Regulatory Properties of an Inorganic Pyrophosphatase

Transcription

1 Proc. Nat. Acad. Sci. USA Vol. 68, No. 4, pp , April 1971 Regulatory Properties of an Inorganic Pyrophosphatase from the Photosynthetic Bacterium Rhodospirillum rubrum JOBST-HINRICH KLMM* AND HOWARD GST Department of Microbiology, Indiana University, Bloomington, Ind Communicated by Martin D. Kamen, January 18, 1971 ABSTRACT In Rhodospirillum rubrum, inorganic pyrophosphatase activity is observed in both the cytoplasmic and membrane fractions. The soluble enzyme accounts for about 80% of the total activity in crude extracts, and is the subject of this report. Zn2+ is required for both activity and stability of the enzyme, which has a molecular weight of approximately 90,000 (gel-filtration determinations). The substrate is MgP2072-, and free pyrophosphate (PO74-) is a strong inhibitor. Kinetic experiments indicate homotropic interactions between substrate-binding sites; these interactions are influenced by Mg2+, which is an activator. At low concentrations of Zn2+, the pyrophosphatase is inhibited by NADH, NADPH, and MgATP; 50% inhibition occurs at mm. These effects are reversed by high concentrations of Zn2+ ( M). The nucleotides appear to inhibit activity of the "native" enzyme through an effect on Zn2+ binding. The R. rubrum enzyme seems to be the first known example of a bacterial inorganic pyrophosphatase subject to allosteric regulation. Inorganic pyrophosphate (PP) is produced in various reversible ATP-dependent reactions, which presumably are "pulled" in the biosynthetic direction in vivo by enzymatic hydrolysis of PP to inorganic phosphate (Pi) (1). This hydrolysis, catalyzed by inorganic pyrophosphatase (PPase), also replenishes Pi to the energy-converting systems, whatever they may be. Accordingly, PPase plays an important role in the "phosphate cycle" of living cells, and it seems very likely that the activity of such enzymes must be controlled in some way by chemical species involved in both energy metabolism and biosynthesis. Judging from our expanding knowledge of regulation of other kinds of enzymes, it is also reasonable to anticipate that several alternative control schemes will be found in the wide spectrum of existing organisms. In this communication, we report on the regulatory properties of a pyrophosphatase present in the purple bacterium Rhodospirillum rubrum. The enzyme shows allosteric kinetic behavior and, under certain conditions, its activity is sensitive to NADH, NADPH, and ATP. MATRIALS AND MTHODS Growth of bacteria R. rubrum strain S1 was grown photoheterotrophically at 34 C with saturating light intensity (approximately 550 Abbreviations: MgADP, Mg complex of ADP; MgATP, Mg complex of ATP; MgPP, MgP2O72-; Mg2PP, Mg2P207; PPase, inorganic pyrophosphatase (C ); V., relative elution volume; VO, void volume. * On leave from the Institut fur Mikrobiologie der Universitat Gottingen (Fed. Rep. of Germany); recipient of an Ausbildungsstipendium from the Deutsche Forschungsgemeinschaft. 721 foot-candles) in a synthetic medium containing 0.4% DLmalate as the carbon source and 0.1% ammonium sulfate as the nitrogen source (2). Preparation of cell-free extracts and particles Cells were harvested at the end of the logarithmic growth - phase, resuspended in 50 mm Tris HC1 buffer (ph 7.6), and disrupted in a French pressure cell. The crude extract was centrifuged (4 C) at 15,000 X g for 15 min to remove residual intact cells and large cell fragments, and then at 140,000 X g for 90 min to separate the "chromatophores" from the soluble cell extract. Protein concentrations in the soluble extract were in the range of 8-12 mg/ml. For experiments on localization of PPase activity, the chromatophore fraction was washed once with the buffer noted. PPase assay PPase activity was assayed by colorimetric determination (3) of the Pi produced from the enzymatic hydrolysis of PP; activities are expressed in terms of umol of Pi produced per min. Reaction mixtures (1.0-ml final volume) contained 40 mm Tris HCl buffer (ph 9.0), mm Na4PP, mm MgCl2, and an appropriate amount of the soluble enzyme preparation. The reaction was initiated by addition of enzyme or, in experiments involving preincubation of enzyme with nucleotides, by addition of PP to the PP-free reaction mixture. Assay mixtures were incubated for 5-10 min at 37 C; PPase activity was terminated by the addition of 1 ml of molybdate reagent (50 g ammonium molybdate tetrahydrate in 1000 ml of 4 N H2S04). Since the protein concentrations used were relatively low (25-60 ug/ml), visible precipitation did not occur upon addition of the molybdate reagent. Subsequently, 1 ml of lon reagent (1% p-methylaminophenol sulfate in 3% NaHSO3) was added to each mixture and, after 10 min, absorbances at 660 nm were measured (for the blank, molybdate reagent was added immediately after the addition of enzyme). In assays of chromatophore PPase, the reaction was terminated by adding 1 ml of 10% trichloroacetic acid, the precipitated protein was removed by centrifugation, and Pi was estimated in 1 ml of the supernatant fluid. With the assay conditions used, reaction rates are linear with time (0-15 min) and protein concentration (0-100 ;g/ml). ADP, ATP, NADH, NADPH, cytochrome c, type III (horse heart), and lactate dehydrogenase, type II (rabbit muscle) were purchased from Sigma Chemical Co., St. Louis. Sephadex G-200 was obtained from Pharmacia Fine Chemicals Inc., Piscataway, N.J. All other chemicals used were reagent grade.

2 722 Biochemistry: Klernme and Gest FIG. 1. total [MgCI2](mM) Activity of the R. rubrum soluble PPase as a function of MgCl2 and PP concentrations. 30 Ayg of extract protein was used for each assay. stimation of molecular weight of the soluble PPase by gel filtration on Sephadex G-200 The soluble PPase was partially purified from a 140,000 X g supernatant fluid, supplemented with 0.2 mm ZnSO4, by first removing proteins denatured during 10 min of incubation at 600C, followed by ammonium sulfate fractionation (the PPase precipitates at 40-60% saturation). These two steps increased the specific activity about 10-fold, to 11 MUmol Pi/min per mg protein. 2 ml of this PPase preparation (6 mg protein) was applied to a Sephadex G-200 column (2 X 45 cm), together with 0.6 mg of lactate dehydrogenase (LDH) and 4 mg of horse-heart cytochrome c as protein standards. The column had previously been equilibrated with 50 mm Tris HCl buffer (ph 7.6), containing 0.2 mm ZnSO4, and was developed with the same buffer at 4VC. Fractions of 5 ml were collected and assayed for PPase and LDH activity. Cytochrome c was determined by measuring absorbance at 420 nm. The ratio Ve/Vo (V0 was estimated using blue dextran as a marker) was used as the index of molecular weight. Calculation of concentrations of Mg2 + and pyrophosphate species In solutions containing MgCl2 and PP, the following relevant molecular and ionic species occur at ph 9: Mg2+, P2074-, MgP2072-, and MgP207. For calculation of their concentrations, the stability constants for MgP2072- (MgPP) and Mg2P207 (Mg2PP) given by Lambert and Watters (4) were used as described below. /CMaPp = ~~[MgPP] [Mg2+] X [P2074-] (1) kmg~pp = [Mg2PP] = k~~p=[mg2+] X [Mgpp] It was assumed that equilibria (1) and (2) can be calculated independently, and values derived from (1) were employed for the estimation of equilibrium concentrations using (2). (2) Proc. Nat. Acad. Sci. USA 68 (1971) The following related equations can also be written: [Mg"+]* = [MgCl2hotai - [MgPP]* (3) [P2074-] = [PP]tt.l- [MgPP]* (4) -where asterisks apply to data from equilibrium equation (1). By inserting (3) and (4) in (1), a quadratic equation with [MgPP]* as the unknown results. One of the two solutions is greater than the total MgCl2 concentration and is thus rejected. The solution of this first quadratic equation will be designated as [MgPP]*. In connection with equilibrium (2), the following equations may be written, and used for determination of the Mg2+ and MgPP concentrations in assay mixtures. [Mg2+] = [Mg"+]* - [Mg2PP] (6) [MgPP] = [MgPP]* - [Mg2PP] (6) By inserting (6) and (6) in (2), another quadratic equation is obtained, which can be solved for the unknown, [Mg2PP]; solution of this equation yielded equilibrium concentrations of Mg2PP. Concentrations of the other ionic species were then calculated using equations (3)-(6). RSULTS Distribution of PPase activity in cell-free fractions With the assay conditions described (MgCl2, 1 mm; PP, 1 mm; ph 9), approximately 15-20% of the total PPase activity present in crude cell-free extracts (15,000 X g supernatant fluid) appears to be associated with the pigmented chromatophore fraction [membrane-bound PPase in R. rubrum has previously been reported by Baltscheffsky et al. (5)]. It should be noted, however, that the apparent distribution of the activity may be influenced by the ph used for assay, because the ph optima of the soluble and particulate PPases differ. Optimal activity of the soluble enzyme is observed at ph 9 (in Tris * HCl), while the optimum of the membrane-bound PPase is at 8. The experiments described were all conducted with the soluble enzyme. Preliminary experiments on the role of Zn2+ In the crude cell-free extracts, the enzyme activity was found to be very labile. As much as 70% of the original activity 0.6 free[p204 (mm) FIG. 2. Inhibition of PPase activity by free P pg of extract protein was used for each assay. In B, data from A are plotted as 1 /V vs. [P

3 Proc. Nat. Acad. Sci. USA 68 (1971) was lost during 24 hr of storage at 40C. Since addition of 0.5 mm DTA markedly accelerated the inactivation (half time, less than 10 min), a metal requirement for the maintenance of catalytic activity was suggested. To explore this point, the PPase in an extract was inactivated by incubation with 0.2 mm DTA, the metal chelates were removed by passage of the extract through a column of Sephadex G-25, and various metals were tested for reactivation capability. Among 13 metals tested, at a concentration of 0.1 mm (Na+, K+, and the divalent cations of Ca, Mg, Sr, Ba, Mn, Co, Sn, Zn, Fe, Cu, and Ni) only Zn2+----either as the sulfate or chloride-was capable of reactivating and stabilizing the PPase activity. Reactivation in such experiments was observed with Zn2+ concentrations of M. After Zn2+ was shown to be essential for maintenance of catalytic activity, 0.1 mm ZnSO4 was routinely added to the crude cell-free extracts. It should be noted that, in addition to Zn2+, Mg2+ is required for activity. In common with PPases from a number of other organisms (e.g., see refs. 6-8), the actual substrate of the R. rubrum cytoplasmic enzyme is MgPP. Requirements for maximal activity Activity of the soluble R. rubrum PPase is absolutely dependent on the presence of substrate quantities of a divalent metal. MgCl2 supports optimal activity, while the chlorides of Zn2+, Co2+, and Mn2+ are about one sixth as effective. To determine the Mg2+ requirement in the assay system, reaction rates were measured at fixed concentrations of total PP (0.5, 1.0, and 2.0 mm) and varying concentrations of MgCl2 (Fig. 1). Several characteristics of the curves obtained should be noted: (a) no activity is observed in the absence of MgCl2; (b) the curves are strongly sigmoidal, and increasing the total concentration of PP causes a shift to the right; and (c) maximal reaction rates are seen with MgCl2/PP ratios greater than 1, indicating that Mg2+ is an activator. The sigmoidal character of the curves in Fig. [P2 04 ]/ [Mg P2 07] FIG. 3. Inhibition of PPase activity by free P2074-, expressed as a function of the ratio [P2074-] /[MgPP]. Reaction rates were measured with fixed concentrations of MgPP and varying concentrations of the inhibitor P When the results were plotted according to Fig. 2B, the vm,, values (i.e., for [P2074-] 0) could = be estimated and, thus, percentage inhibition calculated. MgPP concentrations used: 0.28 (0), 0.39 (A), and 0.89 mm (0). R. rubrum Pyrophosphatase Regulation 723 -$ mM ll, 1 O mm mM 0.7M I/[MgPp] (mm-) I/Q1gPP)2 (mm-2) FIG. 4. Lineweaver-Burk type plots of PP inhibition data. 30 jg of extract protein was used for each assay. A. As indicated, three different concentrations of free PP were used. B. Data from A, plotted as 1/V vs. 1/[MgPP] 2. 1 does not necessarily indicate allosteric behavior of the enzyme. In this instance, the sigmoid shape seems to be a consequence of the circumstances that the true substrate is MgPP, and free PP is an inhibitor. Inhibition of PPase activity by free P207'- In designing experiments to characterize the inhibitory action of free PP, we carefully prepared assay mixtures, using the calculations described in Methods, so as to avoid the presence of free Mg2+ (which, as shown in Fig. 1, is an activator). Fig. 2A shows the effect of uncomplexed PP (i.e., Na4P207) on the PPase activity at several fixed concentrations of MgPP. Obviously, uncomplexed PP is a potent inhibitor of the enzyme; it is also clear that with a given free PP concentration, reaction velocity is increased as the concentration of substrate, MgPP, increases. In order to test the effect of Mg2PP on activity, reaction rates were measured at fixed concentrations of MgPP (0.15, 0.52, 0.82 mm) with various concentrations of Mg2PP ( , , and mm, respectively). With the MgCl2 and PP combinations used in these experiments, sufficiently high concentrations of free Mg2+ were present to ensure maximal activation by this ionic species. It was found that Mg2PP did not influence the PPase activity, possibly because of a very low affinity of the enzyme for this complex. When the data of Fig. 2A were plotted according to Dixon (9) (1/V vs. i), in order to estimate the inhibitor constant K, for free PP, a set of paraboloid-like curves was obtained. However, a plot of 1/V against j2 gave straight lines (Fig. 2B), which intersected at a common point. The distance between this point and the 1/V axis equals K 2, from which it follows that the K, for free PP is 5 X 10- M. Treatment of the experimental results as in Fig. 2B permits estimation of Vmax values (free PP = 0), and in Fig. 3 the data are replotted to show the % inhibition of maximal activity by free PP as a function of the ratio of inhibitor to substrate concentrations (i/s). Fig. 4A presents a Lineweaver-Burk plot of kinetic data obtained with fixed concentrations of the inhibitor P2074-

4 724 Biochemistry: Klemme and GestP Proc. Nat. Acad. Sci. USA 68 (1971) C.) C a 0o [Mg P2 0-o] (mm) FiG. 5. Substrate-saturation curves. A. In the absence of appreciable amounts of free Mg'+ (MgCl2/PP = 1); 54 ug of extract protein was used in each assay. B. "Theoretical" curve obtained from curve A data, by correcting for inhibition due to free PP using Fig. 3. C. In the presence of free Mg2 + (MgCl2/PP 4); 27 ug of extract protein was used in each assay..-_ 0 -a x So -II ol I MgATP-,..b,& -- - B NADH and various concentrations of substrate (MgPP). Normal Michaelis-Menten kinetics are not obeyed, but when 1/V is plotted against 1/s2, straight lines are obtained that intersect at a common point beyond the 1/V axis (see Fig. 4B); this signifies "mixed type" inhibition of the PPase by uncomplexed PP. Reaction kinetics with different MgCI2/PP ratios Reaction rates were measured at two fixed ratios, namely, MgCl2/PP = 1 and 4 (see Fig. 5). With a ratio of 1 -that is, at very low concentration of free Mg2+-the substrate saturation curve has a distinctly sigmoid shape (curve A). The corresponding Lineweaver-Burk plot is not a straight line, but rather is paraboloid (Fig. 6A, curve a). When the MgCl2/PP ratio is 4 (thus, with excess free Mg2+, but virtually no free PP), the reaction rates are obviously higher than A B bb I/Ngi Pzo073 (mm-') 07;] (mm2) FIG. 6. Lineweaver-Burk type representations of data from Fig. 5. A. 1/Vvs. 1/8 for: (a) MgCl2/PP = 1; (b) as in (a),but data corrected for inhibition by free PP; (c) MgCl2/PP = 4. B. 1/V vs. 1/82 for: (a) MgCl2/PP 1; (b) = as in (a), but data corrected for inhibition by free PP. 0 50,M ZnSO4 FIG. 7. Inhibition of the soluble R. rubrum PPase by nucleotides. In each assay, 60 Mg of extract protein was incubated for 5 min at 370C with MgCl2 and nucleotides (also ZnSO4 in B); ph 8.2. With NAD (P)H, 2 mm MgC12 was used, with ADP and ATP, 16 mm MgCl2. The reaction was then initiated by adding 1 mm Na4PP. 100% activity corresponds to 2.3 Amol of Pi/min per mg of protein. A. ffect of nucleotide concentration; in this experiment, the Zn2+ concentration was only 4 pm. B. Dependence of nucleotide effects (1 mm NADH or 2 mm ATP) on Zn2+ concentration. with the ratio of 1, and the substrate saturation curve does not appear to be sigmoidal (Fig. 5, curve C); moreover, these results give a straight line in the Lineweaver-Burk diagram (Fig. 6A, c). Sigmoidicity of the substrate saturation curve with a MgCl2/PP ratio of 1 could result from either inhibition of PPase action by uncomplexed PP or homotropic interactions among several binding sites for the substrate, MgPP. In this connection, we now note that from the equilibria discussed in Methods, it can be calculated that, as the concentration of total PP is decreased from 1.0 to 0.05 mm in solutions with a fixed MgCI2/PP ratio of 1, the ratio between the inhibitor P2074- and substrate MgPP (designated as i/8) increases from 0.08 to As shown in Fig. 3, increase of i/s over this range leads to progressively greater inhibition of enzyme activity. When the experimental data from curve A of Fig. 5 are corrected for inhibition of activity by free P2074- (using Fig. 3), curve B of Fig. 5 results. It should be noted that this correction of the data eliminates the sigmoid character of the substrate-saturation curve (and also clearly shows that Mg2+ is an activator; see curves B and C, Fig. 5). Nevertheless, the corrected values still do not obey ordinary Michaelis-Menten kinetics, as is seen in Fig. 6A, curve b. On the other hand, a straight line is obtained in a plot of 1/V versus 1/82 (Fig. 6B, curve b); note that the uncorrected values give a relatively poor fit to a straight line 100

5 Proc. Nat. Acad. Sci. USA 68 (1971) (curve a). The Km values for MgPP are 0.06 mm for the "theoretical" situation with MgCl2/PP = 1 (i.e., corrected for inhibition due to free PP), and 0.03 mm for MgCl2/PP = 4; these values are 6-tolO-fold higher than the affinityfor the inhibitor P2074-(K1 = mm). The foregoing is interpreted as indicating homotropic interactions among substrate-binding sites. Accordingly, the data from Fig. 5 were plotted using the Hill relation [log V/(V.. - V)/ = n log s - log K], which yielded curves having markedly different slopes in their linear portions. These were: for curve A, n = 1.8; curve B, 1.5, and curve C, 1.1. Reaction rates were also measured at the fixed MgPP concentrations of 0.09 and 0.45 mm, with various concentrations of free Mg2+ ( mm, and mm, respectively). Hill plots of these data showed slopes of n = 2.1 and 1.2, respectively. The significance of the Hill coefficients is considered in the Discussmon. Reversible inhibition of the soluble PPase by nucleotides A number of nucleotides were tested for possible effects on the enzyme, and, as shown in Fig. 7A, NADH, NADPH, and MgATP were found to be inhibitory at relatively low concentration. In contrast, MgADP inhibited only slightly, while (not shown in Fig. 7) AMP and the oxidized forms of the pyridine nucleotides did not influence activity. The inhibitions by NADH and ATP were studied in particular, and it was established that these nucleotides do not disappear in a detectable amount during the "inactivation" process. It is of special interest that the extent of the inhibitions is related to the Zn2+ concentration. From Fig. 7B, it is evident that as Zn2+ concentration is increased, inhibition progressively decreases, until at 100,uM Zn2+ full activity is observed despite the presence of NADH or MgATP. The following experiment demonstrated the reversible nature of the nucleotide inhibition (inactivation) in another fashion. An enzyme preparation was passed through a column of Sephadex G-25 in order to remove excess Zn2+. Then, the enzyme solution was incubated with 1 mm NADH or 2 mm MgATP for 10 min at 37 C. At this point, the enzyme was found to be almost completely inactivated. The extracts were again passed through a fresh column of Sephadex G-25, now for the purpose of removing the nucleotides. Such removal, in itself, does not restore PPase activity. Activity in such experiments, however, is restored by subsequent incubation with M Zn2+ (sulfate or chloride). These observations suggest that the nucleotides may act by somehow displacing Zn2+, which must be properly associated with the protein to maintain a catalytically active enzyme complex. Molecular weight of the PPase The molecular weight of the enzyme was determined by Sephadex G-200 gel filtration, in the presence of 0.2 mm ZnSO4, as described in Methods. The enzyme was eluted at a position corresponding to a Ve/Vo ratio of 1.63, indicating a molecular weight of about 90,000. This is considerably greater than values reported for other bacterial PPases (10-12), except for the scherichia coli enzyme, which has a molecular weight of 118,000 (13). DISCUSSION The soluble inorganic PPase or R. rubrum shows unexpectedly complex properties which, however, are consistent with the R. rubrum Pyrophosphatase Regulation 725 notion that its activity must be closely regulated in vivo by specific low molecular weight effectors. Although the PPase is Zn2+-dependent, a second metal, Mg2+, is involved in catalytic activity in at least two ways-as a component of the substrate, MgPP, and as an activator; excess Mg2+ also can stimulate activity by complexing free P207 -, which is a potent inhibitor. In the absence of appreciable free Mg2+, allosteric behavior is observed, i.e., the kinetic analysis indicates cooperative (homotropic) effects between substratebinding sites (Hill coefficient n = 1.5, from data corrected for inhibition by free PP). When free Mg2+ is present, reaction rates are increased, normal Michaelis-Menten kinetics are observed, and the cooperative interactions between substrate-binding sites are decreased (Hill coefficient = 1.1 when the MgCl2/PP ratio is 4). xcess Mg2+ evidently increases the activity by lowering the Km and increasing Vm. The R. rubrum enzyme shows a number of kinetic similarities with the PPase of mouse-liver cytoplasm (7), but seems to be quite different in this regard from the other bacterial PPases described. In respect to in vivo control possibilities, the effects of the nucleotides are particularly interesting. The reduced pyridine nucleotides and ATP may be regarded as the ultimate products of the energy conversion system and, as such, could logically act as regulatory feedback signals for the inhibition of PPase activity. In this connection, we suggest that PPase can be thought of as a "biosynthetic" enzyme in view of its role in "pulling" biosynthetic reactions, and in that it presumably facilitates ATP synthesis through replenishing Pi to the energy-conversion apparatus. The opposing effects of free Mg2+ and ATP on the PPase could conceivably be interrelated, since ATP strongly chelates the metal. Thus, as net ATP synthesis increases in vivo, the concentration of free Mg2+ would be expected to decrease, at least locally, and vice versa-accordingly, the Mg2+ and ATP effects on enzyme activity could augment each other. We thank Carol Stahl and llen Kauffman for expert technical assistance. This research was supported by Grant GB 7333X from the National Science Foundation. 1. Kornberg, A., in Horizons in Biochemistry, ed. M. Kasha and B. Pullman (Academic Press, New York, 1962), p Ormerod, J. G., K. S. Ormerod, and H. Gest, Arch. Biochem. Biophys., 94, 449 (1961). 3. Fiske, C. H., and Y. SubbaRow, J. Biol. Chem., 66, 375 (1925). 4. Lambert, S. M., and J. I. Watters, J. Amer. Chem. Soc., 79, 5606 (1957). 5. Baltscheffsky, M., H. Baltscheffsky, and L.-V. von Stedingk, in "nergy conversion by the Photosynthetic Apparatus". Brookhaven Symposia on Biology, (1966), Vol. 19, p Josse, J., J. Biol. Chem., 241, 1948 (1966). 7. Horn, A., H. Bornig, and G. Thiele, ur. J. Biochem., 2, 243 (1967). 8. Nayudu, P. R. V., and P. L. Miles, Biochem. J., 115, 29 (1969). 9. Dixon, M., Biochem. J., 55, 170 (1953). 10. Tono, H., and A. Kornberg, J. Biol. Chem., 242, 2375 (1967). 11. Tono, H., and A. Kornberg, J. Bacteriol., 93, 1819 (1967). 12. Ware, D., and J. R. Postgate, Nature, 226, 1250 (1970). 13. Wong, S. C. K., D. C. Hall, and J. Josse, J. Biol. Chem., 245, 4335 (1970).