Thiamin biosynthesis in eukaryotes: Characterization of the enzyme-bound product of thiazole synthase from Saccharomyces cerevisiae and its implications in thiazole biosynthesis

Article abstract of DOI:10.1021/ja061413o

The biosynthesis of thiamin pyrophosphate in eukaryotes is different from the prokaryotic biosynthesis and is poorly understood to date. Only one thiazole biosynthetic gene has been identified (Thi4 in Saccharomyces cerevisiae). Here we report the identification and characterization of a Thi4-bound metabolite that consists of the ADP adduct of 5-(2-hydroxyethyl)-4-methylthiazole-2-carboxylic acid. The unexpected structure of this compound yields the first insights into the mechanism of thiamin thiazole biosynthesis in eukaryotes. Copyright

Full text of DOI:10.1021/ja061413o

Published on Web 05/11/2006
Thiamin Biosynthesis in Eukaryotes: Characterization of the Enzyme-Bound
Product of Thiazole Synthase from Saccharomyces cerevisiae and Its
Implications in Thiazole Biosynthesis
Abhishek Chatterjee, Christopher T. Jurgenson, Frank C. Schroeder, Steven E. Ealick,* and
Tadhg P. Begley*
Department of Chemistry and Chemical Biology, Cornell UniVersity, Ithaca, New York 14853
Received February 28, 2006; E-mail: tpb2@cornell.edu
Thiamin pyrophosphate is an essential cofactor in all living
systems, where it plays a key role in amino acid and carbohydrate
metabolism.1,2 The biosynthesis of thiamin involves separate
syntheses of the thiazole and the pyrimidine moieties, which are
then linked to form the cofactor.3,4 The biosynthesis of the thiazole
moiety in bacteria has been elucidated and involves a complex
oxidative condensation where five different enzymes interact with
three different substrates (1-deoxy-D-xylulose-5-phosphate, cysteine,
and glycine or tyrosine).4 Labeling studies in Saccharomyces
cereVisiae have demonstrated that the thiamin thiazole is biosyn-
thesized from an unidentified five-carbon carbohydrate, glycine,
and cysteine.5-7 Only one thiazole biosynthetic enzyme has been
identified (Thi4).8 This biosynthesis has not yet been reconstituted
in vitro. In this Communication, we report the identification and
characterization of the tightly bound reaction product at the active
site of the S. cereVisiae thiazole synthase. The unanticipated
structure of this metabolite provides the first insights into the
mechanism of thiazole biosynthesis in eukaryotes.
The S. cereVisiae thiazole synthase (Thi4) was overexpressed
and purified. The overexpressed protein failed to catalyze the
formation of the thiazole moiety from cysteine (or sulfide), glycine,
and a variety of C5 carbohydrates. Thi4 also failed to complement
an Escherichia coli thiazole biosynthetic mutant (ThiF-). This
suggested that additional enzymes may be required for thiazole
biosynthesis in S. cereVisiae, or that the correct substrates have
not yet been identified.
In the process of searching for the activity of Thi4, we found
that denaturing the protein by heat releases four major enzyme-
bound metabolites (Figure 1a). Each of these four species had an
absorption maximum at 260 nm. This absorption, combined with
sequence analysis that predicted an adenine dinucleotide binding
motif, suggested that the Thi4 metabolites might be adenylated.
Treating the released metabolites with nucleotide pyrophosphatase
and demonstrating the production of AMP by HPLC analysis
confirmed this.
The component with retention time of 14 min (compound A)
was sufficiently stable for purification by HPLC; the components
eluting after 11 (compound D) and 12 min (compound C) were
not and therefore could not be characterized further. The intensity
A was again treated with nucleotide pyrophosphatase to generate
of compound B (eluting at 12.5 min) was low and varied between
Figure 1. Identification of Thi4-bound small molecules by HPLC (a).
Analysis of the HPLC-purified peak A by UV spectroscopy (b), by 1H NMR
spectroscopy (c), and by negative-mode ESI-MS (d).
m/z ) 297). This demonstrated that the missing atoms from structure
1 (Figure 1d) have a combined mass of 103 Da. Since the thiamin
thiazole is biosynthesized from glycine and contains a sulfur atom,
this extra mass is consistent with structure 2. Extensive fragmenta-
tion studies were carried out on compound A, and the results were
consistent with this structural assignment.
To confirm the presence of the thiazole moiety in 2, compound
experiments. 1H NMR (Figure 1c) and COSY analysis of compound
A confirmed the presence of the adenosyl moiety. Additionally,
two strongly coupled methylene protons (2.7 and 3.8 ppm) and an
isolated methyl (2.05 ppm) were observed. HMBC analysis
demonstrated that the methyl group was connected to one of the
methylene groups by two aromatic/olefinic carbons. This suggested
partial structure 1.
AMP and the putative thiazole 3 (Figure 2a). Treatment of this
reaction mixture with hydroxymethyl pyrimidine pyrophosphate and
thiamin phosphate synthase (ThiE)4,9 should yield 4, provided that
ThiE can process 3 as a substrate. Compound 4 would then undergo
rapid decarboxylation to give thiamin phosphate 5, which can be
detected by HPLC analysis with high sensitivity after its oxidation
to the intensely fluorescent thiochrome phosphate 6. The results of
this experiment are shown in Figure 2c and demonstrate the
production of thiochrome phosphate, confirming the assigned
Negative-mode ESI-MS analysis of compound A identified the
molecular mass to be 596 Da (monoanionic m/z ) 595; dianionic
9
7158
J. AM. CHEM. SOC. 2006, 128, 7158-7159
10.1021/ja061413o CCC: $33.50 © 2006 American Chemical Society

C O M M U N I C A T I O N S
Figure 4. Proposed mechanism for thiazole biosynthesis in eukaryotes.
All attempts to release the product without denaturing the enzyme
have failed, and we do not yet understand the high stability of this
enzyme product complex. The observation that the Neurospora
crassa Thi4 forms a stable complex with a cyclophilin12,13 suggests
that a proline-isomerization-mediated conformational change may
be required for product release.
Thiamin phosphate synthase, another enzyme on the thiamin
biosynthetic pathway in prokaryotes, also contains tightly bound
reaction product.9 These observations suggest that the analysis of
a protein of unknown function for tightly bounds metabolites can
be a productive strategy for functional assignment. The character-
ization of the reaction product bound at the active site of the yeast
thiazole synthase yields the first insights into the biosynthesis of
the thiamin thiazole in eukaryotes and represents yet another
example of a non-redox function for NAD.11,14
Figure 2. Confirmation of the structure 3 by its enzymatic conversion to
thiamin phosphate 5. (a) Scheme showing the reactions performed. (b) HPLC
analysis of the cleavage reaction of 2 to 3 and AMP. The blue trace shows
the control reaction where no pyrophosphatase has been added to 2; the
purple trace shows the cleavage reaction of 2 by the pyrophosphatase. (c)
HPLC detection of the thiochrome phosphate 6 produced from 2. The blue
trace shows the control reaction with no HMP-PP added; the purple trace
shows the actual coupling reaction.
Figure 3. Active-site structure of Thi4 (PDB code 2GJC). (a) The structure
of compound 2 and its interactions with the enzyme. (b) Electron density
confirming the thiazole carboxylic acid moiety of 2. The density is a
composite omit map contoured at the 1σ level.
Acknowledgment. We thank Dr. Glaucius Oliva (Universidade
de Sao Paulo) for providing us with the pdb file of Thi1 prior to
submission to the Protein Data Bank, Cynthia Kinsland (Cornell
Protein Overexpression and Characterization Facility) for cloning
Thi4 and Thi1, and Dr. Joo-Heon Park for running preliminary
assays with Thi4. This research was funded by NIH grants
DK44083 (to T.P.B.) and DK67081 (to S.E.E.).
thiazole structure. It also shows that ThiE can process the
carboxylated thiazole phosphate 3.
The recently solved structure of Thi4 supports this analysis. This
structure shows that 2 is present at the active site of Thi4 and that
the carboxylic acid of the thiazole forms hydrogen-bonding and
electrostatic interactions with Arg301 (Figure 3).
A crystal structure of Thi1 (the Arabidopsis orthologue of Thi4)
has recently been deposited in the Protein Data Bank (1RP0). This
structure also shows a bound adenylated intermediate, but the
electron density does not allow for an unequivocal assignment of
the thiazole structure. Thi1 was overexpressed as a soluble protein
in E. coli, and HPLC analysis revealed the existence of compound
2 as well as the metabolites B and D. This mixture of enzyme-
bound metabolites might explain the ambiguous electron density
corresponding to the thiazole moiety in the structure.
Supporting Information Available: Detailed experimental pro-
cedures for overexpression and protein purification, enzymatic assays,
data from 2D NMR experiments, MS fragmentation patterns, and the
results from additional HPLC assays. This material is available free of
charge via the Internet at http://pubs.acs.org.
References
(1) Butterworth, R. F. Nutr. Res. ReV. 2003, 16, 277-283.
(2) Jordan, F. Nat. Prod. Rep. 2003, 20, 184-201.
(3) Spenser, I. D.; White, R. L. Angew. Chem., Int. Ed. Engl. 1997, 36, 1032-
1046.
(4) Settembre, E.; Begley, T. P.; Ealick, S. E. Curr. Opin. Struct. Biol. 2003,
13, 739-747.
The identification of 2 at the active site of Thi4 suggests that
the ADP-pentose moiety of 2 is probably derived from NAD and
that the chemistry involved in the early steps of thiazole formation
may be similar to the chemistry involved in ADP ribosylation.10,11
One possible mechanism is outlined in Figure 4. Cleavage of the
N-glycosyl bond of NAD 7 followed by ring opening, tautomer-
ization, and loss of water gives 11. Imine formation, tautomeriza-
tion, sulfide addition, and cyclization gives 15. Elimination of two
water molecules followed by a tautomerization completes the
formation of the thiazole moiety.
(5) White, R. L.; Spenser, I. D. J. Am. Chem. Soc. 1979, 101, 5102-5104.
(6) White, R. L.; Spenser, I. D. Biochem. J. 1979, 179, 315-325.
(7) White, R. L.; Spenser, I. D. J. Am. Chem. Soc. 1982, 104, 4934-4943.
(8) Praekelt, U. M.; Byrne, K. L.; Meacock, P. A. Yeast 1994, 10, 481-90.
(9) Reddick, J. J.; Nicewonger, R.; Begley, T. P. Biochemistry 2001, 40,
10095-10102.
(10) Ueda, K.; Hayaishi, O. Annu. ReV. Biochem. 1985, 54, 73-100.
(11) Berger, F.; Ramirez-Hernandez, M. H.; Ziegler, M. Trends Biochem. Sci.
2004, 29, 111-118.
(12) Faou, P.; Tropschug, M. J. Mol. Biol. 2003, 333, 831-844.
(13) Faou, P.; Tropschug, M. J. Mol. Biol. 2004, 344, 1147-1157.
(14) Ziegler, M.; Niere, M. Biochem. J. 2004, 382, e5-e6.
JA061413O
9
J. AM. CHEM. SOC. VOL. 128, NO. 22, 2006 7159