COMMUNICATIONS TO THE EDITOR
3906
Fig. 1.-(a)
+
Co%+; (b), (c), Co%+ NO; (d), Co%; (e). (0, COQZ NO.
+
VOl. 83
oxide, exclude interactions between Co@y+ and nitric oxide. The reduction of Coa2N0 (Fig, IC, f) leads to a complex [CO@~NO](still including nitric oxide inside). This is confirmed by the coulometry of a solution of CoQ2N0made a t -0.50 v. vs. S.C.E., which eliminated proportionally both waves, the oxidation one (Fig. le) and the reduction one (Fig. If). With the coulometry carried out a t -0.75 v. vs. S.C.E., the polarogram finally shows only the C O Q ~ oxidation wave (Fig. Id). This means that a t the latter potential, a further reduction of nitric oxide within the complex is obtained and Co@Z is produced again. The results show that Co92 exhibits with nitric oxide the same “carrier” property that i t shows with oxygen.
( 1 ) J . 2. Hearon, D. Burk and A. L Schade, J . Nafl. Conccr Inst.. havior of CO@Z+ and of Co@*in the presence of 9, 337 (1949). nitric oxide has been followed polarographically. DI CHIMICA GENERALE In phosphate buffer a t PH 7.5, with histidine con- ISTITUTO SILVESTRONI PAOLO UNIVERSIT.~ DI PERUGIA centration 0.1 F and a t 20°, solutions of Coa2+ PERUGIA, LAURACECIARELLI F> free of nitric oxide give (besides the re- ISTITUTODIITALY GENERALE 2e- + Coo 2@- a t U N I V E R S CHIMICA duction wave Co@Z IDIT ~ROMA m/,li -- 1.4 v. VS. S.C.E., which does not interest ROMA,ITALY RECEIVED JUNE16. 1961 us here) a polarographic wave a t W S = -0.21 v. vs. S.C.E. (Fig. 1, a) corresponding to the reduction CO@Z+ e-+ C o @ 2 . I n the presence of nitric OF THIOSTREFTON. DERIVATIVES oxide, two waves are obtained on the polarogram DEGRADATION OF %HYDROXYQUINOLINE (Fig. lb, c), each having practically the same height Sir : as the one obtained in the absence of nitric oxide Earlier communications from these laboratories (Fig la), the half-wave potentials of which are, respectively, -0.16 and -0.42 v. vs. S.C.E. (fur- have described the isolation and general properties and more recently, ther indefinite waves, not of interest here, are of the antibiotic thiostrepton1-2.8 caused by the second reduction of CoQ2NO by the the isolation of three thiazole amino acid derivat i v e ~ . The ~ ~ antibiotic ~ also contains an additional reduction of nitric oxide and of COQZ + Coo). chromophoric system, and we now present evidence Under the above-mentioned temperature, pH from parallel pyrolytic and hydrolytic degradations and [@] conditions, solutions of Coa2 show on the concerning this part of the molecule. polarograph an oxidation wave with XI/, = -0.20 Pyrolysis of thiostrepton a t 250-350’ and 0.2 v. vs. S.C.E. (Fig. Id). I n the presence of nitric mm. pressure yielded, in addition to the diketooxide, this oxidation wave shifts its half-wave poten- piperazine of alanine and isoleucine, two crystalline tial to -0.16 v. (Fig. le) and a new reduction wave phenolic bases characterized by deep green ferric appears with m/, = -0.40 v. vs. S.C.E. (Fig. If), chloride reactions. The first, m.p. 90-95’, A%? its height being practically equal to that of the 242 mp (42,500) and 314-316 mp (3,500) was oxidation wave of Co% (Fig. Id) in the absence of not obtained completely pure, but its molecular nitric oxide. The oxidation wave ( m l S= -0.16 v.) formula CllHllON was assured by high resolution (Fig. le) is always smaller, though not quantita- mass spectrometry (Found: mol. wt., 173.142; tively reproducible, because of uncontrolled oxida- calcd. for CllHllON: mol. wt., 173.139).6 Its iden-c C O @ ~ + NO-) catalyzed by the tion (CO@~NO tification as a derivative of 8-hydroxyquinoline was electrode surface. These results may be explained suggested by the characteristic ultraviolet spectrum, as follows: the Coat oxidation wave and the CO@Z+solubility in hexane, ferric chloride test, formation reduction wave in the presence of nitric oxide (Fig. of an orange chloroform-soluble copper complex, le, b), both with rill ‘V -0.16 v., are relative to and a close similarity in infrared spectrum with that the reaction CoQz+ NO e- Ft [CO@ZNOI. of 4-methyl-8-hydroxyquinoline. It was identified The reduction waves (Fig. If, c), obtained from as 4-ethyl-8-hydroxyquinoline (I) by comparison solutions of CoQ1 and Co@z+ in the presence of with an authentic sample, m.p. 92-93’; A d . nitric oxide, are attributed to the reduction [Co% (1) J. Vandeputte and J. D. Ihtcher, “Antlbiotica Annual, 1955e- + [CoQ2NO]-. I n the case of the reduc- 1956,” NO] Medical Encyclopedia, Inc., New York, N . Y., p. 660. tion of CoQZ+ in the presence of nitric oxide, this (2) J. P. Pagano, M . J. Weinstein, H. A. Stout and R . Donovick, wave (Fig. IC) is caused by the reduction of CO@Zibid.# 1955-1958, p. 654. (3) B. A. Steinberg, W. P. Jambor and Lyda 0. Suydam, ibid., NO formed a t the electrode surface as a reaction p. 562. product of nitric oxide contained in the solution, 1955-1956, (4) G. W. Kenner, R. C. Sheppard and C. E. Stehr, Tcfrahcdron with CoQz obtained in the first reduction process Ldlcrs, 23 (1960). ( 5 ) M. Bodaoszky, J. T. Sheehan, J . Frled, N. J. Williams and C . A. Co@z+ e- --.c Co@*.In fact, measurements of the partial pressure of nitic oxide on a solution of CO@Z+, Birkhirner. J . Am. Chcm. SOC.,85, 4747 (1960). (6) We are grateful to Mr. J. H. Beynon and Mr. A. E. Williams of and spectrophotometric measurements on a solu- Dyestuffs Divislon, Imperial Chemical Industries, Ltd., Blackley. tion of C O ~in+ the presence and absence of nitric Manchester, England, for this determination.
+
+
+
+
+
+
+
+
COMMUNICATIONS TO THE EDITOR
Sept. 20, 1961
3907
5.70. The acid I11 is decarboxylated smoothly a t 180', yielding 4-acetyl-8-hydroxyquinoline (11) identical in all respects with the product obtained by pyrolysis, The above data strongly support the formulation of 111 as 4-acetyl-8-hydroxyquinaldic acid. Confirmation was obtained by synthesis of the methyl ester methyl ether (IV), as described below. From the mother liquors of I11 a second crystalline phenolic acid (VI), m.p. 190-200' (dec.) was obtained. Anal. Calcd. for ClZHllOaN: C, 66.35; H, 5.10; N, 6.45. Found: C, 65.82; H, 4.67; N, 6.47; A& 255 mp (36,000); 357 mp (3,400). Compound VI had properties very similar to those of I11 except that the ketone function was missing (only On heating, one carbonyl band a t 1740 an.-'). decarboxylation occurred leading to 4-ethyl-8hydroxyquinoline (I), and the parent acid therefore was formulated as 4-ethyl-8-hydroxyquinaldic acid (VI), analogous to the acetyl acid (111). The phenolic acids VI and I11 (the latter as its dimethyl derivative (IV)) were synthesized : 2methyl - 4 - ethyl-8-methoxyquinoline, m.p. 69.5' [anal. Calcd. for C13H16NO: C, 77.58; H, 7.51; N, 6.96. Found: C, 77.47; H, 7.42; N, 7.121 prepared by reaction between o-anisidine and hex-2-end-one, was condensed with 2.5 equivalents of formaldehyde. Oxidation of the crude methylol derivative with potassium permanganate yielded, after R R esterification with diazomethane and chromatography on silica gel, 4-ethyl-8-methoxyquinddic acid methyl ester, m.p. 97.598' [anal. Calcd. for N COOY Cl4HI6O3N: C, 68.55; H, 6.16; N, 5.71. Found: _ . _ _ C, 68.44; H, 6.3; N, 5.81 and also 4-acetyl-8-methOH ox oxyquinaldic acid methyl ester [anal. Found: C, I, R = CHaCI-T? 111, R = CHaCO, X = Y H 64.G6; H, 4.91; N, 5.371. The latter was identical 11, R CHaCO IV, R = CHsCO, X = Y CHs CHaCO, X = CHI, Y = H V, R with the product (IV) derived from thiostrepton. VI. R CHsCH,. _ ~ ~ X _ _ Y~ ~ H The 4-ethyl-8-methoxyquinaldic acid methyl ester VII; R = C H ~ C H ~ O Hx-= ) , Y-= C H ~ was demethylated by potassium iodide in refluxing VIII, R CHaCH(OH), X = Y = H phosphoric acid and yielded the phenolic acid (VI), (1n.p. 233' dec.) and gave a positive iodoform test m.p. 189-199' (dec.) [ a w l . Found: C, 66.41; H , indicating the presence of a CHsCO- grouping. 5.17; N, 6.17; 256 m p (41,000);358 nip (2,600)], This was confirmed by the n.m.r. spectrumQof its the infrared spectrum of which was very similar to methyl ester methyl ether (IV), which showed a but not identical with that of the acid V I derived band at 7.45 7 (CH,CO-) but no bands in the alde- from thiostrepton.'l hyde region.l0 IV was prepared by methylation of In a subsequent experiment, tliiostrepton was I11 with diazomethane, m.p. 165-167'; 263 mp hydrolyzed with a mixture of equal volumes of (30,000); 366 mp (2,700). Anal. Calcd. for C14- formic and hydrochloric acids for 24 hours a t 105'. H1304N: C, 64.86; H , 5.05; N, 5.40; OCHa, 23.95; After evaporation of the acids the product was exmol. wt., 259. Found: C, 65.09; €-I, 5.43; N, 5.58; tracted into ether and methylated with diazoOCHa, 24.16; mol. wt. (Rast), 270. Dinitrophenyl- methane. Crystallization afforded an optically hydrazone, m.p. 243' (dec.). Saponification of IV active dimethyl derivative (VII), m.p. 175-177', with dilute alkali gave the methyl ether (V), m.p. [a]%'-78' (c, 1.6 in ethanol), Ag:x 254 mp (39,000); Anal. Calcd. for C13H1104N : C, 63.67; 347 nip (3,000). Anal. Calcd. for C14H1604N:C, 128-130'. H, 4.52; N, 5.71. Found: C, 63.86; H, 4.68; N, 64.36; H, 5.79; N , 5.36. Found: C, 64.50; H, 6.09; N, 5.58. The infrared spectrum of VI1 (7) Add I11 wan mentioned in footnote 15 of ref. 5. Its presumed showed only one carbonyl band (1750 cm.-l) but in volatility refers to the decarboxylation product 11. (8) A. Butenandt. U. Schiedt and E. Biekert, Ann., 686, 229 addition a sharp band a t 3370 cm.-l indicating the (1954). presence of a hydroxyl group. VI1 was identified (9) For the n.m.r. spectrum and ita interpretation we are indebted to as the methyl ether methyl ester of 4-(a-hydroxyDr. Harold Conroy of the Mellon Institute. The s p e c t n m also shows etliyl)-8-hydroxyquinaldic acid (VIII), by its chrothat three of the four aromatic hydrogens are vicinal, the fourth is separated from them. mic acid oxidation to IV. Catalytic hydrogenation (IO) Evidence for substitution in position 4 was obtained by nitric of I V with the uptake of one mole of hydrogen also acid oxidation of 11, I11 and V, which yielded, after decarboxylation, a yielded racemic VII. mixture shown by paper chromatography to be composed of nicotinic, Calcd. for CllHllON: C, iG.27; H, 6.4; N, 8.09. Found: C, 76.22; H, 6.60; N, 7.89; obtained by reaction between o-aminophenol and l-chloropentanone-3. The second rolysis roduct CIIHg0 2 N ; m.p. 110-113', AZ!J256 mp &6,500), 335 mp (3,500) and 355 mp (3,000); $,:21678 cm.-' (aromatic ketone); anal. Calcd. for C1lHvOIN: C, 70.58; H, 4.85; N, 7.48. Found: C, 70.63; H, 5.13: N, 7.50; formed crystalline dinitrophenylhydrazone and semicarbazone derivatives. Wolff-Kishner reduction of the latter again yielded 4-ethyl-8-hydroxyquinoline (I), and the ketone must therefore be formulated as the corresponding 4-acetyl derivative (11), in agreement with results simultaneously obtained by study of the products of acidic hydrolysis of thiostrepton. Hydrolysis of thiostrepton with boiling N HC1 for 24 hours followed by extraction with ether furnished in the ether extract a phenolic acid (111)' m.p. 190200 (dec.), A%" 265 mp (32,000); 375 mp (3,500); YE$''1730, 1700 cm.-l. Anal. Calcd. for C12H004N: C, 62.34; H, 3.92; N, 6.06; mol. wt., 231. Found: C, 62.61; H, 4.03; N, 6.10; neut. equiv., 225 (alkali, not titrable with anhydrous HC104). The ultraviolet spectrum again is similar to those of 8-hydroxyquinoline derivatives, and the infrared spectrum resembles that of a methylated xanthurenic acid! The acid formed a dinitrophenylhydrazone
qJh
-
3
I I
laonlcotinic and cinchomeronic adds. An identical mixture (infrared and paper chromatography) was obtained by pyrolysis of cinchomernnic acid.
(11) The discrepancies in the two spectra are most likely due to eontamination of the acid from thiostrepton hp the keto a d d 111, which is difficult to remove by crystallization.
3908
COMMUNICATIONS TO THE EIIITOK
Vol. 83
No positive evidence is yet available to indicate whether the 4-acetyl and 4-ethylquinaldic acids, I11 and IV, are genuine constituents of the antibiotic, or are formed from VI11 during the process of hydrolysis and isolation
toiiiary or alternatively with the sign of the net rotational strength associated with some excited state, say (S)T,.On the other hand, if the absolute configuration is known, there is a geometrical method of identifying the enantiomers based on DEPARTMENT OF C C DREY accepted conventions which consists in associating ORGANIC CHEMISTRY G W KENNER with the chelate ring a segment of a helix or screw. LIVERPOOL UNIVERSITY 11 r) LAW In this way the enantiomer depicted in Fig. 1 is asLIVFRPOOI,, ENGLAND R C SHEPPARD sociated with a left-handed helix. Therefore, we TIiE S Q U I R INSTITUTE ~ MIKLOSBODANSZKY FOR MEDICAL RESEARCH JOSEF FRIED propose that this enantiomer be designated AWe NEWBRUNSWICK, NEWJLRSEY NINAJ WILLIAMS C0en8+~ rather than D - C O ~ ~ ~ + use ~ . the Greek JOHN T SHEEHAN letters A and A to avoid confusion with the earlier RECEIVED AUGUST10, 196 t convention. -__ In a study of the rotatory dispersion of these complexes Moffitt3 assumed that the rotation is PARTIAL CHROMATOGRAPHIC RESOLUTION, due to an admixture of 4p orbitals with the 3d orbitROTATORY DISPERSION, AND ABSOLUTE CONFIGURATION OF OCTAHEDRAL COMPLEXES als involved in the transition under the asymmetric CONTAINING THREE IDENTICAL BIDENTATE trigonal field in Y3*3. He predicted that the rotaLIGANDS tion should take its sign from the asymmetric potenSw: tial. However Suganoa has shown that this firstWe have obtained a partial resolution of the order theory cannot account for the rotational acetylacetonates of chromium(II1) and cobalt(lI1) strength on symmetry grounds; hloffitt's error lay by chromatography1 on d-quartz or alumina treated in the phase. While the exact mechanism of the with d-tartaric a c d 2 The rotatory dispersion rotation remains unknown, we may expect that shows a strong Cotton effect associated with the the sign of the rotation should correlate to the sign spin-allowed transitions near 600 nip. For both of the trigonal field since this is based on symmetry.' compounds the rotational strength3 R is negative Our studies of crystal spectra of acetylacetonatesR for the enantiomer less strongly adsorbed on alu- and oxalatesg have afforded values of the trigonal mina-d-tartaric acid. We have attempted to field parameterlo K . These are recorded in Table establish the absolute configurations of the acetyl- I. Notice that for every Cr-Co pair the signs of acetonates and corresponding oxalates4 as well by the trigonal field parameter correlate to the signs referring them to the known absolute configurationb of the rotation as expected. of the ethylenediamine derivative ( + ) ~ - C o e n ~ + ~ TABLEI (see Fig. 1)-but first some pedagogy concerning Sign sign conventions. of R Complex ion
or compound
(600 mp K hand) (cm
Method of rewlntion
Crent +a Less soluble Cociia+3 Chlorotartraten Cr(Cz04)a-8 Less solublc C O ( C ~ O ~ ) Strycliniiie ~-~ salt" Cr(OzCiH,h Less readily adsorbed
All
-I)
+ + + + 45* + +2i0' fSh
-
-looc
--
+600"
