Molybdopterin biosynthesis: Trapping an unusual purine ribose adduct in the MoaA-catalyzed reaction

Article abstract of DOI:10.1021/ja4041048

MoaA/MoaC catalyze a remarkable rearrangement reaction in which guanosine-5′-triphosphate (GTP) is converted to cyclic pyranopterin monophosphate (cPMP). In this reaction, the C8 of GTP is inserted between the C2′ and the C3′ carbons of the GTP ribose. Previous experiments with GTP isotopomers demonstrated that the ribose C3′ hydrogen atom is abstracted by the adenosyl radical. This led to a novel mechanistic proposal involving an intermediate with a bond between the C8 of guanine and C3′ of the ribose. This paper describes the use of 2′,3′-dideoxyGTP to trap this intermediate.

Full text of DOI:10.1021/ja4041048

Communication
pubs.acs.org/JACS
Molybdopterin Biosynthesis: Trapping an Unusual Purine Ribose
Adduct in the MoaA-Catalyzed Reaction
Angad P. Mehta,† Sameh H. Abdelwahed,† and Tadhg P. Begley*
Department of Chemistry, Texas A&M University, College Station, Texas 77842, United States
S
* Supporting Information
ABSTRACT: MoaA/MoaC catalyze a remarkable rear-
rangement reaction in which guanosine-5′-triphosphate
(GTP) is converted to cyclic pyranopterin mono-
phosphate (cPMP). In this reaction, the C8 of GTP is
inserted between the C2′ and the C3′ carbons of the GTP
ribose. Previous experiments with GTP isotopomers
demonstrated that the ribose C3′ hydrogen atom is
abstracted by the adenosyl radical. This led to a novel
mechanistic proposal involving an intermediate with a
bond between the C8 of guanine and C3′ of the ribose.
This paper describes the use of 2′,3′-dideoxyGTP to trap
this intermediate.
oaA/MoaC catalyzes the first step in molybdopterin
M
biosynthesis, converting GTP, 1, to cyclic pyranopterin
monophosphate (cPMP, 2).1 MoaA/MoaC catalyzes a
remarkable rearrangement reaction in which the C8 of GTP,
1, is inserted into the ribose C2′−C3′ bond (Figure 1).2
Figure 2. Mechanistic proposal for the MoaA/MoaC-catalyzed
reaction.
radical-generating antibiotics, addition of a C3′ centered radical
to C8 of a purine has never been reported.5−7 The successful
trapping of an analog of 8 would provide a critical test of the
proposed mechanism. In this communication, we describe the
use of 2′,3′-dideoxy-GTP 14 as a substrate analog to achieve
this trapping.
LCMS analysis of the MoaA/2′,3′-dideoxyGTP 14 reaction
mixture demonstrated the formation of a new product eluting
at 19.6 min (Figure 3A). ESI-MS (positive mode) analysis of
this product demonstrated that its mass [M+H]+ was 490.0 Da,
2 Da less than the [M+H]+ of the substrate. We considered two
possible structures for this product (16 and 20, Figure 3D). To
test for the formation of 16, the enzymatic product was
dephosphorylated by phosphatase treatment and compared to
an authentic sample of 17 (synthesis described in Supporting
Information (SI)) by LCMS. Figure 3C and Figures S28 and
S29 demonstrate that the retention time of 17 and its MS/MS
spectrum are different from those of the enzymatic product.
This excludes the possibility that MoaA catalyzes the formation
of 16 from 14.
Figure 1. Early reactions in molybdopterin biosynthesis: MoaA/
MoaC-catalyzed transformation of GTP (1) to cyclicpyranopterin
monophosphate (2). Color shows the atom transfer pattern derived
from previous labeling studies.2,4,8
Previous experiments with GTP isotopomers demonstrated
that the ribose C3′ hydrogen atom is abstracted by the adenosyl
radical. This led to a mechanistic proposal involving an
intermediate with a bond between the C8 of guanine and C3′
of the ribose (Figure 2).3 In this proposal, the 5′-dA radical 3,
generated by reductive cleavage of S-Adenosyl methionine
(AdoMet), abstracts the 3′ hydrogen atom from GTP to give 4,
which then undergoes cyclization to give 6. Reduction of this
radical by the purine liganded iron sulfur cluster to 7 followed
by hydrolysis to 8 and a benzilic-like rearrangement to 9
completes the insertion of the purine carbon into the ribose.
Ring opening of 9 followed by dehydration of 10 and a
conjugate addition gives 12. Cyclization to 13 followed by a
final tautomerization completes the formation of the reaction
product 2.3 While ribose and deoxyribose radicals have been
extensively studied in the context of enzymes such as
ribonucleotide reductase and DNA damage by radiation or
The MS/MS spectrum of 17 showed fragmentation of the N-
glycosyl bond to form guanine (152.0 Da, Figure S23). This
fragmentation was not detected for the dephosphorylated
enzymatic product (Figure S24) suggesting an additional bond
Received: April 24, 2013
© XXXX American Chemical Society
A
dx.doi.org/10.1021/ja4041048 | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Communication
Figure 3. Analysis of the product formed when 2′,3′-dideoxyGTP 14 is treated with MoaA. (A) LC analysis of the MoaA reaction mixture and
controls. Red trace is the full reaction where all the components are present. Green, pink, black, and blue traces are for reaction mixtures where either
2′,3′-dideoxyGTP, dithionite, SAM, or MoaA is absent respectively. The unidentified signal at 19.9 min has the same [M+H]+ as 2′,3′-dideoxyGTP.
(B) MS of the compound eluting at 19.6 min. The [M+H]+ is 490.0 Da, 2 Da less than the [M+H]+ of the substrate 14. (C) Extracted ion
chromatograms for [M+H]+ = 250.0 Da demonstrate that compound 17 is different from the dephosphorylated enzymatic product. (D) Mechanistic
analysis to suggest possible products formed from 2′,3′-dideoxyGTP 14. (E) Extracted ion chromatograms for [M+H]+ = 268.1 Da demonstrate that
the [M+H]+ = 268.1 Da signal is seen only in reaction mixtures with all the components present after phosphatase treatment and acid hydrolysis.
The two compounds with [M+H]+ = 268.1 Da are most likely a consequence of hemiacetal isomerization during the acid hydrolysis. (F) MS of the
compound formed by acid hydrolysis of the dephosphorylated enzymatic product ([M+H]+ = 268.1 Da). (G) Extracted ion chromatograms for [M
+H]+ = 268.1 Da demonstrate that the derivatized enzymatic product has the same mass and coelutes with compound 22.
between the sugar and the base consistent with 20.
Unfortunately, NMR characterization of the enzymatic product
was not possible. Only small quantities were formed because
2′,3′-dideoxyGTP 14 is a poor substrate (there is no evidence
for enzyme modification by MS analysis). An identification
strategy, involving comparison with a reference compound, was
therefore adopted. If 19 is the product structure, oxidation and
dephosphorylation, followed by N-glycosyl bond cleavage,
would give 22 which can be synthesized as a reference (see SI).
To test this, the enzymatic product was derivatized by treating
the MoaA reaction mixture with phosphatase followed by acid
at 65 °C for 2 h. The extracted ion chromatograms for 268.1
Da (22 + H+ + H2O) are shown in Figure 3E and demonstrate
that the acid treatment generates isomers of the product as
expected from ring opening of the hemiacetal. The derivatized
enzymatic product coelutes with the synthetic reference (also
generates isomers on acid treatment (see SI) and is identical by
LCMS analysis). Thus, the enzymatic reaction of 2′,3′-dideoxy-
Figure 4. Mechanistic proposal for the MoaA catalyzed reaction and
GTP 14 supports the formation of the remarkable reaction
subsequent derivatization of 2′,3′-dideoxyGTP 14.
intermediate 7 in the MoaA-catalyzed reaction.
A mechanistic proposal for the formation of 22 is described
in Figure 4. The 5′-dA radical 4 abstracts the 3′ hydrogen atom
B
dx.doi.org/10.1021/ja4041048 | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Communication
from 2′,3′-dideoxyGTP 14 to give 15, which then undergoes
cyclization to give 18. Reduction of this radical by the purine
liganded iron sulfur cluster gives 19 which on aerobic oxidation
results in the formation of 20. Phosphatase treatment of 20
followed by acid hydrolysis gives 22 as a mixture of isomers.
While this paper was under review, an important new paper
on the mechanism of MoaA/MoaC appeared in the literature.9
This paper suggests that compound 7 is the product of MoaA
and demonstrates that 7 is a substrate for MoaC. Both papers
provide support for the unprecedented bond formation
between C3′ of the ribose and C8 of the purine using different
but complementary approaches.
ASSOCIATED CONTENT
* Supporting Information
The experimental procedures are described in the Supporting
Information. This material is available free of charge via the
Internet at http://pubs.acs.org.
■
S
AUTHOR INFORMATION
Corresponding Author
begley@chem.tamu.edu
■
Author Contributions
†A.P.M. and S.H.A. contributed equally.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
This research was supported by a grant from the National
Institutes of Health (DK44083) and by the Robert A. Welch
Foundation (A-0034).
REFERENCES
■
(1) Fisher, M.; Thony, B.; Leimkuhler, S. Comprehensive Natural
Products II Chemistry and Biology 2010, 7, 628−640.
(2) Johnson, J.; Wuebbens, M.; Rajagopalan, K. J. Biol. Chem. 1989,
264 (23), 13440−13447.
(3) Mehta, A.; Hanes, J.; Abdelwahed, S.; Hilmey, D.; Hannzelmann,
̈
P.; Begley, T. Biochemistry 2013, 52, 1134−36.
(4) Lees, N. S.; Hazelmann, P.; Hernandez, H. L.; Subramanian, S.;
̈
Schindelin, H.; Johnson, M. K.; Hoffman, B. M. J. Am. Chem. Soc.
2009, 131 (26), 9184−91850.
(5) Stubbe, J.; Nocera, D.; Yee, C.; Chang, M. Chem. Rev. 2003, 103
(6), 2167−2202.
(6) Chatgilialoglu, C.; Ferreri, C.; Terzidis, M. Chem. Soc. Rev. 2011,
40, 1368−1382.
(7) Greenberg M. D., Ed. Radical and Radical Ion Reactivity in
Nucleic Acid Chemistry. Wiley Series of Reactive intermediates. Chem.
Biol. 2009, 1−443.
(8) Wuebbens, M.; Rajagopalan, K. J. Biol. Chem. 1995, 270, 1082−
1087.
(9) Hover, B.; Loksztejn, A.; Ribeiro, A.; Yokoyama, K. J. Am. Chem.
Soc. 2013, 135 (18), 7019−7032.
C
dx.doi.org/10.1021/ja4041048 | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX