Mutational analysis of the interaction between active site residues and the loop region in mammalian purple acid phosphatases.

Mammalian purple acid phosphatases (PAPs) can be divided into two groups, which exhibit distinct spectroscopic and kinetics properties: PAPs that consist of a single 36 kDa polypeptide, and PAPs that have undergone limited proteolysis to give two fragments with masses of 16 and 20 kDa, respectively. Proteolysis results in an increase in enzymatic activity, an increase in the optimal pH for activity, and a change in the g(z)() value of the characteristic EPR spectrum of the mixed-valence binuclear iron center. It has been proposed that these changes are due to the loss of interactions between Asp146 in an exposed loop region and active site residues upon proteolysis. In the present study, site-directed mutagenesis of Asp146 in recombinant rat bone PAP (recRPAP) has confirmed this hypothesis. Conversion of Asp146 into Ala, which eliminates the interaction of the side chain with the active site, resulted in an enzyme with properties typical of PAPs isolated in proteolytically cleaved forms. The Asp146Asn and Asp146Glu mutants were also prepared and examined to assess the effects of altered electrostatic interactions and side-chain length. Limited proteolysis of all three mutant enzymes with cathepsin L resulted in a significant increase in catalytic activity. Thus, although the interaction between Asp146 and (an) active site residue(s) is the major factor responsible for the low catalytic activity of uncleaved PAPs, other interactions are also important. Since both p-nitrophenyl phosphate and osteopontin, a potential in vivo substrate, show the same level of activation, the observed increase in catalytic activity upon proteolysis is likely to be due to electrostatic rather than steric effects. EPR spectra of FeZn-recRPAP before and after cleavage by cathepsin L suggest that cleavage primarily affects the divalent metal site. The observation that pK(es,1) is also sensitive to changes at the divalent site is consistent with the proposal that the nucleophilic hydroxide is that bridging the divalent and trivalent metals.

The highly exposed loop region in mammalian purple acid phosphatase controls the catalytic activity.

Recombinant human purple acid phosphatase (recHPAP) provides a convenient experimental system for assessing the relationship between molecular structure and enzymatic activity in mammalian purple acid phosphatases (PAPs). recHPAP is a monomeric protein with properties similar to those of uteroferrin (Uf) and other PAPs isolated as single polypeptide chains, but its properties differ significantly from those of bovine spleen PAP (BSPAP) and other PAPs isolated as proteolytically "clipped" forms. Incubation of recHPAP with trypsin results in proteolytic cleavage in an exposed region near the active site. The product is a tightly associated two-subunit protein whose collective spectroscopic and kinetics properties resemble those of BSPAP. These results demonstrate that the differences in spectroscopic and kinetics properties previously reported for mammalian PAPs are the result of proteolytic cleavage. Mass spectrometry shows that a three-residue segment, D-V-K, within the loop region is excised by trypsin. This finding suggests that important interactions between residues in the excised loop and one or more of the groups that participate in catalysis are lost or altered upon proteolytic cleavage. Analysis of available structural data indicates that the most important such interaction is that between Asp 146 in the exposed loop and active-site residues Asn 91 and His 92. Loss of this interaction should result in both an increase in the Lewis acidity of the Fe(II) ion and an increase in the nucleophilicity of the Fe(III)-bound hydroxide ion. Proteolytic cleavage thus constitutes a potential physiological mechanism for regulating the activity of PAP in vivo.

Evidence for nonbridged coordination of p-nitrophenyl phosphate to the dinuclear Fe(III)-M(II) center in bovine spleen purple acid phosphatase during enzymatic turnover.

The pH dependence of the catalytic parameters k(cat) and K(M) has been determined for the Fe(III)Fe(II)- and Fe(III)Zn(II)-forms of bovine spleen purple acid phosphatase (BSPAP). The parameter k(cat) was found to be maximal at pH 6.3, and a pK(a) of 5.4-5.5 was obtained for the acidic limb of the k(cat) vs pH profile. Two different EPR spectra were detected for the phosphate complex of the mixed-valent diiron enzyme; their relative amounts depended on the pH, with an apparent pK(a) of 6. The EPR spectra of Fe(III)Fe(II)-BSPAP.PO(4) and Fe(III)Zn(II)-BSPAP.PO(4) at pH 5.0 are similar to those previously reported for Fe(III)Fe(II)-Uf.PO(4) and Fe(III)Zn(II)-Uf.PO(4) complexes at pH 5.0. At higher pH, a new Fe(III)Fe(II)-BSPAP.PO(4) species is formed, with apparent g-values of 1.94, 1.71, and 1.50. The EPR spectrum of Fe(III)Zn(II)-BSPAP does not show significant changes upon addition of phosphate up to 30 mM at pH 6.5, suggesting that phosphate binds only to the spectroscopically silent Zn(II). To determine whether the phosphate complexes were good structural models for the enzyme substrate complexes, these complexes were studied using rapid-freeze EPR and stopped-flow optical spectroscopy. The stopped-flow studies showed the absence of burst kinetics at pH 7.0, which indicates that substrate hydrolysis is rate limiting, rather than phosphate release. The EPR spectrum of Fe(III)Fe(II)-BSPAP.p-NPP is similar, but not identical, to that of the corresponding phosphate complex, both at pH 5 and pH 6.5. We propose that both phosphate and p-NPP bridge the two metal ions at low pH. At higher pH where the enzyme is optimally active, we propose that hydroxide competes with phosphate and p-NPP for coordination to Fe(III) and that both phosphate and p-NPP coordinate only to the divalent metal ion.

Fluoride inhibition of bovine spleen purple acid phosphatase: characterization of a ternary enzyme-phosphate-fluoride complex as a model for the active enzyme-substrate-hydroxide complex.

Purple acid phosphatases (PAPs) employ a dinuclear Fe(3+)Fe(2+) or Fe(3+)Zn(2+) center to catalyze the hydrolysis of phosphate monoesters. The interaction of fluoride with bovine spleen purple acid phosphatase (BSPAP) has been studied using a combination of steady-state kinetics and spectroscopic methods. For FeZn-BSPAP, the nature of the inhibition changes from noncompetitive at pH 6.5 (K(i(comp)) approximately K(i(uncomp)) approximately 2 mM) to uncompetitive at pH 5.0 (K(i(uncomp)) = 0.2 mM). The inhibition constant for AlZn-BSPAP at pH 5.0 (K(i) = 3 microM) is approximately 50-70-fold lower than that observed for both FeZn-BSAP and GaZn-BSPAP, suggesting that fluoride binds to the trivalent metal. Fluoride binding to the enzyme-substrate complex was found to be remarkably slow; hence, the kinetics of fluoride binding were studied in some detail for FeZn-, AlZn-, and FeFe-BSPAP at pH 5.0 and for FeZn-BSPAP at pH 6.5. Since the enzyme kinetics studies indicated the formation of a ternary enzyme-substrate-fluoride complex, the binding of fluoride to FeZn-BSPAP was studied using optical and EPR spectroscopies, both in the presence and absence of phosphate. The characteristic optical and EPR spectra of FeZn-BSPAP. F and FeZn-BSPAP.PO(4).F are similar at pH 5.0 and pH 6.5, indicating the formation of similar fluoride complexes at both pHs. A structural model for the ternary enzyme-(substrate/phosphate)-fluoride complexes is proposed that can explain the results from both the spectroscopic and the enzyme kinetics experiments. In this model, fluoride binds to the trivalent metal replacing the water/hydroxide ligand that is essential for the hydrolysis reaction to take place, while phosphate or the phosphate ester coordinates to the divalent metal ion.

The activity of oxidized bovine spleen purple acid phosphatase is due to an Fe(III)Zn(II) 'impurity'.

Bovine spleen purple acid phosphatase (BSPAP) is a dinuclear iron protein with two stable redox states. The Fe3+Fe2+ state is the active state, while the fully oxidized protein (BSPAPox) has been reported to retain 5-10% activity, corresponding to a kcat of ca. 150 s-1 [Dietrich, M., Münstermann, D., Suerbaum, H., and Witzel, H. (1991) Eur. J. Biochem. 199, 105-113]. Here we show that this activity does not originate from Fe3+Fe3+-BSPAP, but rather from an 'impurity' of FeZn-BSPAP. The FeZn form of BSPAP was prepared from apo-BSPAP following a new procedure, and its kinetic properties were carefully determined for comparison to those of BSPAPox. For the hydrolysis of p-NPP at pH 6.00, both kcat and KM were affected by the Fe2+-to-Zn2+-substitution [Fe3+Fe2+-BSPAP, kcat = (1.8 +/- 0.1) x 10(3) s-1 and KM = 1.2 +/- 0.2 mM; Fe3+Zn2+-BSPAP; kcat = (2.8 +/- 0.2) x 10(3) s-1 and KM = 3.3 +/- 0.4 mM]. The KM of BSPAPox was the same as that of FeZn-BSPAP. pH profiles of BSPAPox and FeZn-BSPAP were both shifted to lower pH compared to that of BSPAPred. FeZn-BSPAP, FeZn-BSPAP.PO4, and FeZn-BSPAP.MoO4 all showed characteristic EPR spectra similar to the corresponding complexes of FeZn-Uf. The same species could also be observed in concentrated samples of native BSPAP. Spin integration of these spectra showed a quantitative relation between the spin concentration of the FeZn-BSPAP 'impurity' and the residual phosphatase activity after oxidation. Since all activity found after oxidation of BSPAP could be attributed to FeZn-BSPAP, there is no direct evidence that Fe3+Fe3+-BSPAP is catalytically active. These results set an upper limit to the possible catalytic activity of the Fe3+Fe3+ form of

Ga3+ as a functional substitute for Fe3+: preparation and characterization of the Ga3+Fe2+ and Ga3+Zn2+ forms of bovine spleen purple acid phosphatase.

A general method has been developed that allows the specific substitution of both iron atoms in the enzyme bovine spleen purple acid phosphatase (BSPAP), which possesses a dinuclear iron center at the active site. The approach is demonstrated by the preparation and characterization (atomic absorption spectrometry, enzyme kinetics, optical spectroscopy, and electron paramagnetic resonance spectroscopy) of two metal-substituted forms in which the ferric iron has been replaced by Ga3+: Ga3+Fe2+-BSPAP and Ga3+Zn2+-BSPAP. Both forms are colorless but exhibit enzymatic activity comparable to that of the native Fe3+Fe2+-BSPAP. Small but consistent changes in kinetics parameters and pH profiles were detected both upon substitution of Fe3+ by Ga3+ and upon substitution of Fe2+ by Zn2+. These results constitute the first evidence that the diamagnetic Ga3+ ion can serve as a functional analogue of Fe3+ in an enzyme, and suggest a novel approach for the study of the role of Fe3+ in other iron enzymes.

Purification and properties of the native form of the purple acid phosphatase from bovine spleen.

The purple acid phosphatase (PAP) from bovine spleen has been shown to exist as a single ca. 36-kDa polypeptide in intact spleen tissue. The previously isolated microheterogeneous complex of 15-kDa and 23- or 21-kDa subunits appears to arise from proteolytic cleavage of an exposed, highly variable loop in the polypeptide chain. Small amounts of a single polypeptide form, presumed to be the native form of the enzyme, have been obtained; this has permitted its optical and EPR spectra and fundamental kinetic properties to be determined. The most notable difference between the native and two-subunit forms of PAP is a ca. 3-fold higher enzymatic activity for the latter, which is due to a simple increase in Vmax. The two forms are very similar spectroscopically and chemically and appear to differ only in the loss of a highly antigenic ca. five amino acid segment of the polypeptide between positions 155 and 160 but not in NH2-terminal sequence or in carbohydrate content. Analysis of published sequence data suggests that the existence of an exposed highly antigenic loop at positions corresponding to 155-161 of the spleen PAP sequence is a relatively general feature of PAP's. Trypsin and chymotrypsin cleave both bovine spleen PAP and uteroferrin, apparently in this region, with significant enhancement of enzymatic activity.

Characterization of the structural gene encoding a copper-containing nitrite reductase and homology of this gene to DNA of other denitrifiers.

A copper-containing nitrite reductase gene (nirU) from Pseudomonas sp. strain G-179 was found in a 1.9-kb EcoRI-BamHI DNA fragment. The coding region contained information for a polypeptide of 379 amino acids. The encoded protein had 78% identity in amino acid sequence to the nitrite reductase purified from Achromobacter cycloclastes. The ligands for type 1 copper- and type 2 copper-binding sites found in A. cycloclastes were also found in Pseudomonas sp. strain G-179, suggesting that these binding sites are conserved. Upstream from the promoter, two putative fnr boxes were found, suggesting that an FNR-like protein may be involved in regulation of the nitrite reductase gene under anaerobic conditions. When the 1.9-kb clone was used to probe Southern blots for similar sequences in DNAs from different denitrifiers, hybridization bands were seen for 15 of 16 denitrifiers known to have nitrite reductase containing copper. Except for Pseudomonas stutzeri JM300, all denitrifiers tested that have nitrite reductases containing heme c,d1 showed no or weak hybridization to this probe. Thus, this structural gene may be useful as a probe to detect denitrifiers with copper-containing nitrite reductases.

Characterization of Tn5 mutants deficient in dissimilatory nitrite reduction in Pseudomonas sp. strain G-179, which contains a copper nitrite reductase.

Tn5 was used to generate mutants that were deficient in the dissimilatory reduction of nitrite for Pseudomonas sp. strain G-179, which contains a copper nitrite reductase. Three types of mutants were isolated. The first type showed a lack of growth on nitrate, nitrite, and nitrous oxide. The second type grew on nitrate and nitrous oxide but not on nitrite (Nir-). The two mutants of this type accumulated nitrite, showed no nitrite reductase activity, and had no detectable nitrite reductase protein bands in a Western blot (immunoblot). Tn5 insertions in these two mutants were clustered in the same region and were within the structural gene for nitrite reductase. The third type of mutant grew on nitrate but not on nitrite or nitrous oxide (N2O). The mutant of this type accumulated significant amounts of nitrite, NO, and N2O during anaerobic growth on nitrate and showed a slower growth rate than the wild type. Diethyldithiocarbamic acid, which inhibited nitrite reductase activity in the wild type, did not affect NO reductase activity, indicating that nitrite reductase did not participate in NO reduction. NO reductase activity in Nir- mutants was lower than that in the wild type when the strains were grown on nitrate but was the same as that in the wild type when the strains were grown on nitrous oxide. These results suggest that the reduction of NO and N2O was carried out by two distinct processes and that mutations affecting nitrite reduction resulted in reduced NO reductase activity following anaerobic growth with nitrate.

Electron paramagnetic resonance studies on the high-salt form of bovine spleen purple acid phosphatase.

The EPR spectra of the high-salt form of reduced bovine spleen purple acid phosphatase (BSPAPr) and its complexes with inhibitory tetrahedral oxyanions, AMP, and fluorine have been examined in the 4-30 K temperature range. The EPR spectrum of the high-salt form of BAPAPr is identical to that previously reported for the low-salt form (Averill et al. (1987) J. Am. Chem. Soc. 109, 3760-3767), indicating that the substantial differences in conformation of the two forms result in undetectable alterations in the electronic structure of the binuclear iron center. Phosphate, AMP, and arsenate all result in broadened, highly anisotropic EPR spectra with decreased values of the antiferromagnetic coupling constant, -2J, while molybdate and tungstate produce a sharp axial or slightly rhombic spectrum, respectively, and fluoride produces an anomalous spectrum with an inverted g-tensor. These results are consistent with binding of the two classes of oxyanions (and AMP) to distinct sites at or near the binuclear iron center, while fluoride binds in yet a third mode. EPR spectra of the BSPAPr complex with molybdate show altered relaxation behavior in the presence of phosphate, consistent with a 50% decrease in the magnitude of -2J, suggesting that phosphate binds to the molybdate complex to produce a ternary complex analogous to that proposed for molybdate inhibition on the basis of kinetics studies.

Evidence that the type 2 copper centers are the site of nitrite reduction by Achromobacter cycloclastes nitrite reductase.

Methods have been developed for selective depletion and reconstitution of the Type 2 Cu (non-blue) sites in the nitrite reductase from A. cycloclastes, resulting in preparations ranging from 0.5 to 2.6 Type Cu per trimer; the Type 1 Cu content is invariant at 3.0 per trimer. The activity of the enzyme is directly proportional to the Type 2 content as measured by direct metal determination or by analysis of the EPR spectra. These results indicate that an earlier report that the A. cycloclastes enzyme contains only Type 1 Cu sites is incorrect, and that the Type 2 Cu centers constitute the site at which NO2- is reduced. Furthermore, they suggest that other Cu nitrite reductases that are reported to contain only Type 1 Cu sites and exhibit relatively low activity may actually be largely Type 2 Cu-depleted forms of the enzymes.

1H NMR and NOE studies of the purple acid phosphatases from porcine uterus and bovine spleen.

The diiron active sites of the purple acid phosphatases from porcine uterus (also called uteroferrin, Uf) and bovine spleen (BSPAP) and their complexes with tungstate are compared by 1H NMR and NOE techniques. The paramagnetically shifted features of the 1H NMR spectrum of reduced BSPAP are similar to those of reduced Uf, while the spectra of the tungstate complexes are almost identical. These observations suggest that the two active sites are quite similar, in agreement with the greater than 90% sequence homology found in the two enzymes. Nuclear Overhauser effect (NOE) experiments on the His N-H resonances show that the Fe(III)-His residue is N epsilon-coordinated, while the Fe(II)-His is H delta-coordinated in both enzymes. On the basis of the above NMR and NOE results, our previously proposed model for the dinuclear iron active site of Uf [Scarrow, R. C., Pyrz, J. W., & Que, L., Jr. (1990) J. Am. Chem. Soc. 112, 657-665] is corroborated, refined, and found to represent the diiron center of BSPAP as well.

Mutants of Pseudomonas fluorescens deficient in dissimilatory nitrite reduction are also altered in nitric oxide reduction.

Five Tn5 mutants of Pseudomonas fluorescens AK-15 deficient in dissimilatory reduction of nitrite were isolated and characterized. Two insertions occurred inside the nitrite reductase structural gene (nirS) and resulted in no detectable nitrite reductase protein on a Western immunoblot. One mutant had Tn5 inserted inside nirC, the third gene in the same operon, and produced a defective nitrite reductase protein. Two other mutants had insertions outside of this nir operon and also produced defective proteins. All of the Nir- mutants characterized showed not only loss of nitrite reductase activity but also a significant decrease in nitric oxide reductase activity. When cells were incubated with 15NO in H2(18)O, about 25% of the oxygen found in nitrous oxide exchanged with H2O. The extent of exchange remained constant throughout the reaction, indicating the incorporation of 18O from H2(18)O reached equilibrium rapidly. In all nitrite reduction-deficient mutants, less than 4% of the 18O exchange was found, suggesting that the hydration and dehydration step was altered. These results indicate that the factors involved in dissimilatory reduction of nitrite influenced the subsequent NO reduction in this organism.

Multiple binding sites for tetrahedral oxyanion inhibitors of bovine spleen purple acid phosphatase.

The theory of multiple inhibition kinetics has been extended to enzymes for which one inhibitor is noncompetitive and the other exhibits mixed inhibition. Plots of reciprocal velocity versus the concentration of either inhibitor at various fixed concentrations of the second inhibitor are predicted to give parallel lines if binding of the inhibitors is mutually exclusive and intersecting lines if the inhibitors interact at different sites on the enzyme. Application of this analysis to the purple acid phosphatase from bovine spleen in the presence of molybdate (a noncompetitive inhibitor) and phosphate (which exhibits mixed inhibition) results in parallel lines in the reciprocal velocity plots, indicating that phosphate and molybdate compete for a common site; since molybdate is a noncompetitive inhibitor, this site is inferred to be distinct from the site at which substrate binds and is hydrolyzed. Extension of these ideas suggests that phosphate ester substrates should be capable of binding to the molybdate-binding site as well as to the active site, and evidence for substrate inhibition at high substrate concentrations has been obtained. The implications of these findings for interpretation of previous spectroscopic studies of purple acid phosphatase complexes with tetrahedral oxyanions are discussed.

Hydrolysis of phosphate monoesters: a biological problem with multiple chemical solutions.

Formation of phosphate esters by kinases has long been recognized as an important process in biochemistry, but the reverse reaction, hydrolysis of phosphate esters by phosphatases, has attracted less attention. Recent work suggests that phosphatases are as important as kinases in regulatory processes, and that they constitute a diverse group of enzymes that utilize a variety of chemical means to accelerate phosphate ester hydrolysis.

Evidence for a NO-rebound mechanism for production of N2O from nitrite by the copper-containing nitrite reductase from Achromobacter cycloclastes.

Reduction of NO2- by the Cu-containing nitrite reductase from Achromobacter cycloclastes produces NO as the primary product initially, but as NO accumulates, NO production levels-off and N2O production becomes significant. Reaction of the enzyme with NO2- in the presence of NO increases the amount of N2O product significantly, while trapping the NO product as nitrosylhemoglobin or rapid removal of NO by sparging results in no detectable N2O production. Reaction of the enzyme with 15NO2- in the presence of 14NO results in rapid formation of the mixed isotope product (14N, 15N)O in ca. 45% yield. In contrast, the presence or absence of NO has no effect on N2O production by a prototypical heme cd1-containing nitrite reductase. These results are consistent with formation of a labile Cu(+)-NO+ species in the copper enzyme, which normally decomposes to NO. Production of N2O requires that the released NO must rebind to the enzyme to combine with a second NO2- or a species derived therefrom.

Evidence for a phosphoryl-enzyme intermediate in phosphate ester hydrolysis by purple acid phosphatase from bovine spleen.

The possibility of the existence of a covalent enzyme-phosphoryl intermediate, E-PO3, during catalysis of phosphate ester hydrolysis by the purple acid phosphatase (PAP) from bovine spleen has been examined. Transphosphorylation experiments show that up to 22% of the phosphoryl group from p-nitrophenyl phosphate (PNPP) can be transferred to primary alcohols. Burst experiments at high pH (9.1 or 8.1 for reduced or oxidized PAP, respectively), where hydrolysis of a phosphoenzyme intermediate is expected to be rate-limiting, show clear evidence for stoichiometric bursts of p-nitrophenolate from PNPP. The formation of base-stable, acid-sensitive adducts between PAP and the 32PO3 group of [gamma-32P]ATP has been demonstrated. The pH dependence of the kinetics parameters for reduced PAP has been determined over the range pH 3-8; a feature with a pKa of approximately 6.75 that is attributable to the enzyme-substrate complex is observed. Taken together, the present results are consistent with a two-stem pseudo Uni Bi mechanism that utilizes a covalent enzyme-phosphoryl intermediate, possibly a phosphohistidine.

H218O isotope exchange studies on the mechanism of reduction of nitric oxide and nitrite to nitrous oxide by denitrifying bacteria. Evidence for an electrophilic nitrosyl during reduction of nitric oxide.

Reduction of NO and NO2-by whole cells of eight strains of denitrifying bacteria known to contain either heme cd1 or copper-containing nitrite reductases (NiRs) has been examined in the presence of H218O. All organisms containing heme cd1 NiRs exhibited relatively large extents of exchange between NO2- and H218O (39-100%), as monitored by the 18O content of product N2O. Organisms containing copper NiRs gave highly variable results, with Achromobacter cycloclastes and Pseudomonas aureofaciens exhibiting no 18O incorporation and Rhodopseudomonas sphaeroides and Alcaligenes entrophus exhibiting complete exchange between NO2- and H218O. Organisms containing heme cd1 NiRs exhibited significant but lower levels of exchange between NO and H218O than between NO2- and H218O, while organisms containing copper NiRs gave significantly higher amounts of 18O incorporation than observed for the heme cd1 organisms. These results demonstrate the existence of an NO-derived species capable of undergoing O-atom exchange with H218O during the reduction of NO. Trapping experiments with 15NO, 14N3-, and crude extracts of R. sphaeroides support the electrophilic nature of this intermediate and suggest its formulation as an enzyme nitrosyl, E-NO+, analogous to that observed during reduction of NO2-. The observation of lower levels of 18O incorporation with NO2- than with NO as substrate for A. cycloclastes and P. aureofaciens indicates that, for these organisms at least, a sequential pathway involving free NO as an intermediate is significantly less important than a direct pathway in which N2O is formed via reaction of two NO2- ions on a single enzyme.