Biosynthesis of Vitamin B12: Identity of Fragment Extruded During Ring Contraction to the Corrin Macrocycle (Acetic Acid/Cobyrinic Acid/Isobacteriochlorins) ALAN R

Proc. Nati Acad. Sci. USA Vol. 78, No. 1, pp. 13-15, January 1981 Chemistry

Biosynthesis of vitamin B12: Identity of fragment extruded during ring contraction to the corrin macrocycle (acetic acid/cobyrinic acid/isobacteriochlorins) ALAN R. BATTERSBY*, MICHAEL J. BUSHELL, CHRISTOPHER JONES, NORMAN G. LEWIS, AND ARMIN PFENNINGER University Chemical Laboratory, Lensfield Road, Cambridge CB2 lEW, England Communicated by V. Prelog, September 30, 1980

ABSTRACT Incorporation experiments with labeled sirohy- A mixture of [20-'4C]sirohydrochlorin ester with an appropriate drochlorin and trimethylisobacteriochlorin demonstrate that ring amount of [2,7-methyl-3H]sirohydrochlorin ester was then hy- contraction in vivo to the corrin macrocycle of vitamin B12 lib- erates acetic acid. The C-20 atom of the precursors becomes the drolyzed to produce [20-14C, 2,7-methyl-3H]sirohydrochlorin acetate carboxyl carbon. (structure 8), and the esters of 14C-labeled forms 5 and 6 similarly gave [20-14C, 2,7-methyl-'4C]sirohydrochlorin (structure The trimethylisobacteriochlorin isolated (1, 2) from the vitamin 9). B12 producer Propionibacterium shermanii was proved (3, 4) Incubation ofthe former precursor (structure 8)with a cobalt- to have structure 11 in which, surprisingly, the third methyl containing cell-free enzyme system (8) from P. shermanii (Table group to be inserted is located at C-20 (Fig. 1). The surprise 1, Exps. 1 and 2) yielded first the trimethylisobacteriochlorin stems from the fact that the carbon atom at C-20 of structure (structure 10) in dihydro form,* followed by further biological 11 has to be extruded at some stage during the demonstrated transformation into cobyrinic acid (structure 13). From the in- conversion of the trimethyl system (structure 11)* into coby- cubation mixtures were isolated cobyrinic acid (structure 13) as rinic acid (structure 13) by using cell-free enzyme preparations its ester (structure 14), acetic and formic acids as their p-bro- from P. shermanii (ref. 7 and refs. therein) or Clostridium te- mophenacyl esters, and acetaldehyde and formaldehyde as their tanomorphum (1). Double-labeling experiments further proved dimedone derivatives. In each case, radioinactive carrier ma- (4, 7) that the C-20 methyl group is also lost during the con- terial was added to assist the isolation. The results in Table 1 version of 11 to 13 and, hence, study of the nature of the ex- (Exps. 1 and 2) demonstrate reproducible incorporation of si- truded fragment is of considerable interest; the necessary ex- rohydrochlorin (structure 8)* into cobyrinic acid (structure 13) periments are outlined here. with formation of labeled acetic acid. The isolated formic acid There are, in principle, several approaches to this problem (Exp. 2) and formaldehyde and acetaldehyde (Exps. 1 and 2) based on different known precursors ofcobyrinic acid (structure carried negligible activity (<2 X 10-3 equiv.), and this trace also 13). [20-14C, 2,7-methyl-3H]Sirohydrochlorin (structure 8), [20- appeared in the corresponding materials from the blank run 14C, 2,7-methyl-'4C]sirohydrochlorin (structure 9), and [2,7,20- (Exp. 3). This parallel blank also showed clearly that the pro- methyl-14C]trimethylisobacteriochlorin (structure 12) were se- duction of acetic acid in Exps. 1 and 2 is an enzymic process. lected to avoid formation of formaldehyde, which can arise These results were confirmed with triply 14C-labeled sirohy- chemically from uroporphyrinogen-III (structure 2). [20- drochlorin (structure 9; Exp. 4), in which a remarkably high 14C]Sirohydrochlorin (structure 5) was biosynthesized from [20- conversion of sirohydrochlorin (structure 9)* into cobyrinic acid 14C]uroporphyrinogen-Ill (structure 3) by using a cell-free en- (structure 13) was achieved. The acetic acid isolated from Exps. zyme system (lacking cobalt) from P. shermanii (see ref. 7) and 1, 2, and 4 corresponded to 65-77% of the stoichiometric it was rigorously purified as its octamethyl ester. [20- amount relative to the cobyrinic acid biosynthesized (Table 1). 14C]Uroporphyrinogen-III (structure 3) was obtained by hy- If this labeled acetic acid arises specifically from C-20 and drolysis and reduction of [20-'4C]uroporphyrin-III octamethyl its attached methyl group of a macrocycle lying on the pathway ester, which was synthesized by MacDonald's method from the beyond the trimethyl system* (structure 10), then it should be [I4Clpyrromethane (structure 1). The [2,7,20-methyl- labeled solely at the carboxyl group. This was proved by hy- 14C]trimethylisobacteriochlorin (structure 12) together with drolysis of the labeled p-bromophenacyl acetate (at constant [2,7-methyl-'4C]sirohydrochlorin (structure 6) were prepared specific activity) followed by Schmidt degradation of the re- biosynthetically from [methyl-'4C]methionine with resting P. sultant acetic acid (9); the methylamine, isolated as its phthal- shermanii cells; similarly, [2,7-methyl-3H]sirohydrochlorin oyl derivative, was essentially radioinactive (Table 1, Exp. 1). (structure 7) was obtained from [methyl-3H]methionine. All these labeled products were purified as their octamethyl esters. * There have been strong indications (5, 6) that the isobacteriochlorins (structures 4 and 11) are biosynthesized as dihydro forms, and this The publication costs of this article were defrayed in part by page charge fits mechanistic expectation. When the labeled aromatic system (e.g., payment. This article must therefore be hereby marked "advertise- structure 8) is used for incorporation experiments, biological reduc- ment" in accordance with 18 U. S. C. §1734 solely to indicate this fact. tion presumably occurs to regenerate the dihydro form. 13 Downloaded by guest on September 24, 2021 14 Chemistry: Battersby et al. Proc. Natl. Acad. Sci. USA 78 (1981)

CO2 Me C02H of enzyme (blank, Exp. 6). Schmidt degradation of the acetic MeO2C acid gave methylamine carring 99% of the 14C activity present H02C 5 CO2H in the p-bromophenacyl acetate (Exp. 5), so proving specific NH NH HN C02H labeling of the acetate methyl group.

- 20 10 Thus, C-20 and its attached methyl group are extruded as MeO2C H H02C NH HN acetic acid from amacrocycle lying beyond the trimethyl system (structure 11)* on the biosynthetic pathway to cobyrinic acid 315 / C02'H (structure 13) and the original C-20 becomes the acetate car- CO2Me I/ C02H CO2H' boxyl carbon. This result contrasts with the earlier report (10) (1) A = ,IC (2) A,= 12c that C-20 is eliminated as formaldehyde. The partial structures 15 and 16, which show rings A and D (3) A = 1"C of possible late intermediates, are plausible ones from which CO2H the acetyl group could be lost; structure 15 or 16 could react C02H C as such or as a metal complex. There are excellent analogies for H Me*, cH A \. C02H H02C the second structure (16) in Eschenmoser's seminal work on

C0 synthetic model systems (11, 12), which has also demonstrated H02C N 15H *Me N HN \CO2H 20 Me 1 ready and varied formation of corrins. H02C /NH IN... HO2C / NH N- The accompanying paper from D. Arigoni's group (13) per- fectly complements our findings on the biological system, and 15 02H 15 CO2H that from A. Eschenmoser's group (14) beautifully demonstrates CO2H CO2H CO2H COH the chemical feasibility of the ring contraction.

A (4) = * =12C (10) A ="'C. * 5 CH2T (5) A =4C, =12C (6) A = 12C C = 14C We thank our colleagues Drs. E. McDonald and G. W J. Matcham A (7) = 12C, .= CH2T Enzymic MeC02H-.MeNH2 for their help. This work was supported by the National Research Coun- (8) A = "Cc, 0= CH2T fur auf Chemie I + CO2 cil of Canada and Stiftung Stipendien dem Gebiete der (9) A = 0 = 14C (to N.G.L. and A.P.) and the Nuffield Foundation, Science Research Council, and Roche Products Ltd. C02H a CO2R a Me~ ~ ~ Me H02C H C 0 2 H~~C0R0C22HCCw.C O eC02R *'2A~~~~~~~INN 2 Enzymic Me, 1. Bergmann, K.-H., Deeg, R., Gneuss, K. D., Kremler, H.-P. & ~e Muller, G. (1977) Z. PhysioL Chem. 358, 1315-1323. 2. Battersby, A. R. & McDonald, E. (1978) Bioorg. Chem. 7, MeCO2H 161-173. 15 C2H eOC.MeM 3. Battersby, A. R., Matcham, G. W. J., McDonald, E., Neier, R., CO2H CO2H CO2R C02R Thompson; M., Woggon, W.-D., Bykovsky, V. Y. & Morris, H. (11) * = C MeNH2+ CO2 (13) R = H R. (1979) J. Chem. Soc. Chem. Commun., 185-186. 4. Muller, G., Gneuss, K. D., Kriemler, H.-P., Scott, A. I. & Irwin, (12) * = l4C (14) R = Me A. J. (1979) J. Am. Chem. Soc. 101, 3655-3657. C02H 5. Muller, G., Deeg, R., Gneuss, K. D., Gunzer, G. & Kriemler, CO2H H.-P. (1979) in Vitamin B12, eds. Zagalak, B. & Friedrich, W. (de Gruyter, Berlin), pp. 279-291. HO2C 6. Battersby, A. R. (1979) in Vitamin B12, eds. Zagalak, B. & Fried- Me; MeMe'- rich, W. (de Gruyter, Berlin), pp. 217-246. 7. Lewis, N. G., Neier, R., Matcham, G. W. J., McDonald, E. & Mew, N Battersby, A. R. (1979) J. Chem. Soc. Chem. Commun., 541-542. H02iC~ H02Ct 8. Battersby, A. R., McDonald, E., Hollenstein, R., Ihara, M., Sa- toh, F. & Williams, D. C. (1977) J. Chem. Soc. Perkin Trans. 1, CO2H CO2H 166-178. 9. Battersby, A. R., Hirst, M., McCaldin, D. J., Southgate, R. & (15) (16) Staunton, J. (1968) . Che/n. Soc. C,' 2163-2172. 10. Kajiwara, M., Ho, K. S., Klein, H., Scott, A. I., Gossauer, A., FIG. 1. Structures. Engel, J., Neumann, E. & Zilch, H. (1977) Bioorg. Chem. 6, 397-402. The experiments described above were then repeated using 11. Pfalz, A., Buchler, N., Neier, R., Hirai, K. & Eschenmoser, A. [2,7,20-methyl-'4C]trimethylisobacteriochlorin (structure 12), (1977) He/v. Chim. Acta, 60, 2653-2672. which has C-20 is unla- 12. Eschenmoser, A. (1979) in Vitamin B12, eds. Zagalak, B. & Fried- the complementary labeling; namely, rich, W. (de Gruyter, Berlin), pp. 89-117. beled but the C-20 methyl group carries 14C. Again, incorpo- 13. Mombelli, L., Nussbaumer, C., Weber, H., Muller, G. & Arigoni ration into cobyrinic acid (structure 13) was accompanied by D. (1981) Proc. NatL Acad. Sci. USA 78, 9-10. formation of 59% of one equivalent of labeled acetic acid (Exp. 14. Rasetti, V., Pfaltz, A., Kratky, C. & Eschenmoser, A. (1981) Proc. 5, Table 1), which was not formed significantly in the absence NatL Acad. Sci. USA 78, 16-19. Downloaded by guest on September 24, 2021 Chemistry: Battersby et al. Proc. Natl. Acad. Sci. USA 78 (1981) 15 Table 1. Incorporation experiments withP. shermanii Incorporation Acetic acid Relative molar into cobyrinic formedc activity of Exp. Precursor Conditionsa acid (13)b, % equiv. methylamined 1 [20-14C, 2,7-methyl-3HI- Enzymice .21 >0.65' <0.1 Sirohydrochlorin (8) 2 [20-14C, 2,7-methyl-3HI- Enzymic, only 8 >0.65' - Sirohydrochlorin (8) endogenous cofactors 3 [20-14C, 2,7-methyl-3H]- Blank (no - <0.0008g Sirohydrochlorin (8) enzyme)e 4 [20-14C, 2,7-methyl-14C]- Enzymice 56 0.77 - Sirohydrochlorin (9) 5 [2,7,20-methyl-14C]_ Enzymice 37 0.59 99 Trimethylisobacteriochlorin (12) 6 [2,7,20-methyl-'4C]- Blank (no - <0.008h - Trimethylisobacteriochlorin (12) enzyme)e a For experimental method, see ref. 7 and references therein. b Isolated as its heptamethyl ester (structure 14, cobester). 'Isolated as p-bromophenacyl ester. The number of equivalents formed are calculated relative to 1.00 eqivalent for cobyrinic acid produced. d Derived from Schmidt degradation of acetic acid from the previous column and calculated relative to 100 for the acetic acid. 'The full set of cofactors (8) was added. f Endogenous acetic acid was not directly determined; the minimal values quoted are based on endogenous acetic acid found in the equivalent Exps. 4 and 5. g Calculated relative to values from Exp. 1. h Calculated relative to values from Exp. 5. Downloaded by guest on September 24, 2021