Search tips
Search criteria 


Logo of jbacterPermissionsJournals.ASM.orgJournalJB ArticleJournal InfoAuthorsReviewers
J Bacteriol. 2008 March; 190(6): 2056–2064.
Published online 2008 January 11. doi:  10.1128/JB.01094-07
PMCID: PMC2258874

Functional Analysis of the Mycobacterium tuberculosis FAD-Dependent Thymidylate Synthase, ThyX, Reveals New Amino Acid Residues Contributing to an Extended ThyX Motif[down-pointing small open triangle]


A novel FAD-dependent thymidylate synthase, ThyX, is present in a variety of eubacteria and archaea, including the mycobacteria. A short motif found in all thyX genes, RHRX7-8S, has been identified. The three-dimensional structure of the Mycobacterium tuberculosis ThyX enzyme has been solved. Building upon this information, we used directed mutagenesis to produce 67 mutants of the M. tuberculosis thyX gene. Each enzyme was assayed to determine its ability to complement the defect in thymidine biosynthesis in a ΔthyA strain of Escherichia coli. Enzymes from selected strains were then tested in vitro for their ability to catalyze the oxidation of NADPH and the release of a proton from position 5 of the pyrimidine ring of dUMP. The results defined an extended motif of amino acids essential to enzyme activity in M. tuberculosis (Y44X24H69X25R95HRX7S105XRYX90R199 [with the underlined histidine acting as the catalytic residue and the underlined serine as the nucleophile]) and provided insight into the ThyX reaction mechanism. ThyX is found in a variety of bacterial pathogens but is absent in humans, which depend upon an unrelated thymidylate synthase, ThyA. Therefore, ThyX is a potential target for development of antibacterial drugs.

The capacity to synthesize thymidine has been recognized as an essential function in all free-living organisms for many decades (13, 14, 19). This dependence has led to the development of a large number of inhibitors of thymidine synthesis. For most of this time, research focused on the enzyme that is common in eukaryotes and many eubacteria, the thymidylate synthase enzyme ThyA (EC, which reductively methylates dUMP in tandem with dihydrofolate reductase DHFR (EC (3). More recently, a novel FAD-dependent thymidylate synthase, ThyX (EC, which transfers the methyl group without oxidation of the tetrahydrofolate carrier, has been identified in a variety of eubacteria and archaea (9, 22, 23). Different from ThyA proteins, ThyX proteins use a redox-activated FAD cofactor to form dTMP (Fig. (Fig.1A),1A), but details of the reaction remain poorly understood. This is exemplified by the fact that four different reaction mechanisms have been proposed (Fig. (Fig.1B).1B). Structural data exist for several ThyX proteins; however, since active site configurations in different structures vary we undertook an exhaustive mutagenesis study to identify amino acid residues indispensable for Mycobacterium tuberculosis ThyX activity.

FIG. 1.
ThyX reactions and mechanisms. (A) Reactions catalyzed by ThyX enzymes. (B) Four currently proposed reaction mechanisms for ThyX, as outlined in reference 4.

In most organisms either the ThyA/DHFR couple or ThyX is utilized for thymidine biosynthesis. Like most eukaryotes, humans depend upon ThyA, so there has been special interest in the enzyme of ThyX-dependent pathogens such as Helicobacter pylori, Campylobacter jejuni, Borrelia burgdorferi, and Treponema pallidum (see reference 22 for a review). ThyX in these organisms is an attractive drug target (4, 6, 15, 23). Eubacterial species in the corynebacterium group are unusual because they have retained the genes for both the ThyA/DHFR and the ThyX pathways (22, 23). Within this group the sequences of the thyA genes are very highly conserved, and the thyX genes show almost complete identity (30). According the high-density transposon mutagenesis study of M. tuberculosis H37Rv by Sassetti et al., thyX is an essential gene in M. tuberculosis whereas thyA is not (28).

The structures of the ThyX protein from Thermotoga maritima (16), M. tuberculosis (27), and Paramecium bursaria chlorella virus-1 (10) have been solved recently, and the basic structures are very similar. All three enzymes are homotetramers, and they are structurally and evolutionarily unrelated to any of the ThyA enzymes that have been studied (22, 23). The genes that encode the putative ThyX enzyme have been identified in organisms from several hundred species, representing more than 70 genera, and there are some highly conserved amino acids that have been termed the ThyX motif (RHRX7-8S) (17, 18). Moreover, a limited study of essential residues in the M. tuberculosis ThyX enzyme confirmed that they were conserved in the orthologous H. pylori enzyme (17, 27).

Our goal in this study was to use directed mutagenesis of the M. tuberculosis thyX gene to identify a much wider range of amino acid residues essential to ThyX enzymatic function in cells and in two distinct in vitro assays. Essentially, we wanted to supplement the crystallographic studies with in-depth knowledge about the importance of individual amino acid residues. The various crystal structures available for ThyX proteins demonstrate considerable structural plasticity at the active configuration and also provide a static image of the active site during catalysis. However, in vivo catalysis is a dynamic process wherein configurational changes rearrange the active site, allowing different amino acid residues to access the substrates. We wanted to focus on identification of amino acid residues that are crucial to catalysis, which may or may not have been identified in the crystal structure. The strong conservation of the structure and the sequence of the thyX genes between M. tuberculosis and human pathogens that depend only upon ThyX suggest that the key residues identified in this work can be usefully extrapolated to the ThyX enzymes in those pathogenic bacteria as well.


Expression vector cloning.

thyX (Rv2754c) from M. tuberculosis H37Rv was cloned into a pUC18 vector for complementation analysis and into a pET24d vector with a six-His tag for high-level protein expression and purification as described previously (27). The cloned vectors are designated pUC18-ThyX and pET24d-ThyX, respectively.


To perform mutagenesis, primers that flank each of the target sites were designed (see Table S1 in the supplemental material). The primers were designed to anneal to the coding strand of DNA, and as such only one primer was necessary for each point mutation. Mutations were generated using a QuickChange multisite-directed mutagenesis kit (Stratagene) according to the manufacturer's instructions. Potential mutant plasmids were sequenced completely on both strands to verify the point mutations. All mutations were made in thyX genes cloned into both the pUC18-ThyX and pET24d-ThyX vectors.


An Escherichia coli strain lacking endogenous ThyA activity, χ2913 (ΔthyA572 recA56), was graciously provided by R. Thompson, University of Glasgow, Scotland, United Kingdom (21). This thyA-negative E. coli strain is incapable of growth on M9 minimal medium plates without supplementation with thymidine. The χ2913 strain was transformed with each pUC18-ThyX vector, containing either wild-type or mutated thyX, by use of the CaCl2 method (5). Transformants were selected on LB agar plates with 100 μg/ml of ampicillin and grown at 37°C. Colonies from each of the thyX mutants were streaked onto LB agar plates with ampicillin and grown overnight at 37°C. Each of the different ThyX mutated strains was again streaked onto two different M9 minimal medium plates, one without thymidine and the other with 50 μg/ml of thymidine. Colonies were counted after 2 days at 37°C.

Protein expression and purification.

E. coli BL21(DE3)/pLysS strains containing the six-His-tagged plasmids carrying either the wild-type or mutant M. tuberculosis thyX genes were grown overnight at 37°C in 50 ml of LB medium containing 30 μg/ml of kanamycin and 35 μg/ml of chloramphenicol. Five milliliters of this overnight culture was added to 500 ml of LB containing the same antibiotics and grown to an absorbance of 0.6 at 600 nm. Cells were induced with 1 mM IPTG (isopropyl-β-d-thiogalactopyranoside) for 3 h at 37°C. Cells were pelleted and lysed with lysozyme (1 mg/ml) and 1 mg/ml Pefabloc SC (1 mg/ml) (both from Sigma) in 1× lysis equilibration wash buffer (3 ml solution/100 ml cells) from a Protino Ni-TED protein purification kit (Macherey-Nagel). Following sonication, lysates were spun at 18,000 × g for 1 h at 4°C. Supernatants were extracted and loaded into preequilibrated columns, and proteins were eluted from the Ni-TED columns provided in the Protino kits according to the manufacturer's instructions. To remove the imidazole from the eluates, eluted protein fractions were dialyzed overnight at 4°C using D-tube dialyzers (Midi molecular weight cutoff 3.5 kDa; EMD Biosciences) in a volume of 50 mM Tris-HCl, 10% glycerol, and 200 mM NaCl at pH 7.5, with a 1,000-fold excess of the protein sample.

Determination of higher-order protein complexes.

To investigate whether mutant proteins were able to form higher-order complexes similarly to wild-type ThyX, 2- to 3-μg aliquots of purified His-tagged ThyX proteins were analyzed on 4 to 20% gradient gels in a buffer containing 25 mM Tris-HCl and 250 mM glycine at pH 8.0 without any detergents or reducing agents added. The predicted pI of the M. tuberculosis ThyX protein is 6.51. We used a high-molecular-weight calibration kit for native electrophoresis purchased from Amersham as markers on the native gels. Each isolate was also analyzed on reducing 8% sodium dodecyl sulfate (SDS)-polyacrylamide gels to determine the size of the monomer. Precast polyacrylamide gels were purchased from NuSep.

NADPH oxidation assays.

To measure the capacity of the mutant ThyX enzymes to oxidize NADPH, reactions were carried out in a total volume of 200 μl with 14 to 15 μg eluted ThyX protein added to a 50 mM HEPES, pH 7.5, solution containing 1 mM MgCl2, 50 μM dUMP, and 12.5 μM FAD. Reactants were incubated for 5 min at room temperature before the oxidation reaction was started by addition of 0.2 mM β-NADPH. Oxidation of NADPH was measured by a decrease in absorbance at 340 nm in a Labsystems Multiscan 96-well plate reader. The amount of NADP+ produced by each ThyX enzyme was determined after the reaction proceeded for 3 min by use of an extinction coefficient (epsilon340) of 6,220 M−1 cm−1 (24).

Tritium release assays.

To measure “deprotonation” of [5-3H]dUMP in vitro, tritium release assays were performed as described previously (17) by using 5 to 10 μg ThyX protein, incubating at 37°C for 20 min, and then stopping the reactions with activated charcoal. Reactions were performed in a total volume of 50 μl, and the mixtures contained 0.2 mM 5,10-methylene-5,6,7,8-tetrahydrofolate, 0.4 mM FAD, 0.1 mM β-NADPH, 0.1 mM β-NADH, 20 μM nonradioactive dUMP, 0.6 μCi (0.92 μM) [5-3H]dUMP (specific activity, 13.6 Ci/mmol; Moravek), 10 μM MgCl2, and 50 mM HEPES, pH 7.5.


All chemical reagents were purchased from Sigma except for the methylenetetrahydrofolate, which was generously provided to us by R. Moser (Merck Eprova AG).


To provide mechanistic insight into the NADPH oxidation and deprotonation activities of ThyX proteins, we examined in depth the previously defined ThyX motif RHRX7-8S (17, 18). We extended this analysis by testing residues within and surrounding the previously defined motif to identify amino acids required for full enzyme function and for oxidation and deprotonation activity in vitro. Since the structure of M. tuberculosis ThyX has been solved previously (27), we focused our attention on highly conserved amino acids and those shown to interact in crystals with the substrates and cofactors required for this complex reaction (Table (Table1).1). Figure Figure22 shows the alignment of this extended ThyX motif from several diverse eubacterial and archaeal species and highlights specific residues that are highly conserved.

FIG. 2.
Extended ThyX motif alignments. The alignment of the extended ThyX motif among ThyX proteins from several organisms, some of which are pathogenic, including representative species from both eubacteria and archaea, was made using psi-BLAST. Conserved amino ...
M. tuberculosis ThyX data


To test the function of certain amino acids within the extended ThyX motif, we performed site-directed mutagenesis on 51 amino acid residues in the coding region of M. tuberculosis thyX, creating 67 mutants, each containing one amino acid substitution (Table (Table1).1). We chose to test amino acid residues based on two criteria. These residues were either conserved among diverse eubacterial and archaeal species or implicated in forming hydrogen bonds with FAD or dUMP as shown from the crystal structure of M. tuberculosis ThyX (Table (Table1).1). It is surprising that not all of the FAD and dUMP binding residues in M. tuberculosis ThyX are conserved, which supports the need for further functional studies.

The enzymatic function of each mutant enzyme was tested by cloning into a pUC18 plasmid and transformation into E. coli χ2913, a strain that does not contain a gene for ThyX and lacks endogenous ThyA activity (21). Following transformation, the number of transformants observed on minimal medium plates without thymidine supplementation was counted. Mutants were placed in one of three categories comparing the level of complementation of growth in the E. coli system with that observed with the wild-type plasmid. We defined full complementation as at least 80% of the wild-type number of colonies, low complementation as less than 5% of the number of wild-type ThyX colonies on minimal medium plates, and no complementation as those mutants that gave no colonies at all on minimal medium. Among the 67 mutants tested, 25 mutants complemented as efficiently as the wild-type M. tuberculosis ThyX, 35 failed to complement, and seven complemented at less than 5% the level of wild-type ThyX. No mutant enzymes that we tested had between 5 and 80% of the wild-type number of colonies. Table Table11 shows the results of these studies.

Several amino acid residues of M. tuberculosis ThyX not included in the previously defined ThyX motif are of special interest. For example, Y44 makes one hydrogen bond to FAD via a water molecule. Two changes were made at this residue, one to phenylalanine to conserve the shape, hydrophobicity, and size of the residue and the other to alanine. Neither the Y44A nor the Y44F mutant complemented growth. The H69 residue has been suggested to function as a key catalytic residue and is highly conserved in all ThyX proteins, but its configuration varies in different structures. The H69 residue had previously been changed to glutamate to reverse the charge of the residue and failed to complement growth of the χ2913 E. coli strain, which supports the catalytic role for this histidine residue (27). The residues H91 and E92 form one and two hydrogen bonds to dUMP, respectively. Proteins with either of the alanine mutations complemented at a very low level.

Structural data have suggested that some amino acid residues that interact with several substrates or cofactors during ThyX catalysis are likely to undergo conformational changes during catalysis. The R95 residue interacts in crystals of M. tuberculosis ThyX with FAD via six hydrogen bonds and dUMP by three hydrogen bonds, and it is the first residue in the previously described ThyX motif. Since this arginine occupies a key area in the active site of the enzyme, three separate mutational changes were made to this residue. R95 was changed to a lysine to conserve both the charge and most of the shape of the residue, to an aspartic acid to reverse the charge, and finally to an alanine as a noncharged residue. As previously shown, the three mutant enzymes containing R95K, R95D, or R95A all failed to complement growth (27). Two other residues, Q103 and R199, interact in crystals with dUMP and FAD (see Fig. S1 in the supplemental material). The Q103 residue is the least conserved among the three residues that bind both dUMP and FAD. It forms one hydrogen bond to dUMP and two hydrogen bonds to FAD. The R199 residue is highly conserved and forms one hydrogen bond to FAD and three hydrogen bonds to dUMP. Proteins containing either alanine mutation did not complement.

A pair of histidines, H96 and H98, both form one hydrogen bond each to FAD. The H96 residue is highly conserved, but H98 is not conserved within the ThyX motif. Complementation still occurred when either histidine was changed to alanine; however, when both residues were converted to alanines complementation was abolished. This observation is analogous to a similar observation by Leduc et al. (17) indicating that at least one serine residue in the vicinity of the ThyX motif is required for H. pylori ThyX catalysis. The double mutant data shown above support the essentiality of at least one histidine residue in the ThyX motif, as demonstrated in previous work (11, 17). The R97 residue forms two hydrogen bonds to FAD and is highly conserved among ThyX proteins, and the R97A mutant protein did not complement growth. The S105 residue is a universally conserved nucleophile and forms two hydrogen bonds to dUMP. ThyA enzymes also transfer a methylene group to dUMP, and a cysteine functions as the nucleophile, so to test whether cysteine might function similarly in ThyX, S105 was changed to cysteine. The S105 residue had also previously been changed to glutamate to engineer a mutant with a less conserved residue (27). Both the S105C and S105E mutant proteins failed to complement. Both Q106 and R107 form two hydrogen bonds each to dUMP, but Q106 is not conserved whereas R107 is highly conserved among ThyX enzymes. The Q106A and the R107A mutant proteins also failed to complement.

Seven residues within the M. tuberculosis ThyX motif, F99, S100, Y101, S102, L104, Y108, and V109, have no known interactions with any substrate. Mutations had not previously been made in these residues, with the exception of the Y108F mutant (discussed briefly in reference 27). This Y108 residue is the only conserved amino acid among ThyX proteins from these seven residues, which is a phenylalanine in some species. ThyX mutant enzymes containing F99A, S100A, or L104A failed to complement. S100 was also changed to threonine to conserve the hydroxyl group; the S100T mutant was able to complement. Two mutants, the Y101A and V109E mutants, complemented at a level lower than wild-type ThyX but higher than the H91A and E92A mutant enzymes. The S102A mutant complemented growth. Two changes were made to Y108, one to a serine to conserve the hydroxyl group, and this Y108S enzyme failed to complement. However, this position is a phenylalanine in some species, and a Y108F mutant was able to complement growth.

This detailed analysis allowed us to identify many amino acids that are essential to ThyX enzyme activity. As one might expect, highly conserved amino acids were likely to be essential to enzyme function, and those residues that interacted closely with FAD and BrdUMP (5-bromo-dUMP) in the M. tuberculosis ThyX crystal structure were also essential. These data extend the list of amino acids that define an enzymatically active ThyX protein.

Formation of higher-order complexes.

In the active enzymatic state, wild-type M. tuberculosis ThyX protein forms tetramers. A mutant protein that prevented tetramer formation would certainly fail to be enzymatically active because residues from three of the four monomers contribute to each active site (17, 27). To determine whether noncomplementing mutant proteins were able to form tetramers, genes encoding each mutant protein were separately cloned into the pET24d protein expression vector and transformed into the χ2913 E. coli strain. Proteins were expressed with six-histidine tags and isolated on nickel columns, and their migration on both native and reducing polyacrylamide gels was analyzed. High levels of protein were isolated from the wild type and all 67 mutant proteins. Figure Figure3A3A shows a representative SDS-polyacrylamide gel, and Fig. Fig.3B3B shows a representative native polyacrylamide gel with wild-type protein and five mutant ThyX proteins containing the most extreme amino acid changes. On the native polyacrylamide gel, the five ThyX mutant proteins migrate similarly to the wild-type protein, which forms tetramers under native conditions. The high-molecular-mass band on the native polyacrylamide gels was probed with an antibody to the six-histidine tag, and Western blotting confirmed that this band was ThyX (data not shown). All 35 mutant ThyX proteins that failed to complement in the E. coli assay showed similar patterns on native polyacrylamide gels. We conclude that failure of complementation did not result from poor expression of the protein or from failure to form higher-order complexes, as expected for a tetrameric enzyme.

FIG. 3.
Mutant ThyX tetramer formation. (A) Eight-percent SDS-polyacrylamide gel showing, from left to right, wild-type M. tuberculosis ThyX and mutant proteins R95A and H98E, respectively, migrating at the correct monomeric size of M. tuberculosis ThyX (27.6 ...

Oxidation assays.

In ThyX enzymes, the conversion of dUMP to dTMP involves two half reactions, oxidation and transfer of the methylene group from dUMP to form the dTMP final product (1). To test the ability of the ThyX mutants to execute these essential half reactions, two biochemical assays were performed. First, the oxidation of NADPH was assayed for a subset of ThyX mutants (Table (Table1).1). We measured the sustained rate of oxidation of NADPH over a 3-min period with saturating amounts of the cofactor FAD (12.5 μM) and substrates dUMP (50 μM) and NADPH (0.2 mM). Overall, 14 of 39 mutant ThyX proteins oxidized NADPH at a rate greater than 50% that of wild-type ThyX. In fact, a subset of these mutant ThyX proteins—9 of 39—was able to oxidize NADPH at greater than 100% of the rate of wild-type ThyX (140.5 pmol NADP+ produced/μg enzyme/min). Figure Figure44 depicts the oxidation rates of NADPH by mutants within the extended ThyX motif grouped by their complementation status. Among the noncomplementing mutants, the Y44F, S100A, Y108S, and L104A mutants oxidized at 69%, 56%, 40%, and 134% of the wild-type rate, respectively. Rates for the other mutants were below 40%. Among the mutants that complemented, only the H96A mutant had a low oxidation rate at 38% that of the wild type whereas all other mutants oxidized NADPH in the same range as the wild type. Notably, the Y108F mutant had an oxidation activity two to three times higher than the wild-type level (252%). Among the low-complementing group, the rates of the H91A, E92A, and Y101A mutants were considerably less than the wild-type rate (between 25 and 36%); however, the V109E mutant oxidized NADPH at a rate greater than the wild-type rate (173%).

FIG. 4.
Bar graph of NAPDH oxidation rates among ThyX mutants. NADPH oxidation rates of a select group of ThyX mutants were assayed, and the rates of oxidation compared to that of the wild type (wt) are shown on the graph. Wild-type M. tuberculosis ThyX oxidized ...

Tritium release assays.

The second assay used for the characterization of mutants measured the release of 3H from the 5-carbon position of the dUMP pyrimidine ring, a prerequisite for the transfer of the methyl group from methylenetetrahydrofolate to the dUMP substrate (1, 2, 23). In general, we selected mutant proteins that failed to complement in the E. coli ThyA knockout strain. These mutant proteins were then assayed for their ability to catalyze the release of a proton from dUMP. This reaction requires a nucleophilic attack on the 6-carbon position of dUMP that liberates the hydrogen at the 5-carbon position, thus preparing dUMP for the addition of a methylene group from methylenetetrahydrofolate to form dTMP. The tritium release assay measures the formation of tritiated water resulting from ThyX catalysis. Tritium release assays were performed on 15 mutant ThyX proteins which had failed to complement the χ2913 E. coli strain (Table (Table1).1). Thirteen of these 15 mutant enzymes were also defective in this basic enzymatic function. To assay the ability of a cysteine to function as the nucleophilic residue in M. tuberculosis ThyX, tritium release was measured for the S105C mutant to be 4% of the wild-type rate. In general, the mutants that did not complement the E. coli ThyA knockout strain had tritium release rates less than 15% of the wild-type rate. Two noncomplementing mutants, the Y44F and Y108S mutants, and two low-complementing mutants, the Y101A and V109E mutants, were all able to oxidize NADPH at modest or, in the case of the V109E mutant, substantial rates. These enzymes were also measured to have low tritium release rates of 8%, 9%, 13%, and 7%, respectively, compared with the wild-type rate. Finally, two noncomplementing mutant proteins, the S100A and L104A mutants, were both able to oxidize quite well and to perform tritium release at substantial rates (89% and 165% of the wild-type rate, respectively). Since they are unable to complement the E. coli ThyA knockout strain, their defect could lie for instance in the transfer of C1 units or other later steps.


Thymidine biosynthesis is essential in virtually all organisms, but two different pathways have evolved to accomplish this critical task. The two thymidylate synthesis enzymes ThyA and ThyX are structurally and evolutionarily unrelated (16, 22, 23), but clearly they perform the same function. This similarity is illustrated most dramatically by the fact that the M. tuberculosis and H. pylori ThyX enzymes can both complement ThyA-deficient E. coli (17, 27). ThyA is used commonly in eukaryotes and in many microorganisms, including most eubacteria. ThyX has been identified more recently, and it is common among the archaea and some eubacteria, particularly those that are anaerobic or occupy extreme environments. The corynebacteria, the group that includes the mycobacteria, are unusual in having both ThyA and ThyX in their genomes. The extreme conservation of the thyX and thyA sequences within this group suggests that both genes serve essential functions for these organisms, but the roles for two parallel pathways are not yet defined. This study of the key residues in the M. tuberculosis ThyX enzyme is the first step toward understanding the function of this enzyme in the growth and various physiological states of this pathogen in its active and latent phases.

The classic pathway uses the ThyA enzyme to methylate dUMP in the final step of thymidine biosynthesis (3). In the alternative pathway, ThyX performs the transfer without oxidation of tetrahydrofolate, utilizing both dUMP and methylenetetrahydrofolate as reactants and FAD and NADPH in the redox phase of the reaction (1, 10, 17, 22, 23). Whereas ThyA requires a companion enzyme, dihydrofolate reductase, to regenerate tetrahydrofolate that has been oxidized in the transfer reaction, ThyX requires dUMP, NADPH, FAD, and methylenetetrahydrofolate (3). Currently, four reaction mechanisms of ThyX have been proposed, as shown in Fig. Fig.1B1B (1, 10-12). Our data have identified mutants (e.g., the Y44F, Y108S, V109E, R190A, F195V, and R199A mutants) that are still able to oxidize NADPH but are unable to catalyze the deprotonation reaction. This behavior is consistent with NADPH oxidation taking place before deprotonation of dUMP, as shown in Fig. Fig.1B.1B. Therefore, these analyses allowed identification of several residues that normally participate in coupling the NADPH oxidation and deprotonation reactions of ThyX. Some of these amino acid residues have interactions with FAD and/or dUMP, as shown in the crystallographic image of the M. tuberculosis ThyX active site (see Fig. S1 in the supplemental material).

We focused this study on amino acids that had been identified in two ways. First, alignment of the many eubacterial and archaeal thyX genes showed a set of highly conserved residues; these had been used to define a ThyX signature motif, RHRX7-8S. The structure of the M. tuberculosis ThyX protein was solved with both FAD and BrdUMP bound, and this structure detailed important interactions of the protein and these molecules in crystals. Our goal was to test the validity of the ThyX motif for the M. tuberculosis enzyme and the importance of the molecular interactions observed in crystals.

A number of interesting mutants provided more guidance on the details of the enzyme activity. For example, several substitutions of amino acid residues with similar side groups successfully complemented the E. coli χ2913 strain (E74Q, S100T, Y108F, and T181S) whereas the less conservative substitutions failed to complement (E74A, S100A, Y108S, and T181L). Moreover, there were mutants that failed to complement the E. coli χ2913 strain that did catalyze the NADPH oxidation reaction in vitro (the Y44F, Y108S, F195V, and R199A mutants). There were no examples of mutants that catalyzed the 3H release but not the oxidation. The simplest explanation for this asymmetry is that the oxidation step of the reaction precedes deprotonation, which provides further support to the reaction mechanisms wherein NADPH oxidation precedes deprotonation (Fig. (Fig.1B)1B) (10).

When the structures of T. maritima, M. tuberculosis, and P. bursaria chlorella virus-1 ThyX were solved, FAD was identified as being firmly bound within these ThyX enzymes (10, 16, 27). Furthermore, dUMP has been cocrystalized with T. maritima and M. tuberculosis ThyX (20) and NADPH was observed in the active site of M. tuberculosis ThyX (27); however, the position of the methylenetetrahydrofolate within the enzyme is not known. Under low-pH conditions (pH 4.5 to 5.5), NADPH was observed to displace FAD from preformed crystals in M. tuberculosis ThyX (26). Nevertheless, the fact that two of the residues (R190 and H194) that interacted with NADPH in the crystals also interacted with tightly bound FAD in the crystals (26, 27) indicates that the location of NADPH binding might not be biologically relevant. Two mutant proteins (the S100A and L104A mutants) were able to catalyze both the oxidation and the 3H release in vitro but still failed to complement the E. coli χ2913 strain. This failure to complement is probably due to a defect in a later stage of the reaction possibly involving the transfer of C1 units or the release of dTMP. These data provide the first clues to residues essential for the final steps of methylene transfer to the dUMP and/or release of the tetrahydrofolate.

In ThyA enzymes, a cysteine residue acts as the nucleophile by covalently binding the position 6-carbon of dUMP and stabilizing the deoxyribose ring (3, 25, 29). In all ThyX enzymes sequenced to date, a serine residue appears to function as the nucleophile and forms an integral part of the ThyX motif (see Table S2 in the supplemental material). The S105C mutant was unable to function as the nucleophile in ThyX, and this further underscores the fundamental differences between the two thymidylate synthase enzymes.

Several positions in the X7-8 portion of the previously defined ThyX motif are not highly conserved among the thyX genes so far identified. It is likely that these serve to position key neighboring residues for their enzymatic function.

The previously defined motif was defined principally by alignment of the thyX genes so far identified (see Table S2 in the supplemental material), and our mutagenesis study has added to our understanding of the role of particular amino acids in the enzyme activity. Despite our work, it is still unclear which residues contribute to binding of methylenetetrahydrofolate and NADPH. Nonetheless, the combination of structural and mutagenesis studies reported here has provided a clearer picture of FAD and dUMP binding residues. Our analyses permit us to propose an extended range of the ThyX motif, shown here with the underlined histidine acting as the catalytic residue and the underlined serine as the nucleophile: (Y/W)X19-40HX25-28RHRX7-8SXR(Y/F)X68-114R.

The strong evolutionary conservation of the thyX gene suggests that our study of the M. tuberculosis enzyme will be useful for analysis of this enzyme with a wide range of human pathogens, including those that depend entirely on ThyX for biosynthesis of thymidine. Since this reaction is clearly essential in those organisms, the identification of inhibitors is a first step in developing drugs effective against them. When ThyX inhibitors are available, this work will aid in understanding their inhibitory effects. Because mycobacteria have both ThyA and ThyX pathways for thymidine synthesis, any effective drug might need to inhibit both enzymes. However, ThyA requires concomitant dihydrofolate reductase to regenerate tetrahydrofolate, so a combination of ThyX and DHFR inhibition might limit M. tuberculosis growth. Promising inhibitors of M. tuberculosis DHFR have been identified previously (7, 8). Our detailed determination of the function of key residues in the ThyX enzyme should facilitate further study of this potential drug target and will help to better understand the complex reactions of this interesting enzyme.

Supplementary Material

[Supplemental material]


This work was supported by grants from the Puget Sound Partners for Public Health (to C.H.S.) and from INSERM and Fondation Bettencourt-Schueller (to H.M.).

We thank Josh Hunter and Pradip Rathod for help with the 3H release assay and Ho Gun Rhie and members of the Sibley lab for their advice and review of the paper.


[down-pointing small open triangle]Published ahead of print on 11 January 2008.

Supplemental material for this article may be found at


1. Agrawal, N., S. A. Lesley, P. Kuhn, and A. Kohen. 2004. Mechanistic studies of a flavin-dependent thymidylate synthase. Biochemistry 4310295-10301. [PubMed]
2. Armstrong, R. D., and R. B. Diasio. 1982. Improved measurement of thymidylate synthetase activity by a modified tritium-release assay. J. Biochem. Biophys. Methods 6141-147. [PubMed]
3. Carreras, C. W., and D. V. Santi. 1995. The catalytic mechanism and structure of thymidylate synthase. Annu. Rev. Biochem. 64721-762. [PubMed]
4. Chernyshev, A., T. Fleischmann, and A. Kohen. 2007. Thymidyl biosynthesis enzymes as antibiotic targets. Appl. Microbiol. Biotechnol. 74282-289. [PubMed]
5. Cosloy, S. D., and M. Oishi. 1973. Genetic transformation in Escherichia coli K12. Proc. Natl. Acad. Sci. USA 7084-87. [PubMed]
6. Costi, M. P., S. Ferrari, A. Venturelli, S. Calo, D. Tondi, and D. Barlocco. 2005. Thymidylate synthase structure, function and implication in drug discovery. Curr. Med. Chem. 122241-2258. [PubMed]
7. da Cunha, E. F., T. de Castro Ramalho, R. Bicca de Alencastro, and E. R. Maia. 2004. Interactions of 5-deazapteridine derivatives with Mycobacterium tuberculosis and with human dihydrofolate reductases. J. Biomol. Struct. Dyn. 22119-130. [PubMed]
8. Gerum, A. B., J. E. Ulmer, D. P. Jacobus, N. P. Jensen, D. R. Sherman, and C. H. Sibley. 2002. Novel Saccharomyces cerevisiae screen identifies WR99210 analogues that inhibit Mycobacterium tuberculosis dihydrofolate reductase. Antimicrob. Agents Chemother. 463362-3369. [PMC free article] [PubMed]
9. Giladi, M., G. Bitan-Banin, M. Mevarech, and R. Ortenberg. 2002. Genetic evidence for a novel thymidylate synthase in the halophilic archaeon Halobacterium salinarum and in Campylobacter jejuni. FEMS Microbiol. Lett. 216105-109. [PubMed]
10. Graziani, S., J. Bernauer, S. Skouloubris, M. Graille, C. Z. Zhou, C. Marchand, P. Decottignies, H. van Tilbeurgh, H. Myllykallio, and U. Liebl. 2006. Catalytic mechanism and structure of viral flavin-dependent thymidylate synthase ThyX. J. Biol. Chem. 28124048-24057. [PubMed]
11. Graziani, S., Y. Xia, J. R. Gurnon, J. L. Van Etten, D. Leduc, S. Skouloubris, H. Myllykallio, and U. Liebl. 2004. Functional analysis of FAD-dependent thymidylate synthase ThyX from Paramecium bursaria Chlorella virus-1. J. Biol. Chem. 27954340-54347. [PubMed]
12. Griffin, J., C. Roshick, E. Iliffe-Lee, and G. McClarty. 2005. Catalytic mechanism of Chlamydia trachomatis flavin-dependent thymidylate synthase. J. Biol. Chem. 2805456-5467. [PubMed]
13. Hardy, L. W., J. S. Finer-Moore, W. R. Montfort, M. O. Jones, D. V. Santi, and R. M. Stroud. 1987. Atomic structure of thymidylate synthase: target for rational drug design. Science 235448-455. [PubMed]
14. Hitchings, G. H., Jr. 1989. Nobel lecture in physiology or medicine-1988. Selective inhibitors of dihydrofolate reductase. In Vitro Cell Dev. Biol. 25303-310. [PubMed]
15. Kompis, I. M., K. Islam, and R. L. Then. 2005. DNA and RNA synthesis: antifolates. Chem. Rev. 105593-620. [PubMed]
16. Kuhn, P., S. A. Lesley, I. I. Mathews, J. M. Canaves, L. S. Brinen, X. Dai, A. M. Deacon, M. A. Elsliger, S. Eshaghi, R. Floyd, A. Godzik, C. Grittini, S. K. Grzechnik, C. Guda, K. O. Hodgson, L. Jaroszewski, C. Karlak, H. E. Klock, E. Koesema, J. M. Kovarik, A. T. Kreusch, D. McMullan, T. M. McPhillips, M. A. Miller, M. Miller, A. Morse, K. Moy, J. Ouyang, A. Robb, K. Rodrigues, T. L. Selby, G. Spraggon, R. C. Stevens, S. S. Taylor, H. van den Bedem, J. Velasquez, J. Vincent, X. Wang, B. West, G. Wolf, J. Wooley, and I. A. Wilson. 2002. Crystal structure of thy1, a thymidylate synthase complementing protein from Thermotoga maritima at 2.25 A resolution. Proteins 49142-145. [PubMed]
17. Leduc, D., S. Graziani, G. Lipowski, C. Marchand, P. Le Marechal, U. Liebl, and H. Myllykallio. 2004. Functional evidence for active site location of tetrameric thymidylate synthase X at the interphase of three monomers. Proc. Natl. Acad. Sci. USA 1017252-7257. [PubMed]
18. Liu, X. Q., and J. Yang. 2004. Bacterial thymidylate synthase with intein, group II intron, and distinctive ThyX motifs. J. Bacteriol. 1866316-6319. [PMC free article] [PubMed]
19. Marsh, S. 2005. Thymidylate synthase pharmacogenetics. Investig. New Drugs 23533-537. [PubMed]
20. Mathews, I. I., A. M. Deacon, J. M. Canaves, D. McMullan, S. A. Lesley, S. Agarwalla, and P. Kuhn. 2003. Functional analysis of substrate and cofactor complex structures of a thymidylate synthase-complementing protein. Structure 11677-690. [PubMed]
21. Michaels, M. L., C. W. Kim, D. A. Matthews, and J. H. Miller. 1990. Escherichia coli thymidylate synthase: amino acid substitutions by suppression of amber nonsense mutations. Proc. Natl. Acad. Sci. USA 873957-3961. [PubMed]
22. Myllykallio, H., D. Leduc, J. Filee, and U. Liebl. 2003. Life without dihydrofolate reductase FolA. Trends Microbiol. 11220-223. [PubMed]
23. Myllykallio, H., G. Lipowski, D. Leduc, J. Filee, P. Forterre, and U. Liebl. 2002. An alternative flavin-dependent mechanism for thymidylate synthesis. Science 297105-107. [PubMed]
24. Patel, M. P., and J. S. Blanchard. 1999. Expression, purification, and characterization of Mycobacterium tuberculosis mycothione reductase. Biochemistry 3811827-11833. [PubMed]
25. Phan, J., E. Mahdavian, M. C. Nivens, W. Minor, S. Berger, H. T. Spencer, R. B. Dunlap, and L. Lebioda. 2000. Catalytic cysteine of thymidylate synthase is activated upon substrate binding. Biochemistry 396969-6978. [PubMed]
26. Sampathkumar, P., S. Turley, C. H. Sibley, and W. G. Hol. 2006. NADP(+) expels both the co-factor and a substrate analog from the Mycobacterium tuberculosis ThyX active site: opportunities for anti-bacterial drug design. J. Mol. Biol. 3601-6. [PubMed]
27. Sampathkumar, P., S. Turley, J. E. Ulmer, H. G. Rhie, C. H. Sibley, and W. G. Hol. 2005. Structure of the Mycobacterium tuberculosis flavin dependent thymidylate synthase (MtbThyX) at 2.0A resolution. J. Mol. Biol. 3521091-1104. [PubMed]
28. Sassetti, C. M., D. H. Boyd, and E. J. Rubin. 2003. Genes required for mycobacterial growth defined by high density mutagenesis. Mol. Microbiol. 4877-84. [PubMed]
29. Stroud, R. M., and J. S. Finer-Moore. 1993. Stereochemistry of a multistep/bipartite methyl transfer reaction: thymidylate synthase. FASEB J. 7671-677. [PubMed]
30. Ulmer, J. E. 2007. Analysis and studies of inhibition of the two divergent thymidine biosynthesis pathways in Mycobacterium tuberculosis. Dissertation.University of Washington, Seattle.

Articles from Journal of Bacteriology are provided here courtesy of American Society for Microbiology (ASM)