Functional identity of the active sites of crustacean and viral thymidylate synthases.

Thymidylate synthase (TS) catalyzes the synthesis of deoxythymidine monophosphate (dTMP), which is an essential precursor for DNA synthesis. The rationale underlying drug design is to identify compounds that differentially inhibit a viral or parasite enzyme vs. the host homologue. We studied the TS of the white spot syndrome virus (WSSV TS) and the corresponding TS from the host, the marine invertebrate shrimp Litopenaeus vannamei. TS is the only de novo source of dTMP and is essential for host and viral DNA replication. To establish proof of principle, we cloned a full-length TS cDNA from the white shrimp L. vannamei (shrimp TS) that corresponds to a deduced sequence of 289 amino acids and over-expressed it to study inhibition of both shrimp and viral TSs. Steady-state kinetic parameters for both TSs are similar, and dissociation (K(d)) or half maximal inhibitory concentration constants (IC(50)) did not show differential inhibition between the folate analogues. Differences in their amino acid sequence are not reflected in theoretical molecular models of both TSs, since both appear to have identical active sites. These results suggest that the eukaryotic TS active site is very constrained into the functional residues involved in reductive methylation of 2'-deoxyuridine-5'-monophosphate (dUMP).

Role of an invariant lysine residue in folate binding on Escherichia coli thymidylate synthase: calorimetric and crystallographic analysis of the K48Q mutant.

Thymidylate synthase (TS) catalyzes the reductive methylation of deoxyuridine monophosphate (dUMP) using methylene tetrahydrofolate (CH(2)THF) as cofactor, the glutamate tail of which forms a water-mediated hydrogen bond with an invariant lysine residue of this enzyme. To understand the role of this interaction, we studied the K48Q mutant of Escherichia coli TS using structural and biophysical methods. The k(cat) of the K48Q mutant was 430-fold lower than wild-type TS in activity, while the K(m) for the (R)-stereoisomer of CH(2)THF was 300 microM, about 30-fold larger than K(m) from the wild-type TS. Affinity constants were determined using isothermal titration calorimetry, which showed that binding was reduced by one order of magnitude for folate-like TS inhibitors, such as propargyl-dideazafolate (PDDF) or compounds that distort the TS active site like BW1843U89 (U89). The crystal structure of the K48Q-dUMP complex revealed that dUMP binding is not impaired in the mutant, and that U89 in a ternary complex of K48Q-nucleotide-U89 was bound in the active site with subtle differences relative to comparable wild-type complexes. PDDF failed to form ternary complexes with K48Q and dUMP. Thermodynamic data correlated with the structural determinations, since PDDF binding was dominated by enthalpic effects while U89 had an important entropic component. In conclusion, K48 is critical for catalysis since it leads to a productive CH(2)THF binding, while mutation at this residue does not affect much the binding of inhibitors that do not make contact with this group.

Role of Y94 in proton and hydride transfers catalyzed by thymidylate synthase.

Thymidylate synthase (TS) catalyzes the substitution of a carbon-bound proton in a uracil base by a methyl group to yield thymine in the de novo biosynthesis of this DNA base. The enzymatic mechanism involves making and breaking several covalent bonds. Traditionally, a conserved tyrosine (Y94 in Escherichia coli, Y146 in Lactobacillus casei, and Y135 in humans) was assumed to serve as the general base catalyzing the proton abstraction. That assumption was examined here by comparing the nature of the proton abstraction using wild-type (wt) E. coli TS (ecTS) and its Y94F mutant (with a turnover rate reduced by 2 orders of magnitude). A subsequent hydride transfer was also studied using the wt and Y94F. The physical nature of both H-transfer steps was examined by determining intrinsic kinetic isotope effects (KIEs). Surprisingly, the findings did not suggest a direct role for Y94 in the proton abstraction step. The effect of this mutation on the subsequent hydride transfer was examined by a comparison of the temperature dependency of the intrinsic KIE on both the wt and the mutant. The intrinsic KIEs for Y94F at physiological temperatures were slightly smaller than those for wt but, otherwise, were as temperature-independent, suggesting a perfectly preorganized reaction coordinate for both enzymes. At reduced temperatures, however, the KIE for the mutant increased with a decrease in temperature, indicating a poorly preorganized reaction coordinate. Other kinetic and structural properties were also compared, and the findings suggested that Y94 is part of a H-bond network that plays a critical role at a step between the proton and the hydride transfers, presumably the dissociation of H4folate from the covalently bound intermediate. The possibility that no single residue serves as the general base in question but, rather, that the whole network of H-bonds at the active site catalyzes proton abstraction is discussed.

Significance of mutations on the structural perturbation of thymidylate synthase: implications for their involvement in subunit exchange.

Wild-type thymidylate synthase (WT-TS) from Escherichia coli and several of its mutants showed varying degrees of susceptibility to trypsin. While WT-TS was resistant to trypsin as were the mutants C146S, K48E, and R126K, others such as Y94A, Y94F, C146W, and R126E were digested but at different rates from one another. The peptides released from the mutants were identified by mass spectrometry and Edman sequence analysis. The known crystal structures for WT-TS, Y94F, and R126E, surprisingly, showed no structural differences that could explain the difference in their susceptibility to trypsin. One explanation is that the mutations could perturb the dynamic equilibrium of the dimeric state of the mutants as to increase their dissociation to monomers, which being less structured than the dimer, would be hydrolyzed more readily by trypsin. Earlier studies appear to support this proposal since conditions that promote subunit dissociation in solutions of R126E with other inactive mutants, such as dilution, low concentrations of urea, and elevated pH, greatly enhance the rate of restoration of TS activity. Analytic ultracentrifuge studies with various TSs in urea, or at pH 9.0, or that have been highly diluted are, for the most part, in agreement with this thesis, since these conditions are associated with an increase in dissociation to monomers, particularly with the mutant TSs. However, these studies do not rule out the possibility that conformation differences among the various TS dimers are responsible for the differences in susceptibility to trypsin, particularly at high concentrations of protein where the WT-TS and mutants are mainly dimers.

Chloroviruses encode a bifunctional dCMP-dCTP deaminase that produces two key intermediates in dTTP formation.

The chlorovirus PBCV-1, like many large double-stranded DNA-containing viruses, contains several genes that encode putative proteins involved in nucleotide biosynthesis. This report describes the characterization of the PBCV-1 dCMP deaminase, which produces dUMP, a key intermediate in the synthesis of dTTP. As predicted, the recombinant protein has dCMP deaminase activity that is activated by dCTP and inhibited by dTTP. Unexpectedly, however, the viral enzyme also has dCTP deaminase activity, producing dUTP. Typically, these two reactions are catalyzed by proteins in separate enzyme classes; to our knowledge, this is the first example of a protein having both deaminase activities. Kinetic experiments established that (i) the PBCV-1 enzyme has a higher affinity for dCTP than for dCMP, (ii) dCTP serves as a positive heterotropic effector for the dCMP deaminase activity and a positive homotropic effector for the dCTP deaminase activity, and (iii) the enzymatic efficiency of the dCMP deaminase activity is about four times higher than that of the dCTP deaminase activity. Inhibitor studies suggest that the same active site is involved in both dCMP and dCTP deaminations. The discovery that the PBCV-1 dCMP deaminase has two activities, together with a previous report that the virus also encodes a functional dUTP triphosphatase (Y. Zhang, H. Moriyama, K. Homma, and J. L. Van Etten, J. Virol. 79:9945-9953, 2005), means that PBCV-1 is the first virus to encode enzymes involved in all three known pathways to form dUMP.

Structure of the Y94F mutant of Escherichia coli thymidylate synthase.

Tyr94 of Escherichia coli thymidylate synthase is thought to be involved, either directly or by activation of a water molecule, in the abstraction of a proton from C5 of the 2'-deoxyuridine 5'-monophosphate (dUMP) substrate. Mutation of Tyr94 leads to a 400-fold loss in catalytic activity. The structure of the Y94F mutant has been determined in the native state and as a ternary complex with thymidine 5'-monophosphate (dTMP) and 10-propargyl 5,8-dideazafolate (PDDF). There are no structural changes ascribable to the mutation other than loss of a water molecule hydrogen bonded to the tyrosine OH, which is consistent with a catalytic role for the phenolic OH.

Crystal structures of thymidylate synthase mutant R166Q: structural basis for the nearly complete loss of catalytic activity.

Thymidylate synthase (TS) catalyzes the folate-dependent methylation of deoxyuridine monophosphate (dUMP) to form thymidine monophosphate (dTMP). We have investigated the role of invariant arginine 166, one of four arginines that contact the dUMP phosphate, using site-directed mutagenesis, X-ray crystallography, and TS from Escherichia coli. The R166Q mutant was crystallized in the presence of dUMP and a structure determined to 2.9 A resolution, but neither the ligand nor the sulfate from the crystallization buffer was found in the active site. A second structure determined with crystals prepared in the presence of dUMP and the antifolate 10-propargyl-5,8-dideazafolate revealed that the inhibitor was bound in an extended, nonproductive conformation, partially occupying the nucleotide-binding site. A sulfate ion, rather than dUMP, was found in the nucleotide phosphate-binding site. Previous studies have shown that the substitution at three of the four arginines of the dUMP phosphate-binding site is permissive; however; for Arg166, all the mutations lead to a near-inactive mutant. The present structures of TS R166Q reveal that the phosphate-binding site is largely intact, but with a substantially reduced affinity for phosphate, despite the presence of the three remaining arginines. The position of Cys146, which initiates catalysis, is shifted in the mutant and resides in a position that interferes with the binding of the dUMP pyrimidine moiety.

Hydride transfer versus hydrogen radical transfer in thymidylate synthase.

The nature of a H-transfer in the thymidylate synthase catalyzed reaction was investigated by comparison of the wild-type enzyme with the W80M mutant. The nature of the H-transfer was not affected, as indicated by intrinsic isotope effects and their temperature dependence. These findings support a single-step hydride transfer instead of a two-step radical transfer.

Binding and repression of translation of the cognate mRNA by Trichinella spiralis thymidylate synthase differ from the corresponding interactions of the human enzyme.

Thymidylate synthase (TS) of Trichinella spiralis, a parasitic nematode causing trichinellosis, was found to bind its own mRNA and repress translation of the latter, similar to its human counter-part [Chu, Koeller, Casey, Drake, Chabner, Elwood, Zinn and Allegra (1991) Proc. Natl. Acad. Sci. U.S.A. 88, 8977-8981]. However, in striking contrast with human TS, the parasite enzyme's interaction with mRNA was not affected by any of the substrate (deoxyuridylate or N(5,10)-methylenetetrahydrofolate) nor by the inhibitor (fluorodeoxyuridylate; used alone or in the presence of N(5,10)-methylenetetrahydrofolate) similar to that shown for the bifunctional enzyme from Plasmodium falciparum [Zhang and Rathod (2002) Science 296, 545-547]. Moreover, repression of the translation of the parasite enzyme was enhanced by the same ligands that were shown by others (Chu et al., 1991) to prevent human TS from impairing its translation. On comparing the capacity of TS to bind to its cognate mRNA, relative to its ability to inhibit its translation, the same enzyme preparation was active as translational repressor at a considerably lower protein/mRNA ratio, suggesting the two phenomena to be disconnected. Of interest is the fact that the presence of the enzyme protein N-terminal methionine proved to be critical for binding, but not for repression of its translation, indicating that mRNA binding requires a methionine or an adduct (i.e. methionine-histidine) at the N-terminus of TS, but that the translational repression effect does not. Notably, chicken liver dihydrofolate reductase, which is incapable of binding to T. spiralis TS mRNA, repressed the translation of TS.

Three-dimensional structure of the R115E mutant of T4-bacteriophage 2'-deoxycytidylate deaminase.

2'-Deoxycytidylate deaminase (dCD) converts deoxycytidine 5'-monophosphate (dCMP) to deoxyuridine 5'-monophosphate and is a major supplier of the substrate for thymidylate synthase, an important enzyme in DNA synthesis and a major target for cancer chemotherapy. Wild-type dCD is allosterically regulated by the end products of its metabolic pathway, deoxycytidine 5'-triphosphate and deoxythymidine 5'-triphosphate, which act as an activator and an inhibitor, respectively. The first crystal structure of a dCD, in the form of the R115E mutant of the T4-bacteriophage enzyme complexed with the active site inhibitor pyrimidin-2-one deoxyribotide, has been determined at 2.2 A resolution. This mutant of dCD is active, even in the absence of the allosteric regulators. The molecular topology of dCD is related to that of cytidine deaminase (CDA) but with modifications for formation of the binding site for the phosphate group of dCMP. The enzyme has a zinc ion-based mechanism that is similar to that of CDA. A second zinc ion that is present in bacteriophage dCD, but absent in mammalian dCD and CDA, is important for the structural integrity of the enzyme and for the binding of the phosphate group of the substrate or inhibitor. Although the R115E mutant of dCD is a dimer in solution, it crystallizes as a hexamer, mimicking the natural state of the wild-type enzyme. Residues 112 and 115, which are known to be important for the binding of the allosteric regulators, are found in a pocket that is at the intersubunit interfaces in the hexamer but distant from the substrate-binding site. The substrate-binding site is composed of residues from a single protein molecule and is sequestered in a deep groove. This groove is located at the outer surface of the hexamer but ends at the subunit interface that also includes residue 115. It is proposed that the absence of subunit interactions at this interface in the dimeric R115E mutant renders the substrate-binding site accessible. In contrast, for the wild-type enzyme, binding of dCTP induces an allosteric effect that affects the subunit interactions and results in an increase in the accessibility of the binding site.

Protection of hematopoietic stem cells from pemetrexed toxicity by retroviral gene transfer with a mutant dihydrofolate reductase-mutant thymidylate synthase fusion gene.

Myelosuppression is one of the major side effects of most anticancer drugs. To confer myeloprotection, our laboratory generated drug-resistant mutants of select target human enzymes for gene transfer to the bone marrow. Mutants of two of these enzymes, dihydrofolate reductase (DHFR F/S) and thymidylate synthase (TS G52S), were previously shown to confer resistance to methotrexate and 5-FU, respectively, and recently a fusion cDNA of both mutant enzymes (DHFR F/S-TS G52S) was shown to confer dual resistance to both antimetabolites. In this study, we examined the sensitivity of the DHFR F/S-TS G52S fusion protein to the multitargeted antifolate, pemetrexed (LY231514, Alimta), which targets both DHFR and TS and is currently in phase III trials for the treatment of solid tumors and in combination with cisplatin has been shown to be an advance in the treatment of mesothelioma. The K(i) for the DHFR F/S portion of the purified fusion protein to pemetrexed was increased by greater than 9000-fold when compared to wtDHFR (8000 versus 0.86 nM), while the K(i) for the TS G52S portion of the fusion protein to pemetrexed was similar to that of wtTS (2.8 versus 3.1 nM). When the fusion gene was retrovirally transduced into NIH 3T3 fibroblasts, the IC(50) to pemetrexed was three- to four-fold higher than cells transduced with DHFR F/S or TS G52S alone (163 versus 53 and 45 nM, respectively). Similarly, expression of the DHFR F/S-TS G52S fusion gene in retrovirally transduced mouse marrow cells resulted in an increased survival of CFU-GM colonies when compared to cells transduced with either of the mutants alone. Co-expression of mutant DHFR and TS enzymes has additive effects in conferring resistance to pemetrexed-induced toxicity. This construct may be useful for conferring myeloprotection to patients receiving this drug.

A trojan horse approach for silencing thymidylate synthase.

In this paper we present a new and possibly more effective way of inhibiting thymidylate synthase (TS) in cells than through the use of substrate analogue inhibitors. An inactive double mutant of TS (DM), Arg(126)Glu/Cys(146)Trp, is shown to progressively impair the reactivation of native Escherichia coli TS when the two are denatured together in vitro. The individual single mutant proteins Arg(126)Glu and Cys(146)Trp showed little or no inhibition. When the DM is introduced into E. coli and induced from an expression plasmid, the mutant subunits act as a decoy in deceiving newly formed native TS subunits to fold with them to yield inactive heterodimers. As a consequence of the depletion of TS, the cells die a "thymineless" death when grown in medium devoid of thymine. Addition of thymine to the medium enables the cells to grow normally, although only very low levels of TS activity could be detected in those cells containing induced DM. The individual single-site mutations of the DM, Arg(126)Glu and Cys(146)Trp, did not inhibit growth, as might be expected from the in vitro studies. However, when a nontoxic level of 5-fluoro-2'-deoxyuridine 5'-monophosphate (FdUMP) is added to growing DM-transformed cells, the combination is lethal to the cells. These experiments suggest that a similar dominant-negative response to the DM of TS could be affected in tumor cells, for which preliminary evidence is presented. This technique, either alone or combined with other modalities, suggest a new approach to targeting cells for chemotherapy.

Role of cysteine amino acid residues on the RNA binding activity of human thymidylate synthase.

The role of cysteine sulfhydryl residues on the RNA binding activity of human thymidylate synthase (TS) was investigated by mutating each cysteine residue on human TS to a corresponding alanine residue. Enzymatic activities of TS:C43A and TS:C210A mutant proteins were nearly identical to wild-type TS, while TS:C180A and TS:C199A mutants expressed >80% of wild-type enzyme activity. In contrast, TS:C195A was completely inactive. Mutant proteins, TS:C195A, TS:C199A and TS:C210A, retained RNA binding activity to nearly the same degree as wild-type human TS. RNA binding activity of TS:C43A was reduced by 30% when compared to wild-type TS, while TS:C180A was completely devoid of RNA binding activity. In vitro translation studies confirmed that mutant proteins TS:C43A, TS:C195A, TS:C199A and TS:C210A, significantly repressed human TS mRNA translation, while TS:C180A was unable to do so. To confirm the in vivo significance of the cysteine sulfhydryl residue, mutant proteins TS:C180A and TS:C195A were each expressed in human colon cancer HCT-C18:TS(-) cells that expressed a functionally inactive TS. A recombinant luciferase reporter gene under the control of a TS-response element was co-transfected into these same cells, and luciferase activity increased in the presence of the TS:C195A mutant TS protein to a level similar to that observed upon expression of wild-type TS protein. In contrast, luciferase activity remained unchanged in cells expressing the TS:C180A mutant protein. Taken together, these findings identify Cys-180 as a critical residue for the in vitro and in vivo translational regulatory effects of human TS.

Human cytomegalovirus requires cellular deoxycytidylate deaminase for replication in quiescent cells.

We have previously observed that the expression of two thymidylate biosynthesis enzymes, dihydrofolate reductase and thymidylate synthase (TS), is upregulated in quiescent human fibroblasts infected with human cytomegalovirus (HCMV). Here, we have demonstrated that HCMV increases expression of the cellular deoxycytidylate deaminase (dCMP deaminase), which provides the substrate for TS by converting dCMP to dUMP. We observed an increase in dCMP deaminase protein levels, whereas deoxyuridine triphosphatase (dUTPase), another cellular enzyme that may provide dUMP by hydrolysing dUTP, was undetectable. The essential requirement of cellular dCMP deaminase for productive HCMV replication was further emphasized by showing that a precursor of a potent dCMP deaminase inhibitor, zebularine, suppressed virus replication and DNA synthesis. These results suggest that HCMV exploits the host's dCMP deaminase activity to replicate in quiescent cells.

Modification of Escherichia coli thymidylate synthase at tyrosine-94 by 5-imidazolylpropynyl-2'-deoxyuridine 5'-monophosphate.

Evidence is presented that 5-imidazolylpropynyl-2'-deoxyuridine 5'-monophosphate (IP-dUMP) is a mechanism-based, irreversible inactivator of Escherichia coli thymidylate synthase (TS), which covalently modifies Tyr94 at the active site of the enzyme. The inactivation of TS was time and concentration dependent and did not require the folate cofactor. Due to the rapidity of the inactivation process, accurate kinetic parameters could be determined only in the presence of saturating concentrations (1000K(M)) of the competing substrate, dUMP. Under these conditions, a K(I) of 0.36 +/- 0.09 microM and an inactivation rate constant (k(inact)) of 0.53 +/- 0.15 min(-1) were obtained from Kitz-Wilson plots. Electrospray ionization-mass spectrometry (ESI-MS) determined a 412 amu mass increase of TS after inhibition by IP-dUMP with no mass difference being detected for the TS mutants Tyr94Phe or Cys146Ala, thus indicating the importance of these residues for complex formation. The change in WT-TS mass was consistent with covalent modification by IP-dUMP, which was confirmed by proteolytic digestion of the modified protein followed by ESI-MS. By these means, a 43-residue trypsin peptide (residues 54-96), a 16-residue endoAspN peptide (residues 89-104), and an 8-residue endoAspN/endoLysC peptide (residues 89-96), each containing the IP-dUMP adduct, were observed. MS/MS analysis of the IP-dUMP-endoAspN peptide identified a modified 3-residue daughter ion, YGK (residues 94-96). A mechanistic scheme requiring the participation of Cys146 is proposed for the covalent modification of IP-dUMP by Tyr94, which, unlike an earlier proposal [Kalman, T. I., Nie, Z., and Kamat, A. (2001) Nucleosides Nucleotides Nucleic Acids 20, 869-871], does not require the release of imidazole for the activation of the inhibitor.

Retroviral transduction of a mutant dihydrofolate reductase-thymidylate synthase fusion gene into murine marrow cells confers resistance to both methotrexate and 5-fluorouracil.

Gene transfer-based myeloprotection strategies against chemotherapy require the development of effective drug resistance genes or gene combinations. Our laboratory has previously generated drug-resistant mutants of dihydrofolate reductase (DHFR F/S) and thymidylate synthase (TS G52S) for myeloprotection against methotrexate (MTX) and 5-fluorouracil (5-FU), respectively. For the purpose of conferring dual myeloprotection against both MTX and 5-FU, we have generated two retroviral constructs encoding both DHFR F/S and TS G52S as a fusion protein (DHFR F/S-TS G52S) or as individual proteins from a bicistronic gene. The DHFR F/S-TS G52S fusion protein is functional and exhibits kinetic properties similar to that of the individual mutant enzymes. NIH 3T3 cells and mouse bone marrow progenitors retrovirally transduced with the fusion DHFR F/S-TS G52S cDNA provided similar levels of resistance to MTX and 5-FU as cells expressing the individual mutant enzymes and higher levels of resistance to MTX than cells expressing DHFR F/S from the 3' end of a bicistronic gene. As MTX and 5-FU are used in combination therapy for diseases such as breast and colon cancer, this fusion gene may be useful in the clinic to reduce myelosuppressive toxicity associated with this drug combination.

Mechanistic characterization of Toxoplasma gondii thymidylate synthase (TS-DHFR)-dihydrofolate reductase. Evidence for a TS intermediate and TS half-sites reactivity.

This study describes the use of rapid transient kinetic methods to characterize the bifunctional thymidylate synthase-dihydrofolate reductase (TS-DHFR) enzyme from Toxoplasma gondii. In addition to elucidating the detailed kinetic scheme for this enzyme, this work provides the first direct kinetic evidence for the formation of a TS intermediate and for half-sites TS reactivity in human and Escherichia coli monofunctional TS and in T. gondii and Leishmania major bifunctional TS-DHFR. Comparison of the T. gondii TS-DHFR catalytic mechanism to that of the L. major enzyme reveals the mechanistic differences to be predominantly in DHFR activity. Specifically, TS ligand induced domain-domain communication involving DHFR activation is observed only in the L. major enzyme and, whereas both DHFR activities involve a rate-limiting conformational change, the change occurs at different positions along the kinetic pathway.