[COXTRIBUTIONFROM THE DEPARTMENTS OF CHEMISTRY AND P a T H O L O G Y O F THE UNIVERSITY OF CHICAGO]
6'-METHOXYRUBANOL-9' MORTON KLEIMAN2
AND
SIDNEY WEINHOUSES
Received August 18, 1 9 4
In connection with some of our work on the synthesis of antimalarials, we became interested in the preparation of the analogs of the quinine alkaloids which lack the vinyl substituent in the 3 position (Formula I, R = H). Since minor changes in the quinine molecule which involve this substituent do not
H
I appreciably affect the chemotherapeutic activity (l), it seemed likely that a compound having the same fundamental structure but lacking the vinyl side chain might possess antimalarial properties. The vinyl-free compound, 6'-methoxyrubanol-9, can exist in four stereoisomeric, optically active forms which may be designated (++), (- -), (+-), and (- +),and in two optically inactive racemic mixtures, (++) ( - -), and (+-) (-+). The four optically active isomers were first prepared by Rabe and coworkers (2), and were reported by these investigators to be therapeutically inactive (3). However, the (++) (- -) racemic mixture4 prepared by Prelog et aE. (4)was reported to be fully as active as quinine itself (5). The verification of one or the other of these apparently conflicting claims was the object of the present investigation. The synthesis of the desired compound involves the following series of reactions: 1
This paper was submitted t o the Committee on Medical Research June 17, 19-14, but
was not released for publication until July 30, 1945. 2 Semour Coman Fellow in Medical Chemistry a t the University of Chicago, 1942-1943. Present address: The Velsicol Corporation, Chicago, Illinois. a Present address : Catalytic Development Corporation, Marcus Hook, Pennsylvania. 4 While Prelog did not designate his diastereoisomer as the (++)(--) one, i t was undoubtedly this isomer with which he was dealing, since the (++)(- -) racemate of Rabe (9) was also active, and the dihydrochlorides of each melted a t 239-240" and ca. 242", respectively.
562
6'-METHOXYRUBANOL-g
563
I. Synthesis of ethyl N-benzoyl-@-(4-piperidy1)propionate: A. Chloral hydrate HnSO4 ,Chloral. B. Chloral y-Picoline 4 y-Chloral-picoline [1-(4-pyridyl)-2-hydroxy3,3,3-trichloropropane]. C. y-Chloral-picoline KOH in EtOH, p-(4-Pyridyl)acrylic acid. D. p- (4-Pyridy1)acrylic acid Hz ,p-(4-Piperidyl)propionic acid. E. p-(4-Piperidyl)propionic acid Et OH dry HCl+ Ethyl @-(4-piperidyl)propionate (6). F. Ethyl p-(4-piperidyl)propionate Benzoyl chloride_, Ethyl N-benzoyl-@-(4piperidy1)propionate (6).
+
+
11. Synthesis of ethyl quininate (7): A. p-Anisidine Acetoacetic ester+ Acetoacet-p-anisidide. B. Acetoacet-p-anisidide HBOt ,6-Methoxy-4-methylcarbostyril. C. 6-Methoxy-4-methylcarbostyril POCls , 6-Methoxy-4-methyl-2-chloroquinoline. D, 6-Methoxy-4-methyl-2-chloroquinoline Hydrogenation , 6-Methoxylepidine. E. 6-Methoxylepidine CsHsCHO ,4-Styryl-6-methoxyquinoline. F. 4-Styryl-6-methoxyquinoline KMnOi Quininic acid. G. Quininic acid EtOH HzSO~+ Ethyl quininate.
+
,
111. Condensation of ethyl N-benzoyl-p-(4-piperidyl)propionatewith ethyl quininate, and the subsequent steps completing the synthesis: A. Claisen type condensation of the two esters, and hydrolysis. B. Bromination of the @-(4-piperidyl)ethyl6-methoxy-4-quinolylketone produced in step A above. C. Ring closure by dehydrobromination of the bromo ketone hydrobromide produced in step B above. D. Hydrogenation of the 6'-methoxy-9-ketoruban obtained in step C. E. Isolation of the two racemic mixtures of 6'-methoxyrubanol-9 as the dihydrochlorides. This approach to the synthesis of the quinine structure was first proposed in 1919 by Rabe (8) and was subsequently applied to the preparation of 6'-methoxyrubanol-9. We have repeated the synthesis of this compound and have devised alternative procedures in the several instances in which we found the original methods to be unsatisfactory. We have obtained two products: the dihydrochloride of the (++) (- -) racemic pair, and the dihydrochloride of the (4- -) (- +) racemic pair. The former compound melted a t 238-239', a value which is in close agreement with the melting points reported by Prelog (4) (239-240') and by &be (9) (ca. 242'). The latter compound melted unsharply, sintering a t ca. 110" and melting a t ca. 140" (decomp.); no literature melting point for this salt is available for comparison. It is likely that the failure of the (+-) (- +) dihydrochloride to melt sharply is due to its extreme hygroscopicity, for it deliquesces very rapidly upon exposure to the atmosphere. The two diastereoisomeric racemic mixtures, as the crystalline dihydro-
564
MORTON KLEIMAN AND SIDNEY WEINHOUSE
chlorides, were dissolved in physiological saline solution and submitted for antimalarial tests. Preliminary results5 indicate that the (++)(- -) racemate is one-fourth to one-half as active as quinine. The other racemic mixture was found t o be inactive when administered a t the same level. These results are in good agreement with those reported by Prelog, for a quantitative comparison of the antimalarial activities obtained in each instance is not significant when the tests are conducted upon different test animals or by different methods. These conclusions are further confirmed in a recent publication by Rabe (9) which appeared while our manuscript was in preparation. Here, Rabe corroborates the activity of the (++)(- -) racemate, a t the same time reiterating the former claim that each of the four optically active isomers is therapeutically inactive when tested individually. This seems anomalous, for it would be expected that the (- -) form would be comparable in activity to quinine, the (++) form to quinidine, and the (++) (- -) racemate to a mixture of quinine and quinidine. Further investigation of this apparent anomaly had been planned but could not be realized before the interruption of this work became necessary. EXPERIMENTAL
Chloral (Z-A). Chloral hydrate (400 g.) and concentrated sulfuric acid (200 cc.) were mixed in a l-liter Erlenmeyer flask and the mixture heated to ca. 30" on the steam-bath. The layer of chloral was separated and washed twice with fresh acid, using 50-cc. portions for each wash. The chloral thus obtained was distilled at 35 mm. pressure in an all-glass apparatus, b.p. 26". The yield of pure chloral was 316 g. (89%). The chloral was stabilized with a trace of hydroquinone and was found to keep in the dark without appreciable decomposition or polymerization. l-(.&Pyridyl)-I-hydrozy-S,S,J-trichloropropane( I - B ) . Attempts to effect the condensation of chloral and 7-picoline as described by Rabe were entirely unsatisfactory; instead of the reported 6% yields, only traces of the desired product were obtained. The method of Alberts and Bachman (10) was also found t o be unsuitable for the preparation of significant quantities of the 7-chloral-picoline. The following procedure, evolved after a comprehensive series of experiments, gave the desired product in good yield and purity: 7-picoline (62 cc., 0.64 mole) was placed in a l-liter Erlenmeyer flask and mixed with 3 g. of decolorizing carbon (Nuchar) and 3 g. of "Filter-Cel" (the use of these adsorbents greatly facilitated the subsequent solution of the reaction product in the dilute hydrochloric acid) ; chloral (58.5 cc., 0.6 mole) was now added slowly t o the vigorously agitated suspension. The flask was securely stoppered with a cork and kept in an oven at 40" for 4 days, then at 70" for one day. During this time, the white, crystalline addition product first obtained was transformed into a dark, tarry mass. The latter was taken up in a mixture of 250 cc. of water and 75 cc. of concentrated hydrochloric acid, digested for 2 hours on the steam-bath t o effect complete solution, then boiled for 5 minutes and filtered. Upon neutralization of the resultant black solution with concentrated aqueous sodium carbonate, the crude reaction product was precipitated. This was filtered with suction, washed with water, and dried at 70". The crude material was crystallized from 400 cc. of boiling dibutyl ether (ethyl acetate or benzene may be used instead), using decolorizing carbon; the insoluble residue and carbon which remained behind yielded further quantities of product by repeats We gratefully acknowledge the cooperation of Professor W. H. Taliaferro and Mr. William Cantrell, Department of Parasitology, University of Chicago, who tested these compounds as part of work done under a contract recommended by the Committee on Medical Research between the Office of Scientific Research and Development and the University of Chicago.
6'-METHOXYRUBANOL-g
565
ing the crystallization procedure with the mother liquor of the crystalline material already obtained. I n this manner there was obtained 95-100 g. of slightly yellowish, cryatalline l-(4-pyridyl)-2-hydroxy-3,3,3-trichloropropane, m.p. 162-164" (65-7Gy0 yield, based upon the chloral used). This product is satisfactory for the subsequent hydrolysis without further purification. fl-(4-Pyridyl)acryZicacid (I-C). The procedure used for the hydrolysis of y-chloralpicoline was essentially that of Rabe and Kindler (8) ; attempts to effect the desired hydrolysis in aqueous solution, in alcohol which was not entirely free from water, or inalcoholic sodium ethoxide, gave much lower yields. y-Chloral-picoline (96 g.) was dissolved in 700 cc. of hot, absolute ethanol [dried by the method of Smith (11) and Manske (12)] in a 3-liter round-bottom flask equipped with a n efficient, large capacity reflux condenser, then cooled t o about 20". T o this was added a cold solution of potassium hydroxide (135 9.) in 700 cc. of absolute ethanol; the solutions were mixed by swirling, then warmed to ca. 40" or 50". I n the vigorous reaction which followed, the solution boiled violently and potassium chloride was precipitated. I n several experiments, i t was found necessary t o cool the reaction mixture in an ice-water bath during this stage of the reaction t,o avoid loss of mat>erialthrough the condenser; the use of a 3-neck flask equipped with three condensers helped t o avoid this difficulty. After the initial reaction had subsided, the mixture was heated at 60" for 2 hours, filtered, and the filtrate distilled to dryness at reduced pressure. The solid residue was taken up in 400 cc. of water, the solution heated to boiling, acidified t o incipient precipitation with acetic acid, boiled with decolorizing carbon, and filtered. The hot filtrate was acidified with acetic acid and cooled. The precipitate of fl-(4-pyridyl) acrylic acid was filtered, washed with cold water, and air-dried. The crude, brown product thus obtained (44.1 g., 74% yield, m.p. 296" (decomp.), was contaminated with a tenacious impurity which poisoned the catalyst upon attempted hydrogenation, necessitating thorough purification before the reduction could be achieved catalytically. Recrystallization of the crude product from various solvents and solvent mixtures proved unsatisfactory. Likewise, solution and reprecipitation from either dilute hydrochloric acid or dilute sodium hydroxide was not fully satisfactory. Good purification of the crude acrylic acid was finally achieved by dissolving the material in the minimum amount of hot, very dilute ammonium hydroxide, boiling with decolorizing carbon, and reprecipitating with acetic acid. The white product thus obtained was pure and Satisfactory for hydrogenation; over 90% of pure material was recovered. &(4-PiperidyZ)propionic acid (I-D). The pure pyridylacrylic acid (14.9 g., 0.1 mole) was dissolved in a mixture of 83 cc. of water and 17 cc. of concentrated hydrochloric acid and hydrogenated a t 60" and 3 atmospheres pressure, using 0.3 g. of platinum oxide catalyst. The hydrogenation was complete after 6 hours and the theoretical amount of hydrogen was consumed. After filtration and distillation of the filtrate to dryness under reduced pressure, the pure hydrochloride of fl-(4-piperidyl)propionic acid, m.p. 241-242", was obtained in quantitative yield. The compound may be crystallized from absolute ethanol. Ethyl N-benzoyl-@-(4-piperidyl)propionate ( I - E ,F). Piperidylpropionic acid hydrochloride (22.6 g., 0.12 mole) was dissolved in 400 cc. of absolute ethanol (11,12) containing 11 g. of dry hydrogen chloride, and boiled under reflux for 4 hours. The solvent and acid were then distilled off in a bath a t 1 0 0 - l l O o , and the last traces removed with suction in a n anhydrous system. The white residue of the solid ester hydrochloride was dissolved in 80 cc. of cold chloroform (if the esterification has been complete, the material is completely soluble) and the solution transferred t o a 500-cc. 3-neck round-bottom flask equipped with a sealed stirrer, a reflux condenser, and a dropping-funnel. After diluting the solution t o 200 cc. with chloroform, 96 g. of anhydrous potassium carbonate and 10 cc. of water were added; the mixture was vigorously stirred and a solution of 16.8 g. of benzoyl chloride in a n equal volume of chloroform was immediately added dropwise at a rate sufficient t o keep the mixture refluxing gently. After the addition of the acid chloride had been completed, the mixture was stirred and heated for another one-half hour in a bath at 80", To ensure com-
566
MORTON KLEIMAN AND SIDNEY WEINHOUSE
plete benzoylation, a further quantity of benzoyl chloride (2 9.) in an equal volume of chloroform was slowly added while stirring and refluxing the mixture for another half-hour period. The reaction mixture was then cooled, filtered, and the solid washed with chloroform. The combined chloroform solution was dried over anhydrous potassium carbonate. After removal of the solvent by distillation, the crude product-a light yellow oil-was dis(25.6 g., 75.7% yield) tilled at 1mm. pressure. Ethyl N-benzoyl-~-(4-piperidyl)propionate was obtained as a light yellow oil, b.p. 184-185"/1 ma. Ethyl quininate (11). The quininic ester was prepared essentially by methods already described in the literature (7), with slight modifications in some instances; the dechlorination of 6-methoxy-2-chlorolepidineis given in some detail, however, since our procedure for carrying out this reaction is new. 6-Methoxylepidine (11-D) . The reduction of 6-methoxy-2-chlorolepidinet o 6-methoxylepidine proceeded smoothly with Raney nickel in the presence of alkali at 3 t o 4 atmospheres hydrogen pressure. The reaction was carried out in batches of one-tenth mole. A solution of 20.8 g. of 6-methoxy-2-chlorolepidine and 8 g. of potassium hydroxide in 200 cc. of 95% ethanol was shaken with hydrogen a t a pressure of about 50 lbs./sq. in., using about 3 cc. of a Raney nickel suspension in methylcyclohexane. The temperature was kept at 60-70" during the reduction; the requisite amount of hydrogen was consumed in 3 to 4 hours. A t the end of this time, the catalyst was allowed t o settle, and the solution was decanted through a filter into a distillation flask. The same catalyst was reused two or three times without appreciably affecting the rate of reduction. The combined solution from several such batch reductions was distilled t o a small volume and poured into ca. 500 cc. of cold water. The oil thus obtained soon solidified t o a solid melting at 40-50'; i t was purified by vacuum distillation, b.p. 145-150"/18 mm. A colorless solid, m.p. 52", was obtained. The yield of crude, dry 6-methoxylepidine averaged about 95%. &(.4'-Piperidyl)ethyl 6-methoxy-4-quinolyl ketone (111-A). The condensation of ethyl N-benzoyl-& (4-piperidy1)propionate with ethyl quininate was carried out under rigorously anhydrous conditions. D r y sodium ethoxide (0.211 mole), prepared in the usual manner, was suspended in 200 cc. of dry benzene in a 3-neck round-bottom flask equipped with a sealed stirrer, a dropping-funnel, and a reflux condenser carrying a drying tube. A solution in 50 cc. of dry benzene of the freshly distilled esters, ethyl N-benzoyl-&(4-piperidyl) propionate (25.6 g., 0.088 mole), and ethyl quininate (40.5 g., 0.176 mole), was now added, and the vigorously stirred mixture was heated at 80" for one hour. At the end of this time, the apparatus was arranged for distillation and the benzene, together with the ethyl alcohol formed in the reaction were distilled from the mixture while stirring and heating at 100". By evacuation of the apparatus while heating for an additional 2 hours, the last traces of solvent and alcohol were removed. The reaction mixture was cooled, ice-water and ether were added, and after complete solution had been obtained, the brown aqueous layer was separated and extracted twice with ether. The aqueous layer was acidified with concentrated hydrochloric acid until the precipitate first formed had all dissolved. The solution thus obtained was mixed with a n equal volume of concentrated hydrochloric acid and refluxed for 5 hours. Benzoic acid crystallized from the solution upon cooling, and was separated by filtration. The filtrate was cooled,, covered with ether, and made strongly alkaline with concentrated, aqueous sodium hydroxide. A small amount of black, tarry material was removed, and the solution extracted exhaustively with ether. The combined ether extracts were dried over anhydrous potassium carbonate, filtered, and the ether removed by distillation. The dark, reddish brown methoxyrubatoxanone thus obtained used). weighed 18.7 g. (71% yield, based upon the ethyl N-benzoyl-p-(4-piperidyl)propionate The aqueous, alkaline solution remaining after removal of the toxanone was made almost neutral with hydrochloric acid and acidified weakly with acetic acid. In this manner, 16.6 g. of quininic acid was recovered. 6'-Methosy-g-kelo7uban (1114'). The rubatoxanone (4.0 g., 0.013 mole) was dissolved in
6'-METHOXYRUBANOL-g
567
20 CC. of 40% hydrobromic acid, heated at 70°, and a solution of bromine (2.35 g., 0.015 mole) in 30 cc. of 40% hydrobromic acid added a t the rate of one drop every two seconds. The resulting solution, after filtering from a minute amount of tar, was distilled to dryness i n oacuo (bath temperature at 50-60'). The residue, a brilliant, pale yellow, crystalline solid, was transferred with a very small amount of water t o a 500-cc. round-bottom 3-neck flask equipped with a sealed stirrer, reflux condenser, and dropping-funnel. Ether was added, the mixture cooled in a n ice-water bath, and, with vigorous stirring, 100 cc. of 5% aqueous sodium carbonate was slowly added. This was followed by the slow addition of 25 cc. of 1 N aqueous sodium hydroxide. Finally, the reaction mixture was allowed to come t o room temperature while stirring 20 minutes longer. The ether layer was now separated, and the aqueous phase extracted exhaustively with ether. After drying the combined extracts over anhydrous potassium carbonate and removal of the ether, there remained a viscous, dark brown oil. The crude product was dissolved in a small volume of 95% ethanol; dibutyl ether was added t o the solution to incipient precipitation, the solution was boiled with Nuchar decolorizing carbon, filtered, and diluted further with dibutyl ether. Upon cooling, a light yellow solid (0.25 g.), m.p. 270" (decomp.), was obtained. This substance corresponds with the product. reported by Prelog et al. (4) t o be 6'-methoxyd-(?) -bromorubatoxanone-9. Impure 6'-methoxy-9-ketoruban was recovered as a brown, viscous oil from the filtrate upon removal of the solvents under reduced pressure. The product desired was purified by Rabe through a series of extractions with ligroin followed by treatment with decolorizing carbon, and recrystallization of the material separating at this point from ether. The melting point 89" was reported for the pure material. This method of purification did not give satisfactory results with the small portion of the impure product at hand. Consequently, the purification was effected as follows: the picrate was prepared by treatment of a n alcoholic solution of the material with a saturated alcoholic solution of picric acid. The impure picrate thus obtained could not be recrystallized satisfactorily a t this stage; hence i t was triturated with hot acetone, and filtered. By dissolving the resultant, partially purified picrate in nitrobenzene and extracting this solution with dilute hydrochloric acid, the original ketone was recovered in the aqueous phase. The latter was extracted thoroughly with ether to remove nitrobenzene and picric acid, was made alkaline with a concentrated sodium hydroxide solution, and the free rubanone now recovered by extraction with ether. Since the product still could not be crystallized, it was again purified oia the picrate and recovered by the same procedure given above. The picrate which was obtained the second time crystallized well from 95% ethanol, and melted sharply at 211", as reported by Prelog. The recovered ketone (0.695 g.), however, could be obtained only as a light yellow, viscous glass. B'-Methozyrubanol-9 (IZZ-D, E). The methoxyrubanone (0.695 g.) was dissolved i n absolute ethanol (11,12) and hydrogenated a t room temperature at 25 Ibs./sq. in. pressure, using platinum oxide catalyst. The requisite amount of hydrogen was absorbed in ea. 20 minutes. The alcoholic solution of the four stereoisomeric methoxyrubanols obtained upon filtration was treated with a slight excess of dry hydrogen chloride and concentrated at a low temperature under reduced pressure. The solution was cooled in ice, and the white, crystalline substance which separated (product A) was filtered and washed with ice-cold absolute ethanol. The filtrate and washings were combined and again vacuum distilled a t low temperature. The residue was a slightly brownish, extremely hygroscopic, crystalline material (product B). Product A (0.341 g.) melted at 238-239'; i t is the (++)(--) racemic mixture, corresponding to the dihydrochloride for which Prelog reported the melting point 239-240", and Rabe, the melting point ca. 242'. Product B (0.345 g.), the hydrochloride of the other racemic pair, did not melt sharply; it sintered at ca. 110" and decomposed slowly as the temperature was raised to 140".
568
MORTON KLEIMAN AND SIDNEY WEINHOUSE ANTIMALARIAL TESTS
For testing purposes, solutions of products A and B in 0.7% aqueous sodium chloride were prepared; each contained the equivalent of 20 mg. of free alkaloid per cc. of solution (250 mg. of the dihydrochloride in 10 cc. of saline solution). The drugs were tested against blood-induced infections of Plasmodium gallinaceum in chickens. The infections were initiated by an intravenous dose of one million parasites. In the first series of tests, the drugs were given a t doses of 10 mg. per kg. each day, beginning one day after infection and continuing until six doses had been given. At this dosage, no activity was detected. A second series of tests was carried out in which a daily intravenous dose of 20 mg. per kg. was used until six doses had been given. As indicated in Table I, subTABLE I ANTIMALARIAL TESTS PARASITES PER 10,oClLl RED BLOOD CELLS’ DRUG
DOSE, YG./KILO
1
4th day
5th day
6th day
Quinineb
20 10 5
4 16 98
4 27 378
2 51 2265
Substance Ab
20
27
77
100
Substance Bb
20
244
1256
3591
Untreated controls, Group I
223
2231
3887
Untreated controls, Group I1
352
1952
3304
a
b
Mean of five determinations. Dihydrochloride; dose calculated as free alkaloid.
stance B did not suppress the infection as compared t o the untreated controls Substance A, however, suppressed the infection to a less extent than 10 mg. of quinine per kg., but to a greater extent than 5 mg. of quinine per kg. From these preliminary tests, it is concluded that the antimalarial activity of substance A is between one-fourth and one-half that of quinine in the chicken. SUMMARY
1. The preparation of 6’-methoxyrubanol-9 has been reinvestigated and improvements have been made in the synthesis. 2. The statement in the literature that this compound (as the (++) (- -) racemic mixture) possesses anti-malarial activity has been confirmed; preliminary tests indicate that the activity is between one-fourth and one-half that of quinine.
6’-METHOXYRUBANOL-9
569
REFERENCES
(1) HENRY,T. A., “The Plant Alkaloids,” 3d Edition, Blakiston’s Sons and Co., Inc., Philadelphia, 1939, page 442. (2) RABEAND HAGEN,Ber., 74, 636 (1941). (3) RABE,HAQEN,AND Voss, Naturwiss., 29, 44 (1941). SEIWERTH, HEIMBACH-JUH~SZ, AND STERN, Ber., 74, 647 (1941). (4) PRELOQ, (5) PRELOG, STERN,SEIWERTH, AND HEIMBACH-JUH~SZ, Naturwiss., 28, 750 (1940). G.,Dissertation, Hamburg, 1930: page 73. (6) VOLGER, (7) AINLEYAND KING,Proc. Roy. Soe. (London), Bl26, 83 (1938). (8) RABEAND KINDLER, Ber., 62, 1842 (1919). Ber., 76, 318 (1943). (9) RABEAND SCHULER, (10) ALBERTS AND BACHMAN, J. A m . Chem. SOC.,67, 1285 (1935). (11) SMITH,J . Chem. SOC.,1288 (1927). (12) MANSKE,J . A m . Chem. SOC.,63, 1106 (1931).
