5044
COMMUNICATIONS TO THE EDITOR
Acknowledgement. - This research was supported by the A.E.C. under Grant COO-1087-84, and by the N.S.F (GP-1471).
CH3
\ /
VOl. 86
x=o I
oc'
z
DEPARTMENT O F CHEMISTRY NORTHWESTERN UNIVERSITY EVANSTON, ILLINOIS
I
z
ROGERJ MAWBY FREDBASOLO RALPHG PEARSOS
RECEIVED SEPTEMBER 11, 1964
CH3 I \
i,
7;Eo
'\ +--OC-?!Mn-CO OC'J
gqco (b) migration
CH~-M&CO
oc'
I z
+ co
New Lincomycin-Related Antibiotics
Sir:
Chemical studies' have elucidated the structure of lincomycin2-6 (Ia). We now report the isolation and the structure of four new lincomycin-related antibiotics designated U-21,699, LT-11,921, U-11,973, and U20,943. Antibiotic U-21,699 normally occurs in lincois by methyl migration (see Fig. 1, 2 = P(Ph)3). mycin fermentations, while the production of LTThis would require that the rate of disappearance of the acetyl band in t r a n ~ - C H ~ C O l t n ( C 0 ) ~ P ( P beh ) ~ 11,921, U-11,973, and U-20,943 is induced by addition of DL-ethionine, methyl a-thiolincosaminide (IIa), equal to the rate of change of the arrangement of terand ethyl a-thiolincosaminide (IIb), respectively, to minal C-0 stretching frequencies from that for transmedia of S. Lincolnenszs. The antibiotics CH$20Mn(C0)4P(Ph)3to that for G Z S - C H ~ M ~ ( C O ) ~ Pfermentation were recovered by adsorption on carbon followed by (Ph),. Since previous work' on the related reaction elution with aqueous acetone and crystallization as shown in eq. 1 had shown that in n-hexane the formahydrochloride salts. tion of intermediates, which could complicate kinetics Antibiotic U-21,699 (Ib) hydrochloride, CI7Ha2and make a decision on a mechanism more difficult, NzOeS.HCl.0.5H~0,7[ a ] ? ' D +147', contains one was avoided, n-hexane was used as the solvent for the -NCH3, one -SCH3, two C-CH3, and one basic funcpresent work. tion, pKa' 7.68. Infrared absorptions (3400, 1(j9O. The reactions were followed by thermostating soluand 1590 cm.-l) indicated OH, N H , and amide functions of ~ ~ U ~ S - C H ~ C O M ~ ( C O in )dry ~ P n-hexane (P~)~ tions. The n.rn.r.*agspectrum of I b is very similar under nitrogen. Samples were withdrawn a t intervals to that of Ia, except that it shows eight hydrogens in and their spectra recorded in 1-mm. cells on a Beckman the 30-90-C.P.S.region, while the spectrum of Ia shows IR9 spectrophotometer. Solution concentrations were ten hydrogens in the same area. Both display the 0.6 X M for following changes in terminal C-0 doublet at 67 and 75 C.P.S. (3H) assigned to the CH3stretching frequencies, and 1.5 X M for following CHO grouping present in methyl a-thiolincosaminide the disappearance of the acetyl C-0 band. I The change from the terminal C-0 frequencies of (Ira) and a triplet a t 45, 52, and 59 C.P.S. (3H) assigned ~ T U ~ ~ - C H ~ C O M ~ ( C to O )the ~ P arrangement (P~)~ for to the CH3CH2- of the side chain on the hygric acid cis-CH3Mn(C0)4P(Ph)3, as calculated from the rate of nucleus (IIIa). This triplet, along with two hydrogens appearance of the bands in the product a t 2055 and 1983 in the -CH,- region, is indicative of an ethyl group cm. -I, was a first-order process with a rate constant of replacing the propyl group of lincomycin. Thus, I b 3.83 (+0.05) X lop5set.-'. The disappearance of the represents the structure of U-21,699. This conclusion acetyl band of ~ ~ u ~ s - C H ~ C O M ~ ( C O was ) ~also P ( Pa ~ ) ~was substantiated by hydrazinolysis of U-%1,699 first-order process with a rate constant of 3.71 (+0.08) affording authentic methyl a-thiolincosaminide (1Ia) , X lO+ set.-'. and the hydrazide of a new amino acid which was hyThe agreement between these two figures is good drolyzed to the crystalline hydrochloride C8Hl5NO2 evidence for the direct conversion of trans-CH3COMnHCl (IIIb). Infrared spectra and the positive rota(CO) 4P(Ph) to cis-CH3Mn(CO)4P(Ph) by methyl tional shift a t lower pH suggested an a-L-amino acid. migration (see Fig. 1, 2 = P(Ph)3). The n.m.r. spectra of IIIa. and I I I b differ only in the I t is possible to devise more complicated mechanisms, area from 40 to 120 C.P.S. in that IIIa has a triplet involving two or more steps, one of which is rate con(1) H . Hoeksema, B. Bannister, R . D. Birkenmeyer, F . A . MacKellar, trolling and the others very rapid, which are compatible F . Kagan, B . J. Magerlein, W . Schroeder, G. Slomp, and R . R . Herr, J . A i n . with our kinetic results. However, we regard methyl Chem. Soc., 86,4423 (1964). (2) T.incocin is the trademark of T h e Upjohn Company for lincomycin migration as the simplest and by far the most likely hydrochloride. mechanism for this reaction. This is believed to be the (3) D. J . Mason, A . Dietz, and C. DeBoer, Anfimicrobiai A g e n f s Chemo&ropy, 554 (1962) first real evidence for such a migration in these systems. ( 4 ) I
