturated acid. Similar decarboxylations have been reported by other workers (152). 3-Hydroxy-2, 3-di(p-methoxyphenyl)pentanoic acid. To 0. 11 mole of isopropylmagnesium chloride (from 2.43 g. , O. 11 mole of magnesium and 8.63 g., 0.11 mole of isopropyl chloride) in 100 ml. of ether there was added over a period of 2 hrs. , 10. 32 g. (0.055 mole) of sodium p—methoxyphenylacetate as a slurry in 70 m1. of dry ether. The sodium p—methoxyphenylacetate was prepared by the method of Shaw and Warrener (165). After this solution had been refluxed for 4 hrs., the reaction mixture was cooled to 100 and 18.1 g. (0.11 mole) of p-methoxypropio- phenone in 30 ml. of dry ether was added at such a rate that the temperature of the reaction mixture did not exceed 100. This addition required about 30 min. The reaction mixture was allowed to come to room temperature and was then refluxed for 3 hrs. The reaction mixture 140 was cooled to 100 and 200 m1. of saturated ammonium chloride was slowly added with vigorous stirring. Five hundred milliliters of benzene was added to dissolve the product which had separated during the hydrolysis. The water layer was separated and was extracted with three 100 ml. portions of benzene. The combined ether-benzene layer was washed with dilute hydrochloric acid and with saturated sodium chloride and was finally extracted with five 50 m1. portions of 10% sodium carbonate solution. The latter were cooled and were acidified to Congo red with cold 6N hydrochloric acid. The acidic solution was extracted with chloroform. After the chloroform was removed with a Rinco evaporator, 8. 2 g. of light yellow solid was obtained. Recrystallization from methanol produced three crops of solid. The total yield of 3-hydroxy—2, 3-di(p-methoxyphenyl)pentanoic acid was 7. 53 g. (42%). Crop Grams m. p. , c)C. 1 4.75 194-196 2 2.06 192-196 3 0. 72 189- 195 residue . 32 oil Several recrystallizations of the first crop from ethanol gave an analytical sample of 3—hydroxy-2, 3-di(p-methoxyphenyl)pentanoic acid which melted at 194-195. 5°. Analysis: Calculated for C19HZZO5: C, 69.07; H, 6.71. Found: C, 68.96; H, 6.69. The neutralization equivalent of this acid was found to be 328. 1 The calculated value for CHI-12204 is 330.4. Attempted dehydration of 3—hydroxy-2, 3-di(p-methoxyphenyl)- pentanoic acid. The seven procedures described for the dehydration of methyl 3-hydroxy-3, 4—di(p-methoxyphenyl)hexanoate were attempted with 3—hydroxy-2, 3-di(p-methoxyphenyl)pentanoic acid. In these cases the procedures were carried out employing much less material, but the 141 equivalent quantities were the same. In no case was an acidic product isolated, other than the starting Iii—hydroxy acid. Procedure Isolated a. Iodine in refluxing toluene Starting acid b. p=Toluenesulfonic acid in Starting acid r efluxing benz ene c. Phosphorous pentoxide in Starting acid refluxing benzene d. Oxalic acid in refluxing benzene Ci-Ethyl-4, 4“—di(p—me(;choxy)- stilbene, m.p. 84585 . Reported m.p. 85 (132). e. Sulfuric acid Red oil f . Thionyl chloride Red oil g. Acetyl chloride Red oil The red oil which was obtained in procedures e, f and g above distilled at 140-1650 (0. 03 mm.). It retained its brillant red color throughout the distillation. The structure of this compound was not investigated. Attempted preparation of 3-acetoxy-2, 3-di(p-methoxyphenyl)- pentanoic ethanoic anhydride. A solution of 41. 5 g. (0. 25 mole) p—methoxyphenylacetic acid in 500 ml. of dry ether was added over a period of 4 hrs. to 0. 5 mole of isopropylmagnesium chloride (from 14. 6 g., 0.6 mole of magnesium and 43. 3 g., 0. 55 mole, of isopropyl chloride) in 450 m1. of ether. After this solution had been refluxed for 18 hrs. , 41. 1 g. (0. 25 mole) of p—methoxypropiophenone in 500 ml. of dry benzene was added to it over a period of 5 hrs. As the latter was added 400 ml. of the ether was removed with a distillate collector. When all of the p—methoxypropiophenone had been added, the reaction mixture had become a thick, tan slurry. This was refluxed at 510 for 5 hrs. and was then allowed to cool to room temperature. Redistilled acetyl chloride 142 (42. 6 g. , 0. 55 mole) in 100 ml. of dry benzene was added over a period of 1 hr. without external cooling. After being stirred at room tempera- ture for 12 hrs. the viscous yellow—tan mixture was poured over 2 kg. of ice and the mixture acidified to Congo red with dilute hydrochloric I acid. The water layer was separated and was extracted with three 50 ml. portions of benzene, which were combined with the organic layer from the hydrolysis. The bright red ether-benzene solution was washed with water and was extracted with four 50 ml. portions of 10% sodium carbonate solution. The combined light yellow basic extract was extracted with benzene (discarded), was treated with Norite, and was acidified to Congo red with 6N hydrochloric acid. The yellow semi=solid which separated on acidification was extracted into methylene chloride. After this was dried over anhydrous sodium sulfate, the methylene chloride was removed with a Rinco evaporator and left 23. 2 g. of yellow solid. Recrystallization of the latter from 75% methanol gave a first crop of 6. 0 g. , m.p. 174-1890 and a second crop of 15. 3 g. , m.p. 192—195. 50. The melting point range of the first crop was not improved by further recrystallization. The mixed melting point of the material which was obtained in the second crop and pure 3-hydroxy—2, 3-di(p-methoxyphenyl)pentanoic acid, m.p. 194— 195.50, was 193-194.50. Concentration of the red, ether-benzene solution with the Rinco evaporator left a red oil which was similar in appearance to that obtained in the attempted dehydrations of the fi-hydroxy acid with thionyl chloride. Attem ted re aration of 2—(p-methox’ phen l)—3- N,N-bis(2- hydroxyethyl) amino I— p-methoxypropiophenone . a) By the Mannich reaction of deox anisoin and diethanolamine. The viscous sirup obtained by the evaporation of 10. 9 g. (0. 1 mole) of diethanolamine which had been neutralized with 6N hydrochloric acid was dissolved in 50 ml. of absolute ethanol. This solution was added, 143 in one portion, to a slurry of 1. 8 g. (0. 2 mole) of trioxymethylene in 100 ml. of absolute ethanol containing 25.6 g. (0.1 mole) of deoxy- anisoin and 1 drop of concentrated hydrochloric acid. The mixture became homogenous after being heated on the steam bath for 15 min. The heating was continued for 3 hrs. with the addition of two 0. 9 g. (0. 1 mole) portions of trioxymethylene at the end of the first and second hours. The resulting clear solution was concentrated to 100 ml. with a Rinco evaporator and was stored at -100 for 48 hrs. The solid which had separated was recrystallized once from methanol, giving 24. 8 g. (96% recovery) of deoxyanisoin. This is essentially the procedure of Levy and Nisbet (166) for the preparation of 2(3—[N, N=bis(2-hydroxyethyl)amino]propanoyl)furan hydrochloride which was obtained in an unspecified yield. The above reaction was repeated using (a) a twofold excess of diethanolamine hydrochloride, (b) free diethanolamine, (c) aqueous formaldehyde, and (d) anhydrous formaldehyde. Unchanged deoxyanisoin was the only material isolated in every attempt. b) B the addition of diethanolamine to a-meth lenedeox anisoin. A solution of 26. 8 g. (0. 1 mole) of a-methylenedeoxyanisoin and of 10. 9 g. (0. 1 mole) of diethanolamine in 50 ml. of ethanol containing one drop of acetic acid was heated to 400 for 12 hrs. This solution was cooled and was concentrated to 40 ml. The concentrate was cooled at —100 for 5 hrs. One recrystallization of the solid from methanol gave 24.6 g. (95% recovery) of a-methylenedeoxyanisoin. This procedure was successfully employed by Fiesselinann and Ribka (140) for the preparation of Mannich bases from a series of sub— stituted benzoins. Attempted preparation of 2, 3=di(p—methoxyphenyl)—2-pentenylamine hydrochloride. A solution of 0.88 g. (0.0027 mole) of 3, 4-di(p-methoxy- phenyl)—3-pentenoic acid and O. 22 g. (0. 0027 mole) of pyridine in 20 ml. of 144 dry benzene was cooled to 100. As this cold solution was vigorously stirred with a magnetic stirrer, a solution of 2 ml. of freshly distilled thionyl chloride in 10 ml. of benzene was added over a period of 30 min. The resulting light yellow solution was stirred at room temperature for 3 hrs. After cooling to 100 for 1 hr. , the solution was filtered into a dry flask and was concentrated with a Rinco evaporator. The residual thionyl chloride was removed by repeatedly dissolving the sirupy residue in 10 ml. of dry benzene and then by evaporating the latter after each addition. A solution of this acid chloride in 25 m1. of anhydrous acetone was cooled to —50 in an ice—salt bath. Then a cold solution of 1. 25 g. (0. 02 mole) of sodium azide in 9 ml. of water was added and the mixture was stirred for 30 min. at 0 to 100. The ice bath was removed and 50 m1. of water was added to the solution. The stirring was continued for an additional 10 min. The aqueous solution was decanted and the semi—solid azide was dissolved in 50 ml. of benzene. A Dean and Stark water trap was placed on the flask and the solution was refluxed for 12 hrs. Three milliliters of concentrated hydrochloric acid was added to this solution and the refluxing continued for another 4 hrs. . No solid separated from the solution during this period. The solution was concen— trated to 10 ml. with a Rinco evaporator and the solid was filtered from the solution. ,One recrystallization from benzene gave 0. 72 g. of colorless solid, m.p. 176-1780. This solid proved to be soluble in dilute sodium hydroxide. When the basic solution was acidified with dilute nitric acid, free 3, 4—di(p-methoxyphenyl)—3—hexenoic acid separated from solution. . After the acid was filtered off, the filtrate gave a pre- cipitate with silver nitrate. This indicates that the solid isolated may have been the unchanged azide, but the identity of this compound was not investigated. This reaction sequence was repeated using toluene and xylene in place of the benzene to effect the rearrangement. The same solid, m.p. 176-1780, was isolated in both cases. 9) 10) 11) 12) 13) 14) 15) LITERATURE CITED . H. A. Marshall and W. A. Jolly, Phil. Trans. Roy. Soc., . 198, 99 (1905).: 171'?! E. Allen and E. A. Doisy, J. Am. Med. Assn., 81, 819 (1923).. E. Allen and E. A. Doisy, Am. J. Physiol., _6_9, 577 (1924). E. A. Doisy, C. D. Valer and S. Thayer, Am. J. Physiol., 22, 329 (1929). A. Butenandt, Naturwiss., E, 879 (1929). J. A. Hogg and J. Korman, in F. F. Blicke and C. M. Suter, "Medicinal Chemistry, " Vol. II, John Wiley and Sons, New York, 1956, p. 34. A. Horeau, VIIIe Cong. chim. biol., _1_, 147 (1948). C. W. Emmens, J. Endocrinol., 3, 64, 168, 174 (1942). W. E. Bachmann, W. Cole and A. L. Wilds, J. Am. Chem. Soc., Q, 824 (1940). W. E. Bachmann and D. W. Holmes, J. Am. Chem. Soc., E, 2750 (1940). W. E. Bachmann and A. L. Wilds, J. Am. Chem. Soc., E, 2084 (1940). W. E. Bachmann and W. S. Horton, J. Am. Chem. Soc., Q, 58 (1947). A. L. Wilds and L. W. Beck, J. Am. Chem. Soc., g, 1688 (1944). A. L. Wilds and W. J. Close, J. Am. Chem. Soc., 69, 3079 (1947). V. Prelog, R. Ruzicka and L. Wieland, Helv. Chim. Acta, Q, 250 (1945). 145 16) 17) 18) 19) 20) 21) 22) 23) 24) 25) 26) 27) 28) 29) 30) 31) 32) 33) 146 A. Koebner and R- Robinson, J. Chem. Soc., 1994 (1938). B. C. Dodds, L. Goldberg, W. Lawson, and R. Robinson, Nature, 141, 247 (1938). A. Carayon-Gentil and J. Cheymol, Bull. soc. chim. biol., 28, 136 (1946). E. C. Dodds, L. Goldberg, W. Lawson, and R. Robinson., Nature, 142, 34 (1938). J. M. Robson and A. Schonberg, Nature, 150, 22 (1942). J. M. Robson and A. Schonberg, Nature, 140, 196 (1937). J. M. Robson, A. Schonberg and H. A. Fahim, Nature, 142, 292 (1938). E. W. Blanchard, A. E. Stuart and R. C. Tallman, Endocrinol., 32, 307 (1943). K. Miescher, Helv. Chim. Acta, _2_7, 1727 (1944). A. Horeau and J. Jacques, Compt. rend., 2_24_,862 (1947). R. B. Brandbury and D. E. J. White, J. Chem. Soc., 3447 (1951). G. Giacomello and E. Bianchi, Gazz. chim. ital., 7_l, 66l“(1944). F. W. Schueler, Science, 1_03_, 221 (1946). J. Heer and K. Miescher, Helv. Chim. Acta, E, 777 (1947). E. Macovske and J. Georgescu, Bull. sect. sci. acad. roumaine, E, 645 (1946); Chem. Abstr., 4_6, 3522 (1952). H. H. Keasling and F. W. Schueler, J. Am. Pharm. Assn., 32, 87 (1950). M. Oki, J. Chem. Soc. Japan, Pure Chem. Sect., 7_3, 252 (1952); Chem. Abstr., g, 3522 (1952). A. Serini and W. Logemann, Chem. Ber., 71, 186 (1938). 34) 35) 36) 37) 38) 39) 40) 41) 42) 43) 44) 45) 46) 47) 48) 49) 50) 51) 147 G. A. Jeffrey, H. P. Koch and S. C..Nyberg, J. Chem. Soc., 1118 (1948). H. P. Koch, Nature, 161, 309 (1948). U. V. Solmssen and E. Wenis, J. Am. Chem. Soc., 7_0, 4197 1948). A. L. Wilds and R. E. Sutton, J. Org. Chem., :6, 1371 (1951). J. Grundland, Compt. rend., £2_1, 671 (1945). J. Grundland, Compt. rend. soc. biol., 142, 941 (1948). E- K. Rideal and J. H. Schulman, Nature, _144, 100 (1939). E. K. Rideal, Endeavour, :1, 83 (1945). P. Ehrlich, in F. Himmelweil, ed., "Paul Ehrlich, Gesammelte Arbeiten," Vol. I, Pergamon Press, London, 1956, p. 590. A. Gilman and F. S. Philips, Science, 103, 409 (1946). R. B. Ross and P. E. Swartzentruber, "Literature Survey of Nitrogen Mustards, " Cancer Chemotherapy National Service Center, Bethesa 14, Maryland, 1959. S. M. Buckley, C. C. Stock, M. L. Crossley and C. P. Rhoads, Cancer Research, 12, 207 (1950). J. H. Burchenal, M. L. Crossley and C. C. Stock, Arch. Biochem., 2_6, 321 (1950). C. Heidelberger and M. E. Baumann, Cancer Research, H, 277 (1957). W. C. J. Ross, Ann.» N. Y. Acad. Sci., 6_8_, 669 (1958). W. C. J. Ross, J. Chem. Soc., 2257 (1950). W. C. J. Ross, Adv. Cancer Research, 1, 397 (1953). J. L. Everett and G. A. R. Kon, J. Chem. Soc., 3131 (1950). 148 52) G. M. Timmis, 12th Internat. Cong. Pure and Appl. Chem. Abstr., 334 (1951). 53) H. G... Khorana, G. M. Tener, R. S. Wright and J. G. Moffatt, J. Am. Chem. Soc., 19, 430 (1957). 54) H. McCombie and D. Purdie, J. Chem. Soc., 1217 (1935). 55) K. Ward, J. Am. Chem. Soc., 21, 914 (1935). 56) S. Gabriel and G. Eschenbach, Ber., 2, 810 (1897). 57) V. Prelog and V. Stephan, C011. Czechoslov. Chem. Comm., 1, 93 (1935); Chem. Abstr., £23 4013 (1935). 58) C. Golumbic, J. S. Frocton and M. Bergman, J. Org. Chem., 1, 518 (1946). 59) P. D. Bartlett, J. D. Ross and C. D. Swain, J. Am. Chem. Soc., 11, 1415 (1948). 60) A. G. Ogstom, Trans. Faraday Soc., :54, 45 (1948). 61) R. M. Herriott, J. Gen. Physiol., _3_2_, 221 (1948). 62) S. Revell, Heredity, 6(supp1.), 107 (1953). 63) D. M. Needham, Biochem- Soc. Symposia, Cambridge, Eng., 2, 16 (1948). 64) F. S. Philips, J. Pharmacol. Exptl. Therap., 22, 281 (1950). 65) S. Farber, e_t a_._1., Cancer Research, 4, 2 (1956). 66) F. Bergel and J. A. Stock, J. Chem. Soc., 2409 (1954). 67) F. Bergel and J. A. Stock, Brit. Empire Cancer Campaign Rep., 3_1_, 6 (1953). 68) J. M. Luck, Cancer Research, a, 1071 (1957). 69) W. C. J. Ross and G. P. Warwick, Nature, 176, 298 (1955). 70) W. C. J. Ross and G. P. Warwick, J. Chem. Soc., 1364 (1956). 71) 72) 73) 74) 75) 76) 77) 78) 79) 80) 81) 82) 83) 84) 85) 86) 87) 88) 89) 90) 149 W. C. J. Ross and G. P. Warwick, J. Chem. Soc., 1724 (1956). H. Arnold and F. Bourseaux, Angew. chemie, 19, 539 (1958). H. Host and R. Nissen—Meyer, Cancer Chemotherapy Reports, No. 9, 47 (1960). G. A. R. Kon and J. J. Roberts, J. Am. Chem. Soc., 1&9 978 (1950). L. Varghn, L. Toldy and O. Feher, J. Chem. Soc., 805 (1957). H. R. Ing, Trans..Faraday Soc., 3_9, 372 (1943). F. Berge1, Ann. N. Y. Acad. Sci., 6L8, 1238 (1958). L. F..Larinov, Brit. J. Cancer, 1_0, 26 (1956). C. C. Stock, et al., Cancer Research, 11, 432 (1951). R. Jones, e_t 211., Ann. N. Y. Acad. Sci., fl, 1133 (1958).: J. H. Creech, Ann. N. Y. Acad. Sci., 6_8, 868 (1958). E. Hirschberg, A. Gellhorn and W. S. Gump, Cancer Research, g, 432 (1951). D. A. Lyttle and H. Petering, J. Nat. Cancer Inst., 23, 153 (1960). D. A. Lyttle and H. Petering, J. Am. Chem. Soc., 89, 6459 (1960). M. Lane and M. S. Kelly, Cancer Res., 29, 511 (1960). J. D. Evans and G. D. Menyel, Proc. Soc. Exptl. Biol. and Med., _9_9, 620 (1958). W. H. Carothers and R. Adams, J. Am. Chem. Soc., g, 1675 (1924). L. B. Palfray, Bull. soc. chim. France, [2] Z, 407 (1940). D. Davidson and M. T. Bogert, J. Am. Chem. Soc., 51, 905 (1935). T. F. Dankova, L. G- Evodokimova, I. I. Steparov, and N. A. Preobrazhenskii, _1_8, 1724 (1948); Chem. Abstr., Q, 2607 (1949). 91) 92) 93) 94) 95) 96) 97) 98) 99) 100) 101) 102) 103) 104) 105) 106) 107) 108) 150 K. Rorig, J. Am. Chem. Soc., 7_3, 1290 (1951). R. S. Livshits, G. I. Bazilevskaya, M. S. Bainova, O. E. Dobrovinskaya, and N. A. Preobrazhenskii, Zhur. Obshchei Khim. _11, 1671 (1947); Chem. Abstr., 42, 2607 (1948). J. Lee, A. Ziering, L. Berger and S. D. Heineman, Jubilee V01. Emil Barell, 264, (1946); Chem. Abstr., fl, 6252 (1947). K. Rorig, J. D. Johnson, R. W..Hami1ton and T. S. Telinski, Org. Syntheses, 3_6, 50 (1956). J. B. Niederl and A. Ziering, J. Am. Chem. Soc., 6;4_, 885 (1942). E. P. Kohler, Am. Chem. J., g, 386, 399 (1906). S. Wawzonek, J. Am. Chem. Soc., 13, 5746 (1951). S. Wawzonek, J. Am. Chem. Soc., _6_§, 1157 (1946). J. H. Hunter and J. Korman, J. Am.,Chem. Soc., _7_0_, 3424 (1948). R. F. Nystrom, J. Am. Chem. Soc., 3, 2544 (1955). A. F. McKay and H. H. Brownell, J. Org. Chem., 2, 648 (1950). L. Bock and P. deBenneville, Cicular, Rohrn and Haas Co. , Philadelphia, Pa.; E. L. Eliel, in M. S.. Newman, ”Steric Effects in Organic Chemistry, " John Wiley and Sons, New York, 1956, p. 70. S. I. Sergievskaya, K. V. Levshina, A. K. Chizhov, A. I. Gavrilova and A. I. Kravchenko, J. Gen. Chem. (U.S..S.R.), (Eng. Trans.), Q, 1884 (1958). G. G. Hazen, Ph.D. thesis, University of Michigan, 1951,. p. 13. J. H. Burckhalter and J. Sam, J. Am. Chem. Soc., 7_4, 187 (1952). C. R. Hauser, D. L. Lednicer and W. R. Brazen, J. Am. Chem. Soc., 82, 4345 (1958). H. R. Snyder, R. R. Adams and A. V. McIntosh, Jr., J. Am. Chem. Soc., Q, 3280 (1941). R. Paul, Bull. soc. chim. France, [5]_1, 973 (1934). .109) 110) 111) 112) 113) 114) 115) 116) 117) 118) 119) 120) 12m 122) 123) 124) 125) 151 W. G. Dauben and H. L. Bradlow, J. Am. Chem. Soc., E, 559 (1952). K. Rorig, Abstr. of Papers, 119th Meeting, Am. Chem.. Soc., Boston, Mass., Apr. 1-5, 1951, p. 20M. S. Wawzonek, J. Am. Chem. Soc., 13, 5746 (1951). D. L. Cottle, J. Am. Chem. Soc., 6_8, 1380 (1946). F. C. Troper, U.. S. patent 2, 539,679 (1951); Chem. Abstr. , 42,7315(195n. W. Bissinger, U..S. patent 2,529,671 (1951); Chem. Abstr., 4g, 4898(1951L W. J. Genster and G. M. Sherman, J. Org. Chem., _2_3_, 1227 (1958). H. H. Bosshand and H. Zollinger, Angew. Chem., 7_1, 375 (1959). H. H. Bosshand and H. Zollinger, Helv. Chim. Acta, 42, 1653 (1959). P. A. Lasselle and S. A. Sundet, J..Am. Chem. Soc., Q, 2374 (1941). K. B. Wiberg, "Laboratory Technique in Organic Chemistry, " McGraw-Hill, New York, 1960, p. 111° A. J. Vogel, ”Practical Organic Chemistry, “ Longmans, Green and Co., London, 1956, p. 971. W. E. Bachmann and A. L. Wilds, J. Am. Chem. Soc., 62, 824, 2084 (1940). J. Klein and E. D. Bergmann, J. Org. Chem., Q, 1019 (1957). E. Hofstetter and A. E. Smith, Helv. Chim. Acta, 3_6, 1706 (1953L V. Prelog and K. Bozicevic, Ber., 7_2, 1103 (1939). A. J. Vogel, "Practical Organic Chemistry, " Longmanns, Greene and Co., London, 1956, p. 870. 126) 127) 128) 129) 130) 131) 132) 133) 134) 135) 136) 137) 138) 139) 140) 141) 142) 152 H. E. French and A. F. Wirtel, J. Am. Chem. Soc., _4_8_, 1736 (1926). F. R. Duke, Ind. 'Eng. Chem., Ana1.7Ed., 1_7, 196 (1945). W. S. Johnson, R. G. Christiansen and R. E. Ireland, J. Am. Chem. Soc., B, 1995 (1957). R. E. Bowman and W. D. Fordham, J- Chem. Soc., 3945 (1952). K. Hess and H. Frahm, Ber., 71, 2627 (1938). D. A. Foss,. W. Freund, and E. R. Stove, J. Chem. Soc., 5038 (1952). E. C. Dodds, L. Goldberg, W. Lawson and R. Robinson, Proc. Royal Soc. (London) B 127, 140 (1939). I. N. Nazarov and I. L. Kotlyarevskii, Zhur. Obschchei Khim., fl, 1431 (1950); Chem. Abstr., 4_5_, 1963 (1951). M. Bosher, Ber., _1_4, 327 (1881). G. Sumrell, J. I. Stevens, G. E. Goheen, J. Org. Chem., _2_2_, 39 (1957). P. H. Carter, J. C. Craig, R. E. Lack and M. Moyle, Org. Syntheses, :0, 16 (1960). R. H. Martin and A. Stoffyn-Thomas, Bull. soc. chim. Belges, 64, 352 (1955). E. S. Gould, "Mechanism and Structure in Organic Chemistry, " Henry Holt and Co. , New York, 1959, p. 346. M. Silverman and M. T. Bogert, J. Org. Chem., g, 34 (1946). H. Friesselmann and J. Ribka, Chem. Ber., 89, 27 (1956). S. Natelson and S. Gottfried, Org. Syntheses, C011. Vol. III, 381 (1955). R. 0. Clinton and S. E. Lasowski, J. Am. Chem. Soc., 12, 3153 (1948). ’ 143) 144) 145) 146) 147) 148) 149) 150) 151) 152) 153) 154) 155) 156) 157) 158) 159) 153 L. F. Fieser and W. S. Johnson, J. Am. Chem. Soc., Q, 575 (1940). W. E. Bachman and A. L. Wilds, J. Am. Chem. Soc., E, 2086 (1940). K. Brand and O. Loehr, J. prakt. Chem., [2]109, 365 (1925). K. Rorig, J. Am. Chem. Soc., 13, 1290 (1951). E. E. Royals, "Advanced Organic Chemistry, " Prentice Hall, Inc., New York, 1954, p. 787. D. Ivanoff and A. Spassoff, Bull. soc. chim. France, [4] 4_9, 371 (1931). D. Ivanoff and J. Popoff, Bull. soc. chim.. France, [4] fl, 1547 (1931). F. F. Blicke, U. S. patent 2,922,795 (1960). H. Zimmerman and M. D. Traxler, J. Am. Chem. Soc., 7_9, 1920 (1957). B. H. Alexander and W. F. Barthel, J. Org. Chem., Q, 389 (1958). F. F. Blicke, Org. Reactions,_I, 303 (1942). R. C. Noller and R..Adams, J. Am. Chem. Soc., :4_6_, 1892 (1924). R. L. Jacobs, M. S. thesis, Michigan State University, 1955, p. 19. K. Rorig, U. S. patent 2,745,866 (1956); Chem. Abstr., fl, 1272 (1956). A. C. Cope, C. M. Hofmann, C. Wyckoff and E. Hardenbergh, J. Am. Chem. Soc., 63, 3452 (1941). F. S. Prout, J. Org. Chem., _l_8_, 928 (1953). J., Scheiber and F. Meisel, Ber. , 48, 259 (1915). 160) 161) 162) 163) 164) 165) 166) 154 D. M. Cowan and A. I. Vogel, J..Chem. Soc., 1528 (1940). D. Bryce-Smith and E. E. Turner, J. Chem. Soc., 861 (1953). R. Adams and A. F. Thal, Org. Syntheses, C011. Vol. I_, 436 (1941). D. A. Shirley, ”Preparation of Organic Intermediates, “ John Wiley and Sons, New York, 1951, p. 192. J.. Levy and P. Gallis, Bull. soc. chim., France, [4] Q, 862 (1928). G. Shaw and R. N. Warrener, J. Chem. Soc., 153 (1958). G. A. Levy and A. B..Nisbet, J. Chem. Soc., 1053 (1938). ““1111111111111111))1l1111111l1“