Kinetic and Spectroscopic Analyses of Mutants of a Conserved Histidine in the Metallophosphatases Calcineurin and λ Protein Phosphatase*

Calcineurin belongs to a family of serine/threonine protein phosphatases that contain active site dinuclear metal cofactors. Bacteriophage λ protein phosphatase is also considered to be a member of this family based on sequence comparisons (Lohse, D. L., Denu, J. M., and Dixon, J. E. (1995)Structure 3, 987–990). Using EPR spectroscopy, we demonstrate that λ protein phosphatase accommodates a dinuclear metal center. Calcineurin and λ protein phosphatase likewise contain a conserved histidine that is not a metal ligand but is within 5 Å of either metal in calcineurin. In this study the conserved histidine in calcineurin was mutated to glutamine and the mutant protein analyzed by EPR spectroscopy and kinetic methods. Parallel studies with an analogous λ protein phosphatase mutant were also carried out. Kinetic studies using paranitrophenyl phosphate as substrate showed a decrease in k cat of 460- and 590-fold for the calcineurin and λ protein phosphatase mutants, respectively, compared with the wild type enzymes. With a phosphopeptide substrate, mutagenesis of the conserved histidine resulted in a decrease ink cat of 1,300-fold for calcineurin. With the analogous λ protein phosphatase mutant, k catdecreased 530-fold compared with wild type λ protein phosphatase using phenyl phosphate as a substrate. EPR studies of the iron-reconstituted enzymes indicated that although both mutant enzymes can accommodate a dinuclear metal center, spectroscopic differences compared with wild type proteins suggest a perturbation of the ligand environment, possibly by disruption of a hydrogen bond between the histidine and a metal-coordinated solvent molecule.

Calcineurin, also known as protein phosphatase 2B, consists of a 58-kDa catalytic subunit, calcineurin A, and a 19-kDa regulatory subunit, calcineurin B. It is a serine/threonine protein phosphatase whose activity is regulated by Ca 2ϩ /calmodulin. Calcineurin is the target of the immunosuppressant drugs cyclosporin A and FK506 (1,2). These drugs bind to intracellular proteins, termed immunophilins; cyclophilin is the binding protein for cyclosporin A, and FK506 binds to the FK506binding proteins. The complex of immunosuppressant drug and immunophilin in turn binds to and inhibits the phosphatase activity of calcineurin. Calcineurin inhibition prevents the transcriptional activation of the interleukin 2 gene in helper T cells, leading to suppression of the immune response.
Calcineurin is a member of the class of serine/threonine protein phosphatases, whose members include protein phosphatases 1 (PP1) 1 and 2A (PP2A), phosphatases essential for a number of signal transduction pathways in eukaryotic cells (3,4). Another protein phosphatase from bacteriophage , PP, also belongs to this family (5). In addition, a number of less characterized enzymes containing the "phosphoesterase" consensus motif of this family, DXH(X) n GDXXD(X) n GNHD/E, have been identified via protein sequence comparisons (6,7). It has been hypothesized that this motif provides a scaffold for an active site dinuclear metal center (7,8), similar to the dinuclear metal centers in PP1 (9, 10) and calcineurin (11)(12)(13). A variety of experimental evidence indicates that this cluster in calcineurin is an Fe 3ϩ -Zn 2ϩ center.
Although little is known about the catalytic mechanism of the serine/threonine protein phosphatases, several pieces of experimental data indicate that the dinuclear metal center is a key component of the active site. First, x-ray crystallographic data of calcineurin and PP1 indicate that the dinuclear metal center has a ligand environment nearly identical to that of mammalian and plant purple acid phosphatases, enzymes that contain either dinuclear Fe-Fe or Fe-Zn centers that have been demonstrated to be essential for catalytic activity (14,15). Second, these crystallographic studies indicate that the product of the reaction, phosphate, and the product analog, tungstate, directly coordinate both metal ions (10,11,16). Third, redox titrations of either the Fe 3ϩ -Zn 2ϩ (12) or Fe 3ϩ -Fe 2ϩ forms 2 of calcineurin indicate a correlation between enzyme activity and the oxidation state of the bound metal ions.
In addition to the dinuclear metal center, there are several conserved amino acids within the active site which are likely to contribute to catalysis. One of these residues in calcineurin is histidine 151 (numbering based on the rat calcineurin A␣ sequence (18)). His-151 is not a ligand to either metal but is within 5 Å of both metal ions and is conserved in other metallophosphoesterases such as PP1 (histidine 125) and PP (histidine 76) (7). In one crystal structure, His-151 was modeled to participate in a hydrogen bond to a metal-coordinated solvent molecule (13). The importance of this residue has been demon-strated by site-directed mutagenesis of both PP and PP1 which found substantial effects on catalytic activity and/or protein stability.
In this study the conserved histidine residue in calcineurin A, His-151, was changed to glutamine by site-directed mutagenesis. After reconstitution with calcineurin B, the calcineurin H151Q heterodimer was purified to homogeneity. EPR spectroscopy was used to assess whether mutagenesis affected the ligand environment of the dinuclear metal center. Kinetic studies were also carried out using either pNPP or [P]-R II peptide as substrates. The analogous PP mutant, H76N, was characterized by EPR in a similar fashion and assayed using either pNPP or phenyl phosphate. For both enzymes, mutagenesis resulted in decreases in k cat of 10 2 -10 3 . EPR spectroscopy of iron-reconstituted PP confirms that this enzyme accommodates a dinuclear metal center as predicted for members of the metallophosphatase family. EPR also indicates that mutagenesis does not prevent assembly of the dinuclear metal center in either calcineurin H151Q or PP (H76N). Nevertheless, differences observed in the EPR spectra of wild type versus mutant enzymes indicate a perturbation of the ligand environment, possibly by disruption of a hydrogen bond between the histidine and a metal-coordinated water molecule.
[␥-32 P]ATP (ϳ3,000 Ci/mmol) was purchased from Amersham Corp. Calmodulin was prepared from bovine brain (19,20) and coupled to Affi-Gel active ester agarose (Bio-Rad) for use in calmodulin affinity chromatography. PM30 and YM30 membranes and Centricon-30 concentrators were purchased from Amicon (Beverly, MA). Tryptone, yeast extract, and Luria Bertani medium were purchased from Difco. Sephadex NAP-25 columns were purchased from Pharmacia Biotech Inc. The protein expression vectors used were pRCNAT77 (21) and pRCNBT775-3 (22), encoding the genes for calcineurin A and B, respectively, and the plasmid, pBB131 (23,24), encoding the gene for Nmyristoyl transferase. The construction of pT7-7 plasmids containing the genes for wild type and H76N mutant of PP is described elsewhere (5,25). The Wizard Maxipreps and Wizard Minipreps DNA purification kits and T4 polynucleotide kinase were purchased from Promega (Madison, WI). The Geneclean II kit was purchased from BIO 101, Inc. (Vista, CA). The oligonucleotide required for site-directed mutagenesis was synthesized by the Mayo Clinic Molecular Biology Core Facility. R II peptide (DLDVPIPGRFDRRVSVAAE) and [ 31 P]-R II peptide (DLD-VPIPGRFDRRVS(p)VAAE) were synthesized by the Mayo Clinic Protein Core Facility.
Methods-Protein concentrations were determined by the Bradford assay using the Pierce Coomassie Plus Protein Assay Reagent with bovine serum albumin as a standard (26). Alternatively, calcineurin concentrations were determined by UV-visible spectrophotometry (27) using ⑀ 281 ϭ 50,000 M Ϫ1 cm Ϫ1 . Protein concentration values determined using this extinction coefficient agreed within 10% with concentrations determined by amino acid analysis.
Site-directed Mutagenesis-The H151Q mutation of calcineurin A, CN(H151Q), was created using a 5 Prime 3 3 Prime MORPH sitespecific plasmid DNA mutagenesis kit (Boulder, CO). The primer 5Ј-GCCTACATTCCTGGTTTCCAC-3Ј, complementary to the coding strand of calcineurin A, was used for mutagenesis with the underlined bases representing the codon of the mutated residue. This primer was phosphorylated by T4 kinase and used according to the manufacturer's instructions. Half-mutant plasmid DNA was subsequently transformed into Escherichia coli strain BMH 71-18. Plasmid DNA isolated from colonies was screened for the desired mutation by DNA sequencing of the entire calcineurin A gene and yielded the expression plasmid pCNAT77(H151Q).
Expression and Purification of CN(H151Q)-The plasmid pCNAT77(H151Q) was transformed into competent BL21(DE3) cells. The growth of these cells, crude extract isolation, reconstitution with myristoylated calcineurin B, and purification of the calcineurin het-erodimer were performed as described previously (21,22).
Expression and Purification of Wild Type PP-Expression of wild type PP was performed as described (5). All purification steps were performed at 4°C. After growth and induction with isopropyl ␤-Dthiogalactopyranoside, the cells were harvested by centrifugation at 3,400 ϫ g for 30 min, washed with 250 ml of 0.1 M Tris-Cl, pH 7.5, and recentrifuged at 4,200 ϫ g for 20 min. The cells were resuspended in 25 mM Tris-Cl, pH 8.0, 20% glycerol, 1 mM EGTA (TGE buffer) and lysed by three passages through a French pressure cell operating at 16,000 p.s.i. The cell lysate was subsequently centrifuged at 39,000 ϫ g for 3 h. The supernatant (40 ml) was batch adsorbed onto 150 ml of DEAE-Sepharose CL-6B preequilibrated with TGE buffer. The resin was washed in a fritted funnel with 300 ml of TGE and PP eluted with TGE buffer containing 0.1 M NaCl. Fractions containing PP were pooled and precipitated by the addition of ammonium sulfate to 50% saturation. After centrifugation at 34,800 ϫ g, the protein pellet was resuspended in TGE buffer ϩ 0.5 M NaCl and applied to a phenyl-Sepharose column (20 ϫ 1-cm diameter) previously equilibrated with TGE buffer ϩ 0.5 M NaCl. The column was washed with 200 -300 ml of the same buffer and then with 250 ml of 20 mM Tris-Cl, pH 7.5. The enzyme was eluted with 250 ml of 50 mM Tris-Cl in 50% glycerol, pH 7.5. Fractions were assayed using pNPP as a substrate, pooled, and stored at Ϫ70°C in 50 mM Tris-Cl and 50% glycerol, pH 7.5.
Expression of PP(H76N)-The PPT77(H76N) plasmid (25) was transformed into BL21(DE3) cells and single colonies used to inoculate 10 ml of Luria Bertani medium/ampicillin (0.1 mg/ml) for overnight culture at 37°C. Overnight cultures were then used to inoculate 5 liters of 2 ϫ YT/ampicillin medium (10 g/liter yeast extract, 20 g/liter tryptone, 10 g/liter NaCl, 0.05 g/liter ampicillin) in a New Brunswick Bioflo 3000 fermentor. Cells were grown overnight at 22°C maintaining aeration at 30% of air saturation to a cell density that gave an absorbance at 595 nm of ϳ9. Glucose was added to a final concentration of 0.4%, and the cells were induced with 1 mM isopropyl ␤-D-thiogalactopyranoside. Another aliquot of glucose was added to a final concentration of 0.4% when the cell density corresponded to an absorbance at 595 nm of 16. The cells were harvested 20 h postinduction by centrifugation at 3,400 ϫ g for 20 min. The cell pellet was resuspended in ϳ2 ml of 50 mM Tris-Cl, pH 7.5/g of cells, wet weight. To this resuspension, 0.4 mg/ml lysozyme, 23 mM EDTA, and 0.05% Triton X-100 were added sequentially with stirring on ice for 30 min followed by a freeze/thaw process to lyse the cells. To reduce the viscosity, MgCl 2 (20 mM), DNase (0.1 unit/ml final concentration), and 2% protamine sulfate (1/6 total volume) were added sequentially with stirring on ice. After centrifugation at 10,000 ϫ g for 1 h, the protein was purified as described above for wild type PP.
Circular Dichroism Measurements-Circular dichroism spectra were recorded at 25°C on a Jasco J-710 circular dichroism spectrometer. A quartz cell of 0.0202-cm path length was used for all measurements. Mean residue ellipticities were calculated from the relationship m ϭ obs /(10C r l) where m is the mean residue ellipticity, obs is the observed ellipticity in millidegrees, C r is the mean residue molar concentration, and l is the path length of cell in cm. m is measured in degree cm 2 dmol Ϫ1 . Samples of calcineurin (18 M) and CN(H151Q) (24 M) were examined in 10 mM Hepes, pH 7.5, 1 mM MgCl 2 , 0.1 M EGTA, 0.2 mM DTT.
pNPP Assays-Calmodulin-dependent phosphatase activity of calcineurin and CN(H151Q) was measured using pNPP as a substrate at 30°C in 25 mM MOPS, pH 7.0, 1.0 mM MnCl 2 , 0.1 mM CaCl 2 , 1 M calmodulin, and 15-23 nM wild type calcineurin or 710 nM CN(H151Q). Wild type PP and PP(H76N) activities were measured at 30°C in 100 mM Tris-Cl, pH 7.8, 10 mM DTT, 1 mM MnCl 2 , and 0.64 nM wild type PP or 860 nM PP(H76N). After incubation for 5 min at 30°C, reactions were started by the addition of pNPP. Specific activity was measured by following the increase in absorbance at 410 nm with time using ⑀ 410 ϭ 7,180 M Ϫ1 cm Ϫ1 at pH 7.0 and 14,400 M Ϫ1 cm Ϫ1 at pH 7.8 based on a pK a of 7.17 and ⑀ 410 ϭ 17,800 M Ϫ1 cm Ϫ1 for the p-nitrophenolate anion. The concentration of pNPP was varied from 2 to 100 mM, and the kinetic parameters k cat and K m were determined by fitting the data to the Michaelis-Menten equation using a nonlinear least squares analysis method.
[P]-R II Peptide Assays-R II peptide was phosphorylated with [␥-32 P]ATP to a specific activity of 833 Ci/mol using the catalytic subunit of bovine cardiac cyclic-AMP dependent protein kinase and purified as described (28). Assays were done as described (28)  Phenyl Phosphate Assays-Assays were performed by determining the amount of inorganic phosphate released during hydrolysis of phenyl phosphate as described (29). Assays were done in 100 mM Tris-Cl, pH 7.8, containing 1 mM MnCl 2 and 10 mM DTT. Phenyl phosphate concentrations were varied from 1 to 70 mM for PP and 1 to 90 mM for PP(H76N). Wild type protein concentrations ranged from 7.0 to 630 nM, whereas PP(H76N) concentrations ranged from 2.2 to 12 M. Enzyme was incubated 30°C, 5 min, and reactions started by the addition of phenyl phosphate. At various times from 0.5 to 7.0 min, 50 l of the reaction was taken and added to 800 l of a solution containing a 3:1 ratio of 0.045% malachite green hydrochloride to 4.2% ammonium molybdate in 4 N HCl. After 1 min, 100 l of 34% sodium citrate was added and the absorbance at 660 nm measured. Free phosphate was determined from a standard curve prepared using solutions of KH 2 PO 4 . Kinetic parameters for PP and PP(H76N) were determined as described above.
Reconstitution of CN(H151Q) and PP(H76N) with Iron-About 0.4 mg/ml CN(H151Q) in 20 mM Tris-Cl, pH 7.5, 100 mM KCl, 1.0 mM magnesium acetate, 1.0 mM DTT, 0.1 mM EGTA, or 0.6 mg/ml PP(H76N) in 50 mM Tris-Cl, pH 7.5, 100 mM NaCl, 10% glycerol was added to septum-sealed vials and made anaerobic by flushing repeatedly with oxygen-free argon. BME was added to a final concentration of 0.715 M and Fe(NH 4 ) 2 (SO 4 ) 2 ⅐6H2O added dropwise to a final concentration of 0.375 mM. The solutions were flushed with argon and incubated an additional 17 h at room temperature. Samples were concentrated using YM30 membranes in an Amicon filtration cell to ϳ 2 ml; buffer exchanged over Sephadex NAP-25 gel filtration columns equilibrated with 50 mM MOPS, pH 7.0, 1 mM BME; and concentrated using Centricon-30 membranes to ϳ250 l. Samples were then transferred to quartz EPR cuvettes and frozen in liquid nitrogen. Iron-reconstituted forms of wild type calcineurin and PP were prepared in a similar manner except that incubation proceeded at 4°C for 17 h, and both enzymes were desalted into 100 mM Tris-Cl, pH 7.5, 1 mM BME prior to the final concentration step.
Addition of Orthophosphate to CN(H151Q)-To the iron-reconstituted EPR sample of CN(H151Q), a solution of 0.5 M potassium phosphate, pH 7.5, was added anaerobically to a final concentration of 20 mM, incubated for 5 min at room temperature, and frozen in liquid nitrogen.
EPR Analysis-EPR spectra were recorded using a Bruker ESP300E spectrometer operating at 9 GHz (X-band) microwave frequency equipped with an Oxford Instruments ESR 900 continuous flow cryostat for temperature regulation. Background cavity resonances were subtracted from all spectra.
Metal Analysis-Metal analysis was performed by the Mayo Clinic Metals Laboratory using inductively coupled plasma emission spectrometry.

Preparation of Recombinant CN(H151Q) and PP(H76N)
Proteins-The relationship of His-151 in calcineurin relative to the dinuclear metal center can be seen in Fig. 1. His-151 was mutated to a glutamine by site-directed mutagenesis to investigate the effect on enzyme activity and assembly of the dinuclear metal cofactor. The mutant calcineurin A subunit was expressed in E. coli in a fashion identical to that of the wild type calcineurin A subunit and reconstituted with myristoylated calcineurin B to generate the mutant protein CN(H151Q). The presence of the proper codon as well as the lack of inadvertently introduced mutations in the entire calcineurin A gene were confirmed by DNA sequence analysis. Although the yield of CN(H151Q) was less than that obtained for the wild type reconstituted protein, enough material could be obtained and purified to homogeneity for biochemical and spectroscopic (EPR) analyses. A typical purification yielded approximately 1 mg of CN(H151Q) protein/liter of cell culture.
The analogous residue in PP, identified by primary sequence comparisons as His-76 (7), was mutated to an asparagine residue (25). In this study, the PP (H76N) protein was purified to homogeneity as described under "Methods" to yield approximately 17 mg/liter of culture.  Table I. Using either substrate, the values of k cat for the mutant enzyme were significantly lower than the k cat values for wild type enzyme. Thus, the k cat for CN(H151Q) using pNPP as a substrate, 5.6 ϫ 10 Ϫ2 s Ϫ1 , is 460-fold lower than recombinant wild type calcineurin prepared in an identical fashion. Using [P]-R II peptide, the k cat values for wild type calcineurin and CN(H151Q) were 1.2 ϫ 10 1 s Ϫ1 and 9.0 ϫ 10 Ϫ3 s Ϫ1 , respectively, a difference of 1,300-fold. Using [P]-R II peptide, K m values for both forms of calcineurin were the same within the error of the measurement. However, a 10-fold decrease in K m was observed for CN(H151Q) compared with wild type calcineurin using pNPP as a substrate.
Phosphatase Activities of Wild Type PP and PP(H76N)-Kinetic parameters using pNPP and phenyl phosphate as substrates for wild type PP and PP(H76N) are compared in Table II. These parameters were also determined by inclusion of 1.0 mM MnCl 2 in assay buffers. Using pNPP as substrate, the k cat values for wild type PP and PP(H76N) were 3.9 ϫ 10 2 s Ϫ1 and 6.6 ϫ 10 Ϫ1 s Ϫ1 , respectively. This represents a 590-fold difference, which is comparable to the decrease in k cat observed for calcineurin for the analogous substitution. The difference in k cat for PP is less than the Ϸ10 5 fold difference found previously (25) mostly because of a 30-fold higher activity measured for the PP (H76N) protein isolated in the present study. Using phenyl phosphate as substrate, the k cat value for PP was 1.7 ϫ 10 1 s Ϫ1 , compared with 3.2 ϫ 10 Ϫ2 s Ϫ1 for PP (H76N), a difference of 530-fold. The K m values were very similar for both mutant and wild type proteins using either substrate.
EPR Analysis of Iron-reconstituted Wild Type Calcineurin and CN(H151Q)-Although not a metal ligand, H151Q in calcineurin is close enough to either metal ion such that mutagenesis might perturb the environment surrounding the metal cluster. To investigate the effect of mutagenesis on the dinuclear center, wild type calcineurin and CN(H151Q) were reconstituted with iron to generate a mixed valence Fe 3ϩ -Fe 2ϩ cluster as a spectroscopic probe of the active site and analyzed by EPR spectrometry. The EPR spectra of wild type calcineurin, CN(H151Q), and CN(H151Q) in the presence of 20 mM potassium phosphate are compared in Fig. 2. Several features are evident in the EPR spectra including a minor high spin Fe 3ϩ species with g values of 9.2 and 4.3 (about 700 -2,000 Gauss region, not shown for clarity), a minor radical species centered at g av ϭ 2.0, and a component with g av Ͻ 2.0 representing the major paramagnetic species.
Simulations of the EPR signal in Fig. 2A to an S ϭ 1/2 species yielded g values of 1.93, 1.77, and 1.64. This signal is identical to the signal observed previously in bovine brain calcineurin reconstituted with iron and arises from a dinuclear iron center in the mixed valence (Fe 3ϩ -Fe 2ϩ ) oxidation state (12). The spin Hamiltonian H e that describes the magnetic properties of the dinuclear iron center is where H is the external magnetic field, ␤ is the Bohr magneton, J is the exchange-coupling constant, and D i , E i , and g i (i ϭ 1, 2) represent the zero field splitting terms and g tensors for each metal ion, respectively. Antiferromagnetic coupling (J Ͼ 0) between the high spin ferric (S 1 ϭ 5/2) and high spin ferrous (S 2 ϭ 2) ions of the cluster yields a ground state with S ϭ 1/2 which gives rise to this EPR signal. Fig. 2B shows the EPR spectrum of iron-reconstituted CN(H151Q). An EPR signal representative of a dinuclear iron center in the mixed valence state is evident and indicates that a dinuclear metal center can be assembled in the mutant enzyme. This signal, however, is broader than the signal observed from the mixed valence center in wild type enzyme. Metal analyses of the EPR sample found 1.5 mol each of iron and zinc/mol of protein, consistent with the formation of a dinuclear iron center but also indicating the possible presence of a mixed metal Fe-Zn center or adventitious zinc, either of which would contribute to the EPR spectrum in the g av Ͻ 2.0 region.
Further proof that the signal in Fig. 2B results from an active site metal center was demonstrated by adding 20 mM potassium phosphate to the sample. The addition of phosphate led to a noticeable sharpening of the EPR signal (Fig. 2C), whereas none of the other species was affected, suggesting that phosphate coordinates to one or both of the metal ions of the dinuclear iron center.
EPR Analysis of Iron-reconstituted PP and PP(H76N)-In a fashion similar to that for calcineurin, PP and PP(H76N) were reconstituted with iron to generate a spectroscopic probe of the active site metal cluster. PP as purified contains very little iron, zinc, or manganese as determined by metal analysis using inductively coupled plasma emission spectrometry (Յ 0.05 mol of iron, 0.09 mol of zinc, and 0.01 mol of manganese/ mol of protein). Likewise, PP(H76N) also contained low amounts of these metals (Յ 0.3 mol of iron, 0.09 mol of zinc, and 0.01 mol of manganese/mol of protein). Reconstitution of wild type PP and PP(H76N) with iron yielded samples that exhibited low temperature EPR spectra with g av Ͻ 2.0 (Fig. 3, A  and B). Metal analysis of both EPR samples found 1.74 iron/ mol of protein and 0.20 zinc/mol of protein for wild type PP, and 1.7 iron/mol of protein and 0.05 zinc/mol of protein for PP(H76N). Thus both PP and PP(H76N) can accommodate dinuclear metal clusters. As in the case of calcineurin, the EPR spectrum of the PP(H76N) mutant is different from the spectrum of wild type PP.

DISCUSSION
Recent crystallographic models of calcineurin (11,13), PP1 (9, 10), and purple acid phosphatase (16,31) have identified a conserved histidine in each active site which, although not coordinated to either metal of the dinuclear metal center, is within Ϸ 5 Å of both metal ions. In this study we have mutated the corresponding residue of calcineurin (His-151) to glutamine to explore its significance in catalysis and effect, if any, on the active site dinuclear metal center. An analogous histidine in the bacteriophage protein phosphatase has been previously  modified to asparagine (PP(H76)) (25). The H76N mutation resulted in a 10 5 -fold reduction in k cat toward pNPP, a 40-fold increase in the K m for Mn 2ϩ , the divalent metal ion activator used in assay buffers, yet little change in K m for substrate. In a study by Lee and colleagues (32), the comparable residue in PP1 (His-125) was mutated to a number of residues. In that study it was found that most of the substitutions resulted in the production of insoluble protein except for two mutations, H125A and H125S, where a fraction of the protein was soluble and could be purified by affinity chromatography. Although neither PP1 mutant exhibited any detectable phosphatase activity, the upper limit for activity and/or fold reduction relative to wild type PP1 was not reported. Similar to the results noted in the PP1 study, the level of expression of H151Q soluble protein was also lower than that found for wild type calcineurin A. Hence, the yield of the CN(H151Q) heterodimer was lower (Ϸ4-fold) than wild type reconstituted enzyme prepared in an identical fashion. 3 As with wild type calcineurin (21), growth of E. coli expressing the calcineurin A subunit at lower temperatures (23°C) improved the yield of soluble protein in crude extract. Reconstitution of the mutant calcineurin A subunit with calcineurin B allowed purification of sufficient protein for biochemical studies and spectroscopic analysis using EPR.
To assess the affect of mutagenesis on catalytic activity, two different substrates were used for calcineurin, pNPP and [P]-R II peptide. 4 Furthermore, MnCl 2 was included in buffers to obtain the maximum activity for both enzymes (5,33,34). Manganese is known to incorporate into one or both metal sites in PP1 (9,10). Preliminary data indicate that a spin-coupled dinuclear Mn 2ϩ cluster is also assembled in PP. 5 In terms of catalytic activity, mutagenesis of His-151 to glutamine resulted in a 460-fold decrease in k cat using pNPP as substrate. In comparison, the k cat for CN(H151Q) decreased 1,300-fold compared with wild type calcineurin using [P]-R II peptide as substrate. For either substrate, K m was either not affected or slightly improved by mutagenesis. The k cat values for wild type calcineurin measured in this study are comparable or slightly higher than previously reported (24,34,35); these differences may reflect slight variability between rat versus bovine isoforms.
Using pNPP as substrate, PP(H76N) also exhibited little difference in substrate K m and a decrease in k cat of 590-fold relative to wild type PP. This difference is more than 100-fold lower than the 10 5 noted previously (25) primarily because of a 3 P. Mertz, L. Yu, R. Sikkink, and F. Rusnak, unpublished results. 4 Although pNPP is a substrate for calcineurin (33), the K m of 10 -20 mM is significantly greater than the K m of Ϸ100 M for [P]-R II peptide. Furthermore, calcineurin activity using [P]-R II peptide is inhibited by the immunosuppressant drug complexes cyclosporin A⅐cyclophilin and FK506⅐FKBP, whereas it is slightly stimulated when assayed using pNPP (1,40). Nevertheless, calcineurin phosphatase activity toward [P]-R II peptide is progressively inhibited in the presence of increasing concentrations of pNPP with an IC 50 equivalent to the K m for pNPP (data not shown). These results indicate that both pNPP and [P]-R II peptide utilize the same active site, and hence a comparison of their kinetic parameters is valid. 5  30-fold higher activity measured for the PP(H76N) mutant in this study but also because of a slightly lower activity of the wild type enzyme used in this study (Ϸ 5-fold; see Table II). In fact, variability in k cat has been documented for wild type PP (5,25,36); we have also observed about a 2-3-fold difference in activity for several different preparations of wild type PP prepared by similar procedures. Kinetic parameters of wild type PP and PP(H76N) for a second substrate, phenyl phosphate, paralleled the results using pNPP. Thus PP(H76N) exhibited a decrease in k cat of 530-fold and a K m similar to wild type enzyme.
Based on primary sequence homologies, PP is thought to be a member of the family of metallophosphatases including calcineurin, PP1, and PP2A (6). We now show that PP can accommodate a binuclear metal center as predicted for enzymes containing the phosphoesterase motif. Therefore, it is likely that His-76 in PP and His-151 in calcineurin have similar functions in phosphate ester hydrolysis. Proposed roles for this conserved histidine include an active site nucleophile, a role in orienting substrate, a role in general acid catalysis involving protonation of the leaving group, or a role in general base catalysis by deprotonation of an iron-coordinated solvent molecule. All of these would be consistent with the 10 2 -10 3 -fold decrease in activity observed for wild type versus mutant enzymes.
The fold decreases for His-151/His-76 are comparable to the decrease of Ϸ10 3 found for serine-to-alanine and serine-toleucine mutations for the nucleophilic serine residue of alkaline phosphatase (37). However, experiments with purple acid phosphatase, a member of the metallophosphatase superfamily with a strikingly similar active site as calcineurin (8) demonstrated that hydrolysis proceeded by direct transfer of the substrate phosphoryl group to solvent (38). By analogy, therefore, it seems unlikely that this histidine participates in nucleophilic catalysis. Similar substrate K m values for mutant and wild type enzymes also argue against a necessary role in substrate binding.
In the crystal structure of calcineurin with phosphate bound at the active site, His-151 is within H-bonding distance of the most solvent-exposed oxygen atom of phosphate, close enough to assist in leaving group protonation (11). Similar orientations are observed in the crystal structure of human PP1␥ with bound tungstate ion (10), rabbit PP1 ␣ with phosphate modeled in the active site (9), and in the purple acid phosphatase structures with phosphate and tungstate bound at the active site (16). If His-151/His-76 were involved in proton donation to the leaving group, the k cat for each mutant enzyme relative to wild type should show a marked dependence using substrates that have leaving groups of disparate pK a values. In fact, only a ϳ3-fold difference in relative k cat (wild type versus CN(H151Q)) for pNPP versus [P]-R II peptide is observed, even though the products of the reaction have acidities that differ by Ͼ10 6 (p-nitrophenol has a pK a of 7.2 compared with Ϸ14 for serine). With PP, there is no difference between relative k cat values of wild type versus H76N using pNPP and phenyl phosphate (pK a ϭ 9.95). It seems likely therefore that this histidine is not required for protonation of the leaving group.
Although the k cat values for CN(H151Q) and PP(H76N) represent significant decreases compared with wild type enzyme, they are 10 5 -10 7 -fold greater than the noncatalyzed rate of hydrolysis. 6 A significant amount of this remaining catalytic activity is most likely derived from the presence of the active site dinuclear metal center, which has been proposed to lower the pK a of a coordinated water molecule, the putative nucleophile in the reaction. His-151/His-76 could be functioning in concert with this solvent molecule to either position a lone pair on the oxygen atom for optimum in-line attack on the phosphorus atom of the substrate or to serve as a general base to take up a proton concomitant with solvent nucleophilic attack. At least in one crystal structure model of calcineurin, the N⑀ atom of His-151 was H-bonded to one of two solvent molecules coordinated to the iron atom (13). In the crystal structure of PP1 with microcystin bound, the N⑀ atom of the analogous histidine, His-125, was also within H-bonding distance of a water molecule, but that water was modeled Ն 3.2 Å away from the metal ions (9). Further evidence for this model is provided by mutagenesis studies of PP1 examining the influence of a conserved aspartic acid residue, Asp-95, on catalytic efficiency. This conserved aspartate residue is part of the phosphoesterase consensus motif (6,7). PP1 residue Asp-95 is within H-bonding distance of the conserved histidine, and mutagenesis to asparagine resulted in a 71-fold decrease in activity compared with wild type using phosphorylase a as substrate (32). The analogous mutant in PP, D52N, resulted in a 36-fold reduction in activity using pNPP as substrate (25). The corresponding residue in calcineurin, Asp-121, is also within H-bonding distance of His-151 (Fig. 1). Thus, the interaction of this conserved histidine/aspartate pair with a solvent molecule is analogous to the catalytic aspartate/histidine/serine motif of serine proteases and could be thought of as a "catalytic tetrad" with the metal ion serving as a Lewis acid to lower the pK a of the nucleophile.
If His-151 participates in a hydrogen bond with a metalcoordinated solvent, mutagenesis will disrupt this interaction and is likely to affect the spectroscopic properties of the dinuclear metal center. We have reconstituted calcineurin and PP with iron to generate an active site dinuclear iron center for use as a spectroscopic probe of the active site. The Fe 3ϩ -Fe 2ϩ oxidation state of this cluster gives rise to a signature EPR resonance with g av Ͻ 2.0 (14,15) which is sensitive to changes in the metal environment via perturbation of zero-field splitting (D i , E i i) and spin coupling (J) constants in Equation 1 (39). The EPR spectrum of iron-reconstituted CN(H151Q) exhibited g values consistent with the formation of a Fe 3ϩ -Fe 2ϩ center, indicating that the H151Q mutant enzyme is still able to support a dinuclear metal center. However, the overall shape of this spectrum was quite different from that of wild type calcineurin. In comparison, the EPR spectrum of the ironreconstituted PP(H76N) also exhibited the characteristic g av Ͻ 2.0 signal with a shape distinct from the corresponding spectrum of wild type PP.
The fact that phosphate addition to iron-reconstituted CN(H151Q) caused a change in the shape of the EPR resonance indicates that it arises from an active site metal center. Interestingly, in both wild type calcineurin and purple acid phosphatase, phosphate binding to the mixed valence cluster led to a broadening of the corresponding EPR signal, a result of a decrease in the spin coupling constant, J, caused by the phosphate ion bridging the two metal ions of the cluster (39). 2 With CN(H151Q), on the other hand, phosphate caused a sharpening of the EPR resonance. Further spectroscopic analysis is required to understand the structural basis for these differences.
These results demonstrate that the conserved histidine in the metallophosphatases calcineurin and PP is an essential component of the active site since disruption led to significant decreases in activity. Loss of activity may have resulted from removal of an active site base and the disruption of an essential H bond to a metal-coordinated solvent molecule. Future experiments to confirm this are in progress.