The Two Phosphofructokinase Gene Products of Entamoeba histolytica*

Two phosphofructokinase genes have been described previously in Entamoeba histolytica. The product of the larger of the two genes codes for a 60-kDa protein that has been described previously as a pyrophosphate (PPi)-dependent enzyme, and the product of the second, coding for a 48-kDa protein, has been previously reported to be a PPi-dependent enzyme with extremely low specific activity. Here it is found that the 48-kDa protein is not a PPi-dependent enzyme but a highly active ATP-requiring enzyme (k cat = 250 s− 1) that binds the cosubstrate fructose 6-phosphate (Fru-6-P) with relatively low affinity. This enzyme exists in concentration- and ATP-dependent tetrameric active and dimeric inactive states. Activation is achieved in the presence of nucleoside triphosphates, ADP, and PPi, but not by AMP, Pi, or the second substrate Fru-6-P. Activation by ATP is facilitated by conditions of molecular crowding. Divalent cations are not required, and no phosphoryl transfer occurs during activation. Kinetics of the activated enzyme show cooperativity with Fru-6-P (Fru-6-P0.5 = 3.8 mm) and inhibition by high ATP and phosphoenolpyruvate. The enzyme is active without prior activation in extracts of E. histolytica. The level of mRNA, the amount of enzyme protein, and the enzyme activity of the 48-kDa enzyme are about one-tenth that of the 60-kDa enzyme in extracts of E. histolytica trophozoites.

The pH dependence and apparent substrate affinities of the cloned enzyme were identical to those of the PP i -PFK in trophozoite extracts, indicating that the product of the cloned gene accounts for most if not all of the PFK activity in E. histolytica trophozoites (3).
The smaller gene, which codes for a 48-kDa protein, has been expressed in E. coli as a fusion protein that was found to have a much lower specific activity than that of the larger enzyme (1). Whereas the 60-kDa PFK has been purified from the amoeba, no information concerning the expression of the 48-kDa protein is available. The 48-kDa PFK described in the earlier studies is clearly an expressed product in E. histolytica because it was cloned from a cDNA library (1). It may have been present in extracts of the organism but did not copurify with the 60-kDa product or with the activity of PP i -PFK (3). Furthermore, if a second activity had been present which represented at least 10% of the total PFK activity, it would have been detected in native gel electrophoresis.
The problem in attributing a significant role to the 48-kDa protein in phosphorylation of fructose 6-phosphate (Fru-6-P) is its extremely low specific activity with PP i as a phosphoryl donor. The specific activity of the 60-kDa enzyme is about 2,000 -3,000 times higher than that reported for the smaller PFK (3). Thus, if expressed at the same level in the organism, the smaller PFK would be virtually undetectable under normal assay conditions for PP i -PFK. One possibility is that the smaller PFK has a yet to be determined catalytic activity. Another possibility is that the 48-kDa protein represents a regulatory protein as one observes in the multisubunit structure of plant PP i -PFKs (4). In the instance of the plant enzymes, the catalytic and regulatory subunits copurify. This was shown to be unlikely regarding the two E. histolytica PFKs in the earlier study (3) because no 48-kDa protein was present in the partially purified fractions of the 60-kDa enzyme from E. histolytica.
In the current work, we compare expression of the two forms of E. histolytica PFK in extracts of trophozoites. The 48-kDa PFK has been purified to homogeneity from both native and recombinant sources and has been found to have no detectable activity with PP i as a phosphoryl donor. On the other hand, the enzyme has high activity with ATP as a phosphoryl donor, but only after prior activation with ATP.

EXPERIMENTAL PROCEDURES
Expression Constructs-Two oligonucleotide primers designed on the basis of the sequence at the 5Ј-and 3Ј-ends of the 48-kDa PFK gene and containing additional nucleotides at the 5Ј-ends to generate NdeI and BamHI restriction sites were used to amplify by polymerase chain reaction (PCR) a fragment containing the 48-kDa PFK gene from a genomic clone (2). The PCR fragment was then cloned into the pCR-Script SK(ϩ) plasmid using the PCR-Script cloning kit as directed by the manufacturer (Stratagene, La Jolla, CA). The plasmid construct was digested with NdeI and BamHI to isolate the fragment containing the 48-kDa PFK gene. The digested fragment was then cloned into the complimentary sites of the pJC45 prokaryotic expression vector (5) (a gift from Dr. Iris Bruchhaus of the Bernard Nocht Institute for Tropical Medicine, Hamburg, Germany). The pJC45 expression vector generates a fusion protein with an additional N-terminal sequence that includes a stretch of 10 consecutive histidine residues.
To utilize a second expression system, the above PCR-Script SK(ϩ) plasmid construct containing the 48-kDa PFK gene was digested with NdeI and EcoRI and cloned into the complementary sites of the pALTER-Ex1 plasmid (Promega). The E. histolytica 60-kDa PFK gene cloned into the pALTER-Ex1 has been described previously (3).
Enzyme Preparation-The recombinant 60-kDa PP i -PFK was purified as previously described (1). The enzyme preparation was homogeneous on the basis of 10% SDS-PAGE. The enzyme was stored in 50% glycerol at Ϫ20°C. Before being used for kinetic assays, the enzyme was dialyzed against at least 400 volumes of 150 mM KTes, 1 mM EDTA, pH 7.2.
The pJC45 vector containing the 48-kDa gene was transformed into BL2 (DE3)[pAPlacIQ] E. coli (a gift from Dr. Bruchhaus), and the bacteria were plated onto LB medium agar plates with 100 g/ml ampicillin, 50 g/ml kanamycin, and 2% (w/v) glucose at 37°C. Freshly transformed single colonies were inoculated into LB medium with 100 g/ml ampicillin, 50 g/ml kanamycin, and 2% (w/v) glucose and grown at 37°C until the bacterial culture reached an absorbance of 0.2 at 600 nm. After induction for 3 h in the presence of isopropyl-␤-D-thiogalactoside, the recombinant fusion protein was purified using the His-Bind System (Novagen, Madison, WI) following the manufacturer's recommendations for native purification of cytoplasmic proteins.
The 48-kDa PFK lacking the N-terminal polyhistidine sequence and inserted into pALTER-Ex1 was expressed as follows. The plasmid construct was transformed into DF1020 E. coli, which was grown on LB. After induction by 0.4 mM isopropyl-␤-D-thiogalactoside for 12-24 h at 30°C, the cells were harvested by centrifugation at 5,000 ϫ g for 5 min and resuspended in ϳ2 volumes of ice-cold buffer consisting of 50 mM Tris-HCl, 0.1 mM EDTA, 14 mM ␤-mercaptoethanol, pH 7.4 (extraction buffer). Phenylmethylsulfonyl fluoride was added to the extraction buffer to a final concentration of 1 mM only during the extraction step. The cells were lysed by sonication and centrifuged to remove debris. The supernatant was loaded on a 15-ml column of N-6-aminohexylcarboxymethyl-ATP-Sepharose (ATP-Sepharose) (6) preequilibrated with the extraction buffer. The column was then washed with extraction buffer until the absorbance of the flow-through was below 0.02 at 280 nm. The enzyme was then eluted with extraction buffer plus 1 mM ATP. Elution fractions were pooled and concentrated to a volume of ϳ10 ml, and the enzyme was exchanged simultaneously into 20 mM Tris-HCl, 0.1 mM EDTA, 14 mM ␤-mercaptoethanol, pH 7.2, using a membrane filtration apparatus. The concentrated protein was then applied to a Mono Q HR 5/5 anion exchange column on a fast protein liquid chromatography system (Amersham Pharmacia Biotech) preequilibrated with the same buffer. The enzyme was eluted with a linear gradient of 0 -0.5 M NaCl in the same buffer. The enzyme, which eluted at ϳ100 mM NaCl, was homogeneous on the basis of 10% SDS-PAGE. The purified enzyme was stored in 50% glycerol at Ϫ20°C. Prior to kinetic assays, ultrafiltration was used to exchange the preparation to assay buffer.
The 48-kDa PFK lacking the added N-terminal polyhistidine was also purified by chromatography using Blue Sepharose (Cibacron Blue F3G-A, immobilized on Sepharose CL-6B, Sigma). Harvested bacteria were resuspended in extraction buffer plus 1 mM phenylmethylsulfonyl fluoride. Cells were then lysed as described above for ATP affinity column purification. After centrifugation, the supernatant was loaded onto a 100-ml column of Blue Sepharose preequilibrated with extraction buffer. The column was then washed with extraction buffer until the absorbance of the flow-through was below 0.02 at 280 nm. The enzyme was then eluted with extraction buffer containing 1 mM ATP. Elution fractions were pooled and concentrated to a volume of ϳ10 ml, and the buffer was changed simultaneously to the Mono Q buffer described in the ATP affinity purification section. The concentrated protein was then purified on a Mono Q anion exchange column as described for the ATP affinity purification procedure. The resultant enzyme preparation was homogeneous on the basis of 10% SDS-PAGE. The enzyme was stored and prepared for activity analysis as described above. The Blue Sepharose method yielded ϳ10 fold greater amounts of pure enzyme per unit column volume than the ATP-Sepharose procedure. Using Blue Sepharose, the overall yield from the lysate was ϳ55%.
Assay of the 48-kDa PFK-To measure activity, the 48-kDa PFK was first activated by preincubating at standard activation conditions unless otherwise indicated. The standard activation conditions were 4 M 48-kDa PFK and 2 mM ATP in 150 mM KTes (pH 7.2), 3 mM MgCl 2 , 1 mM EDTA at 30°C for 30 min. Aliquots of the preincubations were then diluted 10-fold in a standard dilution buffer (2 mM ATP and 20 mM Fru-6-P in 150 mM KTes (pH 7.2), 3 mM MgCl 2 , 1 mM EDTA) unless otherwise indicated, and fixed amounts of the dilution were added to assay cuvettes to start the reaction. The reactions were conducted at standard assay conditions (1 mM ATP and 20 mM Fru-6-P in the aforementioned assay medium at 30°C in a 1-ml assay cuvette) unless otherwise indicated. Activity was determined spectrophotometrically by measuring the decrease of absorbance at 340 nm. The measured rate of the first 60 s of the reaction was recorded. For the determination of kinetic constants, one of the two substrates (a nucleoside triphosphate and Fru-6-P) was kept saturated while the other substrate was varied from 0.1 to 10 K m . The magnesium ion concentration was kept 4 mM higher than the concentration of nucleoside triphosphates for all assays containing nucleotide to ensure that virtually all of the nucleotides existed as the magnesium complex. All nucleoside triphosphate solutions were determined to contain less than 0.1% PP i . Nucleoside triphosphate decomposition in the assay cuvette to its nucleoside monophosphate and pyrophosphate constituents was undetectable.
Kinetic estimates in this study were obtained using unweighted linear or nonlinear least squares regressions to the Michaelis-Menten and Hill models using the GraFit graphical analysis program. All assays were repeated at least twice, and standard errors of intercepts and slopes were all less than 10%. The k cat values were calculated assuming one active site existed per subunit and with a subunit mass of 48 kDa.
ATP-dependent phosphorylation of sugar substrates other than Fru-6-P was determined by measuring the generation of ADP. The 48-kDa PFK was first activated using standard activation conditions. The enzyme was then diluted 10-fold in 150 mM KTes (pH 7.2), 3 mM MgCl 2 , 1 mM EDTA, and identical quantities were added subsequently to assay cuvettes containing the same components plus 0.2 mM NADH, 1 mM ATP, 5 mM phosphoenolpyruvate (PEP), 5 units each of lactate dehydrogenase and pyruvate kinase, and 10 mM indicated sugar substrate. The initial velocities were determined spectrophotometrically by measuring the decrease of absorbance at 340 nm.
Antibody Preparation and Purification-Antibodies against 60-kDa PFK and histidine-tagged 47-kDa PFK were raised in New Zealand White rabbits. Approximately 200 g of enzyme with adjuvant was injected at 2, 4, and 8 weeks, and blood was removed 3 days after the last injection. For further purification where required, each preparation was purified by passing the polyclonal antibody-containing serum through a column of the respective PFK linked to CNBr-activated Sepharose 4B. In the case of the preparation of the 48-kDa PFK Sepharose column, the enzyme without the histidine tag was used. The columns were washed extensively with 0.1 M Tris-HCl, 0.3 M NaCl, pH 8.0, until the absorbance of the flow-through was below 0.01 at 280 nm. Specific antibodies were then eluted successively in five steps with buffers of decreasing pH from 7.0 to 2.3 containing 150 mM NaCl. Fractions were neutralized after elution. Specificity of eluted fractions was determined by Western blot analysis using dilutions of the elution fractions as primary antibodies. Antibodies against both 48-kDa PFK and 60-kDa PFK that eluted at pH 5.5 and pH 4.3 had the greatest specificity and were pooled and used in all subsequent analyses.
Northern Blot and Quantitation of the mRNA Level of the Two PFK Genes-E. histolytica total RNA was isolated from an amoebae cell sediment containing 1-2 ϫ 10 8 cells with the Qiagen DNA/RNA isolation kit. Denatured RNA isolated from trophozoites and RNA markers was then separated on 1.2% agarose gel and transferred to a nylon membrane. The membrane was then air dried and exposed to UV light to cross-link the RNA to the membrane. After prehybridization, membranes were hybridized with 32 P-labeled cDNA probes (1 ϫ 10 6 cpm/ml) prepared by restriction enzyme digestion of the plasmids containing the PFK genes. For quantitation of the mRNA level of the two PFK genes, slot blot analysis was performed. A standard was constructed by a series of 2-fold dilutions of the DNA for each of the PFK genes, beginning from 1 ng to 1/64 ng. The DNA standard and 30 -50 g of E. histolytica total RNA were slot blotted onto nylon membranes. The membranes were air dried and UV cross-linked. Northern blot was performed as described above. The content of PFK mRNA within the total RNA was determined by comparing the intensity of the signal from the total RNA with the DNA standard. The ratio of the mRNA level of the two PFK genes was determined.
Western Blot Analysis-Optimal dilution of the affinity-purified an-tisera for Western blot was determined by dot blot. E. coli cell extract and protein molecular weight markers were used as negative controls. When quantitation was required, a series of dilutions of known amounts of the two PFKs ranging from 1 to 100 ng was run in adjacent lanes of the gel electrophoresis. Negative control and protein samples were separated by 10% SDS-PAGE, then transferred onto nitrocellulose membranes in 25 mM Tris, 200 mM glycine, 20% methanol at 24 V. Washing and detection were performed by following the instructions of Amersham Pharmacia Biotech ECL Western blotting protocols using goat anti-rabbit immunoglobin conjugated with horseradish peroxidase. E. histolytica Cultivation-E. histolytica trophozoites (strain HM-1: IMSS) were grown axenically in TYI-S-33 medium (7) at 35°C. Routine cultures were maintained in 15-ml borosilicate glass tubes and transferred every 3 or 4 days. To obtain sufficient cells for 48-kDa PFK purification, trophozoites were cultured in 600-ml Nunclon triple flasks (Fisher Scientific).
48-kDa PFK Purification from E. histolytica-Amoebae from 3-dayold cultures (4 ϫ 10 8 ) were detached from the surface of flasks by chilling at 4°C for 30 min. Cells were then harvested by centrifugation at 500 ϫ g for 5 min, washed twice with phosphate-buffered saline at pH 7.2, and resuspended in 3 ml of the extraction buffer described for ATP affinity purification of the recombinant enzyme. The cells were then lysed by sonication. After centrifugation, the lysate was loaded onto 3 ml of an ATP-Sepharose column preequilibrated with extraction buffer. The column was washed with 20 volumes of extraction buffer, and the enzyme was subsequently eluted using the same buffer plus 1 mM ATP. Each elution fraction was analyzed by Western blot for both PFKs.
Molecular Sizing-Molecular mass determinations were carried out on a fast performance liquid chromatography system fitted with a Superdex 200 HR 10/30 column (Amersham Pharmacia Biotech). A standard curve was constructed by using a mixture containing 200 g each of cytochrome C (12.4 kDa), carbonic anhydrase (29 kDa), glycerol-3-phosphate dehydrogenase (70 kDa), and alcohol dehydrogenase (150 kDa) in a medium of 20 mM Tris-HCl, 1 mM EDTA, and 14 mM ␤-mercaptoethanol, pH 7.2. The standard mixture, E. coli PFK (142 kDa), and E. histolytica PFK samples were chromatographed individually using a Superdex 200 column preequilibrated with the buffered medium plus or minus additions as indicated.
Other Methods-Gel electrophoresis of proteins was carried out using a 10% polyacrylamide support according to the system of Laemmli (8). Protein concentrations were determined by Bradford's dye binding assay with bovine serum albumin as the standard (9). All chemicals and enzymes were purchased from Sigma.

RESULTS
Purification of the 48-kDa PFK-In an attempt to repeat the findings of Bruchhaus et al. (1), who were able to detect very low PP i -PFK activity with a recombinant 48-kDa PFK protein bearing a histidine tag, the 48-kDa PFK was prepared as described in their report. A homogeneous protein with a mass of the predicted 50 kDa as indicated by SDS-PAGE was purified successfully (not shown); however no PP i -PFK activity under the conditions described previously could be detected at any point during the purification. Because the relatively high concentrations of imidazole used to elute the enzyme from the nickel column (400 mM) may have denatured the enzyme, the CD spectrum of the preparation was compared with that of the homogeneous 60-kDa PP i -PFK. The spectra were nearly identical, suggesting that the global structure of the 48-kDa protein was maintained. All attempts at dialyzing the eluted protein into a lower salt buffer resulted in an irreversible precipitation. Several other methods of elution from the nickel column were attempted, including various concentrations of imidazole and gradients of imidazole and EDTA. However, all of these methods also failed to produce an enzyme with detectable PP i -PFK activity. The yield of the 48-kDa PFK fusion protein using this method, however, was sufficient to raise polyclonal antibodies that were used as a means of detection of the native protein expressed without the histidine tag during its subsequent purification.
Conventional PFK purification procedures were attempted to isolate the 48-kDa PFK without the N-terminal histidine tag using the antibody to follow the 48-kDa protein at each step of the procedure. The recombinant enzyme did not bind to phospho-cellulose, which is commonly used for the purification of PP i -PFKs (3,10), under a variety of conditions. No PP i -PFK activity was detected at any point in the purification process or in cell extracts using the assay conditions described by Bruchhaus et al. (1). The inability to duplicate previously reported activity measurements and these results suggested that the 48-kDa PFK gene does not utilize PP i to phosphorylate Fru-6-P and thus prompted the trial of alternative purification methods.
Because this laboratory commonly uses both N-6-aminohexylcarboxymethyl-ATP-Sepharose and Blue Sepharose for the purification of various ATP-dependent PFKs (11)(12)(13), these media were tried with the recombinant 48-kDa PFK. It was found that the protein bound to both ATP-Sepharose and Blue Sepharose. The 48-kDa PFK was eluted from both types of medium by employing 1 mM ATP in the eluting buffer. Subsequent Mono Q anion exchange chromatography of the eluate from either procedure yielded homogeneous enzyme with a size by SDS-PAGE equivalent to the calculated 47.6-kDa mass (not shown). The recombinant enzyme purified by ATP-Sepharose chromatography was identified by the crude antibodies that were raised against the purified, histidine-tagged recombinant 48-kDa PFK. For the isolation of specific 48-kDa PFK antibodies, the ATP-Sepharose-purified enzyme was linked to CNBr-activated Sepharose as described under "Experimental Procedures." Catalytic Properties of the 48-kDa PFK Activation-The affinity chromatography isolation procedure indicated that the 48-kDa PFK interacts with ATP. This observation suggested a reexamination of the activity in the presence of ATP. In such experiments it was observed that when assays with relatively high concentrations of enzyme were allowed to proceed for 30 or more min, a very gradual increase in ATP-dependent activity was observed, suggesting activation in the assay cuvette. This led to preincubation assays of the enzyme with various components of the assay mixture. The testing of the assay components led to the significant finding that ATP-dependent PFK activity can only be detected when relatively high concentrations of enzyme and ATP are preincubated together before adding the enzyme to the assay mixture (details discussed below). Addition of the same amount of enzyme to the assay mixture without prior incubation with ATP resulted in no activity even when ATP concentrations in the assay mixture were high. In such cases no activity is detected because the enzyme concentration in the reaction mixture is too low to become activated. Consistent with this hypothesis, when the enzyme is preincubated with ATP at too low an enzyme concentration, no activity results when adding an equivalent amount of enzyme as above to the assay mixture.
The dependence of the activation process on the concentrations of enzyme and ATP is shown in Fig. 1. The enzyme and ATP concentrations in the preincubation mixtures that result in half-maximal activity are 0.72 M and 0.21 mM, respectively. The time course of activation was measured at saturating concentrations of both ATP and enzyme. Maximal activity is attained after 5 min of preincubation as shown in Fig. 2A. To determine whether the temperature of the preincubation had any effect on the resultant rate of the enzyme, the preincubation mixtures were incubated at various temperatures before the resultant activity was measured. The temperature optimum for the preincubation is 30°C (Fig. 2B). Based on these results, preincubations for all standard kinetic assays were subsequently conducted using 4 M enzyme and 2 mM ATP in 150 mM KTes (pH 7.2), 3 mM MgCl 2 , 1 mM EDTA at 30°C and lasted for at least 30 min.
The enzyme concentration dependence of the activation process was investigated further using polyethylene glycols. PEGs have been shown to have an associative effect on macromolec-ular solutes in aqueous solution without specifically interacting with them (14). Aggregating systems have been shown to be shifted to higher degrees of association by increasing PEG concentration (15). Inclusion of PEG in preincubation mixtures allowed the 48-kDa PFK to be activated at preincubation enzyme concentrations that were too low to become activated in the absence of PEG (Fig. 3). The activation process was enhanced by increasing concentration and size of PEG in the preincubations, with the enhancement effect peaking at 20% PEG. PEG apparently encourages native self-association of the 48-kDa PFK into the activated state by increasing the local protein concentration in solution.
The 48-kDa PFK does not require the MgATP complex for activation because it is activated maximally without Mg 2ϩ in the preincubation buffer. Maximal activation of the 48-kDa PFK can also achieved when it is incubated with other nucleotide triphos-phates (Table I). GTP and ITP as well as the pyrimidines UTP and CTP are all equally as effective as ATP at activating the enzyme for measuring ATP activity in the resultant assay mixture. The nonhydrolyzable ATP analog AMP-PNP and ADP also can activate the enzyme, both being at least 60% as effective as ATP. Incubation with AMP, the cosubstrate Fru-6-P, and the product orthophosphate results in no activation at all. Interestingly, PP i , despite lacking the nucleotide moiety entirely, is quite capable of activating the 48-kDa PFK to achieve ATP-dependent activity, being 75% as effective as ATP in a 30-min preincubation.
Inactivation-Once activated, the enzyme spontaneously inactivates by simple dilution. This inactivation can be seen during the PFK assay, where one observes a decrease in the rate about 100 s after the start of the reaction which is the result of the dilution of activated enzyme from the concentrated preincubation mixture into the assay. The inactivation proceeds as a first order reaction. To characterize the dilution effect, the enzyme was activated by preincubation at the optimal conditions and subsequently diluted in 150 mM KTes (pH 7.2), 3 mM MgCl 2 , 1 mM EDTA, with or without 2 mM ATP. After activation, the activity was measured under standard assay conditions. The data were fitted using the Michaelis-Menten model to estimate the preincubation enzyme concentration that achieves half the maximal rate. A, enzyme concentration dependence of activation. The enzyme was incubated at concentrations from 0.1 to 28.5 M in KTes (pH 7.2) assay buffer containing 10 mM ATP in 20-l volumes for 30 min at 30°C. Fixed amounts of enzyme from the preincubations were then diluted 10-fold in standard dilution buffer, and identical volumes were taken from each dilution and added to assay cuvettes. B, ATP dependence of activation. The enzyme was incubated in KTes (pH 7.2) assay buffer at fixed concentrations of 4 M in separate tubes containing increasing ATP concentrations from 0.1 to 4 mM. Incubations were carried out in 20-l volumes at 30°C. Identical amounts of enzyme were taken after 30 min of incubation from each tube and were assayed under standard assay conditions. Increasing the incubation time an extra 90 min did not increase the activation of the enzyme.

FIG. 2. Time and temperature dependence of PFK activation.
A, time dependence. The enzyme was first prepared at 4 M in assay buffer at 30°C. ATP was added to a final concentration of 2 mM, and aliquots of the activated enzyme were subsequently added to assay cuvettes using the standard dilution method at time points from 1 s to 2 h after the addition of ATP. The reactions were then measured under standard assay conditions. B, temperature dependence. Fixed concentrations of enzyme at 4 M were activated at various temperatures in assay buffer containing 2 mM ATP in 100-l volumes for 30 min.
Aliquots were then taken from each dilution mixture at increasing time points and added to assay mixtures to measure the activity (Fig. 4). The first order inactivation rate constant without additions was 0.09 min Ϫ1 , and it decreased by nearly half to 0.05 min Ϫ1 when the enzyme was diluted in buffer containing 2 mM ATP and all other preincubation and dilution conditions were identical. Diluting in buffer containing both substrates at concentrations that produce maximal PFK activity (2 mM ATP and 20 mM Fru-6-P) substantially decreases the rate of inactivation (not shown). This experiment is complicated by the fact that the reaction is proceeding under these conditions. The rate of inactivation measured at early time before significant reaction has taken place gave a rate constant of 0.016 min Ϫ1 . As a result of these experiments, all kinetic assays were performed by diluting the activated enzyme into assay buffer containing 2 mM ATP and 20 mM Fru-6-P when dilution was necessary.
Aggregation State-The above experiments on activation and inactivation suggested that the state of polymerization of the molecule was the determinant of activity. To determine the aggregation state of the native 48-kDa PFK as well as the activated enzyme, a size exclusion chromatography experiment was performed. The polymerization state was determined for the native enzyme, the enzyme activated with ATP, and the enzyme incubated with the cosubstrate Fru-6-P alone, which does not activate the enzyme. For the enzyme-ATP experiment, the concentrations of ATP and enzyme which were found to activate the enzyme maximally were used in the preincubation mixtures before chromatography. For the Fru-6-P experiment, the concentration of Fru-6-P in the preincubation mixture was 20 mM, which is the concentration determined to achieve maximal ATP-PFK activity in the kinetic assay (see below). The results (not shown) of the molecular sizing experiments indicate that both the unactivated enzyme and the enzyme incubated with Fru-6-P alone eluted as a single peak at a position similar to that of glycerol-P dehydrogenase (70 kDa), indicating that the 48-kDa PFK exists as a dimer under these conditions. The enzyme preincubated with ATP eluted as a single peak at a position near that of the E. coli ATP-PFK (140 kDa), suggesting that it exists as a tetramer after activation. Kinetic Properties-The 48-kDa PFK was found to be a highly active ATP-utilizing enzyme with a k cat value of 250 s Ϫ1 , which is almost three times the maximum activity of the ATPdependent activity of E. coli ATP-PFK (16) and about threefourths the maximum activity of the PP i -PFK activity of E. histolytica 60-kDa enzyme. No PFK activity (0.01% level of detectability) was observed when PP i (at 2 mM) was used as a phosphoryl donor in the assay. Also, ATP activity was not inhibited by this concentration of PP i . To determine if the 48-kDa PFK could phosphorylate other sugars using ATP as a phosphoryl donor, the production of ADP was measured when the enzyme was incubated with other sugar compounds. Fru-1-phosphate, glucose, glucose 1-phosphate, glucose 6-phosphate, mannose, and ribose 5-phosphate could not substitute for Fru-6-P in the kinase assay.
Similar to many other ATP-PFKs, the 48-kDa PFK shows cooperative kinetics with respect to Fru-6-P, with a Hill con-   4. Inactivation by dilution. The enzyme was first activated under standard activation conditions. Aliquots of enzyme were then each diluted 10-fold in assay buffer (150 mM KTes (pH 7.2), 3 mM MgCl 2 , 1 mM EDTA) or assay buffer with 2 mM ATP. At time points from 0 to 120 min, identical amounts of enzyme were assayed under standard conditions. The rate constants (k) were calculated as the negative slope of the first order plot of the natural log of the rate against time. Only time points within the first 20 min of the ATP-buffer diluted mixture were used because the reaction reaches an equilibrium after that time. The reversible first order reaction with ATP was fitted using the first order exponential decay equation stant (n H ) of 2.3 and a relatively high Fru-6-P 0.5 value of 3.8 mM (Fig. 5A). With regard to ATP, the kinetic estimates of the E. histolytica 48-kDa PFK compare favorably with the ATP-PFK from E. coli (16). The 48-kDa PFK has an apparent affinity for ATP (K m ϭ 0.12 mM) (Fig. 5B) which is higher than that for the E. coli ATP-PFK (K m ϭ 0.21 mM). The k cat /K m value with ATP k cat /K m ϭ 2,200) of the 48-kDa PFK is significantly higher than that for the E. coli enzyme (k cat /K m ϭ 390) (16). The presence of PEG in the assay mixture was found to have a modest effect on the apparent affinity for Fru-6-P. The apparent K m value for Fru-6-P was ϳ33% lower with the inclusion of PEG, with the effect consistent under different concentrations and sizes of PEG. There was little effect on cooperativity and on the maximal velocity of the Fru-6-P saturation profiles (data not shown).
Although clearly an ATP-utilizing enzyme, the 48-kDa PFK exhibited many characteristics not typical of ATP-PFKs. The pH optimum was relatively acidic, which is more characteristic of PP i -PFKs. The highest activity in the presence of subsaturating Fru-6-P (2.5 mM) was observed between pH 6 and 7 with only 30% activity at 7.5 and 15% activity at 8.5. Activity measurements below pH 6.0 were compromised by the limited activity of one or more of the auxiliary enzymes used in the coupled assay. In contrast, E. coli and all known mammalian PFKs have alkaline pH optima. Also unusual was the proficiency of the enzyme in using other nucleotides as substrates relative to ATP (Table II). The apparent affinity and the activity at low concentrations of substrate were even higher for GTP than for ATP. The 48-kDa PFK still showed cooperativity with Fru-6-P with each of the nucleotides as cosubstrates, and the apparent affinity for Fru-6-P remained relatively high with each of the nucleotides tested. In comparison, the E. coli ATP-PFK has k cat /K m values for GTP and ITP which are an order of magnitude lower than the value for ATP (16).
Although the 48-kDa PFK did not require Mg 2ϩ for activation, it was required for catalytic activity, similar to other PFKs. Substituting Mn 2ϩ in the assay resulted in only 16% of the observed activity with Mg 2ϩ , whereas no activity was detectable when substituting with Ca 2ϩ and Zn 2ϩ .
The 48-kDa PFK appears to be inhibited by ATP at high ATP concentrations. This inhibition was evident at low concentrations of the cosubstrate Fru-6-P and disappeared at saturating Fru-6-P (Fig. 5B). ATP inhibition has been demonstrated in other ATP-PFKs. Mammalian PFK has a separate ATP inhibitory site (17), whereas E. coli PFK displays mechanism-based, nonallosteric inhibition by ATP (18). The mechanism of ATP inhibition in the 48-kDa PFK remains to be elucidated. Cooperativity in the interaction with Fru-6-P increased at a higher ATP concentration (Fig. 5A). The mechanism of the cooperative interaction appears to be allosteric, but the mechanism needs to be resolved. It may be related to association/dissociation behavior, but the failure of PEG to eliminate cooperativity argues against this interpretation. PEP, which is known to inhibit other PFKs (12,19), is an inhibitor of the 48-kDa PFK (Fig. 6). PEP decreased the apparent Fru-6-P affinity substantially, although it had a limited effect on cooperativity (n) and no effect on ATP binding (Table III). The steady-state PEP concentration in E. histolytica is not known.
We investigated many other compounds for their ability to regulate the activity of the 48-kDa PFK and have tentatively found no other effectors. Activity with each potential effector was measured both at half-saturating (2.5 mM) and saturating concentrations (20 mM) of Fru-6-P. The apparent Fru-6-P affinity of the 48-kDa PFK was not affected by the metabolites AMP, ADP, GDP, cAMP, orthophosphate, sodium ion, ammonium ion, phosphocreatine, citrate, fructose 2,6-bisphosphate, 3-phosphoglycerate, and glucose 6-phosphate (all at 1 mM concentrations), metabolites that have been demonstrated to modulate PFK in other organisms. Other compounds that were examined for their ability to regulate the 48-kDa PFK included phosphoglycolate, lactate, calcium ion, and calmodulin. No effects were seen.
mRNA Levels for 60-kDa and 48-kDa PFKs in E. histolytica Trophozoites-Total RNA isolated from trophozoites was used for Northern blots to determine the expression of the two PFKs  at the transcriptional level. Using a cDNA probe of the 48-kDa PFK gene, blots of the E. histolytica total RNA showed a single band of about 1.3 kilobases. The blot with the 60-kDa gene probe showed a single band of ϳ1.6 kilobases. Both were the expected size as determined from length of the genes and the short untranslated region sequences. A slot blot was used to compare the expression of the two E. histolytica PFK genes at the mRNA level. By comparing the intensity of the signal from the total RNA with that of the signal from the standard DNA series as described under "Experimental Procedures," the amounts of the mRNA of the two PFK genes within total RNA were compared (not shown). In 30 g of E. histolytica total RNA, there were 250 pg of 60-kDa PFK mRNA and only 16 pg of 48-kDa PFK mRNA. Considering the size of the two PFK genes, the mRNA level of the 60-kDa PFK gene is about 10 times higher than that of the 48-kDa PFK gene. PFK Content in E. histolytica Trophozoites-The relative quantities of the two PFK enzymes in amoebal extracts were compared by Western analysis using known amounts of recombinant 48-kDa and 60-kDa PFKs as standards as described under "Experimental Procedures" (not shown). In trophozoites the 48-kDa PFK enzyme was present at about one-tenth the level of the 60-kDa PFK. These data are consonant with the data on the mRNA levels.
The native 48-kDa PFK enzyme was readily isolated from trophozoite extracts using chromatography on ATP-Sepharose as described under "Experimental Procedures." The ATP-Sepharose isolation procedure indicated no apparent association between the two PFKs in trophozoites. The 60-kDa PFK was never detected by Western blot at any point during chromatography except in the initial effluent containing proteins that do not bind to ATP-Sepharose. The search for possible interaction between the two PFKs was motivated by the observation of a multisubunit structure in plant PP i -PFKs (4). In the instance of the plant enzymes, catalytic and regulatory subunits copurify. No copurification was observed, nor was coprecipitation of the two enzymes from trophozoite extracts seen when either specific antibody was used. Furthermore, assays of purified 60-kDa PP i -PFK were not influenced by the presence of an equal amount of purified 48-kDa ATP-PFK, nor was there any effect when the two enzymes were preincubated together. Similarly, no effect of 48-kDa PFK was seen when the reverse experiments were performed. Thus any direct interactions between the two proteins are very unlikely.
A reinvestigation of trophozoite extracts showed that ATP-PFK activity could be detected without prior activation. ATPdependent PFK activity in amoebae is about 11 fold lower than PP i -dependent activity (0.43 unit of ATP activity versus 4.1 units of PP i activity in 100 l of trophozoite extract), corresponding to the relative amounts of the two PFKs enzymes detected by Western analysis. To ensure that the measured ATP-PFK activity was not an artifact of the 60-kDa PFK catalyzing PP i produced in other metabolic pathways, amoebal extracts were dialyzed exhaustively to eliminate all small metabolites. Also, ATP-PFK activity was readily measured at high Fru-6-P concentrations (20 mM) and was totally undetectable at 1.5 mM Fru-6-P, which is a saturating concentration of the sugar phosphate for the 60-kDa PFK. This indicated that the ATP-PFK activity detectable only at the higher Fru-6-P concentration was not measuring the 60-kDa PFK catalyzing contaminating PP i because such contamination would have been detectable at 1.5 mM Fru-6-P. In the study that first identified the PP i -PFK enzyme of E. histolytica, Reeves et al. (20) also detected ATP-PFK activity in trophozoite homogenates. Those investigators were unable to characterize the E. histolytica ATP-PFK activity further because of activity losses during purification. In fact, Reeves later concluded that the observed ATP-PFK activity was an artifact (21). That is clearly not the case as demonstrated here.
An interesting finding was that the amoebal ATP-PFK activity was not increased by preincubation of amoebal extracts with 2 mM ATP even after eliminating the small metabolites by dialysis. In contrast, the trace of ATP-PFK activity in bacterial extracts containing recombinantly expressed E. histolytica 48-kDa PFK was increased dramatically after incubation with 2 mM ATP (data not shown).

DISCUSSION
The two PFKs of E. histolytica display distinct phosphoryl donor specificities. The 60-kDa PFK is a PP i -dependent enzyme and is responsible for all detectable PP i -PFK activity in trophozoite extracts (3). The 48-kDa PFK, contrary to previous reports, demonstrates no detectable PP i -PFK activity when produced recombinantly. The 48-kDa PFK is in fact a highly active ATP-utilizing PFK that is also able to use other nucleotides efficiently for catalysis. However, the apparent K m value for Fru-6-P of the 48-kDa PFK is more than 20-fold greater than previously measured intracellular Fru-6-P concentrations (0.16 Ϯ 0.06 mM) in amoebae (20), indicating that without a positive effector this enzyme may have limited physiological activity unless one invokes compartmentalization. Considering that the 60-kDa PFK has been shown to account for the glycolytic flux in amoebal extracts (21) and that mRNA, protein, and activity levels all indicate that the 60-kDa PFK is present in trophozoites in 10-fold greater amounts than the 48-kDa enzyme, the significance of the ATP-PFK in the glycolysis of trophozoites remains in question. However, E. histolytica has a complex life cycle, and the ATP-PFK may have functions in other stages of that cycle.  The findings in this study as well as results from earlier studies from this laboratory (3) argue against the possibility of the two PFKs of E. histolytica associating or affecting each other in some regulatory manner. The two PFKs do not copurify when isolating either protein, a protein 47 kDa in size was not seen in the active PFK fractions during native PP i -PFK purification, immunoprecipitation of the trophozoite cell extract with antibodies against the 60-kDa PFK did not precipitate a protein close to 47 kDa in mass, and the activity of either the 60-kDa PFK or 48-kDa PFK was unaffected by the presence of its PFK counterpart in the assay mixture or in preincubations (data not shown).
E. histolytica 48-kDa PFK is an unusual ATP-utilizing PFK that is only active after incubation at high enzyme concentrations with ATP. This activation appears to be the result of a change in the state of aggregation of the enzyme upon binding ATP rather than a catalytic event such as ATP hydrolysis or the formation of a phosphoenzyme complex. The activation can be reversed by simply diluting the enzyme-ATP incubation, and the enzyme-ATP preincubation mixture eventually reaches an equilibrium. These results indicate that activation does not involve a permanent alteration in either the enzyme or the ATP molecule. The enzyme can also be activated by other nucleotide triphosphates, AMP-PNP, ADP, and even PP i , indicating that a specific ATP modification is not involved in activation. These observations suggest that the nucleoside moiety is not essential for activation and that the last two phosphoryl groups of ATP are the most critical features. Closer analysis reveals PP i to be a better activator than ADP and AMP-PNP, both of which deviate from ATP in the terminal polyphosphate region. This polyphosphate moiety is completely absent in AMP and orthophosphate, and incubation with these compounds does not activate the enzyme at all. The Michaelis constant value for ATP derived from the ATP-dependent activation assay (K m ϭ 0.21 mM) is similar to that observed from the substrate dependent assay (K m ϭ 0.12 mM), which is consistent with ATP binding at the same site for both activation and catalysis. However, our results show that the adenosine moiety seems to be of little relevance in activation, whereas PP i activates the enzyme to nearly maximum levels. Although PP i can activate the enzyme, it is not a substrate, nor can it inhibit ATP activity. These results introduce the possibility that the 48-kDa PFK may bind PP i for activation at a site other than the substrate binding site. It is also possible that the activators may produce their effects by chelating some unknown inhibitor; however, this situation is unlikely because all activation assays were performed in the presence of 1 mM EDTA. Although the enzyme cannot be activated by the cosubstrate Fru-6-P alone, the sugar phosphate does provide some protection against inactivation when present with ATP.
The dependence of the activation process on protein concentration suggests that the reversible activity loss is associated with association-dissociation behavior of the protein. This was supported by experiments with molecular crowding with PEG which showed that crowding increased the rate of activation. Finally, molecular sizing experiments show that the inactive PFK exists as a dimer that associates into an active tetramer upon incubation with ATP.
The 48-kDa PFK in dialyzed amoebal extracts was found in an activated state. This information introduces many new possibilities for the native activation state of the enzyme. Although it is possible that ATP was not eliminated from the trophozoite extracts by dialysis, it is more likely that another activator exists in the amoeba which is either too large or too tightly associated to be removed by dialysis. Also, the 48-kDa PFK may be activated in trophozoites by some other means such as subcellular localization. Whatever the mechanism of native activation, it is likely that in vivo the 48-kDa PFK displays substantially different kinetic features.
The two PFKs of E. histolytica have a low sequence identity of about 17%, although there are many identical residues in the presumed active site. Phylogenetic studies of the sequences of PFKs place the two E. histolytica PFKs in a large group of proteins, most of which have been described as PP i -PFKs that are distinct from the typical ATP-PFKs such as those found in E. coli as well as all mesozoans. The 60-kDa enzyme falls into a monophyletic subgroup that contains a number of other well characterized PP i -PFKs including those of plants (22,23). On the other hand, the 48-kDa PFK sequence from E. histolytica falls into a monophyletic subgroup within the PP i -PFK group that also contains Treponema pallidum and Borrelia burgdorferi and the peroxisomal ATP-PFK of Trypanosoma brucei (22,23). Of the other three members of the group, the T. pallidum gene product has not been characterized and preliminary studies of the B. burgdorferi product have not found either ATP-or PP i -PFK activity (24). The T. brucei ATP-PFK is a homotetramer with a subunit mass of 50 kDa and is not regulated by the metabolites that modulate the activity of ATP-PFKs in other organisms (25). The members of this group of four proteins have a common sequence in the presumed region where the phosphoryl transfer reaction takes place. The two sequences are GGDG and PKTIDND, which may be contrasted to GGDD and PKTIDND of almost all well characterized PP i -PFKs and GGDG and PGTIDND in ATP-PFKs of E. coli and all mesozoans. Recently we have shown that mutation of the second Asp in the GGDD sequence of the E. histolytica 60-kDa PP i -PFK to Gly changes the specificity to that of an ATP-PFK (26). The last residue in the GGDG sequence would appear to be a particularly important determinant of the phosphoryl donor specificity of all PFKs.