Substitution of a group such as chloro or methyl on the 5-position of the 4quinazolone moiety of the Hydrangea alkaloid has been found to have a favorable effect on the activity and the chemotherapeutic index (1, 2). In order to exploit this lead the fluoro, iodo, ethyl, propyl, trifluoromethyl, nitro, acetamino, methylthio, and methylsulfonyl groups were introduced into the 5-position.The fluoro derivative had an index and activity1 about the same as the 5-C1, whereas the iodo had a very low activity and index. When the methyl group was replaced by the larger ethyl or propyl groups the activity decreased in that order. These data indicate that a large group has an unfavorable action. However, trifluoromethyl had the highest index, 15, yet observed in analogs of the alkaloid, although the quinine coefficient was decreased to 35.These compounds were all prepared by condensation of the appropriately substituted 4-quinazolone with the blocked side chain, l-carbethoxy-2-(7-bromoacetonyl)-3-methoxypiperidine, followed by two-stage hydrolysis of the blocking groups as previously described (3).Six of the requisite 4-quinazolones were synthesized from the known 5-nitro-4quinazolone (4) via the readily available 6-nitroanthranilic acid (I) (5). 6-Iodo-2-nitrobenzoic acid (IV) was obtained in good yield from I, but all attempts to reduce this compound chemically or catalytically to 6-iodoanthranilic acid (VII)were unsuccessful. The steps were then reversed by reduction of 6-nitro-2formylaminobenzoic acid (VI) to VIII. Attempted replacements of the amino group with iodo through the diazonium salt were unsuccessful. The diazonium salt from I when treated with potassium ethyl xanthate gave 2-nitro-6-methylthiobenzoic acid (V) in low yield after methylation. All of the above difficulties were surmounted by the use of the diazonium salt from 5-amino-4-quinazolone (III). The 5-iodo derivative, IX, was obtained in 59% yield by treatment with potassium iodide. The 5-methylthio derivative (XI) was obtained with sodium methyl mercaptide in 71% yield and was oxidized to the sulfone, XII, with permanganate in dilute acetic acid. The use of sodium thiophenoxide gave XIII, but the procedure could not always be duplicated. The 5-fluoro derivative offered some difficulty since the diazonium fluoroborate was water-soluble. However, the diazonium fluoroborate could be isolated when the diazotization was carried out in 42% fluoroboric acid (6) to give an over-all yield of 56% of IX after pyrolysis in xylene.3-Nitrophthalic acid appeared attractive as a starting material for 5-ethyl-(XXIII) and 5-propyl-4-quinazolone (XXXIII). -Methyl hydrogen

Puromycin has been shown to be a derivative of 6-dimethylamino-9-(3'aminoribosyl)purine (1). In a previous paper (2) a method was described for the glycosidation of 6-dimethylaminopurine on the 9-position via the 2-methylmercapto derivative. Since a search of the literature revealed no example of the glycosidation of a purine with an amino sugar, it would be necessary to solve this problem before a total synthesis of the antibiotic could be effected. The proper blocking group and activation for the coupling of n-glucosamine, as a model amino sugar, with a purine has now been successfully completed.A derivative of n-glucosamine (I) suitable for variation of the N-blocking group is 1,3,4,6-tetraacetyl-2-amino-/3-n-glucopyranose (II) prepared by the elegant method of Bergmann and Zervas (3). This derivative was then N-blocked by acetyl (IV), phthalyl (VI), or carbobenzoxy. The introduction of the phthalyl blocking group by the standard method of direct fusion of the base, II, with phthalic anhydride was unpromising. This difficulty was circumvented by treatment of the base, II, with phthalic anhydride in boiling chloroform which formed the crystalline phthalamic acid, III, in 85 % yield. The ring closure to the phthalimido derivative, VI, did not proceed readily with boiling acetyl chloride. A new and milder method for cyclization of phthalamic acids was devised. Treatment of the phthalamic acid, III, in chloroform with triethylamine and ethyl chlorocarbonate at 0°g ave a mixed anhydride (4) which cyclized at room temperature in 2 hours to the desired phthalimido derivative, VI, in 74-81 % yield.2By proper modification of the literature procedure (5), the conversion of n-glucosamine /3-pentaacetate (IV) to a-bromoaceto-n-glucosamine (Va) was increased from 42 to 55 %. Attempts to condense this bromo sugar derivative with the chloromercury salt of 2-methylmercapto-6-dimethylaminopurine (2) in boiling xylene, toluene, or benzene did not lead to any of the nucleoside derivative (Villa), but decomposition of the bromo sugar took place with consequent formation of the free purine base. Similar results were obtained with the crystalline l-bromo-2-phthalimido-3,4,6-triacetyl-D-glucopyranoside (IXa). Since the bromo phthalylglucosamine, IXa, obviously decomposes at its m.p. of 120°and a-bromoaceto-n-glucosamine (Va) has no definite m.p., but gradually decomposes, the difficulties in the coupling reaction were attributed to the instability

No abstract

The general method for the synthesis of the Hydrangea alkaloid (I) involved coupling of 1 -carbethoxy-2-( -bromoacetonyl) -3-methoxypiperidine with 4-quin-O II HO ,C\ NCHsCOCH, CH 2\nZ H.2HC1nZ azolone followed by removal of the blocking groups by two stage strong acid hydrolysis (1). A series of related compounds with the benzene ring replaced by a heterocycle could theoretically be synthesized. However, it is imperative that the heterocycle be stable to strong acid. Secondly, the heterocycle should not be basic such as pyrimidine, pyridine, or imidazole since substitution of a basic group on the benzene ring of the alkaloid (I) caused loss of activity (2). One of the few heterocycles satisfying these conditions is thiophene. This communication describes the syntheses of 6,7-and 7,8-thia-4-quinazolones, two of the three possible thienopyrimidones, and conversion of the 6,7-thia-4-quinazolone to the alkaloid isoster (IX).The most general method of synthesizing 4-quinazolone and its derivatives is by the Niementowski reaction (3), involving fusion of the proper anthranilic acid with formamide (3). This would require the preparation of the proper o-aminothenoic acid, none of which are described in the literature. However, Cheney and Piening (4) have described a synthesis of 2-alkyl-3-amino-4-thenoic esters by a general method. The parent amino ester (IV) has now been synthesized by their procedure.3-Carbomethoxy-4-ketothiophane (II) (5) was converted to the oxime (III) and rearranged with hydrogen chloride (4) to the desired 3-amino-4-carbomethoxythiophene hydrochloride (IV). Saponification gave impure 3-amino-4-thenoic acid (VII) in low yield which was not readily purified due to its apparent instability. It seemed desirable at this point to convert the amino ester base (IV) directly to the thiaquinazolone (V). A test reaction with methyl anthranilate and formamide at 175°g ave only 35% of 4-quinazolone after four hours. However, addition of ammonium formate to the fusion mixture raised the yield of 4-quinazolone to 71%. Application of these conditions to IV free base gave only traces of the desired V along with large amounts of tar. Since aminothiophene bases are notoriously unstable, the molecule was stabilized by formylation to VI.

No abstract