Biochemistry 2001, 40, 15009-15016
15009
Product Release during the First Turnover of the ATP Sulfurylase-GTPase† Sean Sukal and Thomas S. Leyh* The Department of Biochemistry, Albert Einstein College of Medicine, 1300 Morris Park AVenue, Bronx, New York 10461-1926 ReceiVed August 30, 2001; ReVised Manuscript ReceiVed October 12, 2001
ABSTRACT: ATP sulfurylase, from Escherichia coli Κ-12, is a GTPase target complex that catalyzes and couples the chemical potentials of two reactions: GTP hydrolysis and activated sulfate (APS) synthesis. Previous work suggested that the product release branch of the GTPase mechanism might include ratedetermining release and/or isomerization step(s). Such steps are known to couple chemical potentials in other energy transducing systems. Rate-determining, product release step(s) were confirmed in the ATP sulfurylase-GTPase reaction by a burst of product in pre-steady-state, rapid-quench experiments. Classical rapid-quench experiments, which measure total product formation, do not allow the slow steps to be assigned to the release of a specific product, or to slow isomerization, because they do not distinguish solution-phase from enzyme-bound product. Assay systems that exclusively monitor solution-phase Pi and GDP were used to obtain free product progress curves during the first turnover of ATP sulfurylase. Together, the free and total product data describe how the products partition between the enzyme surface and solution during the first turnover. In combination, the data provide the time dependence of the concentrations of specific product intermediates, AMP‚PPi‚E‚GDP‚Pi and AMP‚PPi‚E‚GDP, the rate constants for the release of Pi (4.2 s-1) and GDP (4.8 s-1) from these complexes, respectively, and the equilibrium constant for the enzyme-bound, β,γ-bond cleavage reaction: [AMP‚PPi‚E‚GTP′]/[AMP‚PPi‚ E‚GDP‚Pi] ) 0.7. The data are fit, using global analysis, to obtain a complete kinetic and energetic description of this GTPase reaction.
Sulfate, a nonreactive compound, is chemically activated in metabolism by transfer of the adenylyl moiety of ATP (∼AMP) from pyrophosphate to sulfate to form activated sulfate (adenosine 5′-phosphosulfate or APS) (reaction 1). The chemical potential of the phosphoric-sulfuric acid anhydride bond formed in the transfer is remarkably large (∆G°′hydrolysis ∼ -19 kcal/mol). Consequently, and despite cleavage of the R,β bond of ATP, APS synthesis is extremely unfavorable, Keq ) 1.1 × 10-8 [∆G°′ ) -11 kcal/mol (1-3)]. Enzymes that labor against such equilibria to provide essential metabolites must overcome the unfavorable mass action problem, as well as the attendant kinetic problem that requires it, in this case, to “give up” 1.1 × 10-8-fold in V/K in the forward versus the reverse direction. Thus, if the enzyme is a reasonably good catalyst in the favorable direction, it will likely be quite poor in the unfavorable direction. Nature has solved this problem, in Escherichia coli and other Gram-negative bacteria, by creating a GTPase target complex that couples the chemical potential of GTP hydrolysis to the synthesis of activated sulfate. ATP sulfurylase (ATP:sulfate adenylyltranferase, EC 2.7.7.4), from E. coli Κ-12, catalyzes and conformationally couples the hydrolysis of GTP and the synthesis of activated sulfate (4, 5). The enzyme is a tetramer of heterodimers (6). Each dimer is composed of a GTPase subunit (CysN, 53 kDa) and an APS synthesis subunit (CysD, 35 kDa) (7). The potential energy of these reactions are linked by changing allosteric interactions between the two active sites during †
Supported by National Institutes of Health Grant GM54469. * Corresponding author. Phone: 718-430-2857. Fax: 718-430-8565. E-mail:
[email protected].
the catalytic cycle that oblige the chemistries to proceed in a stepwise fashion, resulting in a well-defined, 1:1, stoichiometry (4). A key energy coupling step in the mechanism is an isomerization that precedes, and rate limits, GTP hydrolysis. The isomerization occurs in response to the opening of the R,β bond of ATP and can also be elicited by the binding of both AMP and PPi. The E‚AMP‚PPi complex mimics extremely well the effects of the native reaction on kcat and kisomerization for the GTP hydrolysis reaction (8, 9) and provides a simplified system for studying the GTPase.
ATP4- + SO42- / APS2- + P2O74-
(1)
GTP4- + H2O / GDP3- + HPO42- + H+
(2)
At a saturating concentration of GTP, the rate-limiting step among first three steps of the E‚AMP‚PPi-catalyzed GTP hydrolysis reaction (i.e., GTP binding, isomerization, and β,γ-bond cleavage) is the isomerization (8, 9). The rate constant for this step, 14 s-1, is considerably larger than kcat, 1.2 s-1. Hence, the mechanism must contain rate-determining steps following β,γ-bond cleavage. Such steps might include release of either or both products and/or slow isomerization(s). Slow product release is characteristic of energy transducing systems and is often associated with the coupling of chemical potentials (10-14). It is possible, if not likely, that slow steps in the product release branch of the ATP sulfurylase mechanism will also be involved in energy coupling. The current study provides a detailed investigation of product release in the E‚AMP‚PPi-catalyzed GTP hydrolysis reaction.
10.1021/bi015735a CCC: $20.00 © 2001 American Chemical Society Published on Web 11/09/2001
15010 Biochemistry, Vol. 40, No. 49, 2001 MATERIALS AND METHODS Materials. The materials and suppliers used in this study are as follows: restriction enzymes (New England Biolabs); Pfu polymerase (Stratagene); PEI-F TLC plates (EM Science); radionucleotides (NEN Life Science Products); QSepharose fast flow and Superdex 200 prep grade resins, rabbit muscle pyruvate kinase, and lactate dehydrogenase (Pharmacia); DNA primers (Oligonucleotide Synthesis Facility, Albert Einstein College of Medicine); BL21(DE3) and BL21(DE3) pLysS E. coli, pET-23a(+) and pET-28a(+) plasmids, His‚Bind resin (Novagen); buttermilk xanthine oxidase (XO), bacterial purine nucleoside phosphorylase (PNP), porcine brain guanylate kinase, and Sephadex G-25 gel filtration resin (Sigma); 7-(diethylamino)-3-((((2-maleimidyl)ethyl)amino)carbonyl)coumarin (MDCC) (Molecular Probes); buffers, media, nucleotides, and solventssthe highest available grades (Aldrich, Sigma, or Fisher). ATP Sulfurylase. Native E. coli Κ-12 ATP sulfurylase was expressed in E. coli Κ-12 (9) and purified to >95% homogeneity as described previously (6). The specific GTPase activity of the E‚AMP‚PPi complex of ATP sulfurylase was 1.2 units/mg (8). PyruVate Kinase and Lactate Dehydrogenase. The enzymes were exchanged into 50 mM Hepes,1 pH 8.0/KOH (the buffer used in the studies described in this paper) using size exclusion chromatography. Initial rate studies (1/V versus 1/[GDP] and 1/[pyruvate]) were performed to obtain kinetic constants for pyruvate kinase and lactate dehydrogenase reactions in 50 mM Hepes, pH 8.0/KOH. The optical assays for these enzymes are described elsewhere (15). Cloning and Mutagenesis of the Phosphate Binding Protein (PBP). The phoS coding region (accession number K01992) encodes PBP and was cloned, using PCR amplification, from a wild-type E. coli Κ-12 genomic DNA library using the following primers: 5′-ATTCCATATGAAAGTTATGCGTACCAC-3′ and 5′-GCGGATCCGTTAGTACAGCGGCTTACC-3′. The NdeI and BamHI restriction sites in the primers (bold font) were used to insert the PBP coding region into the pET-23a(+) expression plasmid. The construct appeared to be lethal in BL21(DE3) but not in strains in which T7 polymerase is either more stringently regulated [e.g., BL21(DE3)pLysS] or absent. The PBP signal sequence was removed, and the A197C mutation needed for fluorescent labeling of the protein was inserted using overlap PCR mutagenesis (16). The mutagenesis primers were (the bold sequences indicate the mutations) 5′-TATGTTGAATATTGTTACGCGAAGCA-3′ and 5′TGCTTCGCGTAACAATATTCAACATA-3′. The PBP coding region was truncated to remove its signal peptide sequence by creating a new 5′-terminus for the gene using the primer 5′-CAGATCCATATGGAAGCAAGCCTGA-3′. The NdeI restriction site (bold) in this primer and the BamHI site discussed above were used to clone the truncated, mutagenized insert into the pET-28a(+) plasmid. This plasmid placed a His tag at the N-terminus of the expressed 1 Abbreviations: EDTA, ethylenediamine-N,N,N′,N′-tetraacetic acid; Hepes, 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid; mant-nucleotides, 3′-O-(N-methylanthraniloyl)-2′-deoxynucleotide; PEI-F, poly(ethylenimine)-cellulose-F; PMSF, phenylmethanesulfonyl fluoride; Tris, tris(hydroxymethyl)aminomethane; unit, micromoles of substrate converted to product per minute at Vmax.
Sukal and Leyh protein that was used for purification. The truncated, mutant construct was not lethal in BL21(DE3) cells, indicating that the lethality may be associated with the N-terminal signal sequence. Purification of A197C PBP. Colonies from an overnight transformation of BL21(DE3) cells were inoculated into LB media containing ampicillin (50 µg/mL) and grown at 37 °C. IPTG was added to 0.5 mM when the cultures reached an OD600 of 0.5-0.6. Four hours later, the cells were harvested and lysed by sonication. The preparation of the cellular extract and His-resin chromatography were performed as recommended by the manufacturer with the exception that PMSF was added to the cell extract to 1.0 mM. The PBP that eluted from the His‚Bind resin was concentrated, and the His‚Bind Wash buffer was exchanged for 10 mM Tris, pH 7.6/HCl and 1.0 mM MgCl2. PBP was then loaded onto a Q-Sepharose fast flow column equilibrated in the same buffer, and a 10 column volume linear gradiant of 0-50 mM NaCl in the same buffer was applied. A single protein peak resulted. NaCl was dialyzed away, and the protein was stored at -70 °C. The protein appeared >95% homogeneous by SDS-PAGE. Approximately 20 mg of pure PBP were obtained per liter of cell culture. Labeling A197C PBP with the Fluorophore N-[2-(1maleimidyl)ethyl]-7-(diethylamino)coumarin-3-carboxamide (MDCC). Efficient labeling of A197C PBP with MDCC requires that trace Pi in the reaction mixture be reduced substantially (17). This is accomplished using purine nucleoside phosphorylase (PNP), which converts Pi and inosine to hypoxanthine and ribose 1-phosphate. Pi is further reduced by the addition of phosphodeoxyribomutase (PDRM), which converts the ribose 1-phosphate to ribose 5-phosphate. PDRM is not commercially available. Our protocols replace PDRM with xanthine oxidase (XO), which is stable and commercially available. XO converts hypoxanthine to xanthine and xanthine to uric acid in what are considered irreversible reactions. The labeling protocol is as follows: PBP (100 µM), inosine (1.0 mM), MgCl2 (1.0 mM), PNP (0.2 unit/mL), XO (0.2 unit/mL), MDCC (500 µM, from a 1.0 mg/mL stock in DMSO), and Hepes, 50 mM, pH 8.0/ KOH. The solution was mixed overnight at 4 ( 2 °C, passed through a 0.2 µm filter to remove particulate MDCC, and loaded onto a Sephadex G-25 column equilibrated in Hepes, 50 mM, pH 8.0/KOH. The labeled protein was pooled and loaded onto a Q-Sepharose fast flow anion-exchange column equilibrated in Hepes, 50 mM, pH 8.0/KOH. The pure protein eluted off the column within 2 column volumes of this buffer; no salt was necessary. The protein was concentrated and dialyzed against the Pi-mop system and stored at -70 °C. Quench-Flow Experiments. The experiments were performed using a KinTek quench-flow apparatus, model RQF3. After quenching, samples were incubated in a boiling water bath for 1 min, and the radiolabeled reactants and products were separated on 10 cm PEI-F TLC sheets using 0.9 M guanidine hydrochloride as the mobile phase (5). The Pi and GTP were quantitated using a phosphoimager. Fluorescence Stopped-Flow Measurements. Successful stopped-flow experiments required that Pi be scrupulously removed from the reagents and the reagent contact surfaces of the instrument. PNP using 7-methylguanosine was used to scavenge Pi (18). Pi was removed from the labeled PBP
GDP and Pi Release in ATP Sulfurylase-GTPase Reaction and ATP sulfurylase by dialyzing the concentrated enzymes against buffer containing PNP (1 unit/mL), 7-methylguanosine (1.0 mM), and MgCl2 (1.0 mM). Buffer was also dialyzed against the mop system by placing PNP inside the dialysis tubing. The fluorescence stopped-flow syringes and sample lines were treated for 30 min with the PNP containing buffer and then rinsed in mop-treated buffer without PNP. PNP was kept at a very low level (0.04 unit/mL) in both syringes. This reduced Pi prior to starting the experiment and was sufficiently low that PNP did not compete with PBP for Pi. An Applied Photophysics SX-18MV stopped-flow spectrometer was used for these experiments. The samples were equilibrated, and the time courses were measured at 25 ( 2 °C. For the PBP-MDCC experiments, the excitation wavelength was 425 nm; light emitted above 455 nm was detected using a cutoff filter. The GDP release and guanylate kinase experiments detected the NADH emission above 400 nm, exciting at 340 nm. In all cases, five to eight scans were averaged. Where possible, the concentration of the nonfluorescent rather that fluorescent species was varied. This strategy improves data quality and simplifies the experimental protocols by allowing the data to be gathered at one instrument setting.
Biochemistry, Vol. 40, No. 49, 2001 15011
FIGURE 1: GDP/Pi formation during the first turnover of ATP sulfurylase. A solution containing ATP sulfurylase [40 µM (b) or 200 µM (O)], AMP (1.0 mM), PPi (100 µM), MgCl2 (2.6 mM), and Hepes (50 mM, pH 8.0/KOH) was mixed rapidly using a quench-flow apparatus with an equal volume of an identical solution that did not contain enzyme but did contain GTP/[γ-32P]GTP (1.0 mM; SA ) 80 µCi/mL). The mixture was aged for a defined period and quenched by rapid mixing with a basic solution of EDTA/ NaOH (100 mM, pH 11.0) (8). The instrument and reagents were equilibrated at 25 ( 2 °C prior to mixing. Radiolabeled reactants were separated using TLC and quantitated (see Materials and Methods). Product formation was determined in triplicate at each time point. The average is shown for the lower concentration experiment (b). The smooth curve passing through the data represents the best fit of the data at 100 µM ATP sulfurylase to an ordered release model (see The Global Fit).
RESULTS AND DISCUSSION A Burst of Product. Rapid-quench experiments were performed to test for the presence of slow steps in the product release portion of the GTPase reaction. The E‚AMP‚PPi complex was mixed rapidly with a saturating concentration of GTP and allowed to react for a defined period before being mixed rapidly with a quench solution (see Materials and Methods). The results reveal a pre-steady-state burst of product, followed by linear steady-state turnover of the enzyme (see Figure 1). To ensure that the burst was linear with enzyme concentration, each time point was performed, in triplicate, at two concentrations of enzyme [20 µΜ (b) and 100 µM (O)]. The signal-to-noise ratio was better in the 100 µM experiments, particularly at the early time points, and these data were used in all subsequent quantitative analyses. The burst amplitude was 0.32 of an enzyme active site, and kcat, 1.22 ( 0.04 s-1, was in good agreement with control steady-state measurements performed at catalytic concentrations of ATP sulfurylase (kcat ) 1.2 s-1). The burst of product confirms that product release and/or an isomerization in the release stage of the reaction is at least partially rate-limiting turnover. A burst of 200 mM (33, 38, 39). For most energy coupling systems that hydrolyze nucleotides, these important constants are, at present, unknown. Mechanism and BehaVior of ATP Sulfurylase. The kinetic scheme for the E‚AMP‚PPi-catalyzed hydrolysis of GTP, and its rate constants, is shown in Figure 6. At 1.0 mM GTP, which is near its presumed physiological concentration (33), GTP binding is fast compared to turnover. The isomerization is partially rate limiting in the forward reaction, while its reverse is sufficiently slow that the system becomes irreversibly committed to completing the catalytic cycle once the isomerization has occurred. β,γ-bond cleavage is rate limited by the isomerization, and the bond-breaking and -forming steps are essentially isoenergetic. To achieve this condition, the enzyme must exhibit a profound, selective stabilization of GTP over GDP‚Pi, compared to solution, that is caused by changes in enzyme-substrate interactions in the vicinity of the β,γ bond. The rate constants for release of Pi and GDP from the E‚GDP‚Pi and E‚GDP‚Pi complexes, respectively, are the slowest steps in the mechanism, and Pi release acts both as a kinetic barrier to prevent reversal of the reaction and as a thermodynamic sink to draw the system through its catalytic cycle. CONCLUSIONS Determining how product partitions between the enzyme active site and solution during the first turnover of ATP sulfurylase has proven a valuable method for characterizing the events that occur in the latter stages of this GTPase reaction. It provides the time dependence of the concentration of each of the product intermediates in the catalytic cycle, the order in which products are released, and the microscopic rate constants for product dissociation from these intermediates. It is also capable of detecting a rate-determining isomerization, if it occurs after product is released. The
Biochemistry, Vol. 40, No. 49, 2001 15015 method, as described here, can be extended to other ATPase/ GTPase systems and is generally applicable to other systems for which appropriate coupled assays exist. The E‚AMP‚PPi-catalyzed GTPase mechanism includes an isomerization that “toggles” the GTPase reaction, a bondbreaking step whose equilibrium constant is near unity, a remarkably unfavorable Pi binding reaction, and two partially rate-limiting product release steps. Slow product release is a characteristic of conformationally coupled systems that hydrolyze nucleotides and is often thought to be a critical energy-linking step (10-14). Given this precedent, it is plausible, if not likely, that this will also be the case in the native reaction catalyzed by ATP sulfurylase. This hypothesis will be tested by determining whether events in the APSforming reaction are concomitant with the product release steps of the GTPase reaction. REFERENCES 1. Robbins, P., and Lipmann, F. (1956) J. Am. Chem. Soc. 78, 6409-6410. 2. Robbins, P. W., and Lipmann, F. (1958) J. Biol. Chem. 233, 686-690. 3. Leyh, T. S. (1993) Crit. ReV. Biochem. Mol. Biol. 28, 515542. 4. Liu, C., Suo, Y., and Leyh, T. S. (1994) Biochemistry 33, 7309-7314. 5. Leyh, T. S., and Suo, Y. (1992) J. Biol. Chem. 267, 542545. 6. Leyh, T. S., Taylor, J. C., and Markham, G. D. (1988) J. Biol. Chem. 263, 2409-2416. 7. Leyh, T. S., Vogt, T. F., and Suo, Y. (1992) J. Biol. Chem. 267, 10405-10410. 8. Wang, R., Liu, C., and Leyh, T. S. (1995) Biochemistry 34, 490-495. 9. Wei, J., and Leyh, T. S. (1998) Biochemistry 37, 1716317169. 10. Jencks, W. P. (1989) Methods Enzymol. 171, 145-164. 11. Hackney, D. D. (1996) Annu. ReV. Physiol. 58, 731-750. 12. Sleep, J. A., Hackney, D. D., and Boyer, P. D. (1980) J. Biol. Chem. 255, 4094-4099. 13. Johnson, K. A. (1985) Annu. ReV. Biophys. Biophys. Chem. 14, 161-188. 14. Gilbert, S. P., and Johnson, K. A. (1994) Biochemistry 33, 1951-1960. 15. Davidson, E. A. (1959) Biochim. Biophys. Acta 33, 238-240. 16. Innis, M. A. (1990) PCR Protocols: A Guide to Methods and Applications, Academic Press, San Diego. 17. Brune, M., Hunter, J. L., Howell, S. A., Martin, S. R., Hazlett, T. L., Corrie, J. E., and Webb, M. R. (1998) Biochemistry 37, 10370-10380. 18. Brune, M., Hunter, J. L., Corrie, J. E., and Webb, M. R. (1994) Biochemistry 33, 8262-8271. 19. Johnson, K. A. (1992) in The Enzymes (Purich, D., Ed.) Vol. XX, pp 1-61, Academic Press, New York. 20. Yang, M., and Leyh, T. S. (1997) Biochemistry 36, 32703277. 21. Wei, J., and Leyh, T. S. (1999) Biochemistry 38, 6311-6316. 22. Jenkins, W. T. (1991) Anal. Biochem. 194, 136-139. 23. Plowman, K. M., and Krall, A. R. (1965) Biochemistry 4, 2809-2814. 24. McClure, W. R. (1969) Biochemistry 8, 2782-2786. 25. Rudolph, F. B., Baugher, B. W., and Beissner, R. S. (1979) Methods Enzymol. 63, 22-42. 26. Cleland, W. W. (1979) Anal. Biochem. 99, 142-145. 27. Easterby, J. S. (1973) Biochim. Biophys. Acta 293, 552-558. 28. Agarwal, K. C., Miech, R. P., and Parks, R. E., Jr. (1978) Methods Enzymol. 51, 483-490. 29. Lionne, C., Brune, M., Webb, M. R., Travers, F., and Barman, T. (1995) FEBS Lett. 364, 59-62.
15016 Biochemistry, Vol. 40, No. 49, 2001 30. Nixon, A. E., Brune, M., Lowe, P. N., and Webb, M. R. (1995) Biochemistry 34, 15592-15598. 31. Gilbert, S. P., Webb, M. R., Brune, M., and Johnson, K. A. (1995) Nature 373, 671-676. 32. Liu, C., Martin, E., and Leyh, T. S. (1994) Biochemistry 33, 2042-2047. 33. Liu, C., Wang, R., Varlamova, O., and Leyh, T. S. (1998) Biochemistry 37, 3886-3892. 34. Cleland, W. W. (1975) Biochemistry 14, 3220-3224.
Sukal and Leyh 35. Mendes, P., and Kell, D. (1998) Bioinformatics 14, 869-883. 36. Mendes, P. (1997) Trends Biochem. Sci. 22, 361-363. 37. Mendes, P. (1993) Comput. Appl. Biosci. 9, 563-571. 38. Amory, A., Goffeau, A., McIntosh, D. B., and Boyer, P. D. (1982) J. Biol. Chem. 257, 12509-12516. 39. Cardon, J. W., and Boyer, P. D. (1978) Eur. J. Biochem. 92, 443-448. BI015735A
