Thymidylate synthase protein expression in colorectal cancer metastases predicts for clinical outcome to leucovorin-modulated bolus or infusional 5-fluorouracil but not methotrexate-modulated bolus 5-fluorouracil.

BACKGROUND: Different 5-fluorouracil (5-FU) schedules and/or biochemical modulators may result in different mechanisms of cytotoxicity, potentially affecting the correlation between thymidylate synthase (TS) expression and the clinical response to the fluoropyrimidine. PATIENTS AND METHODS: TS levels were measured immunohistochemically on archival specimens of colorectal cancer metastases from 124 patients homogeneously treated in a series of clinical trials at our institutions with: (A) leucovorin (LV)-modulated infusional 5-FU (n = 48); (B) LV-modulated bolus 5-FU (n = 41); (C) methotrexate (MTX)-modulated bolus 5-FU (n = 35). RESULTS: A statistically significant correlation between TS levels and the clinical response was observed with the regimens involving continuous infusion and/or LV modulation (response rate in patients with low and high TS: 66% versus 24%, P = 0.003, and 50% versus 0%, P = 0.0001, in group A and B, respectively). Conversely, TS levels failed to predict the clinical response within the group of patients treated with MTX-modulated bolus 5-FU (response rate 21% versus 13%, P = 0.50, with low and high TS, respectively). Consistently, the median time to progression/overall survival time in patients with low and high TS were 9 versus 6 months/19 versus 14 months (P = 0.009/0.035, group A), 8 versus 2 months/12 versus 6 months (P = 0.002/0.0006, group B) and 3 versus 2 months/12 versus 13 months (P = 0.14/0.74, group C). CONCLUSIONS: The correlation between intratumoral TS levels and the clinical response to 5-FU depends strongly on the schedule of administration/biochemical modulators that are used in different 5-FU regimens. These data strengthen the notion that different 5-FU schedules have different mechanisms of cytotoxicity.

Parameters affecting the restoration of activity to inactive mutants of thymidylate synthase via subunit exchange: further evidence that thymidylate synthase is a half-of-the-sites activity enzyme.

In a previous study we demonstrated that Escherichia coli thymidylate synthase activity could be restored completely by incubating basically inactive mutants of this enzyme at room temperature with R(126)E, another inactive mutant [Maley, F., Pedersen-Lane, J., and Changchien, L.-M. (1995) Biochemistry 34, 1469-1474]. Since only one of the enzyme's two subunits possessed a functional active site and the restoration of activity could be titrated to be equivalent to that of the wild-type enzyme's specific activity, it was proposed that thymidylate synthase was a half-of-the-sites activity enzyme. We now provide additional support for this thesis by presenting an in-depth analysis of some conditions affecting the restoration of enzyme activity. For this purpose, we employed two mutants with marginal thymidylate synthase activity, Y(94)A and R(126)E. The parameters that were examined included pH, concentration of protein, temperature, and urea concentration, all of which influenced the rate of activity restoration. It was found, surprisingly, that by maintaining the amount of each protein constant, while increasing the volume of solution, the rate and total activity restored was greatly enhanced. Increasing the pH from 6.0 to 9.0 markedly increased the rate at which the optimal activity was restored, as did increasing the temperature from 4 to 40 degrees C. A similar effect was obtained when the incubation of the mutants was conducted at 4 degrees C in the presence of 1.5 M urea, a temperature at which activity is restored extremely slowly. Raising the pH to 9.0 resulted in an almost instantaneous restoration of activity at 4 degrees C. The manner in which thymidylate synthase activity is restored from the mutants in the presence of varying concentrations of ethanol, ethylene glycol, and glycerol suggests that changes in subunit interaction and enzyme conformation are in part responsible for the observed differences. Most significantly, at solution levels of 10%, ethanol was found to activate, while ethylene glycol inhibited slightly and glycerol was somewhat more inhibitory. At a concentration of 20%, ethanol inhibited rather strikingly, ethylene glycol was slightly more inhibitory than at 10%, and glycerol was strongly inhibitory. Since the net result of these findings is the suggestion that the restoration of thymidylate synthase activity is due to a separation of the mutant dimers into their respective subunits, followed by their recombination to an active heterodimer, evidence for this phenomenon was sought by separating the recombined dimers using nondenaturating polyacrylamide gel electrophoresis. Sequence analysis of the isolated homo- and heterodimers clearly demonstrated that the active enzyme is a product of subunit exchange, one that is very efficient relative to the wild-type enzyme, which did not exchange subunits unless denatured.

Crystal structure of a deletion mutant of human thymidylate synthase Delta (7-29) and its ternary complex with Tomudex and dUMP.

The crystal structures of a deletion mutant of human thymidylate synthase (TS) and its ternary complex with dUMP and Tomudex have been determined at 2.0 A and 2.5 A resolution, respectively. The mutant TS, which lacks 23 residues near the amino terminus, is as active as the wild-type enzyme. The ternary complex is observed in the open conformation, similar to that of the free enzyme and to that of the ternary complex of rat TS with the same ligands. This is in contrast to Escherichia coli TS, where the ternary complex with Tomudex and dUMP is observed in the closed conformation. While the ligands interact with each other in identical fashion regardless of the enzyme conformation, they are displaced by about 1.0 A away from the catalytic cysteine in the open conformation. As a result, the covalent bond between the catalytic cysteine sulfhydryl and the base of dUMP, which is the first step in the reaction mechanism of TS and is observed in all ternary complexes of the E. coli enzyme, is not formed. This displacement results from differences in the interactions between Tomudex and the protein that are caused by differences in the environment of the glutamyl tail of the Tomudex molecule. Despite the absence of the closed conformation, Tomudex inhibits human TS ten-fold more strongly than E. coli TS. These results suggest that formation of a covalent bond between the catalytic cysteine and the substrate dUMP is not required for effective inhibition of human TS by cofactor analogs and could have implications for drug design by eliminating this as a condition for lead compounds.

In vitro selection of an RNA sequence that interacts with high affinity with thymidylate synthase.

Previous studies have shown that the repressive effect of thymidylate synthase (TS) mRNA translation is mediated by direct binding of TS itself to two cis-acting elements on its cognate mRNA. To identify the optimal RNA nucleotides that interact with TS, we in vitro synthesized a completely degenerate, linear RNA pool of 25 nt and employed in vitro selection to isolate high affinity RNA ligands that bind human TS protein. After 10 rounds of selection and amplification, a single RNA molecule was selected that bound TS protein with nearly 20-fold greater affinity than native, wild-type TS RNA sequences. Secondary structure analysis of this RNA sequence predicted it to possess a stem-loop structure. Deletion and/or modification of the UGU loop element within the RNA sequence decreased binding to TS by up to 1000-fold. In vivo transfection experiments revealed that the presence of the selected RNA sequence resulted in a significant increase in the expression of a heterologous luciferase reporter construct in human colon cancer H630 and TS-overexpressing HCT-C:His-TS+ cells, but not in HCT-C18 cells expressing a functionally inactive TS. In addition, the presence of this element in H630 cells leads to induced expression of TS protein. An immunoprecipitation method using RT-PCR confirmed a direct interaction between human TS protein and the selected RNA sequence in transfected human cancer H630 cells. This study identified a novel RNA sequence from a degenerate RNA library that specifically interacts with TS.

Escherichia coli thymidylate synthase expression protects human cells from the cytotoxic effects of 5-fluorodeoxyuridine more effectively than human thymidylate synthase overexpression.

In this study, we compared the relative abilities of human thymidylate synthase (hTS) and Escherichia coli thymidylate synthase (eTS) expression to confer resistance to the cytotoxic effects of treatment with the TS inhibitor 5-fluorodeoxyuridine (FdUrd). G418-selected clones expressing either form of the protein were significantly more resistant than the lacZ-expressing clone, VALZ2, to FdUrd-induced cytotoxicity. Although eTS-expressing clones expressed 2- to 3-fold more TS protein than hTS-overexpressing clones, the representative eTS-expressing clone, VAEG8, and hTS-overexpressing clone, VAHGC, were equally sensitive to an FdUrd-induced loss of clonogenicity; in addition, a large fraction of either form of exogenously expressed TS appeared to be inactive in the intact cell. The clones differed, however, in their responses to leucovorin (LV). Although LV significantly enhanced FdUrd-induced TS inhibition, growth inhibition, and cytotoxicity in VAHGC cells, it had no effect on these parameters in VAEG8 cells. These results suggest that eTS may more efficiently confer resistance to FdUrd plus LV when expressed for the purposes of a "host protection" strategy in vivo.

High-level expression of Escherichia coli and Bacillus subtilis thymidylate synthases.

Procedures are described for the preparation of highly purified thymidylate synthases from Escherichia coli and Bacillus subtilis. The yields in each case are quite high with about 350 mg of pure protein obtained from 1 liter of cells. Basically all that is required to obtain pure enzyme is an induction step from a high-expression vector, followed by a DE-52 column elution. Both enzymes appeared to be fairly stable in that incubation at 43 degrees C for 10 min resulted in the loss of 50% of the E. coli thymidylate synthase activity, while 50 degrees C for 10 min was required to obtain the same effect with the B. subtilis enzyme. In the presence of the substrate, dUMP, each protein was stabilized further by 6 to 7 degrees C, which was increased to 9 to 10 degrees C on addition of dihydrofolate. It was shown also that the E. coli thymidylate synthase could be maintained at 4 degrees C for at least 4 months with little or no loss in activity provided that mercaptoethanol was not present. The presence of the latter led to a progressive loss in activity until little activity could be detected after 18 weeks, which was due, in part, to the formation of a disulfide bond with the active site cysteine. Addition of dithiothreitol restored the enzyme activity to its original state.

A T4-phage deoxycytidylate deaminase mutant that no longer requires deoxycytidine 5'-triphosphate for activation.

A deoxycytidylate (dCMP) deaminase encoded in T4-bacteriophage DNA that is induced on phage infection of Escherichia coli was shown earlier (Maley, G. F., Duceman, B. W., Wang, A. M., Martinez, J. M., and Maley, F. (1990) J. Biol. Chem. 265, 47-51) to be similar in size, properties, and amino acid composition to the T2-phage-induced deaminase. Neither enzyme is active in the absence of dCTP or its natural activator, 5-hydroxymethyl-dCTP. However, on changing the arginine (Arg) at residue 115 of the T4-deaminase to either a glutamate (R115E) or a glutamine (R115Q), the resulting mutant enzymes were active in the absence of dCTP, with each mutant possessing a turnover number or k(cat) that is about 15% that of the wild-type deaminase. When compared on the basis of specific activity, however, the mutants are about 40-50% of the wild-type (WT)-enzyme's specific activity. Molecular weight analysis on the wild-type and mutant deaminases using HPLC size exclusion chromatography revealed that the wild-type deaminase was basically a hexamer, particularly in the presence of dCTP, regardless of the extent of dilution. Under similar conditions, R115E remained a dimer, whereas R115Q and F112A varied from hexamers to dimers particularly at concentrations normally present in the assay solution. Activity measurements appear to support the conclusion that the hexameric form of the enzyme is activated by dCTP, while the dimer is not. Another feature emphasizing the difference between the WT and mutant deaminases was observed on their denaturation-renaturation in EDTA, which revealed the mutants to be restored to 50% of their original activities with the WT deaminase only marginally restored.

Characterization of a cis-acting regulatory element in the protein coding region of thymidylate synthase mRNA.

Thymidylate synthase (TS) functions as an RNA-binding protein by interacting with two different sequences on its own mRNA. One site is located in the 5'-upstream region of human TS mRNA while the second site is located within the protein coding region corresponding to nt 434-634. In this paper, a 70 nt RNA sequence, corresponding to nt 480-550, was identified that binds TS protein with an affinity similar to that of full-length TS mRNA and TS434-634 RNA. In vitro translation studies confirmed that this sequence is critical for the translational autoregulatory effects of TS. To document in vivo biological significance, TS sequences contained within this region were cloned onto the 5'-end of a luciferase reporter plasmid and transient transfection experiments were performed using H630 human colon cancer cells. In cells transfected with p644/TS434-634 or p644/TS480-550, luciferase activity was decreased 2.5-fold when compared to cells transfected with p644 plasmid alone. Luciferase mRNA levels were identical for each of these conditions as determined by RNase protection and RT-PCR analysis. Immunoprecipitation of TS ribonucleoprotein complexes revealed a direct interaction between TS protein and TS480-550 RNA in transfected H630 cells. Treatment with 5-fluorouridine resulted in a nearly 2-fold increase in luciferase activity only in cells transfected with p644/TS434-634 and p644/TS480-550. This study identifies a 70 nt TS response element in the protein coding region of TS mRNA with in vitro and in vivo translational regulatory activity.

Thymidylate synthase protein expression in primary colorectal cancer compared with the corresponding distant metastases and relationship with the clinical response to 5-fluorouracil.

Thymidylate synthase (TS) expression in colorectal cancer metastases has been shown to predict for the clinical response to 5-fluorouracil. Because primary tumors may easily provide accessible sources of tissue for marker analysis, we have investigated the stability of TS expression between primary colorectal cancer and the corresponding distant metastases and compared their relative ability to predict response to chemotherapy on a series of 27 patients homogeneously treated with biochemically modulated fluorouracil for advanced disease. By immunohistochemistry, high levels of TS expression were observed in 19 of 27 (70%) primary tumors and in 13 of 27 (48%) metastatic samples. Overall, TS levels observed in primary tumors did not correlate with those measured in the corresponding metastases (r = 0.30, P = 0.13), with higher TS levels in primary tumors in 8 of 10 discordant cases. Accordingly, the degree of TS immunoreactivity was significantly higher in primary tumors compared with the corresponding metastases (mean TS score 3.8; median, 4 versus 2.8; median 3; P = 0.001). Response rates after chemotherapy for metastatic disease were similar for patients with low and high TS levels in their primary tumors (37% versus 53%, P = 0.47). In contrast, response rates were 71% and 23% in patients with low and high TS in metastatic samples (P = 0.012), respectively. In summary, TS levels measured in primary colorectal cancer do not reflect those observed in the corresponding metastases and cannot be used to predict their response to chemotherapy. The basis for the higher TS content of primary colorectal cancer compared with the corresponding metastases needs clarification.

Immunohistochemical quantitation of thymidylate synthase expression in colorectal cancer metastases predicts for clinical outcome to fluorouracil-based chemotherapy.

PURPOSE: To determine whether immunohistochemical thymidylate synthase (TS) quantitation predicts for clinical outcome in patients with advanced colorectal cancer treated by fluorouracil (FUra)-based chemotherapy. PATIENTS AND METHODS: TS levels were measured immunohistochemically on archival specimens of colorectal cancer metastases from 48 patients homogenously treated by bolus FUra plus methotrexate alternating with continuous-infusion FUra plus leucovorin. These measurements were retrospectively correlated with patient characteristics and clinical outcome. RESULTS: A significant correlation was found between intratumoral TS expression and all the parameters of clinical outcome analyzed. In patients whose tumors had low (n = 27) and high (n = 21) TS levels, the overall response rates were 67% and 24%, respectively (P =.003). The percentage of tumor shrinkage after chemotherapy was linearly related to TS immunoreactivity (r =.56, P =.00004), and its mean values were 65% and 14% with low and high TS levels, respectively (P =.0001). By logistic regression analysis, low TS expression was the single best predictor of response to chemotherapy (relative probability, 5.0). In patients with low and high TS expression, the median time to progression was 9.6 months v 6.2 months (P =.005) and the median survival time 18.4 months v 15.4 months (P =.02), respectively. Two- and 3-year survival rates were 41% v 15% and 19% v 0% (P =.02), respectively. CONCLUSION: In this cohort of homogenously treated patients, intratumor TS content was a major predictor of clinical outcome. Immunohistochemical TS quantitation provides a convenient, low-cost technique for identifying patients unresponsive to TS inhibitors who may be candidates for alternative chemotherapy regimens.

Glycosylation of the overlapping sequons in yeast external invertase: effect of amino acid variation on site selectivity in vivo and in vitro.

Yeast invertase contains 14 sequons, all of which are glycosylated to varying degrees except for sequon 5 which is marginally glycosylated, if at all. This sequon overlaps with sequon 4 in a sequence consisting of Asn92-Asn93-Thr94-Ser95(Reddy et al., 1988, J. Biol. Chem., 263, 6978-6985). To determine whether glycosylation at Asn93is sterically hindered by the oligosaccharide on Asn92, the latter amino acid was converted to a glutamine residue by site-directed mutagenesis of the SUC2 gene in a plasmid vector which was expressed in Saccharomyces cerevisiae. A glycopeptide encompassing sequons 3 through 6 was purified from a tryptic digest of the mutagenized invertase and sequenced by Edman degradation, which revealed that Asn93 of sequon 5 contained very little, if any, carbohydrate, despite the elimination of sequon 4. When Ser and Thr were inverted to yield Asn-Asn-Ser-Thr carbohydrate was associated primarily with the second sequon, in agreement with numerous studies indicating that Asn-X-Thr is preferred to Asn-X-Ser as an oligosaccharide acceptor. However, when the invertase overlapping sequons were converted to Asn-Asn-Ser-Ser, both sequons were clearly glycosylated, with the latter sequon predominating. These findings rule out steric hindrance as a factor involved in preventing the glycosylation of sequon 5 in invertase. Comparable results were obtained using an in vitro system with sequon-containing tri- and tetrapeptides acceptors, in addition to larger oligosaccharide acceptors.

Crystal structure of thymidylate synthase A from Bacillus subtilis.

Thymidylate synthase (TS) converts dUMP to dTMP by reductive methylation, where 5,10-methylenetetrahydrofolate is the source of both the methylene group and reducing equivalents. The mechanism of this reaction has been extensively studied, mainly using the enzyme from Escherichia coli. Bacillus subtilis contains two genes for TSs, ThyA and ThyB. The ThyB enzyme is very similar to other bacterial TSs, but the ThyA enzyme is quite different, both in sequence and activity. In ThyA TS, the active site histidine is replaced by valine. In addition, the B. subtilis enzyme has a 2.4-fold greater k(cat) than the E. coli enzyme. The structure of B. subtilis thymidylate synthase in a ternary complex with 5-fluoro-dUMP and 5,10-methylenetetrahydrofolate has been determined to 2.5 A resolution. Overall, the structure of B. subtilis TS (ThyA) is similar to that of the E. coli enzyme. However, there are significant differences in the structures of two loops, the dimer interface and the details of the active site. The effects of the replacement of histidine by valine and a serine to glutamine substitution in the active site area, and the addition of a loop over the carboxy terminus may account for the differences in k(cat) found between the two enzymes.

Regulation of p53 expression by thymidylate synthase.

Previous studies showed that thymidylate synthase (TS), as an RNA binding protein, regulates its own synthesis by impairing the translation of TS mRNA. In this report, we present evidence that p53 expression is affected in a similar manner by TS. For these studies, we used a TS-depleted human colon cancer HCT-C cell that had been transfected with either the human TS cDNA or the Escherichia coli TS gene. The level of p53 protein in transfected cells overexpressing human TS was significantly reduced when compared with its corresponding parent HCT-C cells. This suppression of p53 expression was the direct result of decreased translational efficiency of p53 mRNA. Similar results were obtained upon transfection of HCT-C cells with pcDNA 3.1 (+) containing the E. coli TS gene. These findings provide evidence that TS, from diverse species, specifically regulates p53 expression at the translational level. In addition, TS-overexpressing cells with suppressed levels of p53 are significantly impaired in their ability to arrest in G1 phase in response to exposure to a DNA-damaging agent such as gamma-irradiation. These studies provide support for the in vivo biological relevance of the interaction between TS and p53 mRNA and identify a molecular pathway for controlling p53 expression.

Structural and functional analysis of surface domains unique to bacteriophage T4 thymidylate synthase.

The bacteriophage T4 genome encodes most of its own enzymes for dNTP synthesis, which form a complex in infected Escherichia coli. The T4 thymidylate synthase (TS) and the T4 deoxycytidylate deaminase (CD) catalyze sequential reactions and are physically linked within this complex [McGaughey, K. M., Wheeler, L. J., Moore, J. T., Maley, G. F. , Maley, F., and Mathews, C. K. (1996) J. Biol. Chem. 271, 23037-23042]. From the crystal structure of T4TS [Finer-Moore, J. S., Maley, G. F., Maley, F., Montfort, W. R., and Stroud, R. M. (1994) Biochemistry 33, 15459-15468], it appears that three regions corresponding to insertions relative to E. coli TS lie on one side of the enzyme surface. We have investigated the residual activity of T4TS in response to complete deletion or substitution mutagenesis of these insertions. Two deletions generated in the small domain (residues 70-103) reduced the TS activity to 0.2% and 0.7% of the wild-type activity, with the latter able to complement growth of a thyA- E. coli strain in the absence of thymidine. By insertion of exogenous sequences variable in length and in sequence into these deletion mutants, enzyme activity increased to 44% that of the wild type. Restoration of the TS activity depended mostly on the hydrophobicity of the inserted residues. The sites of insertions also displayed distinct permissiveness for accommodating the exogenous insertions. Deletions and substitutions near the C-terminus resulted in complete inactivation of the T4TS. Proteolysis experiments revealed that the modified surface loops of the small domain were still accessible and flexible for protein-protein interactions. We have used ELISA to detect a physical association between T4TS and T4CD and compared the binding affinity of WT T4TS for two purified insertion mutants of T4CD. The results obtained showed that the native sequences of the small domain inserts are not required for T4TS/T4CD complex formation.

Crystal structures of rat thymidylate synthase inhibited by Tomudex, a potent anticancer drug.

Two crystal structures of rat thymidylate synthase (TS) complexed with dUMP and the anticancer drug Tomudex (ZD1694) have been determined to resolutions of 3.3 and 2.6 A. Tomudex is one of several new antifolates targeted to TS and the first to be approved for clinical use. The structures represent the first views of any mammalian TS bound to ligands and suggest that the rat protein undergoes a ligand-induced conformational change similar to that of the Escherichia coli protein. Surprisingly, Tomudex does not induce the "closed" conformation in rat TS that is seen on binding to E. coli TS, resulting in inhibitor atoms that differ in position by more than 1.5 A. Several species-specific differences in sequence may be the reason for this. Phe 74 shifts to a new position in the rat complex and is in van der Waals contact with the inhibitor, while in the E. coli protein the equivalent amino acid (His 51) hydrogen bonds to the glutamate portion of the inhibitor. Amino acids Arg 101, Asn 106, and Met 305 make no contacts with the inhibitor in the open conformation, unlike the equivalent residues in the E. coli protein (Thr 78, Trp 83, and Val 262). dUMP binding is similar in both proteins, except that there is no covalent adduct to the active site cysteine (Cys 189) in the rat structures. Two insertions in the rat protein are clearly seen, but the N-termini (residues 1-20) and C-termini (residues 301-307) are disordered in both crystal forms.

Thymidylate synthase protein and p53 mRNA form an in vivo ribonucleoprotein complex.

A thymidylate synthase (TS)-ribonucleoprotein (RNP) complex composed of TS protein and the mRNA of the tumor suppressor gene p53 was isolated from cultured human colon cancer cells. RNA gel shift assays confirmed a specific interaction between TS protein and the protein-coding region of p53 mRNA, and in vitro translation studies demonstrated that this interaction resulted in the specific repression of p53 mRNA translation. To demonstrate the potential biological role of the TS protein-p53 mRNA interaction, Western immunoblot analysis revealed nearly undetectable levels of p53 protein in TS-overexpressing human colon cancer H630-R10 and rat hepatoma H35(F/F) cell lines compared to the levels in their respective parent H630 and H35 cell lines. Polysome analysis revealed that the p53 mRNA was associated with higher-molecular-weight polysomes in H35 cells compared to H35(F/F) cells. While the level of p53 mRNA expression was identical in parent and TS-overexpressing cell lines, the level of p53 RNA bound to TS in the form of RNP complexes was significantly higher in TS-overexpressing cells. The effect of TS on p53 expression was also investigated with human colon cancer RKO cells by use of a tetracycline-inducible system. Treatment of RKO cells with a tetracycline derivative, doxycycline, resulted in 15-fold-induced expression of TS protein and nearly complete suppression of p53 protein expression. However, p53 mRNA levels were identical in transfected RKO cells in the absence and presence of doxycycline. Taken together, these findings suggest that TS regulates the expression of p53 at the translational level. This study identifies a novel pathway for regulating p53 gene expression and expands current understanding of the potential role of TS as a regulator of cellular gene expression.

Probing the folate-binding site of human thymidylate synthase by site-directed mutagenesis. Generation of mutants that confer resistance to raltitrexed, Thymitaq, and BW1843U89.

Human thymidylate synthase (TS) contains three highly conserved residues Ile-108, Leu-221, and Phe-225 that have been suggested to be important for cofactor and antifolate binding. To elucidate the role of these residues and generate drug-resistant human TS mutants, 14 variants with multiple substitutions of these three hydrophobic residues were created by site-directed mutagenesis and transfected into mouse TS-negative cells for complementation assays and cytotoxicity studies, and the mutant proteins expressed and characterized. The I108A mutant confers resistance to raltitrexed and Thymitaq with respective IC50 values 54- and 80-fold greater than wild-type but less resistance to BW1843U89 (6-fold). The F225W mutant displays resistance to BW1843U89 (17-fold increase in IC50 values), but no resistance to raltitrexed and Thymitaq. It also confers 8-fold resistance to fluorodeoxyuridine. Both the kinetic characterization of the altered enzymes and formation of antifolate-resistant colonies in mouse bone marrow cells that express mutant TS are in accord with the IC50 values for cytotoxicity noted above. The human TS mutants (I108A and F225W), by virtue of their desirable properties, including good catalytic function and resistance to antifolate TS inhibitors, confirm the importance of amino acid residues Ile-108 and Phe-225 in the binding of folate and its analogues. These novel mutants may be useful for gene transfer experiments to protect hematopoietic progenitor cells from the toxic effects of these drugs.

Mode of action of site-directed irreversible folate analogue inhibitors of thymidylate synthase.

5,8-Dideazafolate analogues are tight binding but not irreversible inhibitors of thymidylate synthase (TS). However, when a chloroacetyl (ClAc) group is substituted at the N10-position of 2-desamino-2-methyl-5,8-dideazafolate (DMDDF), the resulting compound, ClAc-DMDDF, although still a reversible inhibitor (KI = 3.4 x 10(-3) M), gradually inactivates thyA-TS irreversibly at a rate of 0.37 min-1. The corresponding iodoacetyl derivative alkylated the enzyme somewhat slower (k3 = 0.15 min-1 ) than ClAc-DMDDF but was bound more tightly (KI = 1.4 x 10(-5) M), resulting in a second-order rate constant (k3/KI) of inactivation that was 100-fold greater than that of ClAc-DMDDF. A tryptic digest of the ClAc-DMDDF-inactivated enzyme yielded a peptide on HPLC, which revealed that cysteine-146, the residue at the active site that is intimately involved in the catalytic process, had reacted with ClAc-DMDDF to form a covalent bond. This derivative was confirmed indirectly by Edman analysis and more directly by mass spectrometry. Deoxyuridine 5'-monophosphate, a substrate in the catalytic reaction, protected against inactivation. Similar to previously described Lactobacillus casei TS inhibition studies with sulfhydryl reagents [Galivan, J., Noonan, J., and Maley, F. (1977) Arch. Biochem. Biophys. 184, 336-345], the kinetics of inhibition suggested that complete inhibition occurs on reaction of only one of the two active site cysteines, although sequence and amino acid analysis revealed that iodoacetate and ClAc-DMDDF had reacted with both active site cysteines. These studies demonstrate that a sulfhydryl reactive compound that is directed to the folate binding site of TS may diffuse to the active site cysteine, and form a covalent bond with this residue. How this inhibition comes about is suggested in a stereoscopic view of the ligand when modeled to the known crystal structure of Escherichia coli TS.

High-level expression of human thymidylate synthase.

A method is presented for expressing human thymidylate synthase (TS) to the extent of 25-30% of the protein in Escherichia coli. By this procedure, 200-400 mg of pure enzyme can be obtained from a 2-liter culture of cells. The key to the level of expression appears to be related to the conversion of purine bases in the third, fourth, and fifth codons of the TS cDNA to thymine, without altering the encoded protein product. Conversion of the penultimate proline to a leucine did not diminish expression, but while the isolated native enzyme contained only proline on its amino-terminal end, the mutated enzyme was found to contain methionine on its amino terminus. By contrast, the expression of the unmodified TS cDNA represented only about 0.1-0.2% of the total cellular protein. Unlike recombinant rat and human TSs, the respective enzymes purified to homogeneity from eukaryotic cells were blocked at the amino ends and possessed 2- to 4-fold lower specific activities. To determine at what level the impairment of expression occurred, an in vitro transcription, translation system was employed and the results showed that while transcription was unaffected, the translation of native TS mRNA was reduced by at least 20-fold relative to modified TS mRNA using a rabbit reticulocyte translation system. Thus, it appears that at least for the TS gene, expression is greatly influenced by the GC content of the 5' coding region of the gene in both prokaryote and eukaryote systems.