2390
G. E. MCCASLAND, D. S. TARBELL, R. B. CARLINAND NANCY SHAKESPEARE Vol. 68 TABLE I PROPERTIES OF BRIDGED COMPOUNDS
IV IV IV
IV VI VI VI1 VI1 VI1 XI1 XI1 XVII 0
No. R , R' = CHa R = CHa, R' = CsHi R = CH3, R' = 4-CH:OOH4 R = n-CrHn, RZ CsHi R, R' = CHI R = CHs, R' = CsHi R . R' = CHI R = CHI, R ' = CaHs R := CHa, R' = 4-CHsOCeH4 R = CHa R .= CaHs
-
a From xylene. From toluene.
Substance Methylated carbinol Phenylated carbinol Anisylated carbinol Phenylated carbinol Methylated chloride Phenylated chloride Methylated acetate Phenylated acetate Anisylated acetate Methylated carbinol Phenylated carbinol Anhydride
From n-octane.
e
M. P., OC. 286-288" 229) 258O 180-18 1 235 198-1994 199' 281' 2700 206 223 2 10- 2 12a
From butanol.
* One addition. * Insoluble in reagent.
C 87.4 88.5 86.0 88.0
H 6.3 6.0 6.0 7.3
-
Analyses, % -Found-
-Calcd.-
7
Mol. C1 wt.
C
600 8 6 . 1 87.6 6 . 8 527 6 . 0 588 85.0 6 . 2 550 8 5 . 1 86.3 5 . 9 86.7 84.3 5 . 9 84.3 87.3 6.7 536 8 7 . 0 88.3 6.4 88.2 77.2 5 . 6 77.1
From chloroform-methanol.
Mol. Act. wt. H 2 2 6.3 640e 2 2 7.2 6 . 6 569' 1 6 . 0 50V Oh 6.2 543' 5.8 6.2 6.5 543' l i 6.4 5.6
H 87.5 6 . 5
In benzene.
CI
From cymene.
338; act. H, 1; addn., 0. Found: C, 88.7; H, 6.5; act. H, 1; addn., 0.25; mol. wt., 304 (in CeHs). It was oxidized by chromic acid and the benzophenone 1-pMethoxyphenyl-2,S-dimethyl-3,4-diphenylcyclopen- formed was identified in the usual way.' tadienol, XI, resulted when p-methoxyphenylmagneWhen the phenylated carbinol, (XII, R =E CsHs),.wv sium bromide reacted with the dimer, IX, by the proce- distilled, there was some decomposition. The solid distildure just described; it crystallized well from benzene, late was triturated with methanol, followed by extractoluoene and butanol; and has the melting point of 193tion with hot ligroin (b. p. 90-100"). The latter removed 194 . I t gives a bluish-violet solution, tinged with red, the enol, X, which was identified by a mixed melting point. when dissolved in concentrated sulfuric acid. The insoluble portion was crystallized from benzeneAnal. Calcd. for C ~ ~ H H OC,~ 84.7; : H, 6.5; act. H , 1. ligroin, and identified *by comparison with an authentic specimen of the bimolecular product, IX, and also with Found: C,81.9; H,6.5; act.H, 1. The anhydride, XVII, of 3,6-dimethyl-4,6-diphenyl-7-p-its 2,4-dinitrophenylhydrazone. The dimer, IX, was also methoxyphenyl- 7 - hydroxy-3,6-methano- 1,2,3,6-tetrahy- identified after the other carbinols in the series (XII) had heated, but the remainder of the material failed to drobenzene-1,2-dicarboxylicacid was formed by the addi- been tion of maleic anhydride to the carbinol, XI, equal weights crystallize. Summary of the components being heated for two hours a t 200". I t was also obtained by two hours' refluxing a benzene soluAll three types of bimolecular products formed tion of the components; any acidic material was extracted by shaking with aqueous sodium carbonate. The proper- by the dehydration of anhydroacetonebenzils under acidic conditions have been compared with ties are given in Table I. The product from the action of phenylmagnesium bro- respect to their behavior with Grignard reagents. mide upon the dimer, IX, was distilled in vucuo, b. p. They fall into three groups, each having charac197-201 ' (2 mm.). After recrystallization from i-propyl alcohol or heptane, the white 2,5-dimethyl-3,3,4-tri- teristic properties. The structures which are postulated for the orphenylcyclopentadienol-1,X, has a melting point of 137138'; nothing else could be obtained from the yellow solu- ganomagnesium complex of each group appear to tion. The enol gives a yellow solution, with a slight green- account satisfactorily for the behavior. ish fluorescence, in concentrated sulfuric acid. Anal. Calcd. for Cz5HaO: C, 88.7; H , 6.6; mol. wt., ROCHESTER 4, NEWYORK RECEIVED MAY17, 1946 eighteen hours. The products were handled by suitable operations. The properties of the two bridged -~ products are given in Table I.
[CONTRIBUTION FROM THE
DEPARTMENT OF CHEMISTRY OF THEUNIVERSITY
OF ROCHESTER]
Analogs of Pyridoxine. I. Some Hydroxymethylpyrimidine Derivatives' BY
G. E. MCCASLAND,~ D. S T A N L E Y
TARBELL,
The synthesis of analogs of pyridoxine (I) for testing as antimalarials was suggested by the work of Seeler16who found that large doses of pyridoxine counteracted the effects of atebrine or quinine on (1) The early work on this problem was conducted under a contract recommended by the Committee on Medical Research between the O 5 c e of Scientific Research and Development and the University of Rochester. (2) Present address: Department of Chemistry, University of Illinois. (3) Present address: Department of Chemistry, Carnegie Institute of Technology. (4) Present address: American Cyanamid Company, Stamford, Connecticut. (5) Seeler. Proc. Soc. Ex$. Biol. M e d . , 57, 113 (1944).
R. B.
CARLIN' AND
NANCY SHAKESPEARE*
P.Zophurae in ducks. This might mean that pyridoxine was necessary for growth of the parasites, and hence that its action might be antagonized by a compound differing slightly from it in structure. Accordingly, we attempted the preparation of compounds in which the characteristic groups of pyridoxine are attached to a different heterocyclic rings than the pyridine ring. There are seven different ways in which the four characteristic groups of pyridoxine might be attached to the four carbon atoms in the pyrimidine ring. The most promising route for any of (6) Work on thiazole analogs is now in progress in this Laboratory.
SOME HYDROXYMETHYLPYRIMIDINE DERIVATIVES
Nov., 1946 CH:!OH
OH
I
I
xT-RI CHs-&j--RP
CHa
I
11, R1
Rz = -CHzOH V, RI = -H, R2 = -CHzOH OH (hydrochloride) i TI,Rt = -H, R? = -CHzOCzHs /\ VII,RI = -Br, RP = -CH20C2Hb -" VIII,R1 = -CN, Rz = -CH20CeHs XI, Diethyl ether of V (picrate)
HOCHAJ-R~
111, R1 = --CHa, R2 = -CH*OH IV,R1 = --CH*OH, R1 = -CHI
2391
cleavage gave, instead of the expected bromomethyl hydrobromide, a mixture containing about 75y0 hydroxymethyl hydrobromide. The pure hydroxymethyl hydrochloride was obtained by hydrolysis with water and debromination with silver chloride. When direct hydroxymethylation of V or VI was found to be impossible, the introduction of one of the precursor groups -CH20COCH,, -CH20CH3, -CHZCl, -CHO, and -NO2 into VI was attempted, using a variety of methods and conditions, but without success. The 5-bromo derivative VI1 was easily preparedg from VI, but it in turn proved to be extremely unreactive, and could not be converted to the 5-nitrile. The diethyl ether of V was prepared in the hope that i t would be more suitable than VI for some of the reactions previously mentioned. Because the 4-ethoxy group was found unstable to acids (imino-ether properties), the substance was not used for this purpose. The second attempted synthesis of analog I1 involved the condensation of the previously reported keto-ester (CZHE,~)~CHCOCH (CH20CH3)COOC2H610 with acetamidine. Unfortunately this condensation failed to take place, possibly due to steric hindrance around the keto group. The third approach to the synthesis of I1 was by the condensation of ethyl phenoxyacetylcyanoacetate with acetamidine. It was recognized in advance that competitive reactions might result in the formation of any of three pyrimidines, or of three "half-cyclized" intermediates. Nevertheless, there was reason to believeloa that the reaction might yield the desired pyridimine (VIII). Actually, the only product isolated gave an analysis agreeing only with the half-cyclized intermediate I X (or isomer).
these seven analogs seemed to be the condensation of the appropriate amidine or (thio)urea with a suitable 0-dicarbonyl component (0-keto-ester or 0-diketone). Because pyrimidine has two hetero-atoms and pyridine one, no strictly analogous placement of the four groups on the pyrimidine ring is possible. Since no prediction could be made as to the structure most likely to show activity, attention has arbitrarily been limited to those pyrimidine analogs (11, 111, and IV) derivable from 0-keto-esters. Of these three analogs, I1 would involve the use of acetamidine, and I11 or IV, hydroxyacetamidine. A p-ketoester of the type R2COCHR1COOEt would be required in each case. Theoretically, there were three possibilities as to the Rgroups which might be used: (1) either R1 or Rz, or both, could he the group -CH,OH itself; or ( 2 ) RI and R2 could be groups readily convertible into -CH20H, such as -CH20R, --CHZNH~, -CN, -CH(C)Et)z, -CONH2, or -COOEt; or (3) R1 or Rz, or both, might be hydrogen, and the -CH20H group or a suitable precursor might be introduced a.fter formation of the pyrimidine ring. No very feasible method of preparing a keto-ester such as HOCHzCOCH(CH20H)COOEt OH seemed to be available. NH2 CN I The first synthesis tried for the analog I1 was I I t- CH:C HCCOOCtHI + suggested by a report of Kircher' that 6-methylII uracil on treatment with alkaline or acid formal- CH:f)cN x N CHIOC~HI NH LOCH~OC~HI dehyde solution readily gave the 5-hydroxymethyl VI11 derivative in high yield. Unfortunately, the reacNH-C=" tion failed completely8 when tried with 2-methylI 1 4-hydroxy-6-hydroxymethylpyrimidine(V) CHIC CHCOOCzHs The 6-hydroxymethyl compound V was obI1 I NH COCHzOCsHi tained by cleavage of its ether VI with hydrobroIX mic acid. The ether VI was prepared by condensI n view of the synthetic difficulties encouning ethyl y-ethoxyacetoacetate with acetamidine. It is interesting to note that the hydrobromic acid tered with analog 11, attention was shifted to the
+
.
(7) Kircher, A n n , 386, 293 (1911). (9) Under acid conditions salt formation at one nitrogen atom (8) Work to be described in a subsequent paper indicates that this would deactivate the pyrimidine ring; but under the alkaline conreaction fails also with 2-hydroxymethyl-4-hydroxy-6-methylpyrimi-ditions used in bromination the ring would be activated at the 8position, to electrophilic reagents. dine, a compound even more closely analogous to 6-methyluracil, (10) Prepared by the method of Rugeley and Johnson, Tnrs It seems that the electron-donating effects of ring-hyaroxyls in both JOURNAL, 4?, 3000 (1925). The ethyl diethoxyacetate needed as ortho and para positions are essential to give the ring-carbon B to starting material for ethyl y,y-diethoxyacetoacetate("Organic Renitrogen su5cient nucleophilic properties to condense with formaldehyde. It has also been reported by Johnson and Litzinger [THIS actions" Vol. I , p. 282) is most conveniently prepared from commercial ethyl dichloroacetatc by the one-step method of Cope. THIS JOURNAL. 68, 1940 (1936)) that uracil itself fails to condense with JOUXNAL, 68, 570 (1936). formaldehyde. Apparently the electron-donating inductive effect or (loa) Todd and Bergel, J . Chcm. SOC.,364 (1937); Foldi iliitl the hydroxymethyl of an ortho methyl group is also important; Salamon, Bcr., 74,1126 (1941); C.A , . 86,4825 (1042). alkoxymethyl group would presumably have a smaller effect.
2392
G. E. MCCASLAND, D. S. TARBELL, R. B. CARLINAND NANCY SHAKESPEARE
more promking compounds I11 and IV, and the experimental work on I1 is now described. Experimentalll Ethyl r-EthoxyacetoacetateI* (X) .-This compound was prepared in 11% yield from ethyl ethoxyacetate. Thmethod was similar t o that described for the alpha methyl derivative in PaperlI. I n view of the low yield, numerous other methods of preparing X were investigated, but ell were less satisfactory than the above method.13
2-Methy1-4-hydro~y-6-ethoxymethylpyrimidine~~ (VI). -A 51.0-g. portion of X and a 27.6-g. portion of acetamidine hyclrocliloride (llerck) were dissolved in 118 cc. of 5 A\r sodium hydroxide. The resulting solution was kept a t 25" for niiicty hours, filtered, and glacial acetic acid (about 16 cc.) added until precipitation was complete. The dried product (VI) weighed 38.5 g., m. p. 180-140". I t was dissolved in boiling benzene (4 cc. per g.) and filtcrcd (hcated Funnel). On cooling, the filtrate yielded colorless flat nei.dles; weight 21 g. (49%), in. p . 150-154". .sarnplc I recrystallized again froill henzcne, for analysis, s h ~ ) \ r e da 111. p. of 155-157". Innl 12 Calcd. for C S T - I ~ ~ C, ~ ~ 57.12; S ~ : H, 7.19. Ifoiind: C, 57.62; €1, 7.27. 'The ultraviolet absorption spectrum (absolute ethanol) showed a n extinction maximum of 9300 a t X = 273 mp, a minimum of 4300 a t 240 inp, and strong end-absorption 1)elow 2:iO mp. A comparison with spectra reported by I Ieyroth and LoofbourowI6for various pyrimidines agreed Lvith the assignirtl structure. Slow cvaporation of a'solution of VI iii 6 N hydrochloric acid yic.lllctl a hytlrochloride, stout colorless needles, in. p. 19?-20 1",with dec. A similar procedure gave the hydrolirornitlc, plates, in. p . 190-191 ', with dec. Treatment of VI with cthanolic picric acid gave a picrate, photosensitive y c l l ~ \ \nectlles, . 111. 11. 155-157". 2-Methyl-4-hydroxp-6-hydroxymethylpyrimidine HyA.-A 17.0-g. portion of VI was disdrochloride (V; solved i n 100 nil. of 8.5 .lI hydrohromic acid and the solution refluxed for nine hours. The solvent was removed uiitlcr reduced pressure, the resitlue was dissolved in 100 1111. of water, and aga'in clistillctl to dryness. The residue \vas filtercti, washctl (ethanol, 20 nil.), and dried in vacuo. The yield was : N . 4 g. colorless powder, m . p . 231-232", with decomposition. A sample was rccrystallized for analysis from acetic acid, yielding colorless prisms, i n . p . 251-253 O , with decomposition. The analysis inciicatcd that the sample was not the expected 6-bromomethyl hvdrobroinide. b u t a mixture containing- about 92% . - of the 6:hydroxymethyl hydrobromide. -
Anal.
Vol. 68
Calcd. for 92% mixture: C, 32.0; H, 4.0;
Br, 37.7. Found: C,31.69; H; 4.21; Br, 37.43. B.-Another portion (14.4 8.) of VI was cleaved with hydrobromic acid in the same manner, except that no water (other than that in the concentrated acid) was added. The distillation residue was washed with ethanol, and dried in uacuo. The colorless powder thus obtained weighed 18.0 g. and melted a t 236-237", with decomposition. A sample of this product was recrystallized for analysis from 8.8 AI hydrobromic acid ; the melting point was unchanged. -4nalysis indicated that the product so prepared contained about 75% of the hydroxymethyl hydrobromide. Anal. Calcd. for 75% mixture: C, 30.79; H, 3.79; Found: C, 30.72, 30.91; H, 4.01, 3.95. C.-In order t o obtain V, 38.0 g. of the crude cleavage product from B was added to 58 g. of freshly precipitated silver chloride. The mixture was suspended in one liter of water and boiled about thirty minutes with stirring, cooled, and filtered with suction. The filtrate was evaporated t o dryness in vacuo over sodium hydroxide, and the residue recrystallized by dissolving it in 27 ml. of boiling water and adding 100 ml. of acetone. The colorless crystalline powder which separated weighed, after drying, 16.0 g. (55%, based on VI). The powder melted a t 244-246", with decomposition. A sample was recrystallized for analysis from acetone-water (4: l), yielding colorless needles, m. p . 248-250", with decomposition. Anal. Calcd. for C~H~O.NYHCI: C, 40.80; H , 5.14; C1,20.08. Found: C,40.51; H , 5.32; CI, 19.93. The attempted hydroxymethylation (or chloromethylation) of V with formaldehyde or paraformaldehyde in aqueous solutions of alkali hydroxide or hydrochloric acid yielded unchanged starting material (or the corresponding free base). 2 -Methyl-4 -hydroxy-5-bromo-6-ethoxymethylpyrimidine (VII) .-A 6.8-g. portion of VI was suspended in a solution of 2.1 g. of sodium carbonate in 15 ml. of water, and 6.5 g. of bromine was added dropwise with stirring. The mixture was decolorized with sodium bisulfite, heated just t o boiling, and allowed t o cool. The colorless prisms which separated weighed, after drying, 8.6 g., m. p. 135-138". The crystals (7.3 g.) were recrystallized from 8 ml. of water, yielding 6.7 g. of colorless prisms; m. p. 136-138'. A sample recrystallized again from water, for analysis, showed a m. p . of 136-138". The compound gave a positive sodium-fusion test for halogen, but was completely inert t o hot alcoholic silver nitrate. Likewise all attempts to prepare the analogous nitrile by treatment with potassium cyanide in aqueous ethanol or aqueous diethylene glycol, or with cuprous cyanide in pyridine, were unsuccessful. Anal. Cald. for C8HI1N2Br: C, 38.88; H,4.49; Br, 32.34. Found: C,38.79; H, 4.54; Br, 32.52. 2-Methyl-4-ethoxy-6-ethoxymethylpyrimidine Picrate (XI) .-A 3.3-g. portion of VI was converted t o the crude 4chloro compound (red liquid) according t o the directions of Stein, et a1.I8 A solution of 4.3 g. of the crude chloro compound in 15 ml. of absolute ethanol was added dropwise t o a boiling solution of0.7g.ofsodium in30ml. of absolute ethanolduring thirty minutes. The mixture was cooled, hrought to pH 6-7 with acetic acid, and filtered. T o the hot filtrate was added 3.5 g. of picric acid. On cooling, the solution deposited 4.4 g. of thick orange needles, in. p. 90-82". A second crop of 0.5 g., m. p . 90-92" was obtained: total yield, 4.9 g. (57%). The compound gave a negative sodium-fusion test for halogen. A sample recrystallized twice more from ethanol, for ai~alysiz,gave thick yellow needles, In. p. 91-93'. Anal.1g Calctl. for CleI-Iz~OuS5: C, 45.07; H, 4.74; N, 16.43. Found: C,45.20; H,4.57; N , 16.26.
(11) Analyses by Dr. Carl Tiedcke, except as noted. All melting and boiling points are corrected. (12) Soinmelet, Bull. SOC. chim , [4]29, 553 (1031); Johnson a n d Chernoii, THISJ O ~ R N A L , 36, 1742 (1014). (13) Only one,of these will be mentioned. I t was hoped t h a t ethyl ethuxyacetylacetozLcetate could be split preferentially, following the method of Bouveaiilt and Dongert, Bull. SOL. ( h i m . , [3] 27, 1090 (1002), t o yield acetamide and X ; the cleavage, however, took t h e other cuiirse exclusively, t o give ethyl acetoacetate a n d ethoxyacetamide. (14) This compound is apparently not listed in C. A . I t is mentioned i n (:hem. Z m f r . , 109, I , 037 (1038), referring t o some patents (British 471,416; French 816,432). Ethyl y-ethoxyacetoacetate and acetamidine were condensed with sodium ethoxide t o give a compound melting a t 158'. I n 1041 Stein, el al. (THIS JOURNAL, 63, 2050 (1041)) mentioned preparation of the compound a n d gave t h e above patents ns reference, b u t gave no analysis or further details. (15) hn;ilysir Iiy Luis li. hl;ry, C d u m h i a Uxiiversity. (16) Iieyroth and Louibourorv, TIUSJOURNAL, 66, 1728 (19341. (17) This compc,und is not listed in t h e indices t o either C. il. or the Chciii. Zeiilr. However, its preparation is mentioned in t h e Zcnrrnlblofl patent Abstract above14 IIeating of t h e ether (VI) with hydrochloric w i d raportcdly yave a mixture of t h e 6-hydroxymethyl (18) Stein, el nl., THISJOURNAL, 63, 2059 (1941). T h e compound hv,lrochl
