Formation of the dimethylbenzimidazole ligand of coenzyme B12 under physiological conditions by a facile oxidative cascade

Article abstract of DOI:10.1021/ol034530m

(Matrix presented) Dimethylbezimidazole, the axial ligand of vitamin B ₁₂, is synthesized from riboflavin by a two-electron oxidation, a retro-aldol condensation, and a second two-electron oxidation. This oxidative cascade readily takes place nonenzymatically under physiological conditions.

Full text of DOI:10.1021/ol034530m

ORGANIC
LETTERS
2003
Vol. 5, No. 13
2211-2213
Formation of the Dimethylbenzimidazole
Ligand of Coenzyme B12 under
Physiological Conditions by a Facile
Oxidative Cascade
,†
Lori A. Maggio-Hall,† Pieter C. Dorrestein,‡ Jorge C. Escalante-Semerena,* and
,‡
Tadhg P. Begley*
Department of Bacteriology, UniVersity of Wisconsin, 1710 UniVersity AVenue,
Madison, Wisconsin 53726-4087, and Department of Chemistry and Chemical Biology,
120 Baker Laboratory, Cornell UniVersity, Ithaca, New York 14853
escalante@bact.wisc.edu; tpb2@cornell.edu
Received March 26, 2003
ABSTRACT
Dimethylbezimidazole, the axial ligand of vitamin B12, is synthesized from riboflavin by a two-electron oxidation, a retro-aldol condensation,
and a second two-electron oxidation. This oxidative cascade readily takes place nonenzymatically under physiological conditions.
The formation of the 5,6-dimethylbenzimidazole ligand
(DMB, 4) is an unsolved problem in vitamin B12 biosynthesis,
and no gene or enzyme involved in this process has been
identified. In some bacteria (e.g., Salmonella enterica and
Propionibacterium freundenreichii), DMB is formed from
riboflavin (2), and labeling studies have demonstrated that
the C-2 carbon of DMB (4) is derived from the C-1 carbon
of the ribose moiety (2, Scheme 1).1-3 This remarkable
transformation is likely to require oxygen because S. enterica
appears to synthesize DMB only under aerobic growth
conditions.4,5
While there is no precedent for this complex chemistry,
the observations that riboflavin can be converted, in very
low yield, to a mixture of diaminobenzene (3) and DMB
(4) under strongly basic conditions6 and that diaminoben-
zenes can react readily with aldehydes to give the corre-
sponding substituted benzimidazoles7 suggest that 3 and its
† University of Wisconsin.
‡ Cornell University.
(1) Renz, P. Chemistry and Biochemistry of B12; Banerjee, R.. Ed.; John
Wiley & Sons: New York, 1999; pp 557-575.
(2) Keck, B.; M. Munder, M.; Renz, P. Arch. Microbiol. 1998, 171, 66-
68.
(3) Ho¨rig J. A.; Renz, P. Eur. J. Biochem. 1980, 105, 587-592.
(4) Keck, B.; Renz, P. Arch. Microbiol. 2000, 173 (1), 76-77.
(5) Johnson, M. G.; Escalante-Semenara, J. C. J. Biol. Chem. 1992, 267,
13302-13305.
(6) Renz, P.; Wurm, R.; Ho¨rig J. Z. Naturforsch. 1977, 32C, 523-527.
10.1021/ol034530m CCC: $25.00 © 2003 American Chemical Society
Published on Web 06/03/2003

pyoverdin,9 melanin,10 and styelsamine11 and in the hydroxy-
kynurenine-mediated cross-linking of proteins in the mam-
malian eye lens.12
Scheme 1
Several experiments were carried out that support this
mechanism and differentiate it from alternatives. When
diaminobenzene 10 is treated with ribose phosphate under
the same conditions used to oxidize 3, DMB (yields of
6-10%) is also formed in an oxygen-dependent reaction.
This supports the intermediacy of 6 (Scheme 3). In addition,
imine 5 may be intermediates and that it may be possible to
find relevant nonenzymatic chemistry for the conversion of
3 to 4. Here, we demonstrate the validity of this analysis
and propose that the conversion of 3 to DMB can occur to
a large extent by a facile nonenzymatic oxidative cascade
(Scheme 2).
Scheme 3
Scheme 2
when this reaction is carried out with D-[13C-1]ribose, NMR
analysis of the resulting DMB demonstrated that the 13C was
localized at the C-2 position. This labeling pattern is identical
to the biosynthetic labeling pattern [1] suggesting that the
biomimetic and the biosynthetic conversion of 6 to DMB
are occurring by the same mechanism.
HPLC analysis of the 10 + 11 reaction mixture (Scheme
3) identified DMB and an additional compound as the major
reaction products (Figure 1). Structure 12 was assigned to
DMB (4) was formed when a synthesized sample of 3
was treated with catalytic amounts of potassium ferricyanide,
a mild oxidizing agent known to oxidize amino phenols to
the corresponding quinone imines. This supports the proposal
that DMB can be formed by the nonenzymatic oxidation of
3.8 Aerobic oxidation of 3 under physiological conditions
(pH 7, 37 °C) also gave DMB. However, the reaction was
approximately 30 times slower than the ferricyanide-mediated
oxidation. DMB was not formed after 24 h in argon-purged
buffer, demonstrating that its formation requires oxygen.
Having established the feasibility of converting 3 to 4,
we next turned our attention to the mechanism of this
complex oxidation process. Our current mechanistic proposal
is outlined in Scheme 2. In this proposal, oxidation of the
electron-rich diamine 3 to the bisimine 5 followed by a
tautomerization gives 6. Cyclization to give 7 followed by a
second two-electron oxidation gives 8. Aromatization of 8
by extrusion of 9 gives 4. Analogous oxidative cascades, in
which electron-rich benzenes undergo complex reactions,
have been proposed in the biosynthesis of actinomycin,8
Figure 1. HPLC analysis of the reaction shown in Scheme 3.
this compound on the basis of its MS, UV, and NMR spectra.
Compound 12 was stable and did not undergo conversion
to DMB (Scheme 4). The detection of 12 provides support
(9) Dorrestein, P. C.; Poole K.; Begley T. P. Org. Lett. 2003, 5, 2215-
2218.
(10) Fatibello-Filho O.; Vieira, I. C. Analyst 1997, 122, 345-350.
(11) Skyler, D.; Heathcock, C. H. Org. Lett. 2001, 3, 4323-4324.
(12) Aquilina, J. A.; Carver, J. A.; Truscott, R. J. W. Biochemistry 2000,
39, 16176-16184.
(7) Singh, M. P.; Sasmal S.; Lu, W.; Chatterjee, M. N. Synthesis 2000,
10, 1380-1390.
(8) Barry, C.; Nayar, P.; Begley, T. P. Biochemistry 1989, 28, 6323.
2212
Org. Lett., Vol. 5, No. 13, 2003

can be explained by the observation that this compound
undergoes slow H/D exchange with water under the reaction
conditions (with a rate of 0.013 ( 0.001 h-1).
Scheme 4
The observation of a facile conversion of 3 to DMB does
not exclude the possibility that this conversion is enzyme
catalyzed in vivo, and a previous stereochemical study
supports this. In that study, treatment of a 75:25 mixture of
1′-R/1′-S riboflavin with a cell free extract from Propioni-
bacterium shermanii gave DMB with 27% retention of
deuterium. In the complementary experiment, a 25:75
mixture of 1′-R/1′-S riboflavin gave 46% retention of
deuterium.13 While it is unclear as to why the latter reaction
shows only 46% retention of label, these experiments suggest
that the pro-S hydrogen at C1′ of riboflavin is preferentially
retained in biosynthesized DMB.13 However, the stereose-
lectivity of this reaction does not prove that the formation
of DMB is enzyme catalyzed because the stereochemistry
of the deprotonation of 5 may be controlled by the adjacent
chiral centers on the ribose. To determine if such asymmetric
induction is occurring, we have determined the stereochem-
ical outcome of the nonenzymatic reaction described here.
Aerobic oxidation of a 72:28 mixture of 1′-R/1′-S [2H]-3 gave
DMB with 64% retention of deuterium after 5 h, and
oxidation of a 28:72 mixture of 1′-R/1′-S [2H]-3 gave DMB
with 26% retention of deuterium. These experiments suggest
that the pro-R hydrogen at C1 of 3 is preferentially retained
in DMB generated in the model reaction. This is the opposite
stereochemical preference to that observed for the biosyn-
thesis of DMB and suggests that the in vivo conversion of
3 to DMB, while facile, is likely to be enzyme catalyzed.
One possibility is that the oxidative cascade that converts 3
to 4 takes place in the active site of the putative hydrolase
that converts 2 to 3 (Scheme 1).
for the intermediacy of 8 because 12 is likely to be formed
by the aromatization of 8. When the C-2 proton is replaced
with a methyl group, as occurs in the condensation of 10
with 3-hydroxy 2-butanone, 2-methyl-DMB is the only
isolated product.
We also considered an alternative pathway for the forma-
tion of DMB in which 9 is lost from 5 rather than from 8
(Scheme 5). The detection of 12 as a byproduct argues
Scheme 5
Acknowledgment. This work was supported by NIH
Grant No. GM40313 to J.C.E.-S and by a grant from the
Petroleum Research Fund to T.P.B.
against this mechanism. In addition, when the reaction of
ribose with 10 is run in D2O, there is little incorporation of
deuterium into 4 (incorporation rate of 0.016 ( 0.001 h-1),
and when the reaction is run using 1-[2H]-D-ribose in H2O,
there is little incorporation of protium into 4. The absence
of a high level of exchange with solvent in these two
experiments demonstrates that 6 is not converted to 5 and
therefore that the reaction in Scheme 5 is not the major
pathway to 4. The low level of H/D exchange detected in 4
Supporting Information Available: Synthesis, spectra,
and methods. This material is available free of charge via
the Internet at http://pubs.acs.org.
OL034530M
(13) Lingens, B.; Schild, T. A.; Vogler, B.; Renz, P. Eur. J. Biochem.
1992, 207, 981-985.
Org. Lett., Vol. 5, No. 13, 2003
2213