Xug. 20, 1962
COMMUNICATIONS TO THE EDITOR
:3 18.3
against Gram-positive bacteria and no activity identical with our B component by paper chromaagainst Gram-negative organisms. Compounds tography, ultraviolet and infrared absorption. Compound C crystallized in dark purple-blue B, C, D and E were highly active in vitro against Gram-positive and selected Gram-negative bac- needles which decomposed without melting when teria. When administered subcutaneously or orally, heated: Calculated for CleH~sN30e:C, 55.0; H, a t sublethal doses, B, C, D and E protected 5.5; N, 12.0; 0, 27.5; mol. wt., 349; 4.3 mice infected with Staphylococcus aureus Smith for 1 CHI; active H, 0.87 for 3; found: C, 55.3; and Streptococcus pyogenes C-203. Compounds D H, 5.6; N, 11.9; 0, 27.7; mol. wt. 326 f 6, and E were also active in mice against transplanted OMe, 4.3; NMe, 3.2; CMe, 3.4; active H, 0.88; mammary adenocarcinoma (72j). Compound D [ C X ] ~ ~ D -835O, (C 0.012 in methanol); A$:”:’ was effective against leptospirosis in the chick em- 218 mp 480), 320 mp (E;% 290), 540 mp bryo test. Antibiotics B, C, D and E showed con41.5); :;:A: 2.91, 3.03, 3.12, 3.38, 5.76, siderable toxicity in these tests. The D com- 5.90, 6.03, 6.18, 6.38, 6.92, 7.07, 7.36, 7.31, 7.70, ponent was the least toxic. 7.94, 8.09, 8.32, 8.52, 8.65, 8.80, 9.02, 9.20, 9.34, Compound ABcrystallized in red-purple needles, 9.61, 10.03, 10.18, 10.52, 10.92, 11.39, 11.82, melting a t 124-126’. Calculated for C&19N306: 12.24, 12.79, 13.06, 13.30, 14.12 p. These inC, 55.0; H, 5.5; N, 12.0; 0, 27.5; mol. wt., 349; frared maxima were in good agreement with those 4.3 for 1 CHI; active H , 0.289 for 1, 0.578 published for mitomycin B,2 but the elemental for 2; found: C, 54.8; H, 5.7; N, 11.8; 0, 27.2; analyses and specific absorptions in the ultravioletmol.wt.,346&7; OMe,4.6; NMe,2.6; CMe,3.1; visible region of our C compound were appreciably active H, 0.37. Az:pH: 218 mp 448), 323 mp different. Mitomycin B was not available for (Bi?m 278), 530 mp 34); A:;,” 2.88, 3.11, comparison. Pigments D and E were also isolated as purple 3.17, 3.42, 5.83, 6.06, 6.15, 6.35, 6.79, 6.99, 7.13, 7.34, 7.55, 7.66, 7.77, 7.95, 8.23, 8.31, 8.60, 8.79, crystalline compounds. Direct comparison of the 8.90, 9.08, 9.30 (sh), 9.51, 9.74, 10.00, 10.16, chemical and physical properties, including paper 10.41, 10.69, 11.11, 11.34, 11.61, 11.89, 12.28, chromatographic mobilities, of pigment D with 12.71, 12.81 (sh), 13.22, 13.40 (sh), 14.12, 14.77 those of porfiromycin5proved them to be identical. p. The n.m.r. spectrum in CDC13 indicated a Similarly, crystalline pigment E was compared total of 18 to 19 protons, 2 or 3 of which were directly with mitomycin C3s4 and found to be “active” and readily exchanged with deuterium in identical by paper chromatography, ultraviolet the presence of CD30D. The spectrum further and infrared absorption and by X-ray powder indicated the following functional groups : 1 diffraction. OMe, 1 NMe, 1 CMe, which is in essential agreeWhen compound B was treated with aqueous ment with the data given above. One active hy- ammonium carbonate, a crystalline product was drogen atom was present as NH, the remaining could obtained with characteristics corresponding to not be unambiguously assigned. These data diff- compound E (mitomycin C).6 erentiated the compound from any previously The above compounds were reduced readily reported. It is named mitiromycin A to indicate with sodium hydrosulfite and reoxidized in air. the relationship with the mitomycins and porfiro- This suggests the possibility of a quinoid structure. rnycin. The authors wish to thank Messrs. L. Brancone, Compound B crystallized in purple needles with A. C. Dornbush, W. R . Doughman, W.Fulmor, m.p. 159-160”. Calculated for Cl,,HlgN306: C, R. Hofstader, J. S. Kiser, J. Lancaster, J. N. 55.0; H, 5.5; N, 12.0; 0, 27.5; mol. wt., 349; Porter, G. S. Redin, H . D. Tresner, A. W. Vogel OMe, 8.5 for 2 CH3; CMe, 4.3 for 1 CH3; active and their associates for their assistance in obtainH, 0.86 for 3; found: C, 54.7; H, 5.8; N, 11.8; ing the analytical and microbiological data, and 0, 27.9; mol. wt., 366 f 10; OMe, 8.4; NMe, for the in vivo testing. D none; CMe, 4.0; active H, 0.93; [ C Y ] ~ ~-143O, ( 5 ) R. R. Herr, M . E. Bergy, T. E. Eble and H . K . Jahnke, “Anti(C 0.107 in methanol); 212 mp @itrnmicrobial Agents Annual 1960,” Plenum Press, New York, N. Y., 1961, p. 23. We wish t o thank Dr. G . M. Savage of the Upjohn 500), 319 mp 295), 515 mp 35); Aft;: for t h e sample of porfiromycin. 2.93, 3.03, 3.39, 5.85, 6.08, 6.31, 6.87, 7.08, 7.27, Company (6) J. Patrick, el ol., of our laboratories, independently observed a 7.43, 7.67, 8.15, 8.48, 8.98, 9.31, 10.01, 10.42, similar conversion when pigment B was treated with aqueous ammonia. 11.18, 11.64, 12.02, 12.63, 13.14, 1 4 . 2 5 ~ .Although D. V. LEFEMINE the Xmax in the ultraviolet-visible region were BIOCHEMICAL RESEARCH SECTION M. DANN similar to those reported for mitomycin A,2 the F. BARBATSCHI W K. HAUSMATN specific absorptions of our B component were conDIVISION 17.ZBINOVSKY siderably higher than those published for mito- LEDERLELABORATORIES P. MOSNIKETDAM mycin A. The “mitomycin A-like” compounda AMERICAN CYANAMID COMPANY J . ADAM differed appreciably from our B antibiotic in PEARLRIVER,NEW YORX x. Boliosos elemental analyses. Recently a sample of mitoRECEIVEDJUNE25, 1962 mycin A4 made available to us was found to be (2) T. Hata, Y. Sano, R. Sugawara, A. Matsumae, K . Kanamori, T. Shirna and T. Hoshi, J. Antib., Tokyo, Ser. A, 9, 141 (1956). (3) S. Wakaki. H . Marumo, K. Tomioka, G . Shimizu, E. Kato, H. Kamada. S. Kudo and Y. Fujimoto, Antibiotics & Chemolhempj, 8 , 228 (1958). (4) W e wish t o thank Dr. T. Hata. Kitasato Institute, Tokyo, and Kyowa Fermentation Industry Ltd., Tokyo, Japan, for the ramplea of mitomycin A snd
C.
THE STRUCTURES OF MITOMYCINS A, B AND C AND PORFIROMYCIN-PART I
Sir: On the basis of chemical studies and spectrophotometric data we have concluded that mitomycins A, B and C and porfiromycin ail have tho common
3 186
COMMUNICATIONS TO
structure I1 differing only in minor substituents (Table A). These antibiotics are the first members of a new and unusual structural type, representing the iirst naturally occurring examples of an aziridine,* the pyrrolo [1,2-a]indole ring system, an aminobenzoquinone and a pyrrolizine elaborated by a microorganism. Mitomycins A and B8 have received little attention, but mitomycin C4 has been tested widely in cancer chemotherapy,b although only one brief account of a partial structure study has appeared.6 The a ~ t i v i t y , ~isolation,7b a e ~ a l u a t i o nand , ~ ~assay7d of porhomycin have been reported recently but no ~ ~been available. information as to its ~ t r u c t u r ehas We have studied a group of four antibiotics which were isolated by Lefemine, et a1.,8 of these Laboratories, from three soil isolates of Streptomyces vertuillatus. These compounds were later recognized as mitomycins A, B and C and porfiromycin.9 Most of the present investigation was carried out on mitomycin A (I-A), which was correlated with the other three antibiotics as follows. Reaction of I-A with methanolic ammonia replaced the quinone methoxyl a t C-7 (6.07, CDC13) by an amino group, producing mitomycin C (I-C). Similarly N-methylmitomycin A (I-E) (made by treatment of I-A with CHaI and NaHC08 in aqueous dimethylformamide) reacted with methanolic ammonia t o give porfiromycin (I-D). Although mitomycin B (I-B) has not yet been obtained directly from or converted to any of the other three antibiotics, i t has been degraded to products identical with those obtained directly from N-methylmitomycin A (I-E) and from I-D as described below.
TARLE A I-A I-B I-C I-D I-E
Compound
X
Y
Z
MitomycinA Mitomycin B Mitomycin C Porfiromycin N-MethylI-A
HaCO H,CO H2N HsN HaCO
OCHI OH OCH, OCH: OCHs
H CHa H CHa CH:
THE
Vol. 84
EDITOR
218 mp The chromophore of I-A and I-B 17,400), 320 mp ( e 10,400), 520 mp ( e 1,400)], was identified by its similarity to the spectrum m. of 2-dimethylamino-5-methoxybenzoquinone,10 p. 165-168", 218 mp (e 18,500), 305 mp (e 13,900), 490 mp (e 3,900)]. Likewise the chromophore of I-C and I-D 217 mp (e 24,600), 360 mp (e 23,000), 555 mp (E 209)l was similar to 2,5-bis-dimethylarninoben~oquinone,~~ that of 222 mp ( e 24,000), 365 mp ( e 21,400), 513 mp (e 407)]. Furthermore, I-A in 1 N NaOH was converted to a product (not isolated) with A'.'max NaoH 323-330 (doublet) (e 24,500) which was nearly identical spectrally with 2,5-dihydroxy-Pxyloquinone,12 [ A NaoH~ 332~ (E 26,300)) ~ In 0.1 N HCI at 25' for a few hours I-A underwent a marked change forming one mole of methanol and a new product, apo-mitomycin A (Table B 1I-A),l3 ClsHITNaOe, golden plates from dimethylformamide-water, m.p. 180-200" dec., [a]%n - 10' (3.3% 0.1 N HCI), containing one C-methyl, one 0-methyl, and one amino g r 0 ~ p . l ~ The rnethanol formed did not come from hydrolysis of the quinone methoxyl group, as indicated by the continued presence of the 6.0 T peak in the n.m.r. spectrum of 11-A (D8 dimethyl sulfoxide). The chromophore of 11-A 232 mp (e 20,700), 285 mp (e 14,400), 346 mp (e 3,620), 430 mp (e 1,200)] was unaltered by short term pH changes, but hydrolysis in 0.1 N NaOH a t 26' liberated one moleof methanolresulting in a new product( II-F),l6 C14H17N306, purple needles from dimethylform14.2" (1% 0.1 N HCI) amide-water, [a],::, which was an indicator:,A'[: HC' 235 mp ( e 21,700), 294 mp ( e 15,900), 346 mp (e 3,920), 460 (e 1,050), A''m a r 254 mp (e 19,200), 312 mp (e 11,900), 560 mp (e 1,220)], a characteristic of 2-hydroxyquinones. When oxidized with KMnOd in alkali 11-F showed spectral changes typical of the Hooker oxidationlB of a 3-alkyl-2-hydroxy-l,4-naphthoquinone. An n.rn.r. peak (8.1-8.27, unsplit) likewise indicated the presence of a methyl group (e
+
(10) Prepared from 2-dimethylamino-5-hydroxyhenzoquinone IF. Kehrmann, Bcr.. 23, 897 (lS90)l by methylation with diazomethane. (11) R. Baltzly and E. Lorz, J. A m . Chem. Soc.. 7 0 , 861 (1948). (12) F. Fichter, Ann., S61, 363 (1908). (13) (a) This product seems to be a mixture of CIS- and frans-isomers which are difijcult t o separate. It is always accompanied by a more soluble isomer which is probably the 1-amino-%hydroxy isomer. We (1) T h e systematic names of these compounds are cumbersome. F o r example, mitomycin A is l,la,2,8,8a,8b-hexahydro-8-(hydroxy- propose the trivial name MITOSENE for the structure common t o thls group of compounds, vis., methyl)-6,8a-dimethoxy-5-metbylazirino [2',3'-3,4]pyrrolo [l,2-a]indole d,?-dione carbamate. We propose t h e trivial name MITOSANE for 0 structure I (where X Y Z H ) common to all four antibiotics. C H ~ O C O N H 2.3 ~ dihydro 9 hydroxyT h u s mitomycin A (I-A) becomes 7,9a-dimetboxymitosane. methyl 6 methyl 1H (2) W. R. Roderick, J. Chem. Ed., 39, 2 (1962). pyrrolo[l.2 -a]indole- 5,8(3) T. Hata, et at., J. Antibiotics (Ser. A), 9, 141 (1956). H3C dime, carbamate. (4) S . Wakaki, et al., Anfibiotics & Chemotherapy, 8 , 228 (1958). (5) R. Jones, Jr., et al., "Fourth National Cancer Cmference Pro(b) All compounds gave satisfactory analyses. ceedings, 1960," J. B. Lippincott, Philadelphia, Pa., 1961, p. 175. (6) S. Wakaki. Cancer Chemotherapy ReQorfs, 13, 79 (1961). (14) Under t h e same conditions N-methyl I - A produced a homologous product (11-B),identical by infrared and X-ray comparisons (7) (a) C. DeBoer, et al., P. Gray, B. Tabenkin and S. G. Bradley, with the product obtained by similar treatment of I-B. I-B a n d I-A "Antimicrohial Agents Annual 1960," Plenum Press, New York, N. Y., are thus related. 1961, pp. 17-22. (h) R. R. Herr, ef al., ibid., pp. 23-26. (c) C. Lewis, (15) Also I-C in 0.1 N HCI a t 25O for 24 hours produced a product et d.,ibid., pp. 27-36. (d) L. J. Hanka, ibid., pp. 37-39. (e) Discusidentical with 11-F. I-B on hydrolysis (0.1 N HC1 a t room temp. for sion, ibid., p. 40. 6 hours then in 0.1 N NaOH) gave the homologous product 11-J, (8) D. Lefemine, et al., J . A m . Chem. Soc., 84, 3184 (1902), t o whom CIUHIINSO~, blue, hydrated crystals, +16O (1% 0.1 N HC1). we are indebted f o r the supplies of antibiotics used during this investigation. which was identical with t h a t obtained by hydrolysis of I-Din 0.1 N HC1 for 24 hours a t 2 6'. (9) Because no authentic sample of mitomycin B has been available, i t has not been compared directly with I-B. (16) L. F. Fieser. 6' al., J . Am. Chem. SOC.,68, 1228 (1936).
- - -
-
-
-- -
-
Aug. 20, 1962
COMMUNICATIOXS TO THE EDITOR
on the quinone ring (no adjacent proton) in both I-A and 11-A. 0
I"
xfJ-fz1 c 5
4
N
3187
acetate) suggesting the presence in 11-A of alkyl NH2 and OH, neither of which was present in I-A. One of the two remaining nitrogen atoms and both unassigned oxygen atoms in 11-A and 11-F were part of a -0CONH2 group: hydrolysis of either 11-A or 11-F (and even I-A) in 6 N HC1 at 25' produced one mole each of CO2 and NH4+ and a new compound (11-L) [C1&4N206; [a]260690726 f 5' (1% 0.1 N HCl); tetraacetyl derivative (11-M) m.p. 225-230' 1, but in strong base neither NHI nor COa- were formed from 11-A and 11-F until after brief acidification.a The band a t 5.5 p, absent in 11-L but present in I-A, -B, -C and -D; 11-A, -B, -F and -J, was attributed to a carbamate carbonyl.' By exclusion the remaining nitrogen atom in II-.4 must have been the original nitrogen of the aminobenzoquinone in I-A; in 11-A it was neither hydrolyzable nor basic; there remained the possibility that i t was heterocyclic. In fact the ultraviolet spectrum of 5,6,7,S-tetrahydro-3-hydroxy-2-methyl-l,4carbazoledione6 (111) :,A'[: 237 mp (e 20,200), 293 mp ( E 19,000), 370 mp ( e 4,330), 510 mp (e 1,390)) 246 mp ( E 24,500)) 306 mp ( E 12,*500),36.5 mp ( E 4,620), 595 mp ( E l,GlO)], was nearly identical with that of 11-F. The arrangement of the groups (-OH, -NH2, -CHzOCONH2) on the skeleton of 11-A was estab-
+
z TABLE B X
11-A 11-B 11-F 11-G 11-H 11-J
11-K 11-L 11-M
H,CO HaCO HO HsCO HaCCOz HO HaCC02 HO HzCCOz
2
H CHa H H H CHI CHa H H
RI
CONHz CONHi CONHz CONHz CONHz CONHz COSH, H COCH,
RI
H H H COCHp COCHi H COCHs H COCHa
The foregoing data suffice to establish the interrelations and chromophores of the antibiotics in presents this series ; the following comm~nication'~ additional data sufficient to prove the total structures. (17) J. S. Webb, D. B. Cosulich, J. H. Mowat, J. B. Patrick, R. W. Broschard, W. E. Meyer, R. P. Williams, C. F. Wolf, W. Fulmor, C. Pidacks and J. E. Lancaster, J . Am. Chem. Sac., 84, 3187 (1962).
JOHN S. WEBB ORGANIC CHEMISTRY SECTION DONNA B. COSULICH LEDERLELABORATORIES DIVISION JOHN H. MOWAT AMERICANCYANAMID COMPANY JAMES B. PATRICK PEARL RIVER, NEWYORK ROBERTW. BROSCHARDlished by treatment of i t with HNOz to produce a WALTERE. MEYER compound (IV) C16H14N206,which contained a new RICHARD P. WILLIAMS 5.77 p , was optically inactive, CARLF. WOLF carbonyl group, WILLIAMFULMORand lacked the NH2 and OH groups of IIA. The CHARLES PIDACKS ultraviolet spectrum of IV [AEY~O" 280 mp ( E RESEARCH SERVICEDEPARTMENT CENTRAL RESEARCH DIVISION JOHN E. LANCASTER 41,400) ] indicated probable conjugation of the AMERICAN CYANAMID COMPANY new carbonyl with the indoloquinone chromophore, STAMFORD, CONNECTICUT
Ag;:
RECEIVEDJUNE 27, 1962
THE STRUCTURES OF MITOMYCINS A, B, AND C AND PORFIROMYCIN-PART I1
Sir : In the preceding communication' we report the interrelations and chromophores of the mitomycins and porfiromycin. In the determination of the rest of the skeletal structure and the ariangement of functional groups, the reactions described below were of major significance. apo-Mitomycin A (II-A)2 formed a diacetyl derivative (11-G) m.p. 244-246.5' dec., with carbonyl bands a t 6.44 p (amide 11) and 5.75 p (alkyl (1) J. S. Webb, D. B. Cosulich, J. H. Mowat, J. B. Patrick, R. W. Broschard, W. E. Meyer, R. P. Williams, C. F. Wolf, W. Fulmor, C. Pidacks and J. E. Lancaster, J . Am. Chem. Soc., 84,8185 (1962). (2) All formula rubrics refer to Tables A and B in the preceding Communication.
but was unlike that of ethyl 5-hydroxy-2,6-dimethyl-4,7-dioxo-3-indolecarboxylate~; therefore, the new carbonyl probably was not at the 3-indole position and hence must be a t 2-. The n.m.r. spectrum of I V (De-dimethyl sulfoxide) displayed two widely separated triplets characteristic of an AZXZpattern not shown by 11-A. This suggested the presence of the moiety Y-CH2-CH2-COat the 2-indole position. Y had to be the indole nitrogen: the -CHz0CONH2 group in IV must be at the 3- rather than 1-indole position since strong acid hydrolysis of IV produced no formaldehyde, and acid permanganate oxidation of I V yielded B-alanine, identified by paper chromatography7
(3) T. W. J. Taylor and W. Baker, "Sidgwick's Organic Chemistry of Nitrogen," Oxford University Press, London, England, 1942, p. 272. (4) S. Pinchas and D. Ben-Ishai, 1. Am. Chcm. Soc., 79, 4099 (1957). (5) Prepared by Dr. W. Remers of these Laboratories, unpublished data. (6) H. Teuber and G. Thaler, Bcr., 91, 2253 (1958). (7) E. D. Moffat and R. I, Lytle, A n d . Chcm., 81, 926 (1959).
