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 in-volve a rate-limiting conformational change, the change occurs at different positions along the kinetic pathway.

Thymidylate synthase (TS) 1 and dihydrofolate reductase (DHFR) are essential metabolic enzymes and established targets for anti-cancer and anti-microbial drugs (1)(2)(3). Although, in most species, including humans, TS and DHFR activities reside on separate monofunctional enzymes, several protozoan parasites have these activities expressed on a single polypeptide chain that comprises a bifunctional thymidylate synthasedihydrofolate reductase (TS-DHFR) 2 enzyme (4 -8). One of these protozoa, Toxoplasma gondii, is prevalent and problematic in the United States (8). Opportunistic toxoplasmosis is often associated with the onset of the clinical AIDS syndrome and is a primary cause of suffering and death in AIDS patients (6). Enzymes unique to the parasite, including the bifunctional TS-DHFR, are optimal targets for the development of new anti-parasitic drugs.
As illustrated in Scheme 1, TS catalyzes the only de novo source of deoxythymidine monophosphate (dTMP) from deoxyuridine monophosphate (dUMP). The reaction uses (6R)-L-5,10-methylenetetrahydrofolate (CH 2 H 4 F) as a cofactor for the one carbon transfer reaction generating 7,8-dihydrofolate (H 2 F) in the process. DHFR regenerates the fully reduced form of the folate (6R)-5,6,7,8-tetrahydrofolate (H 4 F) from H 2 F using NADPH as a cofactor and generating NADP ϩ in the process. H 4 F can then by primed for subsequent one-carbon transfer reactions in the cell. In the presence of dUMP, NADPH, and CH 2 H 4 F, the bifunctional TS-DHFR enzymes catalyze the conversion of CH 2 H 4 F directly to H 4 F. Scheme 2 illustrates the detailed catalytic mechanism that has been proposed for the TS reaction (9). Following ordered substrate binding, in which dUMP binds first, a conformational change takes place whereby the C-terminal tetrapeptide of TS closes over the substrates to create an active site cavity shielded from solvent (step 1) (1, 10 -14). It has been suggested that this is followed by the formation of an iminium ion involving the bridge methylene and N-5 of CH 2 H 4 F (step 2) (15).
It is this highly reactive electrophilic iminium ion that has been proposed to be the reactive form of the cofactor. In fact, the structure of a TS mutant lacking a C-terminal valine and crystallized with CH 2 H 4 F and 5-fluorodeoxyuridine monophosphate (FdUMP) in which (6R)-L-5-hydroxymethyltetrahydrofolate (HO-CH 2 H 4 F) is bound at the active site provided structural evidence supporting the formation of this iminium ion during TS catalysis (10). Kinetic isotope effect studies suggest that it is bound CH 2 H 4 F, potentially in its iminium ion form, that accumulates at the active site (9). Several steps in rapid equilibrium ensue leading up to an overall rate-limiting step involving hydride transfer to form the products dTMP and H 2 F (steps 3-6) (9).
All known TS enzymes, with the exception of a class of recently discovered flavin-dependent tetrameric TS enzymes from several non-symbiotic microbes, exist as a TS homodimer (16). It has been suggested that TS is a half-sites-reactive enzyme in which only one TS site of dimeric TS productively binds substrates (17)(18)(19). Until now, however, definitive kinetic evidence was lacking and the issue of whether TS is actually a half-sites-reactive enzyme has remained unresolved (9).
There is structural and kinetic evidence to suggest that bifunctional TS-DHFRs from certain protozoa, including Leishmania major and T. gondii exhibit electrostatic substrate channeling in which H 2 F is directly transferred between the TS and DHFR active sites without release into bulk solvent (5, 20 -22).
Previous kinetic studies from our group provided support for substrate channeling in TS-DHFR from L. major (23). These rapid transient kinetic studies on L. major also provided direct kinetic evidence for domain-domain communication in L. major TS-DHFR in which ligation of the TS active site with FdUMP and CH 2 H 4 F to form the covalent FdUMP⅐CH 2 H 4 F⅐TS-DHFR covalent complex resulted in activation of DHFR chemistry from 14 to 120 s Ϫ1 (23).
In this work, we set out to address fundamental mechanistic questions about the TS-DHFR from T. gondii; a detailed understanding of which might aid in novel anti-parasite drug development against this medically relevant organism. Specifically, we asked: What are the rate-limiting steps in the reactions catalyzed by this bifunctional enzyme? Can we gain any new insight into the catalytic mechanism of TS from this bifunctional enzyme? How does this bifunctional TS-DHFR compare with the previously characterized TS-DHFR from L. major for which a structure is known?
In this report, we describe a transient kinetic analysis using rapid chemical quench and stopped-flow methods to provide the first in-depth characterization of the reaction pathway for the bifunctional TS-DHFR from T. gondii. In addition to elucidating the detailed mechanism for the T. gondii TS-DHFR enzyme, this study provides the first direct kinetic evidence for the formation of a TS intermediate and for half-sites TS reactivity. Subsequent analysis of L. major TS-DHFR, Escherichia coli TS, and human TS provided evidence for an intermediate and for half-sites TS reactivity in these enzymes as well.
Stopped-flow Measurements-Stopped-flow measurements were performed using a Kintek SF-2001 apparatus (Kintek Instruments, Austin, TX) as previously described (23). The data were collected over a given time interval using a PC and software provided by Kintek Instruments. In experiments designed to measure dissociation rate constants, the trapping ligand was used at a concentration of Ͼ5-fold excess over that of the bound ligand to allow analysis as a pseudo-first order rate constant. Coenzyme FRET was utilized in DHFR experiments and in all ligand association and dissociation experiments involving NADPH. For these experiments, a monochromator was set to 287 nm on the input and changes in NADPH FRET were monitored with an output filter at 450 nm. For TS and TS-DHFR experiments, changes in absorbance at 340 nm were monitored. For all other ligand binding experiments, changes in fluorescence with excitation at 287 and emission at 340 nm were monitored. Due to ordered binding to TS in which nucleotide binds first, followed by folate; H 2 F binding to the DHFR domain could be isolated from that to the TS domain by performing the experiments in the absence of nucleotide (dUMP or dTMP). Likewise, binding of H 2 F to TS could be separately assessed by preincubating enzyme with a DHFR-saturating concentration of H 2 F (Ͼ5 M) prior to mixing with a large excess of nucleotide and an equimolar concentration of H 2 F. The combination of rapid chemical quench and stopped-flow methods allowed for an accurate interpretation of fluorescence and absorbance signals.
CCD-array Stopped-flow Measurements-CCD-array stopped-flow measurements were performed using a Kintek SF-2001 apparatus (Kintek Instruments, Austin, TX) and detected with an Ocean Optics PC2000 CCD linear silicon array detector. Absorption measurements were taken for 2.1 s (integration time of 3 ms) from 220 to 450 nm (over 649 elements), and the data were analyzed using a PC and Specfit/32 TM Software (Spectrum Software Associates).
Rapid Chemical Quench-Rapid chemical quench experiments were performed using a Kintek RFQ-3 Rapid Chemical Quench apparatus (Kintek Instruments, Austin, TX). The reactions were initiated by mixing enzyme solution (15 l) with radiolabeled substrate (15 l, ϳ20,000 dpm). In all cases, the concentrations of enzyme and substrates cited in the text are those after mixing and during the reaction. Reactions utilizing radiolabeled folates were terminated by quenching with 67 l of 0.78 N KOH, 10% sodium ascorbate, and 200 mM 2-mercaptoethanol. Ascorbate and 2-mercaptoethanol were added to prevent oxidative degradation of H 4 F after quenching and resulted in a pH of 12.6 for the base quench solution. Because CH 2 H 4 F is more stable under basic conditions, its solutions were maintained at a basic pH (9.5) until mixing with enzyme solution, providing a final pH of 7.8 during the reaction. TS reactions utilizing radiolabeled dUMP were quenched with 67 l of 0.4 N HCl. The quenched reaction solutions were directly collected into argon-purged Waters Wisp autosampler vials, immediately vortexed, and analyzed by HPLC in combination with radioactivity-flow detection. The substrates and products were then quantified as described below. All samples that were not immediately analyzed were stored at Ϫ80°C until just prior to analysis to minimize degradation. To confirm complete quenching of the enzymatic reactions, controls in which substrate was added to a premixed solution of enzyme and quench solution was included with each experiment.
HPLC Analysis-The substrates and products were quantified by radio-HPLC using a BDS-Hypersil C18 reverse phase column (250 ϫ 4.6 mm, Keystone Scientific, Bellefonte, PA) with a flow rate of 1 ml/min. For separation of folates, an isocratic separation using a solvent system of 10% methanol in 180 mM triethylammonium bicarbonate (pH 7.8) was used. The elution times were as follows: H 4 F, 7.5 min; TS intermediate, 12.5 min; H 2 F, 14 min, CH 2 H 4 F, 16 min. For separation of dUMP and dTMP, an isocratic separation using a solvent system of 200 mM triethylammonium bicarbonate was used. The elution times were as follows: dUMP, 11 min; dTMP, 18 min. The HPLC effluent from the column was monitored continuously using a Flo-One radioactivityflow detector (Packard Instruments, Downers Grove, IL). The analysis system was automated using a Waters 712B WISP (Milford, MA) autosampler.
Data Analysis-Rapid chemical quench single turnover and burst data were fit to single-exponential and burst equations, respectively, using the curve fitting program Kaleidagraph. Stopped-flow measurements provided estimates for the association and dissociation rate constants (k on and k off ) and for reaction rate constants. Comparison of rapid chemical quench and stopped-flow reaction time courses allowed for the assignment of observed stopped-flow rates to chemical steps or conformational changes.
Spectrophotometric TS Assay-The K m of CH 2 H 4 F was determined using a steady-state spectrophotometric kinetic assay. T. gondii TS-DHFR enzyme (25 nM) was preincubated with dUMP (100 M) prior to mixing with CH 2 H 4 F (5 to 350 M), and absorbance was monitored at 340 nm using a Hewlett-Packard 8452A spectrophotometer. Initial rates were determined in triplicate using the software provided with the instrument, and these rates were converted to units of specific activity using the reported extinction coefficient for the reaction (⌬⑀ rxn ϭ 6.4 mM Ϫ1 cm Ϫ1 ). Data in this paper are presented as the average of triplicate determination with error bars representing the standard deviation.
Kinetic Simulation-The KINSIM kinetic simulation program was used to model kinetic data presented in this report (31). The data were fit by a trial and error process, maintaining the constraints of constants measured in this study. The focus of this simulation was to validate the minimal kinetic mechanism elucidated in this study. The model and estimated rate constants are described in Chart 1. Half-sites TS reactivity was modeled by defining TS (E) and DHFR (Z) as unique species and defining the modeling parameters such that Z ϭ 2E ϭ concentration of TS-DHFR used.
T. gondii Homology Model-A homology model of T. gondii TS-DHFR was built using the Swiss PDB program in conjunction with the Swiss Model homology modeling link available at the Swiss PDB website. The C-terminal 315 amino acids (residues 295-610) and residues 115-166 were modeled using the PDB file for the L. major TS-DHFR structure. The N-terminal 52 amino acids were modeled using the PDB file for Pneumocystis carinii DHFR (PDB entry 1CD2), which was the highest homology DHFR relative to the N-terminal portion of the T. gondii DHFR domain for which a structure is available.

RESULTS
Overview of the TS-DHFR Reaction-Bifunctional TS-DHFR enzymes catalyze three basic reactions (Scheme 1), the TS reaction, the DHFR reaction, and a bifunctional TS-DHFR reaction. Moreover, each of these reactions can be further subdivided into two classes of events: binding and dissociation of ligands and those events involved in catalysis. We will first address those events involving binding and dissociation of substrates and products, and we will consider an example experiment used to measure rates for each. Those events involved in CHART 1 chemical catalysis will then be elucidated, including the occurrence of conformational changes, formation of intermediates, and the identification of rate-limiting steps. We begin with the characterization of the DHFR reaction, because it is relatively straightforward and has features used to help interpret the TS reaction. We will then consider the TS and TS-DHFR reactions, providing a complete TS-DHFR reaction mechanism and new insight into the mechanism of TS catalysis.
Kinetics of Ligand Binding-The second-order rate constants for the binding of ligands to TS-DHFR were determined by measuring the ligand concentration dependence of the observed binding rate. The apparent first-order rates for the binding of dUMP, dTMP, and H 2 F to the TS domain and for H 2 F, H 4 F, NADPH, and NADP ϩ to the DHFR domain were measured using stopped-flow fluorescence. A representative stopped-flow fluorescence trace for the binding of NADPH to the bifunctional enzyme is shown in Fig. 1a. The trace is biphasic and fits a double-exponential equation with a fast NADPH concentration-dependent phase of 83.0 Ϯ 3.9 s Ϫ1 and a slow NADPH concentration-independent phase of 5.6 Ϯ 0.5 s Ϫ1 . The fast phase represents the apparent first-order binding rate and conforms to the equation, k obs ϭ k on [L] ϩ k off , in which k obs , k on , and k off are the apparent first-order observed binding rate, association rate constant, and dissociation constant, respectively. The plot of k obs versus [L] for NADPH binding is shown in Fig. 1b. Accordingly, the binding and dissociation rate constants are 9.0 Ϯ 0.4 M Ϫ1 s Ϫ1 and 39.8 Ϯ 3.0 s Ϫ1 for the formation and dissociation of the E⅐NADPH complex, respectively, at 25°C and pH 7.8. The slow NADPH concentrationindependent phase is likely to represent a conformational change following NADPH binding. A summary of the association rate constants measured for the various ligands with the bifunctional TS-DHFR enzyme is shown in Table I.
Kinetics of Ligand Dissociation-The rate constants for the dissociation of ligands from TS-DHFR were measured by ligand competition experiments. A representative stopped-flow trace for the measurement of the dissociation of H 2 F from the E⅐H 2 F complex using methotrexate as the trapping ligand is shown in Fig. 2. The resulting trace was fit to a singleexponential equation with a rate of 9.0 Ϯ 0.1 s Ϫ1 , corresponding to the dissociation rate constant, k off , for H 2 F release from the DHFR domain of TS-DHFR. A summary of the dissociation constants obtained from these experiments is shown in Table II.
The DHFR Reaction-The first experiment examining the DHFR activity of the T. gondii TS-DHFR enzyme is a presteady-state burst experiment. As shown in Fig. 3a, a burst in DHFR catalysis is observed at a rate of 180 Ϯ 20 s Ϫ1 . This is followed by slow steady-state product accumulation at a rate of a The error associated with the measurement of each rate constant is Ͻ10% in all cases. b k off as determined from the binding curve (Fig. 1b). The rate of 7 s Ϫ1 for NADPH dissociation from E⅐NADPH likely reflects the rate of release from a different TS-DHFR conformer than that involved in the binding experiment.
c The off-rate of CH 2 H 4 F was estimated from the product of K d and K on and by modeling. 5.7 Ϯ 0.6 s Ϫ1 , corresponding to a rate-limiting step preceding subsequent turnover. The observed burst amplitude was 31.1 Ϯ 1.4 M, within the error of the enzyme concentration used, suggesting that essentially 100% of the DHFR sites are active. The second experiment is a single turnover stopped-flow fluorescence experiment, the time course for which is shown in Fig.  3b. Because enzyme-bound NADPH but not NADP ϩ exhibits coenzyme fluorescence resonance energy transfer, the time course for the fluorescence at 450 nm represents the conversion of NADPH to NADP ϩ at the active site and, hence, the rate of catalysis. The data was fit to a single-exponential equation with a rate (k chem ) of 180 Ϯ 2.7 s Ϫ1 consistent with the rate of chemistry observed in the burst reaction. There was no increase in the rate of catalysis under conditions in which the enzyme concentration was doubled, indicating that substrate binding was not limiting.
T. gondii DHFR Activation Experiment-Experiments designed to examine whether there was domain-domain communication involving DHFR activation in T. gondii TS-DHFR, as previously observed with the L. major enzyme, were performed. Both DHFR single turnover and burst experiments were similar to those described above, except that the enzyme was preincubated with FdUMP and CH 2 H 4 F to form the covalent FdUMP⅐CH 2 H 4 F⅐TS-DHFR covalent complex. In contrast to the DHFR activation observed with L. major TS-DHFR (14 to 120 s Ϫ1 ), no significant change in the T. gondii DHFR rate was observed (data not shown).
The TS and TS-DHFR Reactions: A Burst in TS Activity?-The first experiment examining the TS activity of the T. gondii TS-DHFR enzyme was a pre-steady-state burst experiment in which enzyme was preincubated with excess [ 14 C]dUMP prior to mixing with a large excess of CH 2 H 4 F. The time course for the reaction is shown in Fig. 4a Fig. 4b. In contrast to the linear, steady-state consumption of dUMP, this reaction occurs with a burst in [ 3 H]CH 2 H 4 F consumption. A third TS burst experiment was conducted using stopped-flow absorbance. As shown in Fig. 4c, a burst in absorbance at 340 nm consistent with the rapid chemical quench burst experiment is observed.
The Observation of a TS Intermediate-As shown in   Fig. 5 (top), when a TS-DHFR burst reaction is conducted, the burst in  (Fig. 6a). A shift in the isosbestic point to 337.0 nm was observed when the reaction was repeated using 40 M CH 2 H 4 F (all other conditions were identical) (Fig. 6b). Under single turnover conditions (25 M TS and 10 M CH 2 H 4 F) the isosbestic point was observed at 338.1 nm (Fig. 6c). To determine the isosbestic point for the quantitative conversion of CH 2 H 4 F to H 2 F, the reaction was conducted under steady-state conditions in which E. coli TS (2.5 M) was preincubated with dUMP (500 M) prior to mixing with CH 2 H 4 F (200 M). The resulting isosbestic point was observed at 322.7 nm (Fig. 6d). Similar CCD-array stopped-flow TS experiments were conducted using bifunctional T. gondii and L. major TS-DHFR. Similar shifts in the isosbestic point of the TS reaction spectra for the bifunctional enzymes were observed (data not shown).

DHFR reactions allows one to monitor the bifunctional TS-DHFR reaction and thereby monitor the direct conversion of [ 3 H]CH 2 H 4 F to [ 3 H]H 4 F. As shown in
Isolation of TS Catalytic Rates-To isolate the rates of TS catalytic events, a TS-DHFR burst experiment was conducted, the time course for which, is shown in Fig. 7a. The reaction occurs with a burst in [ 3 H]CH 2 H 4 F consumption at a rate of 130 Ϯ 10 s Ϫ1 coupled to the rapid accumulation of an intermediate species up to a steady-state concentration corresponding to the burst amplitude for the reaction. The burst in [ 3 H]CH 2 H 4 F consumption is followed by a rate-limiting steadystate phase of 6.2 Ϯ 0.6 s Ϫ1 in which [ 3

H]H 4 F is formed without significant [ 3 H]H 2 F accumulation.
Evidence for Half-sites TS Reactivity-Another vital piece of information provided by the burst experiment shown in Fig. 7a is the burst amplitude, representing the concentration of TS active-sites for this reaction. The burst amplitude is 58.9 Ϯ 11.4 M, or roughly 50% the concentration of TS-DHFR used, and of DHFR active sites, consistent with half-sites TS reactivity.
To examine the CH 2 H 4 F concentration dependence of the burst rate and amplitude in the TS burst reaction, a second series of TS bursts, similar to that shown in Fig. 4c, was performed by stopped-flow absorbance at 340 nm. In this series, the bifunctional T. gondii TS-DHFR enzyme (25 M) was preincubated with a large excess of dUMP (1 mM) prior to mixing with excess CH 2 H 4 F (40 -1500 M). Plotting the observed burst rate versus CH 2 H 4 F concentration suggested that the relationship was a linear one, consistent with rate-limiting CH 2 H 4 F association (k on ϳ 1 M Ϫ1 s Ϫ1 ), up to the maximum burst rate of 105 Ϯ 4 s Ϫ1 (data not shown) rather than a hyperbolic one that would be consistent with rapid equilibrium and a weak K d . By contrast, the observed burst amplitude versus CH 2 H 4 F concentration displayed a hyperbolic relationship suggesting that the apparent K d for CH 2 H 4 F binding was 18 Ϯ 5 M (Fig. 8a). This is consistent with the steady-state K m of 17.0 Ϯ 2.0 M determined for the reaction (Fig. 8b).
A TS-DHFR single turnover series was also performed to determine the rate of the rate-limiting chemical step in the TS reaction (k chem ) and to examine the TS-DHFR concentration dependence on the observed single turnover rate. In this series, the bifunctional TS-DHFR enzyme (12.5-100 M) was preincubated with a large excess of dUMP and NADPH prior to mixing with limiting [ 3 H]CH 2 H 4 F. Fig. 8 (c and d) shows that the reaction time course for this series displays a hyperbolic dependence on the concentration of enzyme used. The enzyme concentration at which the observed rate is half-maximal (K d apparent) for this series is 45 Ϯ 8 M, and the maximum rate (k chem ) is 5.5 Ϯ 0.4 s Ϫ1 .
A Kinetic Model for the T. gondii TS-DHFR Reaction-The mechanistic information obtained in this study was used to formulate a minimal kinetic mechanism for the bifunctional TS-DHFR from T. gondii (Scheme 3). The TS reaction is depicted in Scheme 3a, and the DHFR reaction is in Scheme 3b. This mechanism along with the rates obtained in this study was then used to simulate reaction time courses using the program KIN-SIM. The KINSIM model and rate constants used are described in Chart 1.The resulting simulations were consistent with experimental data for the TS-DHFR reactions (Fig. 7b).
Bursts in L. major TS and TS-DHFR-The TS and TS-DHFR bursts observed in T. gondii TS-DHFR represent a marked mechanistic difference from our previous study of L. major TS-DHFR. The L. major TS and TS-DHFR reactions were therefore re-analyzed under identical conditions to the T. gondii reactions described above. The resulting time courses suggested that CH 2 H 4 F binding was rate-limiting under the experimental conditions employed in our previous study in which no burst in TS activity was observed. A representative TS-DHFR burst time course for L. major TS-DFHR is shown in Fig. 9a. This experiment suggests that, like that for T. gondii TS-DHFR, rapid conversion of CH 2 H 4 F to intermediate occurs and that, like for T. gondii, this is followed by an overall rate-limiting chemical step associated with dTMP formation. As with T. gondii TS-DHFR, the burst amplitude of the TS-DHFR burst reactions and the steady state concentration of intermediate that accumulates during the time course are ap- proximately equal to one-half the TS-DHFR concentration used. TS burst experiments were conducted with L. major TS-DHFR at corresponding concentrations and were consistent with the TS-DHFR reaction (data not shown).
Bursts in Monofunctional Human and E. coli TS-The observation of a catalytic burst in the TS reaction of the L. major and T. gondii TS-DHFR bifunctional enzymes also suggested a marked mechanistic difference from previous reports on monofunctional TS enzymes (9). To verify this difference, monofunc-tional human and E. coli TS enzymes were analyzed under identical burst conditions to those used for the bifunctional enzymes. Surprisingly, the time courses for the monofunctional TS enzymes also displayed a burst in CH 2 H 4 F consumption consistent with the rapid formation of a TS intermediate preceding a rate-limiting step in TS chemistry. Fig. 9 (b and c) shows the time courses for the monofunctional E. coli and human TS enzymes, respectively. As with the bifunctional TS-DHFR enzymes, the monofunctional TS burst amplitudes were consistent with half-sites TS reactivity.
Homology Model of T. gondii TS-DHFR-To obtain a possible structural explanation for the difference in DHFR activities observed, specifically the lack of TS-ligand-induced DHFR activation in T. gondii TS-DHFR that is observed in L. major, a homology model of T. gondii TS-DHFR was built as described under "Materials and Methods." As shown in Fig. 10, the homology model suggests that the T. gondii TS-DHFR enzyme lacks an N-terminal tail, linking the DHFR and TS domains, that is present in the crystal structure of L. major TS-DHFR. DISCUSSION In this work, we have characterized the complete kinetic scheme for bifunctional TS-DHFR from T. gondii. In addition to providing the detailed enzymatic mechanism for this important chemotherapeutic target, this study compares the mechanism with that from bifunctional TS-DHFR from L. major for which a structure is known, highlighting both similarities and differences in mechanism. Finally, the use of rapid transient kinetics methods in this study provides the first direct kinetic evidence for half-sites TS reactivity and for the accumulation of an intermediate during TS catalysis.
An Overview of the T. gondii TS-DHFR Reaction-For the T. gondii TS reaction (Scheme 3a), the overall rate-limiting step occurs during chemistry at a rate (k chem ) of 5.5 s Ϫ1 . This is consistent with previous studies with other TS enzymes, which have suggested that a hydride transfer step occurring in the final step of TS catalysis is rate-limiting (9). For the T. gondii DHFR reaction (Scheme 3b), chemistry occurs at a relatively fast rate (k chem ) of 180 s Ϫ1 , compared with an overall rate-limiting conformational change (k ss ) at a rate of 5.6 s Ϫ1 that occurs immediately after NADPH binding.
Comparison of L. major and T. gondii DHFR Mechanisms-As might be predicted on the basis of relative sequence homology, the differences between L. major and T. gondii TS-DHFR reside primarily in the DHFR mechanisms. The first major difference is the lack of DHFR activation in the T. gondii enzyme. The homology model of T. gondii, when compared with the crystal structure of L. major, suggests a possible explanation for this difference. We postulate that the 23-amino acid tail of L. major that is absent in the T. gondii homology model may serve to inhibit DHFR activity in a TS conformation-specific manner. This domain-domain communication mechanism may serve as a TS activity sensor activating DHFR activity during TS catalysis, further coupling the sequential TS and DHFR activities.
A second notable difference between the L. major and T. gondii DHFR mechanisms is found in the location of the ratelimiting step for each. In both cases, the rate-limiting step involves a conformational change; however, this step occurs in different places along the respective DHFR pathways. As shown in Scheme 4, the rate-limiting conformational change for the T. gondii enzyme occurs after NADPH binding, whereas it takes place after NADP ϩ release from E⅐H 4 F⅐NADP ϩ in the L. major enzyme.
A final subtle, but noteworthy, difference between the DHFR mechanisms for these species is that the product release pathway for L. major DHFR is kinetically restricted to one path, whereas product release in T. gondii can occur via multiple kinetically competent paths (Scheme 4). This is not to say that multiple product release pathways are not utilized by the L. major enzyme, but that only one of the pathways is kinetically competent in that it contains no one rate slower than the overall rate-limiting step for the enzyme. In contrast, all the possible product release pathways in the T. gondii enzyme are kinetically competent and therefore likely to be equally utilized.
Evidence for Half-site TS Reactivity-It has been suggested that TS is a half-sites-reactive enzyme in which only one TS monomer of dimeric TS is catalytically active at a time. This proposal is based on structural evidence that suggests that TS binds ligands asymmetrically to each monomer's active site and mutagenesis studies in which an active site and non-active site TS dead mutant are combined to form a heterodimer with fully restored TS activity (17)(18)(19). To our knowledge, however, this study provides the first direct kinetic evidence for half-sites reactivity in TS enzymes.  The pre-steady-state burst amplitude observed for the TS reaction in T. gondii TS-DHFR, which reflects the concentration of TS active sites, is approximately one-half the concentration of TS used, whereas the corresponding DHFR burst experiment suggests that the enzyme contains essentially 100% DHFR active sites. In addition, analysis of the concentration dependence of TS activity under single turnover and burst conditions suggests that this half-sites reactivity is a result of asymmetric substrate binding. Specifically, the observation that the apparent K d for the single turnover series in which the enzyme is in excess and determines binding is approximately twice the K d for the burst series in which CH 2 H 4 F determines binding suggests that only one-half of the TS present in solution can productively bind CH 2 H 4 F. Taken together, these studies provide the first direct kinetic evidence in support of half-sites TS reactivity, in which only one TS monomer binds substrate productively.
It is worth noting that a burst amplitude corresponding to 50% enzyme concentration might also be explained by two other scenarios: (i) 50% misfolded enzyme or (ii) a 50 -50 equilibrium between two forms of the enzyme prior to chemistry. The former is unlikely, because a 50% burst amplitude was observed with multiple enzyme preps and with TS from various species, and the latter scenario is unlikely on the basis of the results of the current as well as previous studies, which are consistent with half-site reactivity resulting from asymmetric binding of TS ligands.
Evidence for a TS Intermediate-It has also been proposed that TS catalysis involves the formation of an iminium ion form of CH 2 H 4 F and that this is the reactive form of the cofactor. Moreover, there is structural evidence involving a mutant of TS lacking a C-terminal valine crystallized with CH 2 H 4 F and FdUMP in which (6R)-L-5-hydroxymethyltetrahydrofolate (HO-CH 2 H 4 F) was found to be bound at the active site suggesting the formation of the putative iminium ion during TS catalysis (10). However, this study provides the first direct kinetic evidence for the accumulation of a TS intermediate during TS catalysis.
The presence of a burst in substrate (CH 2 H 4 F) consumption without a corresponding burst in product formation and the observation of a rapidly formed transient species not attributable to substrate or product in HPLC analyses provide the first direct kinetic evidence for the accumulation a TS intermediate.
The accumulation of a TS intermediate is further supported by the shift in the isosbestic point for the TS reaction in the stopped-flow CCD experiments. Specifically, under steadystate conditions or burst conditions where CH 2 H 4 F is in large excess over enzyme, a single isosbestic point is observed at ϳ322 nm for the TS reaction. Under conditions where enzyme concentration becomes significant relative to CH 2 H 4 F, such as under single turnover conditions, there is a shift in the isosbestic point to ϳ337 nm. Stadman et al. (32) have shown that the wavelength of an isosbestic point may change under varying experimental conditions provided (i) either the molar absorptivity of the substrate changes under the varying experimental conditions or (ii) the fraction of the substrate that is converted to multiple products changes. The observed shift in the isosbestic point during the TS reaction is consistent with the latter, in which there is formation of an enzyme-bound intermediate species that becomes increasingly significant as the reaction conditions approach single turnover conditions.
Considering the proposed TS mechanism (Scheme 2), the results of previous TS mechanistic studies, and the results of this study, we suggest that this TS intermediate is the putative SCHEME 4