The Multiple Activities of Polyphosphate Kinase ofEscherichia coli and Their Subunit Structure Determined by Radiation Target Analysis

Polyphosphate kinase (PPK), the principal enzyme required for the synthesis of inorganic polyphosphate (polyP) from ATP, also exhibits other enzymatic activities, which differ significantly in their biochemical optima and responses to chemical agents. These several activities include: polyP synthesis (forward reaction), nATP → polyPn + nADP (Equation 1); ATP synthesis from polyP (reverse reaction), ADP + polyPn → ATP + polyPn − 1 (Equation 2); general nucleoside-diphosphate kinase, GDP + polyPn → GTP + polyPn − 1(Equation 3); linear guanosine 5′-tetraphosphate (ppppG) synthesis, GDP + polyPn → ppppG + polyPn − 2 (Equation4); and autophosphorylation, PPK + ATP → PPK-P + ADP (Equation 5). The Mg2+ optima are 5, 2, 1, and 0.2 mm, respectively, for the activities in Equations 1, 2, 3, and 4. Inorganic pyrophosphate inhibits the activities in Equations 1 and 3 but stimulates that in Equation 4. The kinetics of the activities in Equations 1, 2, and 3 are highly processive, whereas the transfer of a pyrophosphoryl group from polyP to GDP (Equation 4) is distributive and demonstrates a rapid equilibrium, random Bi-Bi catalytic mechanism. Radiation target analysis revealed that the principal functional unit of the homotetrameric PPK is a dimer. Exceptions are a trimer for the synthesis of ppppG (Equation 4) and a tetrameric state for the autophosphorylation of PPK (Equation 5) at low ATP concentrations. Thus, the diverse functions of this enzyme involve different subunit organizations and conformations. The highly conserved homology of PPK among 18 microorganisms was used to determine important residues and conserved regions by alanine substitution, by site-directed mutagenesis, and by deletion mutagenesis. Of 46 single-site mutants, seven exhibit none of the five enzymatic activities; in one mutant, ATP synthesis from polyP is reduced relative to GTP synthesis. Among deletion mutants, some lost all five PPK activities, but others retained partial activity for some reactions but not for others. Polyphosphate kinase (PPK), the principal enzyme required for the synthesis of inorganic polyphosphate (polyP) from ATP, also exhibits other enzymatic activities, which differ significantly in their biochemical optima and responses to chemical agents. These several activities include: polyP synthesis (forward reaction), nATP → polyPn + nADP (Equation 1); ATP synthesis from polyP (reverse reaction), ADP + polyPn → ATP + polyPn − 1 (Equation 2); general nucleoside-diphosphate kinase, GDP + polyPn → GTP + polyPn − 1(Equation 3); linear guanosine 5′-tetraphosphate (ppppG) synthesis, GDP + polyPn → ppppG + polyPn − 2 (Equation4); and autophosphorylation, PPK + ATP → PPK-P + ADP (Equation 5). The Mg2+ optima are 5, 2, 1, and 0.2 mm, respectively, for the activities in Equations 1, 2, 3, and 4. Inorganic pyrophosphate inhibits the activities in Equations 1 and 3 but stimulates that in Equation 4. The kinetics of the activities in Equations 1, 2, and 3 are highly processive, whereas the transfer of a pyrophosphoryl group from polyP to GDP (Equation 4) is distributive and demonstrates a rapid equilibrium, random Bi-Bi catalytic mechanism. Radiation target analysis revealed that the principal functional unit of the homotetrameric PPK is a dimer. Exceptions are a trimer for the synthesis of ppppG (Equation 4) and a tetrameric state for the autophosphorylation of PPK (Equation 5) at low ATP concentrations. Thus, the diverse functions of this enzyme involve different subunit organizations and conformations. The highly conserved homology of PPK among 18 microorganisms was used to determine important residues and conserved regions by alanine substitution, by site-directed mutagenesis, and by deletion mutagenesis. Of 46 single-site mutants, seven exhibit none of the five enzymatic activities; in one mutant, ATP synthesis from polyP is reduced relative to GTP synthesis. Among deletion mutants, some lost all five PPK activities, but others retained partial activity for some reactions but not for others. polyphosphate polyphosphate kinase guanosine tetraphosphate polyacrylamide gel electrophoresis kilobase(s) nucleoside-diphosphate kinase megarad glucose-6-phosphate dehydrogenase Inorganic polyphosphate (polyP),1 a linear polymer of hundreds of phosphate residues (Pi) linked by phosphoanhydride bonds, is found in all cells in nature (1.Kulaev I.S. The Biochemistry of Inorganic Polyphosphate. John Wiley & Sons, Inc., New York1979: 11-35Google Scholar). The principal enzyme that synthesizes polyP from ATP inEscherichia coli is polyP kinase (PPK), a peripheral, membrane-bound homotetramer of 80-kDa subunits (2.Akiyama M. Crooke E. Kornberg A. J. Biol. Chem. 1992; 267: 22556-22561Abstract Full Text PDF PubMed Google Scholar). PPK is highly conserved in many bacterial species, including some of the major pathogens (e.g. Helicobacter pylori, Mycobacterium tuberculosis, and Neisseria meningitidis) and is a plausible antimicrobial target (3.Tzeng C.M. Kornberg A. Mol. Microbiol. 1998; 29: 381-383Crossref PubMed Scopus (48) Google Scholar). E. coli ppk mutants fail to make the adaptive changes in the stationary phase needed for resistance to various stresses and for survival (4.Rao N.N. Kornberg A. J. Bacteriol. 1996; 178: 1394-1400Crossref PubMed Google Scholar). An initial event in converting the γ-phosphate of ATP into polyP chains of 700–800 residues (Equation 1) is the phosphorylation of critical histidine residues His-435 and His-454 (Equation 5) (5.Kumble K.D. Ahn K. Kornberg A. Proc. Natl. Acad. Sci. U. S. A. 1996; 93: 14391-14395Crossref PubMed Scopus (50) Google Scholar). This is followed by highly processive polymerization with no detectable intermediates; neither ATP, Pi, nor polyP chains prime the reaction (6.Ahn K. Kornberg A. J. Biol. Chem. 1990; 265: 11734-11739Abstract Full Text PDF PubMed Google Scholar). The reverse reaction (Equation 2), in which ADP is converted to ATP by polyP, is kinetically slower than the forward reaction but can be driven to completion by an excess of ADP. More generally, PPK functions as a nucleoside-diphosphate kinase, converting GDP, CDP, and UDP to their respective nucleoside triphosphates (7.Kuroda A. Kornberg A. Proc. Natl. Acad. Sci. U. S. A. 1997; 94: 439-442Crossref PubMed Scopus (84) Google Scholar). Another novel feature of PPK is catalysis of the attack by GDP on a subterminal linkage of polyP, resulting in the transfer of a pyrophosphoryl group to generate the linear guanosine tetraphosphate (ppppG) (7.Kuroda A. Kornberg A. Proc. Natl. Acad. Sci. U. S. A. 1997; 94: 439-442Crossref PubMed Scopus (84) Google Scholar). Radiation target analysis was used to examine all five PPK activities with regard to optimal reaction conditions, the effects of chemical agents, the catalytic mechanism of ppppG synthesis, and the subunit organization needed for each of the activities. Furthermore, site-directed mutagenesis by alanine substitution and deletion mutagenesis were used to identify key residues and fragments within the highly conserved regions of PPK. Purified PPK (100 ng) was added to a reaction buffer (50 mm Hepes (pH 7.5), 50 mmammonium sulfate, 5 mm MgCl2) at 37 °C, as described previously (2.Akiyama M. Crooke E. Kornberg A. J. Biol. Chem. 1992; 267: 22556-22561Abstract Full Text PDF PubMed Google Scholar). Quantification of polyP, ATP, GTP, and ppppG spots on TLC plates was as described previously (7.Kuroda A. Kornberg A. Proc. Natl. Acad. Sci. U. S. A. 1997; 94: 439-442Crossref PubMed Scopus (84) Google Scholar). PolyP accumulationin vivo and enzyme activities in vitro were measured initially in a high throughput, 96-well format and then checked individually. PolyP levels and enzyme activities were measured both by a radioactive method (6.Ahn K. Kornberg A. J. Biol. Chem. 1990; 265: 11734-11739Abstract Full Text PDF PubMed Google Scholar) and by a luciferase-based nonradioactive method (8.Ault-Riché D. Fraley C.D. Tzeng C.M. Kornberg A. J. Bacteriol. 1998; 180: 1841-1847Crossref PubMed Google Scholar). To assay for PPK autophosphorylation, 60–100 ng of an irradiated enzyme was incubated in 50 mm Hepes-KOH (pH 7.2), 40 mm ammonium sulfate, 10 mm MgCl2, and either 5 μm or 1 mm[γ-32P]ATP on ice for 5 min; the reaction was terminated by adding 40 mm EDTA. The complex was precipitated with two volumes of acetone and then washed with ethanol to remove the unreacted ATP. The pellet was resuspended in SDS-PAGE buffer, electrophoresed on a 12% gel, and analyzed with a digital scanner (5.Kumble K.D. Ahn K. Kornberg A. Proc. Natl. Acad. Sci. U. S. A. 1996; 93: 14391-14395Crossref PubMed Scopus (50) Google Scholar). PolyP synthesis or degradation was measured in several activity assays: for polyP synthesis from ATP (Equation 1; for nucleoside-diphosphate kinase (NDK) activities, reverse reaction Equation 2; and GTP synthesis, Equation 3) and for ppppG synthesis (Equation 4). The reactions were terminated by adding 5× urea-PAGE buffer dye (450 mm Tris borate, 15 mm EDTA, 0.125% bromphenol blue, and 50% sucrose). The samples were electrophoresed on 6% urea-polyacrylamide gels at 300 V until the dye migrated 6–8 cm from the top of wells. Labeled polyP was stained with toluidine blue followed by exposure on a PhosphorImager. A 50-μl aliquot of purified PPK (0.2–0.5 μg/ml) in a 0.5-ml microcentrifuge tube was frozen at −80 °C and shipped on dry ice to Taiwan for radiation inactivation. The samples were exposed to γ-rays at −63 °C at 1.5 Mrad/h for various times to obtain the desired dose; nonirradiated samples were assayed as controls. Glucose-6-phosphate dehydrogenase (G6PDHase) was used as an internal standard to measure the functional decay after irradiation by monitoring the rate of NADPH appearance at an absorbance of 340 nm as follows. An assay mixture of 50 mm Tris-HCl (pH 8.0), 1 mm EDTA, 10 mm MgCl2, 3 mm glucose 6-phosphate, and 0.3 mm NADP was incubated with G6PDHase at room temperature. The linear rate of NADPH absorbance during the first 2 min was used to calculate the activity as before (9.Tzeng C.M. Yang S.Y. Yang S.J. Jiang S.S. Kuo S.Y. Hung S.H. Ma J.T. Pan R.L. Biochem. J. 1996; 316: 143-147Crossref PubMed Scopus (27) Google Scholar). The functional size of G6PDHase with a native molecular mass of 104 kDa was 112 kDa as determined by this method. To determine the structural size of PPK, irradiated samples were electrophoresed directly on 12% SDS-PAGE gels (Amersham Pharmacia Biotech Phast System). Functional sizes and D37 values were calculated from the equation, log m = 5.89 − D37,T − 0.0028T where the D37,T is the radiation dosage in Mrad required to inactivate the activity to 37% that of the control at temperatureT (°C); m is the functional size in daltons, and T is the irradiation temperature (10.Kempner E.S. Trends Biochem. Sci. 1993; 18: 236-239Abstract Full Text PDF PubMed Scopus (39) Google Scholar). The kinetic parametersK m, k cat, andK i were determined in duplicate by Lineweaver-Burk plots and time-course titrations (11.Cleland W.W. Methods Enzymol. 1979; 63: 103-138Crossref PubMed Scopus (1929) Google Scholar, 12.Segel I.H. Enzyme Kinetics: Behavior and Analysis of Rapid Equilibrium and Steady-state Enzyme Systems. Wiley-Interscience, New York1975: 506-845Google Scholar). For initial rate studies, the reactions were started by adding 5 μl of purified enzyme (∼100 ng) to 15 μl of reaction mixture (50 mm Hepes, pH 7.5, 50 mm ammonium sulfate, and 0.2 mm magnesium chloride) at 37 °C containing either 1 mm GDP with 1–100 μm polyP or 0.1–2 mm of GDP with 10 μm polyP. For product-analog inhibition experiments, the conditions used were 0–20 μm polyP65 (Sigma) or 0–2 mm GMP (Sigma) as inhibitors. Reactions were spotted onto TLC plates and separated in 0.75 mKH2PO4 as the mobile phase; the products were quantified using a PhosphorImager (Molecular Dynamics). Plasmid pQE30ppk (5.Kumble K.D. Ahn K. Kornberg A. Proc. Natl. Acad. Sci. U. S. A. 1996; 93: 14391-14395Crossref PubMed Scopus (50) Google Scholar) was used as a template for site-directed mutagenesis and the construction of deletion mutants by polymerase chain reaction. The pQE30 (QIAexpressTM, QIAGEN) plasmid vector with the highly efficient isopropyl-1-thio-β-d-galactopyranoside-inducible T5 promoter, and a Hisx6 tag coding sequence was used for overexpression and purification of the wild type and mutant PPK enzymes. E. coli XL2-blue (QIAexpressTM, QIAGEN) was the host strain used for the screening and propagation of plasmids. E. coli CF5802 (MG1655 ΔppkΔppx::kan) (4.Rao N.N. Kornberg A. J. Bacteriol. 1996; 178: 1394-1400Crossref PubMed Google Scholar) was the host strain for overexpression. Site-directed, alanine substitution mutagenesis was performed with the QuikChangeTMsite-directed mutagenesis kit (Stratagene Inc.). The mutagenic oligonucleotide primers were all 27 base pairs in length with the target codon in the center changed to GCT for Ala (A). Alignments of PPK amino acid sequences from 18 microorganisms revealed three highly conserved regions at residues 10–60, 350–480, and 550–630 of the 687-residue E. coli PPK (3.Tzeng C.M. Kornberg A. Mol. Microbiol. 1998; 29: 381-383Crossref PubMed Scopus (48) Google Scholar). Amino acids within the carboxyl-terminal portion were chosen for site-directed mutagenesis, particularly the charged and the hydrophobic residues. Seven polymerase chain reaction oligonucleotide primers were synthesized: 5′-a, CTCGGATCCATGGGTCAGGAAAAG; 5′-b, CTCGCATGCATTACGCCGATTTTA; 5′-c, CTCGCATGCTTCCGCAATGGTTTT; 3′-a, CTCAAGCTTTTATTGAGGTTGTTC; 3′-b, GAAGCATGCGGGCAGCCCTTGCTG; 3′-c, GAAGCATGCTTTATCAAACCAAAT; and 3′-d, GTTGCATGCGTGCTGACGCAGATA with restriction sites added for subcloning into pQE30. These primers were used to generate six deletion mutants (the expected sizes on agarose gels are given in parentheses): PPKΔ135–687 (0.5 kb), PPKΔ327–687 (0.98 kb), PPKΔ532–687 (1.6 kb), PPKΔ1–134 (1.6 kb), PPKΔ135–326 (1.5 kb), and PPKΔ1–326 (1.09 kb), as well as the wild type (2.07 kb). The expected sizes of the 1 mmisopropyl-1-thio-β-d-galactopyranoside-induced proteins (given in parentheses): PPKΔ135–687 (15 kDa), PPKΔ327–687 (36 kDa), PPKΔ532–687 (55 kDa), PPKΔ1–134 (50 kDa), and PPKΔ1–326 (40 kDa), PPKΔ135–326 (58 kDa), and wild type (72 kDa), agreed with the values determined by denaturing gel electrophoresis and Western blotting (data not shown). To verify that the mutations were constructed properly, the entire ppk region of each mutant was sequenced. Standard molecular biology and transformation procedures were used. Wild type and mutant cells were grown in Luria-Bertani (LB) medium with or without 100 μg/ml ampicillin and 25 μg/ml kanamycin at 37 °C and with aeration until the A 600 reached 1.0. Isopropyl-1-thio-β-d-galactopyranoside was then added to a final concentration of 1.0 mm, and the cultures were grown for an additional 2 h at 30 °C. Cells were collected by centrifugation at 6000 × g for 10 min and resuspended to 3 volume/g of wet weight in lysis buffer (50 mmTris-HCl, pH 7.4, 10% (v/v) glycerol, 5 mmMgCl2, 1 mm dithiothreitol, and 250 μg/ml lysozyme). The samples were then incubated at 37 °C for 10 min, subjected to three freeze-thaw cycles and sonication for 1 min, and treated with 25 μg/ml each of DNase and RNase for 30 min at 4 °C. At final concentrations of 1 m KCl, 100 mmNa2CO3, and 0.05% Triton X-100 added sequentially, the mixture was incubated for 2 h at 4 °C and then sonicated for 1 min to solubilize PPK from the membrane. Cell debris was pelleted by centrifugation at 40,000 ×g for 20 min, and the supernatant was applied to a nickel-nitrilotriacetic acid column previously equilibrated with 50 mm Tris-HCl, pH 7.4, 10% (v/v) glycerol, 5 mmMgCl2, 1 mm dithiothreitol, 0.05% Triton X-100, and 100 mm imidazole. After 10 bed-volume washes, bound PPK was eluted with 10 mm EDTA and pooled by passing through a PD-10 desalting column. Samples were verified by SDS-PAGE and Western blotting and further characterized by activity assays. E. coli PPK transfers the γ-phosphate of ATP processively to generate polyP chains of lengths of 700 to 800 residues. nATP→polyPn+nADPEquation 1 As a NDK, PPK catalyzes the transfer of a terminal phosphate residue to ADP (the reverse reaction) or GDP as well as CDP and UDP. ADP+polyPn→ATP+polyPn−1Equation 2 GDP+polyPn→GTP+polyPn−1Equation 3 A GDP attack on the subterminal linkage of a polyP chain generates the linear ppppG. GDP+polyPn→ppppG+polyPn−2Equation 4 Under conditions that favor the reaction in Equation 4 and inhibit the reaction in Equation 3 (see below and TableI), the polyP chains diminish in size progressively (Fig. 1), a distributive reaction in contrast to the highly processive reaction in EquationsEquation 1, Equation 2, Equation 3.Table IEffects of guanidine HCl, Mg 2+, and pyrophosphate on PPK activitiesPPK activitiesGuHCl (5 mm) relative to controlOptimal [Mg2+]PPi (10 mm) relative to control%mm%Synthesis of PolyP (Eq. 1)−504.0−66Synthesis of ATP (Eq.2)02.00Synthesis of GTP (Eq. 3)01.0−75Synthesis of ppppG (Eq. 4)+200.2+100 Open table in a new tab Guanidine HCl at the low concentration of 5 mm inhibited the synthesis of polyP (Equation 1) by 50%, had no effect on the synthesis of ATP and GTP (Equations 2 and 3), and stimulated the reaction in Equation 4 by 20% (Table I). The optimal Mg2+levels differ widely for the activities in Equations Equation 1, Equation 2, Equation 3, Equation 4, with respective values of 5, 2, 1, and 0.2 mm. Whereas 10 mm inorganic pyrophosphate inhibited the activities in Equation 1 by 66% and those in Equation 3 by 75%, it increased those in Equation 4 by 100%. With regard to the synthesis of ppppG at low Mg2+ in the absence of PPi (Fig. 1 C), the removal of polyP is distributive, but the appearance of ppppG is delayed. The basis of these kinetics needs to be explored further. Gel-filtration and sedimentation-velocity measurements indicated that native PPK is a homotetramer (6.Ahn K. Kornberg A. J. Biol. Chem. 1990; 265: 11734-11739Abstract Full Text PDF PubMed Google Scholar), but the functional sizes for each of the four activities was not known. To determine these, radiation inactivation was employed in which the function of the target molecule is destroyed with progressive doses of γ-rays or high energy electrons. The exponential rate of the functional decay is compared with standards in which the mass value has been determined for many proteins by other methods (10.Kempner E.S. Trends Biochem. Sci. 1993; 18: 236-239Abstract Full Text PDF PubMed Scopus (39) Google Scholar). The decay rates observed for the first four PPK activities (Fig.2, A and B; TableII) indicate that a minimal functional size for the synthesis of polyP (Equation 1), of ATP (Equation 2), and of GTP (Equation 3) is 138–156 kDa which corresponds to a dimer (Fig.2). The unusual pattern of the ppppG synthesis decay rate can be interpreted as a two-phase reaction (Fig. 2, A andC) in which an inactive tetramer (306 kDa) decays to a trimeric state as judged by the subsequent decay rate indicative of a trimer (222 kDa). The decay rate of the tetramer in the first phase was calculated by correcting for the rate determined for the subsequent decay of the trimer.Table IIFunctional oligomeric size of PPK activitiesActivitiesD37Functional sizeSubunit organizationMradkDaSDS-PAGE14.2 ± 0.684 ± 5MonomerSynthesis of PolyP7.4 ± 1.0156 ± 18DimerSynthesis of ATP8.5 ± 0.7138 ± 11DimerSynthesis of GTP8.2 ± 0.4149 ± 17DimerSynthesis of ppppG (phase 1)3.9 ± 0.7306 ± 15TetramerSynthesis of ppppG (phase 2)5.8 ± 0.6222 ± 10TrimerAutophosphorylation at 5 μm ATP4.0 ± 0.2293 ± 14TetramerAutophosphorylation at 1 mm ATP7.3 ± 0.8159 ± 13DimerValues are averages of at least two independent experiments. D37 is the radiation dosage in megarad required to inactivate the activity to 37% that of the control. The minimal functional size of G6PDHase, the internal standard, is 110 ± 4 kDa compared to its native size of 102 kDa. Open table in a new tab Values are averages of at least two independent experiments. D37 is the radiation dosage in megarad required to inactivate the activity to 37% that of the control. The minimal functional size of G6PDHase, the internal standard, is 110 ± 4 kDa compared to its native size of 102 kDa. Autophosphorylation of PPK at histidine residues His-435 and His-454 (Equation 5) to a limit of about 0.2/monomer occurs rapidly at an ATP concentration of 5 μm (5.Kumble K.D. Ahn K. Kornberg A. Proc. Natl. Acad. Sci. U. S. A. 1996; 93: 14391-14395Crossref PubMed Scopus (50) Google Scholar, 6.Ahn K. Kornberg A. J. Biol. Chem. 1990; 265: 11734-11739Abstract Full Text PDF PubMed Google Scholar). The reaction is far more extensive (3/monomer) at 1 mm ATP, a concentration near the K m for polyP synthesis. Radiation target analysis revealed that the functional size for PPK phosphorylation is a tetramer (293 kDa) at 5 μm ATP, whereas the active form, as in polyP synthesis at 1 mm ATP, is a dimer (159 kDa) (Fig. 3). When degraded by ionization, purified PPK becomes a monomer (80 kDa) as determined by denaturing SDS-PAGE analysis. The kinetic constants for ppppG synthesis and the order of substrate binding based on the initial rate as a function of both GDP and polyP concentrations in Lineweaver-Burk plots ruled out a ping-pong mechanism for Equation 4. Regression lines through the data of the reciprocal of rate (1/V)versus the reciprocal of GDP concentration and of 1/Vversus the reciprocal of polyP concentration plots do not intersect on the 1/V axis (data not shown). These data were used to determine the following kinetic constants for ppppG synthesis: theK m for GDP was 160 ± 29 μm and the k cat (turnover number) was 28.9 ± 1.6 min−1, whereas the K m for polyP was 35 ± 10 μm and k cat was 62.4 ± 7.4 min−1 (TableIII). The catalytic efficiencies (k cat/K m) were calculated to be 1782 min−1 mm−1 for polyP and 180 min−1 mM−1 for GDP (11.Cleland W.W. Methods Enzymol. 1979; 63: 103-138Crossref PubMed Scopus (1929) Google Scholar).Table IIIKinetics of ppppG synthesisVariable substrateFixed substrateKmk cataThe k cat values were obtained under the stated nonsaturating substrate levels.k cat/K mmmmin −1min −1 mM −1polyP (1–100 μm)GDP (1 mm)0.03562.41782GDP (1–2 mm)polyP (10 μm)0.1628.9180Product inhibitorVariable substrateFixed substrateInhibition patternK iPolyP65 (0–20 μm)polyP750GDP (1 mm)Competitive7.6 μmGDPpolyP750 (10 μm)Competitive1.2 μmGMP (0–2 mm)polyP750GDP (1 mm)Competitive1.4 mmGDPpolyP750 (10 μm)Competitive0.7 mmThe values are averages of two independent experiments. The Mg2+ concentration was 0.2 mm to favor ppppG synthesis.a The k cat values were obtained under the stated nonsaturating substrate levels. Open table in a new tab The values are averages of two independent experiments. The Mg2+ concentration was 0.2 mm to favor ppppG synthesis. To differentiate between ordered and random sequential mechanisms for ppppG synthesis (11.Cleland W.W. Methods Enzymol. 1979; 63: 103-138Crossref PubMed Scopus (1929) Google Scholar, 12.Segel I.H. Enzyme Kinetics: Behavior and Analysis of Rapid Equilibrium and Steady-state Enzyme Systems. Wiley-Interscience, New York1975: 506-845Google Scholar, 13.Hsieh P.C. Kowalczyk T.H. Phillips N.F. Biochemistry. 1996; 35: 9772-9781Crossref PubMed Scopus (28) Google Scholar), the analogs for the products, polyP65 and GMP, were used in inhibition experiments. P65 proved to be a competitive inhibitor of polyP750 and of GDP, both when GDP was held constant (K i = 7.6 ± 0.9 mm) and of GDP when polyP750 was held constant (K i = 1.2 ± 0.3 μm). GMP also proved to be a competitive inhibitor of both polyP750 and GDP, both when GDP was held constant (K i = 1.4 ± 0.3 mm) and of GDP when polyP750 was held constant (K i = 0.7 ± 0.09 μm). These initial-rate and product-analog inhibition studies demonstrate that ppppG synthesis at 0.2 mm MgCl2 occurs by a rapid equilibrium, random Bi-Bi mechanism. The amino acid alignment of 18 prokaryotic PPK sequences demonstrates a high degree of conservation, particularly within specific regions (17% of residues overall and 69% over the 60% most homologous regions). The MEME program (14.Bailey T.L. Elkan C. Fitting a Mixture Model by Expectation Maximization to Discover Motifs in Biopolymers: Proceedings of the Second International Conference on Intelligent Systems for Molecular Biology. AAAI Press, Menlo Park, CA1994: 28-36Google Scholar) revealed 10 motifs of the highest homology among these PPKs. Five of the six most homologous motifs occur in the 300 residues of the carboxyl terminus of E. coli PPK (amino acids ∼360–687) (Fig.4). The motif with the lowest degree of homology is located between residues ∼130 and 320. To determine the critical residues and regions within the 300-residue carboxyl terminus, site-directed mutagenesis was performed based on both sequence-homology analysis and the previous structure-function dissociation results. Initially, 46 single-site alanine-substituted mutants were constructed and verified. Four PPK activities (EquationsEquation 1, Equation 2, Equation 3, Equation 4) in these 46 mutants were assayed at least three times in a 96-well high-throughput format. 2C. M. Tzeng and A. Kornberg, unpublished data. Of the 46 mutants, 34 exhibited PPK activities similar to those of the wild type. Eight mutants with activities that differed from wild type were further purified, checked by SDS-PAGE and Western blotting (data not shown), and characterized in detail (TableIV).Table IVActivities of PPK mutantsin vivoin vitroMutantsPolyP accumulationSynthesis ofAuto-phosphorylationpolyPATPGTPppppG% relative to wild typeR375A2 ± 12 ± 14 ± 26 ± 25 ± 1<1S380A1 ± 1<1<1<13 ± 3<1Y468A45 ± 12144 ± 2136 ± 18146 ± 31122 ± 18144 ± 54F488A<14 ± 25 ± 25 ± 32 ± 24 ± 2P507A<13 ± 16 ± 24 ± 36 ± 35 ± 5R564A<11 ± 12 ± 2<1<118 ± 12R621A3 ± 1<1<1<1<1<1Q674A<15 ± 111 ± 7<1<15 ± 4Δ1–1346 ± 43 ± 19 ± 6<1<1<1Δ135–326<1<1<1<1<1<1Δ1–3261 ± 02 ± 15 ± 1<1<1<1Δ135–6875 ± 26 ± 24 ± 2<1<1<1Δ327–68714 ± 69 ± 211 ± 3<1<1<1Δ532–68717 ± 635 ± 530 ± 8<1<1<1All values are the average of four independent preparations and assays. Open table in a new tab All values are the average of four independent preparations and assays. Single-site mutants R375A, S380A, F488A, P507A, R564A, R621A, and Q674A lost all five PPK activities demonstrating that these amino acids are essential (Table IV). Y468A exhibited high levels (120–140%) of polyP, GTP, and ppppG synthesis and autophosphorylation activities (Equations 1 and 3–5) but diminished (20–50%) ATP synthesis activity (Equation 2). The PPK activities of the deletion mutants were also characterized in detail (Table IV). Only PPKΔ327–687 and PPKΔ532–687 exhibited partial polyP and ATP synthesis activities (Equations 1 and 2), suggesting that the essential fragment or domain for these functions is in the amino-terminal region. None of deletion mutants retained GTP and ppppG synthesis activities (Equations 3 and 4) implying that these require a native structure. In addition, none of the deletion mutants underwent autophosphorylation (Equation 5), indicating that an intact native protein is also required for this activity. In mutants R564A, PPKΔ327–687, and PPKΔ532–687 the polyP synthesis (Equation 1) and autophosphorylation (Equation 5) activities were dissociated; R564A autophosphorylated to 20% that of the wild type but lost all four other PPK activities. PPKΔ327–687 and PPKΔ532–687 did not display any autophosphorylation, but retained residual (10–35%) polyP (Equation 1) and ATP (Equation 2) synthesis activities. Mutant Y468A was examined for its NDK activity in substrate competition experiments (Fig. 5). In the presence of 1 mm ADP and 1 mm GDP, the ratio of ATP to GTP synthesis of the wild type enzyme is about 8:1, whereas for the mutant, the ratio is 1:3, and ppppG synthesis is reduced 40-fold relative to the wild type. The ratio of GTP to ppppG synthesis was also changed from 15:1 in the wild type to 25:1 in the mutant. These findings indicate that the catalytic sites of ATP and GTP synthesis are shared. PPK of E. coli has several discrete functions: the synthesis of polyP specifically from ATP, the reverse reaction to form ATP from polyP and ADP, and the substitution for ADP in the reverse reaction by GDP, CDP, and UDP, in essence a nucleoside-diphosphate kinase activity. The contribution of the reverse reactions in the disposal of polyP compared with that of exopolyphosphatase has not been determined, nor has the synthesis of GTP, CTP, and UTP as auxiliary to the activity of the major nucleoside-diphosphate kinase activity been evaluated. In addition to these activities, PPK autophosphorylates certain histidine residues to generate a putative intermediate and also catalyzes the transfer of a pyrophosphoryl group to GDP to generate the linear ppppG. The aforementioned activities of PPK can be distinguished by several agents, such as the optimal concentration of Mg2+ and the effects of guanidine-HCl and inorganic pyrophosphate (Table II). Inasmuch as the native state of PPK is that of a tetramer of 80-kDa subunits, these distinctive effects may depend in part on its oligomeric state. To this end, the subunit structure of the oligomer for each of the five activities was determined by radiation target analysis. Radiation inactivation of an enzyme activity has been used to measure the target size of the functional unit (9.Tzeng C.M. Yang S.Y. Yang S.J. Jiang S.S. Kuo S.Y. Hung S.H. Ma J.T. Pan R.L. Biochem. J. 1996; 316: 143-147Crossref PubMed Scopus (27) Google Scholar, 10.Kempner E.S. Trends Biochem. Sci. 1993; 18: 236-239Abstract Full Text PDF PubMed Scopus (39) Google Scholar). The dosage of the γ-rays, generated by 60Co required to inactivate a number of enzymes of known size (e.g. glucose-6-phosphate dehydrogenase), provides a scale that can be used to measure the functional size of another enzyme activity. With regard to PPK, the dimeric state best fits the functional size of the forward and reverse activities (Table III; Fig. 2, A and B); the tetrameric state appears to be optimal for autophosphorylation at 5 μm ATP, but the dimer is preferred at 1 mmATP, the concentration needed for the forward and reverse reactions (Table II; Fig. 3, A and B). The functional size for the ppppG synthesis activity is judged to be a trimer (Table II; Fig. 2, A and C) as indicated by the increase in activity as the tetramer is inactivated and the rate of subsequent decay of activity, consistent with the size of a trimer. The PPK of Propionibacterium shermanii like that of E. coli PPK has a subunit mass of 83 kDa and synthesizes polyP processively to a limit of about 750 residues, but unlike E. coli PPK appears to be monomeric (15.Robinson N.A. Clark J.E. Wood H.G. J. Biol. Chem. 1987; 262: 5216-5222Abstract Full Text PDF PubMed Google Scholar, 16.Robinson N.A. Wood H.G. J. Biol. Chem. 1986; 261: 4481-4485Abstract Full Text PDF PubMed Google Scholar). However, the glucokinase of P. shermanii, which utilizes polyP to phosphorylate glucose, is a homodimer of 33-kDa subunits (17.Phillips N.F.B. Horn P.J. Wood H.G. Arch. Biochem. Biophys. 1993; 300: 309-319Crossref PubMed Scopus (45) Google Scholar) by either processive (polyP700) or nonprocessive (polyP30) mechanisms (18.Pepin C.A. Wood H.G. J. Biol. Chem. 1986; 261: 4476-4480Abstract Full Text PDF PubMed Google Scholar). In M. tuberculosis,polyP glucokinase utilizes the long-chain polyP nonprocessively by an ordered Bi-Bi mechanism (13.Hsieh P.C. Kowalczyk T.H. Phillips N.F. Biochemistry. 1996; 35: 9772-9781Crossref PubMed Scopus (28) Google Scholar). The amino acid sequences of NDKs are highly conserved between E. coli and humans (43% identity) and are believed to be essential for DNA and RNA synthesis (19.De La Rosa A. Williams R.L. Steeg P.S. Bioessays. 1995; 17: 53-62Crossref PubMed Scopus (225) Google Scholar), as well as for bacterial growth, virulence, cell-signaling, and polysaccharide synthesis (20.Chakrabarty A.M. Mol. Microbiol. 1998; 28: 875-882Crossref PubMed Scopus (122) Google Scholar). Pyruvate kinase (21.Saeki T. Hori M. Umezawa H. J. Antibiot. (Tokyo). 1974; 28: 974-981Crossref Scopus (10) Google Scholar) and adenylate kinase (22.Lu Q. Inouye M. Proc. Natl. Acad. Sci. U. S. A. 1996; 93: 5720-5725Crossref PubMed Scopus (84) Google Scholar) may also function as NDKs inE. coli. Inasmuch as knockout mutants of E. coli(ndk- and ndk-/pyk-) are still viable (23.Lu Q. Zhang X. Almaula N. Mathews C.K. Inouye M. J. Mol. Biol. 1995; 254: 337-341Crossref PubMed Scopus (98) Google Scholar), PPK by utilizing polyP may provide a backup function as an NDK in vivo. The capacity of PPK to catalyze the attack by GDP on a subterminal linkage of polyP generates ppppG. The activity predominates over the GDP attack on the terminal linkage when the Mg2+concentration is at the low level of 0.2 mm and inorganic pyrophosphate is present. Such a pyrophosphoryl transfer was not observed with ADP. Unlike the forward and reverse reactions, which are highly processive, the synthesis of ppppG is distributive (Fig. 1) and occurs by a rapid equilibrium, random Bi-Bi mechanism (Table III). In yeast, ppppG can be generated by phosphoglycerate kinase (24.Garcia Diaz M. Canales J. Sillero M.A.G. Sillero A. Cameselle J.C. Biochem. Int. 1989; 19: 1253-1264PubMed Google Scholar) and digested by an exopolyphosphatase (25.Kulakovskaya T.V. Andreeva N.A. Kulaev I.S. Biochemistry. 1997; 62: 1051-1052PubMed Google Scholar); ppppG can also stimulate mammalian adenylate cyclase (26.Ignarro L.J. Gross R.A. Biochim. Biophys. Acta. 1978; 541: 170-180Crossref PubMed Scopus (2) Google Scholar). However, unlike ppGpp for which regulatory roles are established (27.Kuroda A. Murphy H. Cashel M. Kornberg A. J. Biol. Chem. 1997; 272: 21240-21243Abstract Full Text Full Text PDF PubMed Scopus (201) Google Scholar, 28.Cashel M. Rudd K.E. Neidhardt F.C. Ingraham J.L. Low K.B. Magasanik B. Schaechter M. Umbarger H.E. Escherichia coli and Salmonella typhimurium : Cellular and Molecular Biology. American Society for Microbiology, Washington, D. C.1987: 1410-1438Google Scholar, 29.Spira B. Silberstein N. Yagil E. J. Bacteriol. 1995; 177: 4053-4058Crossref PubMed Google Scholar, 30.Voelz H. Voelz U. Ortigoza R.O. Arch. Microbiol. 1966; 53: 371-388Google Scholar), the cellular presence of ppppG and its functions are unknown. The availability of high copy number plasmids bearing theppk gene and His tags for PPK should attract proper structural studies of the enzyme to account for its multiple functions. In the meantime, we have carried out some mutational studies and also examined some other bacterial PPKs for comparison with the E. coli enzyme. Site-directed mutagenesis was focused on regions where the amino acid alignments of 18 bacterial PPK sequence show a high degree of conservation. Alanine substitutions at 46 sites were made in five of the six most homologous regions (over 60% identity) that make up the 300 residues at the carboxyl end (360–687). Among these, the PPKs isolated from several of them showed a virtual loss of all five activities. However, in one mutant (Y468A) all the activity levels were enhanced except for the synthesis of ATP, which was diminished. Thus, the active site, altered at this residue, discriminates against GDP far less than in the wild type. Mutants in which 100 or more residues were deleted at the amino- and carboxyl-terminal ends, as well as in between (Fig. 4) retained no significant levels of any of the activities. Efforts to separate the functional domains of PPK by proteolytic digestion were not successful. Individual peptides of PPK generated by treatment with trypsin were isolated by fast protein liquid chromatography and high pressure liquid chromatography, but none contained any of the five PPK activities, although low levels were detected in pools of the digests. The PPKs of H. pylori and Vibrio cholerae have been purified and characterized as homotetramers. 2C. M. Tzeng and A. Kornberg, unpublished data., 3N. Ogawa and A. Kornberg, unpublished results. They also generate polyP with about 750 residues and behave much like that ofE. coli PPK. However, Thek cat/K m of the reverse activity of both PPKs was near 100 times greater than that of E. coli PPK, suggesting that PPK in H. pylori and V. cholerae may play more important roles in generating ATP from polyP. PPKs, partially purified from N. meningitidis and P. shermanii, were reported to have masses of 72 and 83 kDa, respectively, as determined by SDS-PAGE. The values forK m (1.5–2.0 mm) and turnover number (40–60/subunit/second), calculated from the efficiency and yield of the purification, are similar to those published for E. coli. Those enzymes also appeared to be attached to cell membranes (31.Tinsley C.R. Manjula B.N. Gotschlich E.C. Infect. Immun. 1993; 61: 3703-3710Crossref PubMed Google Scholar). An E. coli overproducer of PPK and an isopropyl-1-thio-β-d-galactopyranoside-induced, overexpressed His-tagged PPK are now available for crystallography to explore structure-function relationships. We thank Dr. Rong-Long Pan for providing the60Co radiation apparatus, Dr. Cresson D. Fraley and Leroy Bertsch for critical evaluation of the manuscript, and Dr. Sung-Kay Chiu, Dr. Nobuo Ogawa, and Dr. Dana Ault-Riché for helpful discussions during this study.