Dec. 20, 1963
COMMUNICATIONS TO THE EDITOR
Mass spectra of streptolol and methyl octahydrostreptolate are in good accord with the structures assigned, containing peaks a t M - 18 (m/e 290 and 324, respectively) from loss of water, and major peaks a t 165 and 167, respectively, corresponding to fragmentation along line a-a, together with loss of water. Structure 1 for streptolic acid is well accounted for biogenetically by a combination of the propionate and acetate pathways recently demonstrated for the macrolides erythromycin,8m a g n a m y ~ i nand , ~ methymycin.'O Acknowledgment.-This investigation was supported in part by Public Health Service Research Grant No. AI-01278 from the National Institute of Allergy and Infectious Diseases. We also thank the Upjohn Co. for generous samples of streptolydigin. (8) J . W . Corcoran, T. Kaneda, and J . C. Butte, J . Bioi. Chem., 286, PC29 (1960); H. Grisebach, H . Achenhach, and W. Hofheinz, Z . Natu~forsch., 16b, 560 (1960) (9) H . Grisehach and H. Achenbach, ibid., 17b, 6 (1962). (10) A. J. Birch, E . Pride, R . W . Rickards, P. J . Thomson, J , D. Dutcher, D. Perlman, and C. Djerassi, Chem. I n d . (London), 1245 (1960). (11) Roger Adams Fellow, Standard Oil of California Fellow, and GilletteToni Fellow. (12) Roger Adams Fellow, National Science Foundation Predoctoral Cooperative Fellow.
4037
n.m.r. spectra with those of a synthetic sample prepared from commercial ( =!= )-&methylaspartic acid. The imide unit is shown to be present in ydiginic acid itself by comparison of its characteristic infrared (carbonyl bands a t 1775 and 1700 ~ m . - ' ) ~and , ~ ultraviolet (apparent maximum a t 203 mp, E ca. 10,000)6 spectra to those of authentic samples of N-methylsuccinimide and IV. H "..&Go, I
CHsCHCHOH-
0I
B
N-CHs
I
CHIC@
IV, V, VI,
R R R
= = =
CH3; R' = R" = H CH3; R' = H ; R" = COCHI H , R' = H ; R" = COCHB
I
-
CH2
c
CH~H
The second nitrogen atom (the amino nitrogen of pmethylaspartic acid) in ydiginic acid is present as a DEPARTMENT OF CHEMISTRY AND KENNETHL. RINEHART, JR. tertiary amide, since i t is nonbasic and the infrared CHEMICAL ENGINEERING JAMES R. BECK" spectrum of I contains an amide carbonyl band a t UNIVERSITY OF ILLINOIS WILLIAMW. EPSTEIN URBASA,ILLINOIS LARRYD. SPICER'~ 1665 cm.-l but no amide I1 or N-H stretching absorption.6 RECEIVED NOVEMBER 7, 1963 Acetylation of the methyl ester (11) of ydiginic acid with acetic anhydride and pyridine gives methyl 0acetylydiginate (111), C17Hz4N208, m.p. 27&279', Streptolydigin. 11. Ydiginic Acid [ a l 3 ' D -25' (c 1.13, MeOH) [Anal. Found: C, 53.15; Sir: H , 6.55; N, 7.211. The mass spectrum of I11 contains, We wish to present evidence which assigns structure inter alia, significant peaks a t m/e 384 (M), 324 (M I to ydiginic acid, C14H20N207, m.p. 97-103', [ a I z 5 ~ HOOCCH3), 297 (M - 87), 237 (M - 87 - HOO-37' (c 1.01, 95% EtOH) [Anal. Found: C, 51.05; CCH3], 141, and 126. The relative intensity of the H, 6.501, one of the principal water-soluble compounds last two peaks (126 >> 141) is in agreement with the obtained from ozonolysis of sodium streptolydigin ' tertiary amide assignment, since model secondary or of sodium octahydrostreptolydigin. Together, ydigamides (R' = H ) like V and VI show stronger peaks inic acid ( [ ) and streptolic acid (accompanying comcorresponding to b-b fragmentation (141 and 127 in munication)' account for all of the carbon, oxygen, V and VI, respectively) than those corresponding to and nitrogen atoms of the antibiotic streptolydigin.2 a-a cleavage (126 and 112 in V and VI, respectively). That the ion of mass 297 cannot have lost acetic acid is shown by the metastable ion7 sequence 297 m* 18.:' 237, thus it must have lost the carboniethoxyl group \ and 2s additional mass units. Absence of a strong ,N-CHa peak a t m/e 74 or 88 showsa the 28 mass units can be neither -CH2CHz- nor -CHCHI-, so they must be -CO-, and the 87-mass unit fragment must be CH300CCO-. Since the only carbonyl group in methyl ydiginate besides the imide and carbomethoxyl units (from infrared considerations) is the amide noted above, an I oxamate unit CH,OOCCO-N< is assigned to I1 and CHOR' 111; thus, the peak a t m/e 297 arises from b-b cleavage. O'CH/ I n agreement with this assignment, ydiginic acid ( I ) I has pKa 5 1.70, while the model compound N-cycloCH3 hexyloxamic acidghas pKa 2.16 f 0.03.'O I (ydiginic acid), R = R' = H The formation of I11 establishes in I an alcoholic I1 (methyl ydiginate), R = CHa; K ' = H I11 (methyl 0-acetylydiginate), R = CH3; R' = COCH3) function, which is shown to he secondary by the shift of a one-proton peak a t T 6.4 in the n.m.r. spectrum of Hydrolysis of ydiginic acid in refluxing 4 N aqueous sodium hydroxide gives L-fhreo-P-methylaspartic acid,3 I1 to 7 5.22 in the spectrum of 111." Also in the n.m.r. (4) H . K . Hall, Jr., and R Zbinden, J . A m Chem. S O L . ,80. 6428 (1958). [ c Y ] ' ~ D+12.4' (C 1.01, 5 N HCl), [(rIz7D -10' (C 1.08, (5) C . M . Lee and W . D. Kumler, ibid., 88, 4586 (1961). HzO), and methylamine, while hydrolysis of I in 4 N (6) L. J. Bellamy, "The Infra-red Spectra of Complex Molecules." 2nd hydrochloric acid a t 95' gives the corresponding N'Ed., John Wiley and Sons, Inc., New York, X. Y . , 1958, p . 203. D (c 0.45, methylimide (IV), m.p. 80°, [ c Y ] ~ ~ -152' (7) Cf, J . H . Beynon, "Mass Spectrometry and I t s Applications t o Organic Chemistry," Elsevier Publishing Co., Amsterdam, 1960, p . 251. CHC13), identified by comparison of its infrared and
'
(1) K. L. Rinehart, J r . , J. R . Beck, W. W . Epstein, and L. D . Spicer, J . A m . Chem. Soc., 85, 4036 (1963). (2) T . E . Ehle, C M. Large, W. H. De Vries, G. F . Crum, and J. W. Shell, Anlibiot. A n n . , 893 (1955-1956). (3) H. A. Barker, R. D. Smyth, E. J. Wawszkiewicz, M . N. Lee, and R . M. Wilson, Arch. Biochem. Biophys., 78, 468 (1958).
(8) R . Rybage and E . Stenhagen, J . Lipid Res., 1, 361 (1960). (9) Generously provided by D I . Robert Shapiro, New York University. (10) We are indebted t o D I . H . Boaz, Eli Lilly and Co., for these pKg determinations. (11) L . M . Jackman, "Applications of Nuclear Magnetic Resonanc? Spectroscopy in Organic Chemistry," Pergamon Press, New York, N . Y 1959, p. 5 5 .
4038
Vol. 85
COMMUNICATIONS TO THE EDITOR
spectrum of I1 is a methyl doublet a t T 8.78 (CH3CH-0-), which is replaced by a methyl singlet =C(CH3)-0- a t T 8.30 when I1 is dehydrated with phosphorus oxychloride-pyridine. These data define the unit B in I1 and I. The remaining unassigned atoms (C3H5)of I are found in its n.m.r. spectrum as a single doubly deshielded proton a t characteristically low field, T 5.02 (X-CH-Y), and two overlapping aliphatic methylene groups a t r 8.10 (C-CH2-C and C-CH2-C). The only reasonable arrangement of these groups and unit B is in unit C, and this is confirmed by spin decoupling12experiments. Acknowledgment.-This investigation was supported in part by Public Health Service Research Grant No. AI-01278 from the National Institute of Allergy and Infectious Diseases. We also thank the Upjohn Co. for generous samples of stieptolydigin and Dr. Ragnar Ryhage for the use of mass spectrometric facilities a t the Karolinska Institutet, Stockholm.
spectively) a t pH 8; Xmax 279 mp (Emax 10,500) in 0.01 N ethanolic sulfuric acid. This behavior is remarkable for enols in that the maximum near 280 mp does not shift from acidic to basic solution, but is joined by a new maximum of approximately equal intensity near 240 mp. I t differs from that of simple P-diketones like acetylacetone (Xmax 274 mp, Emax 5700 in 0.01 ethanolic H2S04; Xmax 295 mpj emax 20,000 in 0.01 iV ethanolic KOH), which show a bathochromic shift in base, and is actually the reverse of the behavior of ptriketones like 2-acetyldimedone and triacetylmethane. which give two strong maxima in acid and only one in base (2-acetyldimedone: Ahma, 233 and 274 mp, etmax 11,800 and 11,500, in ’1.01 N ethanolic H2S04; Xmax 275 m p , emax 21,000, in 0.01 Nethanolic KOH). The spectral behavior of octahydrostreptolydigin is, however, nearly exactly that of the chromophore A, exemplified by tenuazonic acid and a synthetic model compound [ A l . Anal. Found: C, 54.36; H, 5.90; N,8.88; mol. wt. (mass spec.), 1551, prepared (12) W. A. Anderson and R . Freeman, J . Chem. P h y s . , 37,8 5 (1962). from ethyl sarcosinate by the general route of Lacey.* (13) Guggenheim Foundation Fellow, 1962. (14) Allied Chemical Fellow, Roger Adams Summer Fellow, National The presence of the acylpyrrolidione (acyltetramic acid) Science Foundation Summer Fellow. chromophore in streptolydigin (A, R = C17H2303; DEPARTMENT O F CHEMISTRY KENNETH L. RIFEHART, JR. l3 R’ = CH(CH3)CONHCH3;R ” = C6HI1O2)indicates AND CHEMICAL ENGINEERING DONALD B. BORDERS’* that ydiginic acid2 is an artifact of the ozonolysis of UNIVERSITY OF ILLIXOIS sodium streptolydigin or sodium octahydrostreptoURBANA, ILLINOIS lydigin, a result of oxidative cleavage of bonds a-a RECEIVED NOVEMBER 7, 1963 and b-b, with imide formation from the newly formed incipient carboxyl group and the methyl amide function. In agreement with this deduction are the lack Streptolydigin. 111. Chromophore and Structure of imide absorption in the infrared spectra of streptoSir: lydigin itself and of other oxidation products obtained from the ozonolysis of sodium octahydrostreptolydigin. In the two preceding communications structures have been established for streptolic acid‘ and ydiginic acid2; The other products (called diginic acids) still retain the open-chain E’-methyl-p-methylaspartamide structhese compounds account for all the carbon atoms and and will be disnitrogen functions of the antibiotic ~treptolydigin.~ ture, HOOC-CHCH-CONHCH3, cussed in the full paper. The present report establishes how these two halves are joined in the intact antibiotic and completes the b OR structure (except for some stereochemical features) of streptolydigin. Streptolydigin is an acidic enol (pKa’ 5.3 in 65% methanol, positive ferric chloride and titanium trichloride t e s t ~ green ,~ cupric salt) and the previously described oxidations of its sodium salt (with sodium periodate to give streptolic acid,l with ozone to give A B ydiginic acid2) occur a t this function, just as the Al, R R ” = CHa; R’ = H B1, R = R ” = CHB; periodate oxidation of 2-methylcyclohexane- 1,3-dione R’ = H -42, R = CHj; R’ = CHCzH5; R ” = H is reported to give glutaric and acetic acids,4 and the 1 periodate oxidation of 2,3-dihydroxy-2-cyclopentenone CH3 is reported to give a-ketoglutaric acid.5 Acetylation of sodium streptolydigin with acetic anhydride-pyriIn principle, ydiginic acid could also arise from cleavdine gives an enol acetate (carbonyl band, 1745 cm.-l) age of an acylpiperidione (chromophore B) a t bonds a-a whose ultraviolet spectrum Xmax 329 mp (Emax 30,000) and b-b. However, the spectral behavior of a synthetic accords well with that expected6 for a trienone model compound [ B l . ilnal. Found: C, 50.36; H , 5.41; N, 7.36 (sodium salt); mol. wt. (mass spec.), 169 I RCH=CCH=CHC=CCO(free acid) ] is distinguished from that of octahydrostrepI I I tolydigin in that there is a definite bathochromic shift of CHJ OAc the maximum near 280 m p (from 274 mp in acid to 286 The complex ultraviolet spectrum of streptolydigin mp in base) and in that the model chromophore B1 shifts markedly with variations in P H . ~ The simpler gives only a single ultraviolet maximum a t pH 8, while ultraviolet spectrum of octahydrostreptolydigin also both octahydrostreptolydigin and the model compound displays shifts with pH : XXmax 278 mp (emax 16,200) and A1 retain both maxima a t pH 8. I n agreement with 246 (14,300) in 0.01 N ethanolic potassium hydroxide; these observations are pKa data in waterg: octahydroXXmax 281 and 246 mp (€emax 13,000 and 12,700, restreptolydigin has pKa 3.25, compound A1 pKa 3.50, compound B1 pKa 7.12. (1) K. L. Rinehart, Jr., J. R. Beck, W. W. Epstein, and L. D . Spicer, Elucidation of the chromophoric unit completes the J . A m . Chem. SOL.,8 6 , 4035 (1963). structure of streptolydigin and assigns the antibiotic (2) K. L . Rinehart, Jr., and D. B. Borders, i b i d . , 86, 4037 (1963). (3) T . E. Eble, C . M. Large, W. H. DeVries, G. F. Crum, and J. W . Shell, formula I. 1
A n f i b i o f .A n n . , 893 (1955-1956). (4) M. L. Wolfrom and J. 1LI. Bobbitt, J . A m . Chem. S o c . , 78, 2489 (1956). (5) G. Hesse and K. Mix, Chem. B e y . , 9 2 , 2427 (1959). (6) Octatrienal absorbs a t 314 mfi, c37,000: A. Smakula, Angew. Chem., 47, 657 (1934).
(7) C. E. Stickings, Biochem. J . , 71, 332 (1959). ( 8 ) R . N . Lacey, J . Chem. Soc., 850 (1954). (9) We are indebted t o Dr. H. Boaz, Eli Lilly and C o . , for pKs determinations.
