Molecular basis of the unusual catalytic preference for GDP/GTP in Entamoeba histolytica 3-phosphoglycerate kinase ´ ´ ´ ´ Rusely Encalada1, Arturo Rojo-Domınguez2, Jose S Rodrıguez-Zavala1, Juan P Pardo3, Hector ´ Quezada1, Rafael Moreno-Sanchez1 and Emma Saavedra1 ´ ´ ´ ´ Departamento de Bioquımica, Instituto Nacional de Cardiologıa, Mexico D.F., Mexico ´ ´ Departamento de Ciencias Naturales Unidad Cuajimalpa and Departamento de Quımica Unidad Iztapalapa, Universidad Autonoma ´ ´ Metropolitana, Mexico D.F., Mexico ´ ´ ´ ´ ´ Departamento de Bioquımica, Facultad de Medicina, Universidad Nacional Autonoma de Mexico, Mexico D.F., Mexico Keywords ATP ⁄ GTP synthesis; glycolysis; nucleotide selectivity; parasite; yeast Correspondence ´ E Saavedra, Departamento de Bioquımica, ´ Instituto Nacional de Cardiologıa, Juan ´ Badiano No Col Seccion XVI, CP 14080, ´ Tlalpan, Mexico D.F., Mexico Fax: +52 55 5573 0994 Tel: +52 55 5573 2911; ext 1298 E-mail: emma_saavedra2002@yahoo.com (Received 14 November 2008, revised 23 January 2009, accepted 28 January 2009) doi:10.1111/j.1742-4658.2009.06939.x Phosphoglycerate kinase (EC 2.7.2.3) catalyzes reversible phosphoryl transfer from 1,3-bisphosphoglycerate to ADP to synthesize 3-phosphoglycerate and ATP during glycolysis Phosphoglycerate kinases from several sources can use GDP ⁄ GTP as alternative substrates to ADP ⁄ ATP; however, the maximal velocities (Vm) reached with the guanine nucleotides are  50% of those displayed with the adenine nucleotides By contrast, Entamoeba histolytica phosphoglycerate kinase (EC 2.7.2.10) is the only reported phosphoglycerate kinase displaying higher activity with GDP ⁄ GTP and lower affinities for the adenine nucleotides To elucidate the molecular basis of the Entamoeba histolytica phosphoglycerate kinase selectivity for GDP ⁄ GTP, a conformational analysis was carried out on a homology model based on crystallographic structures of yeast and pig phosphoglycerate kinases Some amino acid residues involved in the purine ring binding site not previously described were detected Accordingly, Y239, E309 and V311 were replaced by site-directed mutagenesis in the Entamoeba histolytica phosphoglycerate kinase gene for the corresponding amino acid residues present in the adenine nucleotide-dependent phosphoglycerate kinases and the recombinant proteins were purified Kinetic analysis of the enzymes showed that the single mutants Y239F, E309Q, E309M and V311L increased their catalytic efficiencies (Vm ⁄ Km) with ADP ⁄ ATP as a result of both, increased Vm and decreased Km values Furthermore, a higher catalytic efficiency in the double mutant Y239F ⁄ E309M was achieved, which was mainly due to an increased affinity for ADP ⁄ ATP with a concomitant diminished affinity for GDP ⁄ GTP The main Entamoeba histolytica phosphoglycerate kinase amino acid residues involved in the selectivity for guanine nucleotides were thus identified The human parasite Entamoeba histolytica, the causal agent of amebiasis, relies only on glycolysis for its ATP supply because it lacks the Krebs cycle and oxidative phosphorylation pathways [1,2] The glycolytic enzymes of the parasite are highly divergent from the enzymes present in the human host; they include an AMP-inhibited hexokinase [3,4], and the non-allosteric and pyrophosphate-dependent enzymes phosphofructokinase [4,5] and pyruvate phosphate dikinase [4,6,7], which replace the functions of the allosteric enzymes Abbreviations EhPGK, Entamoeba histolytica PGK; PGK, 3-phosphoglycerate kinase; ScPGK, Saccharomyces cerevisiae PGK FEBS Journal 276 (2009) 2037–2047 ª 2009 The Authors Journal compilation ª 2009 FEBS 2037 Entamoeba GDP ⁄ GTP-dependent phosphoglycerate kinase ATP–PFK-1 and pyruvate kinase in the host [8] The importance of glycolysis for parasite survival and the differences found in the glycolytic enzymes compared with those of the human host, make this pathway a suitable target for therapeutic intervention In this regard, another remarkable difference in the amebal glycolytic pathway is found in the first reaction of substrate-level phosphorylation catalyzed by 3-phosphoglycerate kinase (PGK; EC 2.7.2.3) This 43–45 kDa monomeric enzyme, highly conserved during evolution, transfers the acyl-phosphate group from 1,3-bisphosphoglycerate to Mg2+–ADP to produce 3phosphoglycerate and Mg2+–ATP, in a fully reversible reaction under physiological conditions In the majority of PGKs characterized to date, from several organisms, the enzymes show higher or similar affinities for other purine nucleotides such as GTP or ITP to those observed with ADP ⁄ ATP; however, the phosphorylation transfer rates displayed with GTP or ITP are on average 50% lower than with the adenine nucleotides [9–13] In marked contrast, an early study in a partially purified E histolytica PGK (EhPGK; EC 2.7.2.10) [14] demonstrated that this enzyme displayed poor catalysis with ADP ⁄ ATP as substrates; the cause of this behavior was the higher Km values for the adenine nucleotides (at least 10-fold) compared with the Km values exhibited with GDP ⁄ GTP These differences were recently confirmed by our research group with the recombinant purified enzyme, where Km values for ADP and ATP were 12 and 44 times higher than the Km values for GDP and GTP, respectively [4] To our knowledge, the higher selectivity towards guanine nucleotides has only been documented for EhPGK In order to advance our understanding of the molecular basis underlying the kinetic differences found in EhPGK, site-directed mutagenesis analysis was undertaken on the specific amino acid residues interacting with the nitrogen base in the nucleotide-binding pocket Such residues were identified by conformer search of their side chains in a predicted 3D model of the amebal enzyme Our results indicated that one single substitution was able to increase catalysis, R Encalada et al whereas two substitutions were necessary to increase affinity for ADP ⁄ ATP in EhPGK Results Nucleotide specificities of Saccharomyces cerevisiae and E histolytica PGKs in cellular extracts In order to accurately evaluate differences in the nucleotide specificities of the amebal and yeast enzymes in our kinetic assay conditions, and due to the lack of commercially available purified yeast PGK, Vm and Km values were determined under initial velocity conditions in cytosolic cellular extracts of both organisms (Table 1) For the amebal native PGK, the Km values were similar to those displayed by the wild-type recombinant purified enzyme [4] (see Table below) and confirmed the lower Km and higher Vm values reached with the guanine nucleotides, as described previously [4,14] The Km values obtained for the native yeast enzyme (Table 1) were similar to those reported previously at pH 6.9–7.0 (ADP, 0.2–0.4 mm; ATP, 0.11– 0.32 mm) and pH 7.5 (ADP, 0.2; ATP, 0.48; GTP, 0.17 mm) [9,10] Moreover, the yeast PGK in cellular extracts exhibited 2.5 times higher activity with ATP compared with GTP, which is in agreement with previously reported values [9,10] However, such a difference in Vm was not evident when the forward reaction was determined (Table 1); because of the lack of reported kinetic data in the forward reaction, a comparison was not possible These results established substantial differences in the purine nucleotide preferences between the yeast and amebal PGKs 3D predicted model of EhPGK Predicted 3D structures of the EhPGK were obtained by means of modeller software (see Fig S1 for details on these models) and molecular operating environment (moe; http://www.chemcomp.com) packages, using as templates the tertiary structures from yeast [15] and pig [16,17] PGKs, because of their higher levels of Table Nucleotide specificities of E histolytica and S cerevisiae PGKs in cytosolic-enriched cellular extracts Values represent the mean ± SD of titrations made with three independent cellular clarified extracts Amebal Yeast Vm (lmolỈmin)1Ỉmg cellular protein)1) Km (mM) GDP ADP GTP ATP 2038 0.07 7.4 0.016 5.0 ± ± ± ± 0.03 3.9 0.01 1.7 Km (mM) Vm (lmolỈmin)1Ỉmg cellular protein)1) 41 5.5 4.5 1.6 0.37 0.5 0.06 0.42 38 53 4.0 10 ± ± ± ± 18 0.8 0.6 0.06 ± ± ± ± 0.15 0.17 0.01 0.2 ± ± ± ± 15 0.3 FEBS Journal 276 (2009) 2037–2047 ª 2009 The Authors Journal compilation ª 2009 FEBS Entamoeba GDP ⁄ GTP-dependent phosphoglycerate kinase R Encalada et al Table Kinetic parameters for nucleotides of the wild-type and mutant EhPGKs Figures indicate mean ± SEM of titrations made with 3–5 independent batches of purified enzymes GDP ADP )1 Vmf (lmolỈmin Ỉmg protein)1) Wild-type Y239F V311L E309Q E309M Y239F ⁄ E309M 820 1000 969 1250 1530 989 ± ± ± ± ± ± 119 161 10 53b 269 158 )1 Km (lM) Vmf ⁄ Km (LỈmin Ỉmg protein)1) Vmf (lmolỈmin)1Ỉmg protein)1) Km (lM) 112 155 168 158 208 802 7.3 6.5 5.8 7.9 7.4 1.2 280 371 562 774 1411 362 ± ± ± ± ± ± 20 13 23 40 12b 106a GTP Wild-type Y239F V311L E309Q E309M Y239F ⁄ E309M 108 167 115 300 169 348 Vmf ⁄ Km (LỈmin)1Ỉmg protein)1) ± ± ± ± ± ± 14 55 18a 171b 350b 64 2016 2693 1696 1723 1606 280 ± ± ± ± ± ± 210 337 168 240 349 27a 0.14 0.14 0.33 0.45 0.88 1.29 ± ± ± ± ± ± 12a 7b 152 20a 65a 1439 2244 1104 1313 637 587 ± ± ± ± ± ± 215 428 62 46 152b 155b 0.06 0.07 0.11 0.24 0.29 0.65 ATP ± ± ± ± ± ± 12 29 13a 24 69b 151 440 189 291 153 342 ± ± ± ± ± ± 14 160 32 66 24 22a 0.72 0.38 0.61 1.0 1.1 1.0 84 156 120 315 182 384 One-tailed Student’s t-test for nonpaired samples: a P < 0.005; b P < 0.05 versus wild-type identity (59% and 55%, respectively) and similarity (73%) with EhPGK As expected, the resulting models were highly similar to the templates by using either moe (Fig 1A) or modeller (Fig S1) packages Nevertheless, side-chain replacement around the nucleotide-binding site required a finer modeling procedure, exploring the conformational space of sidechain rotamers in the presence of GDP to induce their fitting This procedure was programmed in moe in order to construct and minimize 1000 different combinations of rotamers in the replaced side chains, and evaluate the resulting diversity Although the backbone geometry of the EhPGK models proved to be essentially identical to those of template structures, the variability in side-chain orientations allowed us to propose some mutants which might respond to the changes in the donor ⁄ acceptor of hydrogen bonds in the purinic ring of adenine respect to guanine nucleotides It should be noted that these mutations cannot be detected by a simple replacement method because almost none of the 1000 models have all the non-conserved side chains, which interact with GDP, simultaneously oriented in an optimal position This suggests that the change in specificity from ATP ⁄ ADP to the guanine nucleotides must be acquired by a cooperative effect of several amino acid side chains Amino acid residues interacting with the guanine moiety of GDP in EhPGK The amino acid residues known to interact with the adenine moiety in the ADP ⁄ ATP-binding site in the crystal structures of PGKs from several sources have been identified previously [18] and are illustrated in Fig In yeast PGK, these residues correspond to Gly211, Ala212, Phe289, Leu311, Gly338 and Val339 (Fig 2A), which lie in a hydrophobic binding pocket in the C-terminal domain (Fig 2B) Based on the two predicted structures of the EhPGK obtained using the modeller program (Fig S1), it was found that the only difference in the amino acids that bind the purine ring was the presence of Val instead of Leu at position 311 (Fig 2A) Based on a blast analysis, the frequency of Leu at this position is high, because it was found in almost all PGK amino acid sequences reported in the Protein Data Bank for bacteria, fungi, plants and animals (data not shown) By using a more dynamic method for modeling EhPGK structure with the moe package in the presence of GDP, other putative amino acid side chains interacting with the guanine moiety were detected From this structural analysis it became evident that the amino group at carbon of the guanine ring may interact with the side chain of Glu309, whereas the carbonyl group at position of the guanine ring may interact with the hydroxyl group of the Tyr239 side chain (Figs 1B and 2B) In an extended blast analysis to that shown in Fig 2A, the more frequent amino acid residue at position 309 is Met, although Gln can be found in fungal PGKs, and Glu or Ser in some bacterial PGKs By contrast, a Phe residue at position 239 was present in almost all PGK sequences with some exceptions; Tyr was only found in the PGK structures from Bacillus stearothermophilus (1PHP) [19], Trypanosoma brucei (13PK) [20] FEBS Journal 276 (2009) 2037–2047 ª 2009 The Authors Journal compilation ª 2009 FEBS 2039 Entamoeba GDP ⁄ GTP-dependent phosphoglycerate kinase R Encalada et al ity within 4–7 months for the wild-type, Y239F, Y239F ⁄ E309M and E309Q enzymes Unexpectedly, the E309M mutant exhibited a half diminution in activity only after 12 months (data not shown) A Thermal stability Wild-type, V311L and the double mutant Y239F ⁄ E309Q enzymes incubated at 50 °C for 1–12 did not show drastic reductions in activity, whereas 30% of their initial activities decayed after incubation at 60 °C (data not shown) Thus, inactivation kinetics was carried out at 55 °C (Fig S3) The inactivation constants (kinac) obtained for wildtype, V311L and the double mutant were )0.36, )0.41 and )0.35 min)1, respectively, which indicated no significant thermostability differences B pH activity dependency The PGK activity dependence determined in the reverse reaction for the V311L and double-mutant enzymes showed similar behavior to that of the wildtype enzyme (Fig S4) Fig Predicted 3D structure of EhPGK (A) Overlapped structures of the pig PGK crystal structure with ATP (1KF0; green) and E histolytica PGK predicted model with GDP (red) obtained using MOE software (B) Close up of the purine ring binding site in the C-terminal domain Lines indicate hydrogen bonds interactions Lettering in red corresponds to EhPGK, and green refers to pig PGK and EhPGK Thus, in order to determine the role of Y239, E309 and V311 on the GDP ⁄ GTP preference of EhPGK, the single mutants Y239F, E309M, E309Q and V311L, and a double mutant Y239F ⁄ E309Q were generated by site-directed mutagenesis Biochemical properties of mutant and wild-type EhPGK The mutant proteins were overexpressed as N-terminal histidine-tailed recombinant proteins in Escherichia coli and purified to a high degree (> 98%; Fig S2) No significant differences were found between wildtype and mutant proteins regarding the following properties Storage stability The proteins were highly stable when stored in the presence of 50% glycerol at )20 °C, losing 50% activ2040 Oligomeric structure In a previous study [4], a dimeric oligomeric structure for the recombinant EhPGK was reported; however, in such gel-filtration chromatography experiments the protein fraction recurrently eluted as an entity of intermediate molecular mass between a monomer and a dimer By using a modified protocol, a monomeric, kinetically active protein was determined for wild-type recombinant EhPGK; this quaternary structure was not modified in the V311L and Y239F ⁄ E309M mutant proteins (Fig S5) This structural arrangement agrees with the active monomeric forms of the majority of PGKs from other sources (either native or recombinant forms) reported to date Whether this oligomeric state is preserved within the amebal cells remains to be explored Kinetic parameters in the forward reaction The Vm, Km and catalytic efficiency (Vm ⁄ Km) values for nucleotides were determined in the forward and reverse reactions for the wild-type and the five EhPGK mutants (Table 2) The Vm in the forward reaction (Vmf) and Km values with GDP were not greatly affected in the mutants Y239F, V311L, E309Q and E309M; thus, no significant change in the catalytic efficiency was observed However, the double mutant Y239F ⁄ E309M displayed a sevenfold reduction in its FEBS Journal 276 (2009) 2037–2047 ª 2009 The Authors Journal compilation ª 2009 FEBS R Encalada et al Entamoeba GDP ⁄ GTP-dependent phosphoglycerate kinase A B Fig Amino acid residues surrounding the purine nucleotide pocket in PGKs (A) Partial alignment of the C-terminal domain of several PGKs The EhPGK amino acid sequence alignment was performed with the CLUSTALW2 tool and the indicated primary sequences for the other sources Numbering on right and left correspond to that used in the text for each species Marks indicate amino acid residues known to interact with the nitrogen base groups: circles, conserved residues; down triangles, amino acids changed in EhPGK Residues in italics indicate the three hydrophobic patches (A, B, C) surrounding the purine base (B) Schematic diagram showing the amino acid residues located in the purine base pocket Interactions between functional groups in the adenine base in pig PGK (upper) and guanine base in modeled EhPGK (down) were calculated and represented with MOE Shadows represent solvent accessibility of ligand atoms and protein residues Hydrogen bonds formed between protein and ligand are represented by arrows pointing in the donor to acceptor direction The hydrogen bond established by N6 of adenine is either with Gly312 (pig), Leu313 (horse) or Leu311 (yeast) main chain carbonyl group For yeast PGK numbering, two amino acid residues should be subtracted in the upper figure FEBS Journal 276 (2009) 2037–2047 ª 2009 The Authors Journal compilation ª 2009 FEBS 2041 Entamoeba GDP ⁄ GTP-dependent phosphoglycerate kinase affinity for GDP, which was reflected in a substantial decrease in the catalytic efficiency with this substrate Regarding the values obtained with ADP, the single mutants V311L, E309Q and E309M exhibited a tendency to increase both catalytic capacity (Vmf) and affinity (decreased Km), despite the experimental variability found by using different protein purifications The combined changes in both parameters produced a concomitant 2.4- to 6.3-fold increment in their catalytic efficiencies with this nucleotide Remarkably, the Y239F ⁄ E309M mutant exhibited a significant increase in its affinity for ADP (7.2-fold), which was not attained in its respective single mutants, and which was accompanied by a decrease in its GDP affinity Hence, the double mutant yielded a very strong nonadditive increment of its catalytic efficiency of 9.2 times for the phosphorylation of the adenine nucleotide These observations with the double mutant indicated the existence of a cooperative effect for GDP (and ADP) binding, as predicted by molecular modeling, because the corresponding single mutants did not induce significant changes Kinetic parameters in the reverse reaction In general, the single mutants exhibited no significant differences in the Vm value in the reverse reaction (Vmr) and Km values with GTP compared with the wild-type enzyme, except for a slight increase in the Vmr of the E309Q mutant However, the double mutant displayed a slight increment in Vmr and a significant decrease in the affinity for GTP The kinetic analysis with ATP showed that the single mutants Y239F and E309M and their double mutant Y239F ⁄ E309M significantly increased (1.9–4.6 times) their Vmr values With the exception of the Y239F mutant, which decreased its affinity for ATP compared with the wild-type, the other four mutants displayed a tendency to increase their affinity for ATP, producing a 1.8–10.8 times increment in their catalytic efficiencies Therefore, the increased Vmr and decreased Km values for ATP were in agreement with the results obtained for ADP in the forward reaction It is worth noting that although changes in catalytic efficiency were not significant for GTP in the double mutant, a strong cooperative effect on this parameter was indeed attained for ATP The Km values for the co-substrate 3-phosphoglycerate were not significantly affected in the mutants compared with the wild-type enzyme, with Km values (in lm) of 547 (wild-type), 322 (Y239F), 449 (E309Q), 570 (E309M) and 343 (Y239F ⁄ E309M); these values were obtained by using one protein purification for each substrate titration 2042 R Encalada et al Nucleotide dissociation constants Because the Km value is by definition a kinetic parameter resulting from the Vm ⁄ [Vm ⁄ Km] ratio [21], then the specific nucleotide-binding constant (Kd) was determined in the absence of 3-phosphoglycerate for each nucleotide in the wild-type and the double-mutant enzymes (because the latter displayed the greatest changes in nucleotide affinities, as described above) The true nucleotide substrate of PGKs is the Mg–ATP complex [10,11] Therefore, the Kd values were determined in the presence of saturating concentrations of MgCl2 (Table 3) From these experiments it became evident that wild-type EhPGK indeed displayed higher affinity for the couple GDP ⁄ GTP compared to the couple ADP ⁄ ATP Unexpectedly, this preference was also maintained in the Y239F ⁄ E309M mutant Discussion The primary intermediary metabolism of the amitochondriate parasite E histolytica differs in several aspects from that of its human host, as made evident from the early biochemical studies carried out mainly by Richard E Reeves’ laboratory [1,2] and recently by the genome sequence data analysis [22] The lack of the Krebs cycle and oxidative phosphorylation activities poses glycolysis as the main pathway to generate ATP for cellular work and as a potential target for therapeutic intervention [23] Moreover, the absence of the typical mammalian flux-controlling glycolytic enzymes ATP-PFK and pyruvate kinase [24] and the presence of an AMP-inhibited instead of a glucose 6-phosphate-inhibited hexokinase, emphasizes the important deviations in the control of the glycolytic flux in the parasite compared with its host, as recently demonstrated by metabolic control analysis through kinetic modeling of the entire pathway [8] and pathway reconstitution experiments [25] Table Dissociation constant values for the Mg2+-nucleotide complexes in the wild-type and Y239F ⁄ E309M mutant EhPGKs The Mg2+ concentration was10 mM in the assay Values represent the mean of two experiments performed with two purified protein preparations Kd (lM) Wild-type GDP ADP GTP ATP Y239F ⁄ E309M 11.5 0.38 10.3 0.51 4.5 0.32 7.7 FEBS Journal 276 (2009) 2037–2047 ª 2009 The Authors Journal compilation ª 2009 FEBS R Encalada et al The presence of a guanine nucleotide-dependent PGK in the first reaction of substrate-level phosphorylation of glycolysis most likely changes its interplay with other metabolic pathways and cellular processes in the parasite Saturating concentrations of GDP (0.7 mm) and GTP (0.8 mm) for EhPGK are found in E histolytica trophozoites, whereas non-saturating concentrations of the adenine nucleotides (3.3 mm ADP; mm ATP) are present [8], thus suggesting that the preference for guanine nucleotides of this peculiar amebal enzyme has physiological relevance If EhPGK preferentially synthesizes GTP in vivo, then the presence of a nucleoside diphosphate kinase, identified at the genome level [22], should be able to make ATP readily available from the GTP pool Moreover, because E histolytica lacks a de novo purine synthesis pathway (the parasite uses instead a purine-salvage pathway) [26], GTP for DNA synthesis is perhaps directly supplied by the amebal PGK reaction However, these hypotheses have not been yet explored All PGK tertiary structures experimentally determined are highly conserved throughout the taxonomical groups, mainly because all have a high degree of similarity (> 50%) at the primary sequence level [27] As expected, the predicted 3D structures for EhPGK are very similar to those determined for the yeast and ˚ pig enzymes [15–17], as judged by a RMSD < 1.9 A for alfa-carbon atoms The predicted structures can be considered as high-accuracy models due to the almost 60% identity in their sequences,  95% of their main ˚ chain atoms being expected within 1.5 A, according to Baker & Sali [28] The residues that bind the adenine ring of ADP and ATP, or those from different adenine nucleotide-based analogs (AMP–PNP, adenylylimidodiphosphate, AMP–PCP) or inhibitors (adenylyl 1,1,5,5,-tetrafluoropentane-1,5-bisphosphonate), have been identified in the crystal structures of yeast [15,18], horse [29,30], pig [16,17,31], B stearothermophilus [19], T brucei [20,32] and Thermotoga maritima [33] In all these enzymes, the adenine ring lies in a hydrophobic pocket on the surface of the C-terminus domain where, as described for the horse sequence [29], three highly conserved patches can be identified (Fig 2): Gly212, Gly213, Ala214 (patch A); Gly236,Gly237, Gly238 (patch B); and Val339, Gly340 and Val341 (patch C) The corresponding reported amino acid residues involved in adenine binding in the other crystallized proteins, as well as in the EhPGK identified in the predicted model, were totally conserved (Figs 1B and 2) In the yeast structure, the adenine base is positioned on top of Gly338 (patch C), whereas Gly211 (patch A) and the side chains of Leu311 and Entamoeba GDP ⁄ GTP-dependent phosphoglycerate kinase Val339 (patch C) define the limits of the adenine-binding site [15,18] It has been reported that N6 of the adenine ring establishes a hydrogen bond with the main-chain carbonyl group of Leu311 (yeast) [15], Leu313 (horse) [29] or Gly312 (pig) [31] amino acid residues (Fig 2B) A Gly residue is also present in EhPGK in an equivalent position to that of Gly312 in the pig enzyme On the other hand, the Leu residue present in the horse and yeast enzymes is conserved in the majority of PGK sequences reported to date; however, in the EhPGK, it is substituted by a Val residue (Fig 2A) The presence of the shorter lateral chain of Val as the only evident change in the purine-binding pocket site of the EhPGK suggested that it could accommodate the larger guanine base; however, the single mutant V311L displayed no significant changes in its kinetic parameters with the four purine nucleotides as compared with the wild-type (Table 2) Thus, it was necessary to use combinations of sidechain conformers during modeling to identify other potential amino acid residues involved in the higher preference for the guanine nucleotides of the amebal enzyme As a result, Tyr239 and Glu309 were identified in the predicted 3D structure of the EhPGK They act as hydrogen bond donor and acceptor, respectively, to the carbonyl and amino groups present in the guan˚ ine ring, at distances of  1.6 A (Fig 2B) The carbonyl, a hydrogen-bond acceptor group of guanine, is replaced by a donor amino group in adenine, whereas the guanine amino at position is absent in adenine (Fig 2B) Other close contacts are found with ˚ Phe289 and Val311, at distances  2.9 A from the guanine ring (Fig 2B) Among these residues, Phe289 was conserved in all PGK sequences analyzed The analysis of the predicted structure suggested that the presence of the shorter lateral chain of Val311 instead of Leu allows for the movement of Tyr239 and Glu309 towards the guanine ring, favoring its stabilization by the two hydrogen bonds just described Analysis of the kinetic parameters (Km, Vm, Vm ⁄ Km) with the four purine nucleotides in the single mutants Y239F, E309M, E309Q and V311L EhPGKs, revealed a complex interaction between the mutated amino acid residues and the nucleotides Although the Km values for ADP ⁄ ATP were slightly lower in the mutants compared with the wild-type value, the Vm values showed a tendency to increase, resulting in an increased catalytic efficiency of the single mutants with the adenine nucleotides The Km values for the co-substrate 3-phosphoglycerate were not significantly modified in any mutant, which was in agreement with the rapid equilibrium random bi bi kinetic mechanism FEBS Journal 276 (2009) 2037–2047 ª 2009 The Authors Journal compilation ª 2009 FEBS 2043 Entamoeba GDP ⁄ GTP-dependent phosphoglycerate kinase described for PGK from different sources [10,34,35] Results with the single mutants suggested that the mutations might have produced minor rearrangements in the protein structure (as they were not drastically affected in its storage and thermal stability or optimal pH activity); these putative minor structural changes could have increased the catalysis with the adenine nucleotides at the level of the phosphoryl transfer or the release of products In this regard, it has been well documented that large movements of the two PGK domains have to occur in order to bring the two substrates together for the phosphoryl transfer reaction [20,29,32] In contrast to the single-mutant enzymes, the double mutant Y239F ⁄ E309M showed both an increased affinity for the adenine nucleotides and a decreased affinity for the guanine nucleotide, in addition to a high increase in the Vm value with ATP However, the Kd values for ADP ⁄ ATP obtained in the presence of Mg2+ were not significantly different from those observed in the wild-type enzyme The observed catalytic efficiency changes in the double mutant (Table 2), accompanied by no significant changes in the equilibrium-binding constant (Table 3), might be due to more prominent conformational constraints in the PGK reaction [36] Simultaneous compensating decrements in entropy and enthalpy of binding may be involved in yielding a relatively invariable Kd value Although this compensation has been invoked in protein research [37], it remains controversial [38] and suggests the use of isothermal titration calorimetry to fully characterize the thermodynamic parameters of binding [39] as the subject of a separate investigation In this regard, weak interactions in the adenine ring binding site have been identified in B stearothermophilus [19] and T brucei PGK structures [32] For example, the N6 of the adenine ring establishes a direct hydrogen bond with the main-chain carbonyl groups of Ala292 or Ala314, respectively, and another hydrogen bond mediated by a water molecule is formed with the hydroxyl group of Tyr223 or Tyr245 in the bacterial and trypanosomal enzymes This Tyr residue is equivalent in position to Tyr239 in the amebal enzyme and which was modified in the double mutant However, it has been well documented that there is low activity with GTP (0.2–0.4% of the activity with ATP) in the three trypanosomal PGK isoenzymes [40,41] Also, B stearotermophilus PGK can use GTP or ITP as phosphate donors, although with lesser efficiency than with ATP (27% and 42% of the activity displayed with ATP, respectively) [42] All other PGKs have a Phe residue in this Tyr position and may or may not show activity with GTP Thus, it seems that the presence of a 2044 R Encalada et al Tyr residue at this position is not the only prerequisite for displaying higher catalysis with guanine nucleotides These results show the plasticity of a binding site considered up to now as relatively conserved Thus, although the ability to use guanine nucleotides has been described in the PGK from several sources, to our knowledge it has only been analyzed in detail in the present study of the amebal enzyme Because changes in affinities and rate velocities of the double mutant cannot be explained in terms of a simple additive effect by single mutations, it seems that a cooperative effect of mutations, or interactions, participate in nucleotide selectivity The crystal structures of the wild-type, single- and double-mutant EhPGKs in a closed conformation will certainly help to clarify the molecular arrangement in the guanine ring binding site during the binding of one or the two substrates Experimental procedures Reagents and chemicals Glyceraldehyde-3-phosphate, ADP, ATP, GDP, GTP, 3-phosphoglycerate, NAD+, NADH, EDTA, phenylmethanesulfonyl fluoride, trichloroacetic acid, imidazole, Tris, Mes and Mops were from Sigma (St Louis, MO, USA); glycerol, potassium phosphate, MgCl2 and acetic acid were from JT Baker (Philipsburg, NJ, USA); dithiothreitol was from Research Organics (Cleveland, OH, USA); GAPDH was from Roche (Manheim, Germany) 3D EhPGK structural analysis Only one PGK gene (protein identifier 75.m00170) has been found in the E histolytica genome database (http://www tigr.org/tdb/e2k1/eha1/) A couple of 3D structures from the EhPGK amino acid sequence were predicted by using the program modeller with the PGK structures from S cerevisiae (ScPGK; PDB code 3-phosphoglycerateK), bound to 3-phosphoglycerate and Mg2+–ATP [15], and the pig muscle PGK (PDB code 1HDI), bound to 3-phosphoglycerate and ADP [16], as templates (see Fig S1) In addition, a 3D model was also constructed using moe software and the protein moiety of the PGK structure from pig muscle, bound to 3-phosphoglycerate and Mg2+–AMP– PCP (b,c-methylene-adenosine-5¢triphosphate) as template (PDB code1KF0) [17] Amino acid residues known to interact with the purine ring of ADP ⁄ ATP in PGKs are highly conserved throughout most of the evolutionary lineages [18] Protein sequence alignments were performed by using clustalw2 (http:// www.ebi.ac.uk/Tools/clustalw2/index.html) and blast (http://blast.ncbi.nlm.nih.gov/Blast.cgi) with the EhPGK FEBS Journal 276 (2009) 2037–2047 ª 2009 The Authors Journal compilation ª 2009 FEBS R Encalada et al sequence and PGK sequences from organisms from several taxa to identify the amino acid frequency in relevant positions Most of the conserved positions were also present in the three predicted 3D models of the EhPGK By contrast, non-conserved residues were modeled using the library of side-chain conformations present in moe One thousand different models were constructed, differing only in the particular combination of side-chain conformers, being each of them optimized by energy minimization in the presence of GDP with the CHARMM27 force field [43] Site-directed mutagenesis, protein overexpression and purification Site-directed mutagenesis was performed in the wild-type gene previously cloned in our laboratory [4] by using the PCR-based technique of mega-oligonucleotides [44] and the High Fidelity PCR system (Roche) The 5¢- and 3¢-external oligonucleotides used contained NdeI and BamHI restriction sites; internal oligonucleotides were used to introduce the desired mutations (Table S1) Single mutations were (EhPGK amino acid sequence numbering) Y239F, V311L, E309Q, E309M and a double mutant Y239F ⁄ E309M PCR products were cloned in the pGEM-T-easy vector (Promega, Madison, WI, USA) and sequenced to verify the presence of the mutation and the absence of any other substitution introduced during the PCR Mutated genes were further cloned in the pET28 expression vector (Novagen, Madison, WI, USA) and re-sequenced The wild-type and mutant proteins were overexpressed in E coli BL21DE3pLysS cells (Novagen), fused to a histidine tag at the N-terminus and then purified by metal-affinity chromatography as described previously [4] Purified proteins were concentrated by ultrafiltration to 0.5–2 mg proteinỈmL)1 and stored at )22 °C in the presence of 50% glycerol (v ⁄ v) Purity was determined in silver-stained SDS ⁄ PAGE gels Protein concentration determination was made by the standard Lowry method using trichloroacetic acid-precipitated protein samples to avoid interference by imidazole from the purification buffer E histolytica and S cerevisiae cellular extracts Cytosolic extracts from E histolytica HM1:IMSS trophozoites were obtained as described previously [8] S cerevisiae strain BY4741 was grown in YPD medium (1% yeast extract, 2% peptone and 2% glucose) at 30 °C The cells were harvested at the end of the exponential growth, washed once with ice-cold lysis buffer (25 mm Tris ⁄ HCl pH 7.6, mm EDTA pH 8.0, mm dithiothreitol and mm phenylmethanesulfonyl fluoride) and resuspended in half its wet weight with the same buffer The cells were disrupted with glass beads and a clarified extract was obtained by centrifugation, which was further stored at )20 °C in the presence of 10% glycerol (v ⁄ v) Entamoeba GDP ⁄ GTP-dependent phosphoglycerate kinase Enzymatic assays The Km and Vm values were determined in coupled assays with commercial GAPDH (Roche) at 37 °C by monitoring the absorbance at 340 nm of both the NAD+ reduction in the forward (glycolytic) reaction and the NADH oxidation in the reverse reaction in a spectrophotometer (Agilent, Santa Clara, CA, USA) The forward enzymatic assay contained a buffer mixture adjusted at pH 7.0 (50 mm potassium phosphate and 10 mm each of acetic acid, Mes and Tris), 5–10 mm MgCl2, mm dithiothreitol, mm EDTA, 0.5 mm NAD+, 1.6–3.2 U of GAPDH, varied concentrations of GDP or ADP; mm glyceraldehyde-3phosphate was added just before the assay and the reaction was started by adding 0.04–0.2 lg of wild-type or mutant enzymes The reverse assay components were similar to those of the forward assay except that 50 mm imidazole replaced potassium phosphate and 0.15 mm NADH were used, with varying concentrations of GTP or ATP, in the presence of saturating concentrations of the co-substrate 3-phosphoglycerate (6 mm); the reaction was started by adding 0.4–2 lg of enzyme For Km value determinations, nucleotide concentrations were varied until the following highest concentrations were reached, which were saturating depending on the enzyme: GDP, 2–4 mm; GTP, 3–6 mm; ADP, 4–11 mm; and ATP, 4–10 mm, and using saturating concentration of the co-substrate When determining the Km3-phosphoglycerate of the wild-type and mutant enzymes, the concentration of GTP present in the assay was at least 10 times their respective KmGTP value PGK Km values for nucleotides in cellular clarified extracts were determined in the assay described above in the presence of saturating concentrations of 3-phosphoglycerate (EhPGK mm; ScPGK, mm) and 0.5–5 lg of cellular protein depending on the source and the nucleotide Basal activity in the absence of specific substrates was always subtracted and the activity was calculated under initial velocity conditions One substrate titration was made for each independent protein purification ⁄ clarified extract The concentration of all the substrates was routinely calibrated The nonlinear fitting of the experimental points to the Michaelis–Menten equation was performed by using origin microcal v 5.0 software Kd values Dissociation constants were determined at 37 °C in 50 mm Mops pH 7.0, 10 mm MgCl2 and 0.5 mg of purified protein The decrease in protein intrinsic-fluorescence caused by nucleotide binding was monitored in a spectrofluorometer (SLM Aminco-Bowman, Rochester, NY, USA) at 290 nm excitation and 300–400 nm emission The highest nucleotide concentrations used in the assay were mm for GDP ⁄ GTP and 10 mm for ADP ⁄ ATP FEBS Journal 276 (2009) 2037–2047 ª 2009 The Authors Journal compilation ª 2009 FEBS 2045 Entamoeba GDP ⁄ GTP-dependent phosphoglycerate kinase Acknowledgements ´ This work was supported by CONACyT-Mexico grants 83084 to ES, 80534 to RMS and ‘Acuerdo del Rector General’ UAM to AR References Reeves RE (1984) Metabolism of Entamoeba histolytica Schaudinn, 1993 Adv Parasitol 23, 105–142 McLaughlin J & Aley S (1985) The biochemistry and functional morphology of Entamoeba J Protozool 32, 221–240 Reeves RE, Montalvo F & Sillero A (1967) Glucokinase from Entamoeba histolytica and related organisms Biochemistry 6, 1752–1760 ´ Saavedra E, Encalada R, Pineda E, Jasso-Chavez R & ´ Moreno-Sanchez R (2005) Glycolysis in Entamoeba histolytica: biochemical characterization of recombinant glycolytic enzyme and flux control analysis FEBS J 272, 1767–1783 Reeves RE, Serrano R & South DJ (1976) 6-Phosphofructokinase (pyrophosphate) Properties of the enzyme from Entamoeba histolytica and its reaction mechanism J Biol Chem 251, 2958–2962 Reeves RE (1968) A new enzyme with the glycolytic function of pyruvate kinase J Biol Chem 243, 3202–3204 ´ Saavedra-Lira E, Ramı´ rez-Silva L & Perez-Montfort R (1998) Expression and characterization of recombinant pyruvate phosphate dikinase from Entamoeba histolytica Biochim Biophys Acta 1382, 47–54 ´ Saavedra E, Marı´ n-Hernandez A, Encalada R, Olivos ´ ´ A, Mendoza-Hernandez G & Moreno-Sanchez R (2007) Kinetic modeling can describe in vivo glycolysis in Entamoeba histolytica FEBS J 274, 4922–4940 Krietsch WK & Bucher T (1970) 3-Phosphoglycerate ă kinase from rabbit skeletal muscle and yeast Eur J Biochem 17, 568–580 10 Scopes RK (1973) 3-Phosphoglycerate kinase In The Enzymes, Vol (Boyer PD, ed.), pp 335–351 Academic Press, New York, NY 11 Lee Ch-S & O¢Sullivan W (1975) Properties and mechanism of human erythrocyte phosphoglycerate kinase J Biol Chem 250, 1275–1281 12 Kuntz GWK & Krietsch WKG (1982) Phosphoglycerate kinase from animal tissue Meth Enzymol 90, 103–110 13 Kuntz GWK & Krietsch WKG (1982) Phosphoglycerate kinase from spinach, blue–green algae, and yeast Meth Enzymol 90, 110–114 14 Reeves RE & South DJ (1974) Phosphoglycerate kinase (GTP) An enzyme from Entamoeba histolytica selective for guanine nucleotides Biochem Biophys Res Commun 58, 1053–1057 2046 R Encalada et al 15 Watson HC, Walker NPC, Shaw PJ, Bryant TN, Wendell PL, Fothergill LA, Perkins RE, Conroy SC, Dobson MJ, Tuite MF et al (1982) Sequence and structure of yeast phosphoglycerate kinase EMBO J 1, 1635–1640 ´ 16 Szilagyi AN, Ghosh M, Garman E & Vas M (2001) A ˚ 1.8 A resolution structure of pig muscle 3-phosphoglycerate kinase with bound MgADP and 3-phosphoglycerate in open conformation: new insight into the role of the nucleotide in domain closure J Mol Biol 306, 499– 511 ´ ´ 17 Kovari Z, Flachner B, Naray-Szabo G & Vas M (2002) Crystallographic and thiol-reactivity studies on the complex of pig muscle phosphoglycerate kinase with ATP analogues: correlation between nucleotide binding mode and helix flexibility Biochemistry 41, 8796–8806 18 Joao HC & Williams RJP (1993) The anatomy of a kinase and the control of phosphate transfer Eur J Biochem 216, 1–18 19 Davies GJ, Gamblin SJ, Littlechild JA, Dauter Z, Wilson KS & Watson HC (1994) Structure of the ADP complex of the 3-phosphoglycerate kinase from ˚ Bacillus stearothermophilus at 1.65 A Acta Crystallogr 50, 202–209 20 Bernstein BE, Michels PAM & Hol WGJ (1997) Synergistic effects of substrate-induced conformational changes in phosphoglycerate kinase activation Nature 385, 275–278 21 Northrop DB (1999) So what exactly is V ⁄ K, anyway? Biomed Health Res 27, 250–263 22 Clark CG, Alsmark UCM, Tazreiter M, Saito-Nakano Y, Ali V, Marion S, Weber C, Mukherjee C, Bruchhaus I, Tannich E et al (2007) Structure and content of the Entamoeba histolytica genome Adv Parasitol 65, 51–190 23 Saavedra-Lira E & Perez-Montfort R (1996) Energy production in Entamoeba histolytica: new perspectives in rational drug design Arch Med Res 27, 257–264 ´ 24 Moreno-Sanchez R, Saavedra E, Rodrı´ guez-Enrı´ quez S & Olı´ n-Sandoval V (2008) Metabolic control analysis: a tool for designing strategies to manipulate metabolic pathways J Biomed Biotechnol, doi: 10.1155/2008/ 597913 ´ ´ 25 Moreno-Sanchez R, Encalada R, Marı´ n-Hernandez A & Saavedra E (2008) Experimental validation of metabolic pathway modeling An illustration with glycolytic segments from Entamoeba histolytica FEBS J 275, 3454–3469 26 Albach RA (1989) Nucleic acids of Entamoeba histolytica J Protozool 36, 197–205 27 Fleming T & Littlechild J (1997) Sequence and structural comparison of thermophilic phosphoglycerate kinases with a mesophilic equivalent Comp Biochem Physiol A 118, 439–451 FEBS Journal 276 (2009) 2037–2047 ª 2009 The Authors Journal compilation ª 2009 FEBS Entamoeba GDP ⁄ GTP-dependent phosphoglycerate kinase R Encalada et al 28 Baker D & Sali A (2001) Protein structure prediction and structural genomics Science 294, 93–96 29 Banks RD, Blake CCF, Evans PR, Haser R, Rice DW, Hardy GW, Merret M & Phillips AW (1979) Sequence, structure and activity of phosphoglycerate kinase: a possible hinge-bending enzyme Nature 279, 773–777 30 Blake CCF & Rice DW (1981) Phosphoglycerate kinase Phil Trans R Soc Lond A 293, 93–104 ˚ 31 May A, Vas M, Harlos K & Blake C (1996) 2.0 A resolution structure of a ternary complex of pig muscle phosphoglycerate kinase containing 3-phospho-d-glycerate and the nucleotide Mn-adenylylimidodiphosphate Proteins 24, 292–303 32 Bernstein BE, Williams DM, Bressi JC, Kuhn P, Gelb MH, Blackburn GM & Hol WGJ (1998) A bisubstrate analog induces unexpected conformational changes in phosphoglycerate kinase from Trypanosoma brucei J Mol Biol 279, 1137–1148 33 Auerbach G, Huber R, Grattinger M, Zaiss K, Schurig ă H, Jaenicke R & Jacob U (1997) Closed structure of phosphoglycerate kinase from Thermotoga maritima reveals the catalytic mechanism and determinants of thermal stability Structure 5, 1475–1483 34 Scopes RK (1978) Binding of substrates and other anions to yeast phosphoglycerate kinase Eur J Biochem 91, 119–129 35 Lavoinne A, Marchand JC, Chedeville A & Matray F (1983) Kinetic studies of the reaction mechanism of rat liver phosphoglycerate kinase in the direction of ADP utilization Biochimie 65, 211–220 36 Geerlof A, Travers F, Barman T & Lionne C (2005) Perturbation of yeast 3-phosphoglycerate kinase reaction mixtures with ADP: transient kinetics of formation of ATP from bound 1,3-bisphosphoglycerate Biochemistry 44, 14948–14955 37 Lumry R (2003) Uses of enthalpy–entropy compensation in protein research Biophys Chem 105, 545–557 38 Sharp K (2001) Entropy–enthalpy compensation: fact or artifact? Protein Sci 10, 661–667 39 Ward WH & Holdgate GA (2001) Isothermal titration calorimetry in drug discovery Prog Med Chem 38, 309– 376 40 Misset O & Opperdoes FR (1987) The phosphoglycerate kinases from Trypanosoma brucei A 41 42 43 44 comparison of the glycosomal and cytosolic isoenzymes and their sensitivity towards Suramin Eur J Biochem 162, 493–500 Alexander K & Parsons M (1993) Characterization of a divergent glycosomal microbody phosphoglycerate kinase from Trypanosoma brucei Mol Biochem Parasitol 60, 265–272 Suzuki K & Imahori K (1974) Phosphoglycerate kinase of Bacillus stearothermophilus J Biochem 76, 771–782 MacKerell AD Jr, Feig M & Brooks CL 3rd (2004) Extending the treatment of backbone energetics in protein force fields: limitations of gas-phase quantum mechanics in reproducing protein conformational distributions in molecular dynamics simulations J Comp Chem 25, 1400–1415 Sambrook J & Russell DW (2001) Molecular Cloning A Laboratory Manual Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY Supporting information The following supplementary material is available: Fig S1 (A) Overlapped structures of the EhPGK model based on the pig PGK (1HDI) with the corresponding crystal structure (B) Overlapped structures of the EhPGK model based on the yeast PGK with the corresponding crystal structure Fig S2 Silver-stained SDS ⁄ PAGE gel from purified wild-type and mutant EhPGKs Fig S3 Thermostability of WT and mutant EhPGKs Fig S4 pH dependence on the activity of WT and mutant EhPGKs on the reverse reaction Fig S5 Gel-filtration analysis Table S1 Oligonucleotides used in this study This supplementary material can be found in the online version of this article Please note: Wiley-Blackwell is not responsible for the content or functionality of any supplementary materials supplied by the authors Any queries (other than missing material) should be directed to the corresponding author for the article FEBS Journal 276 (2009) 2037–2047 ª 2009 The Authors Journal compilation ª 2009 FEBS 2047 ... that the amino group at carbon of the guanine ring may interact with the side chain of Glu309, whereas the carbonyl group at position of the guanine ring may interact with the hydroxyl group of the. .. structures of the EhPGK obtained using the modeller program (Fig S1), it was found that the only difference in the amino acids that bind the purine ring was the presence of Val instead of Leu at... due to the almost 60% identity in their sequences,  95% of their main ˚ chain atoms being expected within 1.5 A, according to Baker & Sali [28] The residues that bind the adenine ring of ADP