Dihydrofolate Reductase from Pyrimethamine-resistant Plasmodium

SUMMARY An investigation of the Hz-folate reductases from pyri-methamine-sensitive (Pb/WLTM) and -resistant (Pb/ WLTM/50-63) strains of Plasmodium berghei has revealed that the specific activity of the enzyme from the resistant strain is IO-fold higher than that of the sensitive strain. Since turnover numbers (based on amethopterin binding) are equivalent, the increase in specific activity must be due to an increased number of catalytic sites. Marked differences were found between the two enzymes in several of the properties studied. The K, value for HZ-folate is 12.4-fold higher for the enzyme from the resistant strain. The Pb/WLTM/50-63 Hz-folate reductase is only slightly stimulated by KCI, in contrast to the 3-fold stimulation observed with the enzyme from the sensitive strain. The K; values for pyrimethamine and two other antifolates are higher for the enzyme from the resistant strain. In addition, the inhibition is competitive with Hz-folate for the enzyme from Pb/WLTM and noncompetitive for the enzyme from Pb/WLTM/50-63. It is proposed that Pb/WLTM/50-63 is resistant in vivo to the action of pyrimethamine because of the combined of


SUMMARY
An investigation of the Hz-folate reductases from pyrimethamine-sensitive (Pb/WLTM) and -resistant (Pb/ WLTM/50-63) strains of Plasmodium berghei has revealed that the specific activity of the enzyme from the resistant strain is IO-fold higher than that of the sensitive strain. Since turnover numbers (based on amethopterin binding) are equivalent, the increase in specific activity must be due to an increased number of catalytic sites.
Marked differences were found between the two enzymes in several of the properties studied.
The K, value for HZ-folate is 12.4-fold higher for the enzyme from the resistant strain. The Pb/WLTM/50-63 Hz-folate reductase is only slightly stimulated by KCI, in contrast to the 3-fold stimulation observed with the enzyme from the sensitive strain.
The K; values for pyrimethamine and two other antifolates are higher for the enzyme from the resistant strain.
In addition, the inhibition is competitive with Hzfolate for the enzyme from Pb/WLTM and noncompetitive for the enzyme from Pb/WLTM/50-63.
It is proposed that Pb/WLTM/50-63 is resistant in vivo to the action of pyrimethamine because of the combined effects of the increase in enzyme content and the decrease in inhibitor binding. It is suggested that these alterations in enzymatic properties and enzyme levels reflect mutations in the gene or genes responsible for the coding of this enzyme.
The antifolate properties of the antimalarial agent pyrimethamine' have been well established (1). It has been shown to be a moderately potent inhibitor of Hz-folate reductase (EC 1.5.1.3) from various organisms (2). Recently it was demonstrated that the basis of the chemotherapeutic effectiveness of pyrimethamine in the rodent malarial organism, Plasmodium berghei, is the extremely tight binding of drug to the parasite Hz-folate reductase, in contrast to that to the enzyme from the host (3). Resistance to pyrimethamine has been found in human malaria (4) and can be produced in plasmodial infections in laboratory animals (5,6). With the demonstration of the plasmodial Hzfolate reductase as the locus of action of this drug (3), it became possible to investigate the involvement of this enzyme in resistance. Changes in the amounts and properties of Hz-folate reductases have been observed in some organisms resistant to antifolates (usually close structural analogues of the substrate such as aminopterin and amethopterin) . Antifolate resistance in mammalian cells in most cases has involved increased enzyme levels, with no detectable changes in the enzymatic characteristics (e.g. see Reference 7). On the other hand, qualitative alterations in the enzyme have been observed in several antifolate-resistant bacteria (8-10). Genetic analysis in resistant Diplococcus pneumoniae (8) and Xalmonella typhimurium (10) revealed the presence of multiple loci involved in antifolate resistance.
Evidence is presented in this paper demonstrating changes in the properties of the Hz-folate reductase isolated from a pyrimethamine-resistant strain of P. berghei. It is proposed that the observed decrease in drug binding and the increased level of this enzyme are sufficient to account for the drug resistance of this strain in viva.

METHODS
The reagents used, the spectrophotometric Hz-folate reductase assay, and the enzyme isolation procedure have been described previously (3). The enzyme isolation, in brief, consisted of the lysing of washed, infected erythrocytes with saponin, washing the liberated parasites free from red cell debris and hemoglobin, and then passing the plasmodia through a French pressure cell at 15,000 to 20,000 p.s.i. After centrifugation, freeze-thawing, and a second centrifugation, the crude extracts were passed over a Sephadex G-100 column.
Since the plasmodial Hz-folate reductases were 9-to lo-fold larger than the host enzyme (3), they were eluted at a position immediately after the void volume of the column, whereas any contaminating host enzyme was eluted near two void volumes (and thus was not included in the pool of malarial Hz-folate reductase).
The reciprocal plots were analyzed with the aid of computer programs designed by Cleland (11  from the sensitive strain by passage in the presence of increasing levels of pyrimethamine; after 50 weeks the strain was fully resistant to the maximum tolerated dose (see Table II). This resistance level was retained even when the strain was passed in untreated mice. When inoculated into mice at the Wellcome Laboratories of Tropical Medicine, England, both strains invaded mainly mature erythrocytes, killing the mice in 6 to 7 days. In this laboratory in the United States mainly immature erythrocytes were invaded by the parasites, resulting in average survival times of 14 to 19 days for the mice. Thus, this strain differs from the P. berghei NYU-2 strain previously investigated (3), since the latter developed in mature red cells. Consequently, parasitized blood for enzyme isolation was taken on the 13th to 14th day of infection in this study, in contrast to the 5th to 6th day used with the Pb/NYU-2 strain. In this laboratory, the sensitive and resistant strains were transferred biweekly, and the resistant strain was maintained in the presence of the maximum tolerated dose of pyrimethamine (25 mg per kg, intraperitoneal, on alternate days for six to seven doses). Infected mice for enzyme studies were not treated.
The drug sensitivities in vivo were determined at the Wellcome Laboratories of Tropical Medicine in England. Mice were inoculated intraperitoneally with 5 x lo6 parasitized erythrocytes. Groups of five mice were given seven oral doses of drug starting on the afternoon of the day of infection and were then treated twice a day for the following 3 days. On the 4th day parasitemias were determined from stained tail smears, and the dosage of drug which was required to reduce parasitemia to 50% of the untreated control value was calculated.

RESULTS
Compa,rison of Hz-folate Reductases of Pyrimethamine-sensitive P. berghei Strains-The Hz-folate reductase from Pb/WLTM was essentially identical to the enzyme from the Pb/NYU-2 strain previously reported (3) in most characteristics studied. No significant differences between the two strains could be detected in pH optimum, specific activity, 2-mercaptoethanol stimulation, 50% inhibitory level of several antifolates, and the type of binding of pyrimethamine. The values obtained for these parameters for the enzyme from Pb/WLTM were almost identical with those published for the Pb/NYU-2 Hz-folate reductase (3  both enzymes were similar to certain mammalian (13,14) and avian (15) Hz-folate reductases in their ability to be stimulated by certain reagents (such as KCl), optimum stimulation of activity was found at different KC1 concentrations for the two plasmodial enzymes. The maximum, a-fold stimulation of activity was at 0.2 to 0.5 M KC1 for the Pb/NYU-2 Hz-folate reductase (3), and at 0.15 M KC1 for the Pb/WLTM enzyme. This stimulation is mainly due to the anion, since NH&l and NaCl were as active as KC1 in stimulating the enzyme from Pb/NYU-2, but KBr and KNOJ were only 80% as active, and sodium acetate was less then 60% as effective as KCl. Previous data on Tris buffer salts (3) support this contention.
Comparison of Hz-folate Reductases of Pyrimethamine-sensitive and -resistant P. berghei-"-The Hz-folate reductase isolated from the resistant strain, Pb/WLTM/50-63, differed greatly from the enzyme from the sensitive strain. The K, value for Hz-folate is 12.4-fold higher for the enzyme from Pb/WLTM/50-63 than for that from Pb/WLTM (Table I).
Xpecijic Activity-It was found that the specific activity of the resistant strain Hz-folate reductase was lo-fold higher than that from the sensitive strain (Table I). Since the increased specific activity could be due to an increase in enzyme concentration, or catalytic activity (or both), it was of interest to determine the turnover numbers for the two enzymes. Amethopterin binds Hz-folate reductases stoichiometrically under certain conditions, and Werkheiser (16) used this property to determine turnover numbers in terms of moles of Hz-folate reduced per min per mole of amethopterin bound. Enzyme preparations of the sensitive and resistant strains were titrated with amethopterin, and the activity per drug binding site was calculated for each. The values obtained were essentially identical (Table I). Thus if the number of amethopterin binding sites per molecule are the same for the two enzymes, the difference in specific activities observed is due to an increased amount of enzyme in the resistant strain.
E$ects of Cl--A difference was observed in the response of the two enzymes to KC1 in the assays. The activity of the Hz-folate reductase from Pb/WLTM was stimulated 3-fold at 0.15 M KCl, whereas the resistant strain enzyme was only slightly stimulated at 0.15 M KC1 and was inhibited at higher concentrations (Fig. 1). This difference was not due to the testing of Pb/WLTM/50-63 at Hn-folate concentrations subsaturating for this enzyme.  In Part a the enzyme from Pb/'WLTM was incubated for 5 min at 37" with Hz-folate alone (o), or with 1.05 nM pyrimethamine.
In Part b the enzyme from Pb/WLTM was assayed in the usual manner, with Hz-folate added after the preincubation period ( l , no drug; x, 0.6 nM pyrimethamine).
In Part c the enzyme from Pb/WLTM/50-63 was assayed in the usual manner with Hz-folate added after the incubation period of the enzyme alone (0) or enzyme plus 20 nM pyrimethamine (x). Velocity (v) equals decrease in OD,,, in 4 min.

Pb/WLTM
was stimulated a-fold at 0.15 M KC1 when tested at a Hz-folate concentration which was subsaturating (5 PM), as well as at the usual assay concentration of 50 pM.
Inhibitor Bindings-The binding of several antifolates differed in nature and degree in the Pb/WLTM and Pb/WLTM/50-63 enzymes. The Ki values for trimethoprim and cycloguanil increased 9.6-and 16.6-fold, respectively (Table TI). Moreover, both drugs were competitive with the substrate with the sensitive strain enzyme, but noncompetitive for the enzyme from the resistant strain.
It had been reported that the binding of pyrimethamine to the enzyme from Pb/NYU-2 was affected by the assay conditions (3). Stoichiometric inhibition was observed when enzyme and drug were incubated for 5 min at 37" before the reaction was initiated by the addition of Hz-folate.
If the Hz-folate was included during the incubation (reaction started with NADPH), the binding was reversible.
In the present study, the identical mode of binding was demonstrated in the Hz-folate reductase from Pb/WLTM, but not from Pb/WLTM/50-63. Incubation of drug and enzyme, in the presence or absence of Hz-folate, had no effect on the reversibility of the pyrimethamine inhibition of the Pb/WLTM/50-63 enzyme.
The inhibition was noncompetitive with Hz-folnte (Fig. 2). This was also true of the inhibition of the resistant strain enzyme by the other two antifolates tested. The type of inhibition observed with pyrimethamine against the Pb/WLTM enzyme varied with the assay conditions used. Noncompetitive kinetics resulted when drug and enzyme were incubated prior to the addition of Hz-folate, whereas the inhibition was competitive when the Hz-folate was included in the incubation mixture (Fig. 2).

DISCUSSION
The Hz-folate reductase isolated from Pb/WLTM/50-63 exhibited several properties quite distinct from those of the enzyme from the sensitive strain.
Changes such as increased enzyme content and decreased antifolate binding bear directly upon the ability of the parasite to survive in the presence of otherwise lethal concentrations of drug, and they thus provide a basis for understanding the mechanism or mechanisms of resistance of this strain.
Differences in K, values for Hz-folate and response to KC1 offer further evidence of the dissimilar nature of the two enzymes.
The data strongly suggest the presence of an increased content of a structurally altered Hz-folate reductase in Pb/WLTM/50-63, presumably due to mutations in the gene or genes responsible for the synthesis of this enzyme.
The increase in apparent K, value for Hz-folate and the increases in the Ki values for the antifolates tested could be related phenomena.
A substitution of an amino acid involved in binding ante values obtained are all minimal values based on the ratio of the maximum tolerated doses for mice to the EDSos against the sensitive strain. However, even though no definite values can be placed on the degree of resistance to the three drugs, inspection of Table II reveals that the minimal values calculated are all much greater than the ratios of Ki values for the two enzymes. Thus, the resistance in vivo cannot be due simply to decreased drug binding by the resistant strain enzyme.
The mechanism of resistance of Pb/WLTM/50-63 is more akin to that reported for certain strains of antifolate-resistant bacteria; both decreased affinity for inhibitor and increased enzyme levels are important factors in determining the degree of resistance (8-10).
The specific activity for Hz-folate reductase in extracts of Pb/WLTM/50-63 is II-fold higher than in extracts of Pb/WLTM.
The turnover numbers based on amcthoptcrin titration were essentially identical, indicating that the greater specific activity of the resistant strain enzyme is due to an increased number of catalytic sites and not increased catalytic activity per site. Increase in Hz-folate reductase content has commonly been found in antifolate-resistant cells of mammalian and bacterial origin, frequently associated with another factor such as decreased inhibitor permeability (see Reference 19) or decreased inhibitor binding (8-10).
Antifolate-resistant strains of D. pneumoniae (8) and S. typhimurium (10) have been reported with multiple mutations affecting both the levels and properties of the Hz-folate reductase.
It seems highly likely that multiple genetic loci also are involved in the resistance of Pb/WLTM/50-63. This strain was developed by constant exposure to partially inhibitory doses of pyrimethamine; dosage was increased only when the level used became ineffective.
Since full resistance developed in a stepwise manner over a period of 50 weeks, it is probable that more than one mutation was involved.
The final level of resistance attained appears to be due mainly to the cooperative effects of separate phenomena: increased enzyme content and decreased inhibitor binding.
Although other factors (e.g. drug permeability, substrate levels, end product utilization) might be involved, it seems likely that the differences observed in enzyme levels and properties are sufficient in nature and magnitude to account for the ability of the resistant strain to grow and multiply in the presence of drug levels lethal to the sensitive strain.
of the Hz-folate might also affect inhibitor affinity, since it is felt that substrate and analogues share certain common binding points (2,17). However, the difference in antifolate binding between the two enzymes was not only quantitative.
Differences in the l/v intercept of reciprocal plots of control and inhibitor curves are normally interpreted to mean that the inhibitor combines with a different form of the enzyme than does the substrate (18.). The inability of Hz-folate to completely overcome the antifolate inhibitions of the Pb/WLTM/50-63 enzyme thus indicates that these inhibitors combine with a different form of the enzyme than does Ha-folate.
Since this does not occur with the enzyme from the sensitive strain, it may be inferred that the resistant strain Hz-folate reductase may exist in one state that cannot be attained by the Pb/WLTM enzyme. This could be attributed to different primary structures for the two proteins.
The two distinct patt.erns of chloride effects ( Fig. 1) are also suggestive of different conformational forms for the two enzymes. It has been proposed that activation of Hz-folate reductase by salts and other reagents involves conformational changes in the enzyme protein, resulting in increases in V,,, and possibly in apparent K, values (13)(14)(15).
The lack of stimulation of the Pb/WLTM/50-63 Hz-folate reductase and the higher Hz-folate K, value may indicate that this enzyme exists in an already "activated" state. Thus the inhibition of the enzyme from the resistant strain found at increasing KC1 levels may be similar to the inactivations noted for Hz-folate reductases from various sources at reagent concentrations higher than necessary for optimum activation (13)(14)(15).
The enzymatic characteristics of the Hz-folate reductase isolated from Pb/WLTM were very similar to those reported previously for the Pb/NYU-2 strain (3). The small differences encountered (Km values for Hz-folate and KC1 stimulation patterns) are probably reflections of minor variations of the protein structures of the two enzymes, although unknown in6uences due to the growth of the parasites in erythrocytes of different ages cannot be discounted.
In either case, the intrastrain variations in enzymatic characteristics are minor when compared to the major differences (i.e. molecular weight, substrate K, values, inhibitor binding, urea stability, etc. (3)) reported between the plasmodial Hz-folate reductase and this enzyme from various other organisms.
It is difficult to correlate quantitatively the enzyme changes found with the dcgrcc of resistance observed in viva. The Pb/WLTM/50-63 strain is completely resistant to pyrimethamine and the two other antifolates tested (Table II).
Since the dose levels used were limited by the toxicity to the host, the resists Fig. 2a