His-8 lowers the pKa of the essential Cys-12 residue of the ArsC arsenate reductase of plasmid R773.

The 141-residue ArsC arsenate reductase of plasmid R773 has an essential cysteine residue, Cys-12. The pKa of Cys-12 was determined to be 6.4, compared with a pKa of 8.3 for free cysteine. The possibility of the formation of an ion pair between Cys-12 and a basic residue was investigated. Enzymatic activity was rapidly inactivated by the histidine-modifying reagent diethylpyrocarbonate. The codons for the two histidine residues in ArsC, His-8 and His-88, were changed by site-directed mutagenesis. Cells expressing arsCH88R, arsCH88S, arsCH88W, or arsCH88V genes retained arsenate resistance, and the purified proteins had wild type level of reductase activity. Cells expressing arsCH8P, arsCH8S, arsCH8G, or arsCH8R genes were each sensitive to arsenate, and the purified H8P, H8G, and H8R proteins each lacked enzymatic activity. Using the single histidine proteins it was shown that both histidines react with diethylpyrocarbonate but that only reaction with His-8 resulted in inactivation. The pKa value of Cys-12 was determined to be 6.3 in the H8R enzyme and 8.3 in the H8G enzyme. These results indicate that His-8 is essential for catalytic activity and that a positively charged residue is required at position 8 to lower the pKa of the cysteine thiolate at position 12.

Arsenical resistance (ars) operons of both Gram-positive and Gram-negative bacteria confer resistance to the toxic oxyanions arsenite (As(III)), antimonite (Sb(III)), and arsenate (As(V)) (1). In Escherichia coli resistance to trivalent arsenicals and antimonials conferred by the ars operon of plasmid R773 results from their active extrusion from the cells by an ATPcoupled arsenite pump (1)(2)(3). Resistance to arsenate in both Gram-positive and Gram-negative organisms requires the product of the arsC gene, an arsenate reductase that generates arsenite, the substrate of the resistance pump (4 -6). The reaction catalyzed by the ArsC enzyme of plasmid R773 requires reduced glutathione and glutaredoxin, suggesting that reduction involves sulfhydryl chemistry. Recently Cys-12 was shown to be essential for catalytic activity of enzyme (7). The pK a value of free cysteine in solution is 8.3, considerably higher than pH optimum of the ArsC-catalyzed reaction, which is in the range of 6.3 to 6.6 (6). The pK a value of cysteine residues in the active site of a number of enzymes have been shown to be lowered by the formation of an ion pair with a proximate basic residue (8 -10). ArsC reductase activity was inhibited by treatment with diethylpyrocarbonate (DEPC), 1 suggesting the involvement of histidine residues in ArsC function, perhaps in ion pairing with Cys-12. For that reason the role of histidine residues was investigated.
ArsC has only two histidine residues, His-8 and His-88. The codons for these two residues were individually changed by site-directed mutagenesis. The phenotype of cells expressing arsC H88X mutants (where X can be one of several substitutions) was indistinguishable from wild type, and the corresponding purified gene products each had wild type ArsC properties. Cells expressing arsC H8X mutants lost arsenate resistance, and the purified gene products were each catalytically inactive. When the ArsC enzymes were reacted with DEPC, there was a correlation between the rate of DEPC inactivation and formation of N-carbethoxylated His-8. The pK a of Cys-12 thiolate was nearly 2 units lower in the wild type enzyme than in a H8G enzyme. These results indicate that His-8 is an essential residue of the reductase, with one critical function being to lower the pK a of Cys-12.

EXPERIMENTAL PROCEDURES
Bacterial Strains, Plasmids, and Media-The strains and plasmids used in this study are given in Table I. For protein production cells were grown at 37°C with shaking in 2 ϫ YT medium (11). For phenotypic measurements of arsenate resistance E. coli strain AW10 (⌬ars::cam) (12) carrying the indicated plasmids were grown in a low phosphate medium (5). Sodium arsenate was added at the indicated concentrations. In this strain the chromosomal ars operon was disrupted to produce hypersensitivity to arsenate. Turbidity at 600 nm was measured after 8 -12 h of growth at 37°C with shaking. Appropriate antibiotics were added as required.
Construction of Mutant arsC Genes-Mutations in the sequence of the arsC gene were introduced by site-directed mutagenesis using the Altered Sites in vitro mutagenesis system (Promega). The arsC gene inserted into the multiple cloning site of pALTER-1 vector (Promega) was used as the template (7). Degenerate oligonucleotides were used to introduce mutations in the codons for His-8 and His-88. The identity of the mutations was confirmed by DNA sequencing of each mutant gene (14). Single-stranded plasmid DNA was prepared using a QIAGEN DNA purification system (Qiagen Inc., Chatsworth, CA). Sequencing was performed by using the internal DNA labeling (Cy5 -dATP labeling mix) with the T7 DNA polymerase Cy5 AutoRead Sequencing kit and an ALF-express DNA Sequencer (Pharmacia Biotech Inc).
To construct the double mutants arsC H8R/C106S and arsC H88R/C12S , plasmids containing the arsC H8R , arsC C12S , arsC H88R , or arsC C106S genes were individually digested with restriction enzymes AlwNI and Eco47III. The fragments containing the two desired mutations were ligated together and transformed into E. coli strain JM109, with selection for ampicillin resistance. The mutations were confirmed by DNA * This work was supported by United States Public Health Service Grant GM52216. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18  Purification of the ArsC Protein and Assay for the ArsC Arsenate Reductase Activity-ArsC enzymes were purified and reductase activity assayed as described previously (6). ArsC concentration was determined from the absorbance at 280 nm using a molar extinction coefficient of 3960 M Ϫ1 cm Ϫ1 (7), by a micromodification of the method of Lowry et al. (15) or using a Coomassie protein assay (Bio-Rad).
Modification of ArsC with Diethylpyrocarbonate-Prior to use diethylpyrocarbonate was diluted with anhydrous ethanol and kept under nitrogen. The concentration of DEPC was determined spectrophotometrically with 10 mM imidazole using a molar extinction coefficient of 3200 M Ϫ1 cm Ϫ1 (16,17). Dithiothreitol was removed from the protein prior to reaction with DEPC with a Sephadex G-25 spin column (18). The reaction was initiated by adding varying concentrations of DEPC to the enzyme under nitrogen in a buffer consisting of 50 mM MOPS/MES, pH 6.5, or 10 mM sodium or potassium phosphate buffer, pH 6.7, as noted, at 25°C. At intervals portions of 10 -20 l were removed and diluted into 1 ml of the arsenate reductase assay buffer for estimation of enzymatic activity. The rate of inactivation as a function of pH was fit to the Henderson-Hasselbalch equation with Enzfitter (Elsevier-BIO-SOFT, Cambridge, UK).
To reverse DEPC inactivation DEPC-inactivated ArsC (50 M) was rapidly diluted 2-fold with 0.1 M sodium phosphate buffer, pH 7.0, containing 1 M hydroxylamine. After incubation at 25°C for the indicated times, excess reagents were removed with a spin column, and enzymatic activity was assayed. As a control unmodified ArsC was given the same treatment.
Determination of Modified Amino Acid Residues-The number of N-carbethoxyhistidine residues formed during the reaction with DEPC was determined by difference spectroscopy at 240 nm. Tyrosine modification was estimated from the decrease in absorbance at 278 -280 nm in the presence and absence of 6 M guanidine hydrochloride (19). The number of cysteine residues in ArsC was estimated from the reaction with 5,5Ј-dithiobis(2-nitrobenzoic) acid (DTNB) in 6 M guanidine hydrochloride using molar extinction coefficient of 13,600 M Ϫ1 cm Ϫ1 (20).
Modification of ArsC with DTNB-Dithiothreitol was removed from the enzyme with a Sephadex G-25 spin column before reaction of ArsC with DTNB. Varying amounts of a stock solution of 1 mM DTNB in 10 mM sodium phosphate buffer, pH 7.2, was added to 1 ml of the same buffer at 25°C containing 50 M ArsC, and the absorbance at 412 nm was measured as a function of time. Samples of 10 -20 l were removed at intervals for determination of residual reductase activity. pH titra-tions were conducted in a buffer containing 1 mM each of sodium citrate, sodium borate, and sodium phosphate adjusted to the indicated pH values with NaOH or HCl. The reported pH values were determined at the end of the reaction. The rate of inactivation as a function of pH was fit to the Henderson-Hasselbalch equation with Enzfitter.
Determination of the pK a of Cys-12 Thiolate-Since the thiolate anion has greater absorbance at 240 nm than the unionized thiol group, difference spectroscopy can be used to measure the thiolate concentration as a function of pH (21,22). The pK a was determined from the absorption of ArsC (1-2 mg/ml) in a buffer containing 1 mM each of sodium citrate, sodium borate, and sodium phosphate adjusted to the indicated pH values with NaOH or HCl. Since the C12S ArsC lacks the thiolate of interest, the pK a of Cys-12 could be determined from the difference in absorbance at 240 nm between this protein and ArsC proteins containing Cys-12, fitting the data to the Henderson-Hasselbalch equation with Enzfitter.
Materials-All restriction enzymes and nucleic acid modifying enzymes were purchased from Life Technologies Inc. The Altered Sites in vitro Mutagenesis System was obtained from Promega. Oligonucleotides were synthesized in the Macromolecular Core Facility of Wayne State University, School of Medicine. All other chemicals were obtained from commercial sources.

RESULTS
Inactivation of the ArsC Arsenate Reductase by Diethylpyrocarbonate-Reductase activity was rapidly inactivated by treatment with DEPC, with t1 ⁄2 ϭ 3.5 min (Fig. 1). Inactivation was a pseudo-first order process with respect to DEPC. At 20 mM sodium arsenate gave substantial protection, increasing t1 ⁄2 to 20 min. Phosphate, a weak competitive inhibitor, gave slight protection at 50 mM (t1 ⁄2 ϭ 5.5 min).
Mutagenesis of Histidine Codons in arsC-DEPC is known to modify histidine residues in proteins. His-8 and His-88 was individually changed to several different residues by site-directed mutagenesis. Cells of E. coli strain AW10 bearing the mutated arsC genes on one plasmid and wild type arsA and arsB genes on a compatible plasmid were examined for arsenate resistance (Fig. 2). AW10 (⌬ars::cam) is sensitive to 0.1 mM sodium arsenate (12). Cells expressing the wild type arsC gene in trans with the arsA and arsB genes are resistant to at least 10 mM sodium arsenate (shown only to 1.5 mM in Fig. 2), but cells bearing only arsC gene or only arsA and arsB genes are sensitive to 0.1 mM sodium arsenate (7). Cells bearing the arsC H8S , arsC H8P , arsC H8G , or arsC H8R genes with arsA and arsB genes in trans were sensitive to arsenate ( Fig. 2A). In contrast, cells expressing arsC genes with mutations in the codon for His-88 were as resistant to arsenate as the wild type (Fig. 2B). Two double mutants were constructed by combining individual mutations by molecular cloning: arsC H8R/C106S and arsC C12S/H88R . Cells bearing either double mutant were sensitive to 0.1 mM sodium arsenate (data not shown).
Analysis of the Altered ArsC Proteins-The steady state level of expression of the altered ArsC proteins was approximately the same as a wild type ArsC protein, as was determined by Western blot analysis using ArsC-specific antibody (data not shown). No differences were observed in the mobility on SDSpolyacrylamide gel electrophoresis between wild type and any of the altered ArsC proteins. The purification procedure produced about the same amount of each mutant protein at the same level of purity as wild type ArsC (data not shown). The elution from a molecular sieve column and circular dichroism spectra were essentially identical for all of the altered proteins (data not shown). Although local changes in structure cannot be excluded, there do not appear to be gross alterations in tertiary structure of the ArsC enzymes used in this study.
Arsenate Reductase Activity of the Altered ArsC Proteins-Purified H88W, H88S, H88V, and H88R enzymes each had nearly the same specific activity and pH dependence of arsenate reductase activity as wild type enzyme (data not shown). In contrast, no measurable activity was found for the H8P, H8G, H8R, H8R/C106S, and C12S/H88R enzymes, even at a protein concentration 100-fold greater than wild type. The assay is not sufficiently sensitive to detect activity less than 10 Ϫ3 that of the wild type, so the possibility that these enzymes have very low turnover rates cannot be excluded. These results are consistent with the phenotypic properties of the mutants (Fig.  2) and suggest that His-88 in ArsC protein is not involved in reduction. On the other hand, His-8 appears to be required for activity.
Inactivation of H88R by DEPC-Since H88R exhibited wild type ArsC properties, this protein was used for further characterization of the reaction with DEPC. The slopes of the lines from the concentration dependence of DEPC inactivation as a function of time (Fig. 3A) were used to determine values for the pseudo-first order rate constants of the inactivation, k obs , as described (23). A linear relationship of k obs with DEPC concentration was observed (Fig. 3B).
pH Dependence of DEPC Inactivation of H88R-The pH-dependent increase of k obs at 1 mM of DEPC over the pH range from 5.5 to 8.2 was determined (Fig. 3C). The experimental values fit a theoretical curve calculated for pK a of 7.0. The data indicate that the unprotonated form of an ionizable group reacts with DEPC to inactivate the enzyme. The values of k obs were independent of the buffer concentration over a range of 5 mM to 0.1 M sodium or potassium phosphate at pH 6.5, demonstrating that the observed pK a was not due to titration of buffer components. While a pK a of 7.0 is about 1 pH unit higher than that of free histidine in solution, it most likely reflects the ionization of the imidazolium group of His-8, as the data below indicate.
Kinetics of DEPC Modification-Binding of diethylpyrocarbonate with ArsC was monitored spectrophotometrically. The differences in absorbance between unmodified and DEPC- treated ArsCs was determined over the range of 200 to 300 nm. The peak of absorbance at about 240 nm was used to calculate the concentration of N-carbethoxylated histidine residues (Fig.  4A). For the wild type enzyme two phases were observed (Fig.  4A). In the fast phase of formation of N-carbethoxylated histidine reductase activity was not greatly affected. In the slow phase activity was lost (Fig. 4B). In completely inactivated ArsC 1.9 histidine residues were modified. Modification of tyrosine residues was detected spectrophotometrically from the difference spectra in the range of 278 to 283 nm, and no loss of tyrosine was observed (data not shown). Similarly, the two cysteine residues were unmodified, as determined by titration with DTNB (data not shown). Thus, under these experimental conditions no cysteine or tyrosine residues were modified.
These results indicate that both histidine residues, His-8 and His-88, are modified by DEPC. To determine which histidine residue reacted in the fast phase and which in the slow phase, the time courses of modification were measured in single histidine proteins. The H8R-substituted ArsC, which is inactive and has only His-88, reacted rapidly with DEPC (Fig. 4A). The active H88R enzyme, with only His-8, lost activity at the same rate as the wild type enzyme (Fig. 4B), and His-8 was Ncarbethoxylated slowly (Fig. 4A). Preincubation with the substrate arsenate protected H88R from inactivation by DEPC (data not shown). In contrast, the product, arsenite, had little effect. These results suggest that modification of His-8 leads to loss of reductase activity.
The reductase activity of H88R reacted with a 10-fold excess of DEPC could be fully restored by treatment with neutral hydroxylamine for 2 h (Table II). During the same period the disappearance of a single N-carbethoxylated histidine was observed (data not shown). When the enzyme was incubated with a 100-fold excess of DEPC, it was completely inactivated, and only partial restoration of activity by hydroxylamine with time was observed (Table II). Restoration of activity eliminates possible modification of the lysines and the N terminus of the enzyme in DEPC inactivation, since modification of amino groups are not readily reversed with hydroxylamine. These results are consistent with modification of His-8 being solely responsible for inactivation of reductase activity by diethylpyrocarbonate.
Determination of the pK a of the Cys-12-The rate of inactivation of the C106S enzyme, which has only Cys-12, by DTNB was measured (Fig. 5A). The dependence of k obs on the concentration of DTNB was linear (Fig. 5B). The pH-dependent increase of k obs at 25 M DTNB indicates a group with a pK a of 6.5 is involved in the reaction (Fig. 5C). DTNB is specific for cysteine thiolates. Since the only thiol present in C106S is Cys-12, that thiolate has a pK a of 6.5. Values for k obs were determined from the slopes of semilogarithmic plots of A. C, dependence of DTNB inactivation rate constant on pH. C106S was incubated with 25 M DTNB in a buffer consisting of 1 mM each of sodium citrate, sodium borate, and sodium phosphate adjusted to the indicated pH values with HCl or NaCl, as described under "Experimental Procedures." Pseudo-first order rate constants for the inactivation (k obs ) were calculated from the slopes of semilogarithmic plots of activity against the time of preincubation at varying pH values. The line is a theoretical curve calculated assuming pK a ϭ 6.5.

FIG. 4. Stoichiometry of DEPC modification of ArsC reductases.
ArsCs (50 -500 M) were incubated with 1 mM diethylpyrocarbonate in 10 mM sodium phosphate buffer, pH 6.5. At the indicated times the number of modified histidine residues were determined spectrophotometrically at 240 nm using ⑀ ϭ 3,200 M Ϫ1 cm Ϫ1 (A) and residual enzymatic activity was assayed (B), both as described under "Experimental Procedures." A, wild type (å); H8R (f); H88R (ࡗ). B, wild type (å); H88R (f). a 0.03 mM H88R was inactivated with either 0.3 or 3 mM DEPC for 25 min prior to treatment with hydroxylamine. At the indicated times portions of 50 l of the reaction mixture were passed through a Sephadex G-25 spin column, and reductase activity was assayed as described under "Experimental Procedures." The pK a of Cys-12 could also be determined from the absorbance of the thiolate anion at 240 nm. In these experiments the difference in absorbance between the fully reduced enzymes and the C12S enzyme, which lacks Cys-12, yields the specific absorbance of the Cys-12 thiolate. From the pH dependence of absorbance at 240 nm, the pK a values of Cys-12 in wild type ArsC by this assay were calculated to be 6.4 (Fig. 6). The same pK a for Cys-12 was found in the H8R enzyme, where a positive charge is maintained at residue 8. In contrast, the pK a of Cys-12 in the H8G enzyme, where the positive charge at position 8 is lost, was 8.3. The pK a of His-8 in the H88R C12S enzyme was decreased only 0.4 pH units compared with the H88R enzyme, from 7.0 to 6.6 (data not shown). While reciprocal changes in the pK a values of His-8 and Cys-12 would be expected if those two residues formed a charge pair in a closed system, in the enzyme the ionization of the imidazolium group of His-8 could be affected by other acidic residues.

DISCUSSION
The 141-residue ArsC enzyme reduces arsenate (As(V)) to arsenite (As(III)), the substrate of the ATP-coupled Ars pump that extrudes arsenite from cells of E. coli (6). Coupling of the reductase to the pump expands the range of resistance to include both the oxidized and reduced oxyanions of arsenic. ArsC has a single essential cysteine residue, Cys-12 (7). In this report we demonstrate that reductase activity was inactivated by diethylpyrocarbonate, a histidine modifying reagent. ArsC has just two histidine residues, His-8 and His-88. By sitedirected mutagenesis using degenerate oligonucleotides, mutations were introduced into the codons for His-8 and His-88. Cells expressing four different mutations in the codon for His-88 were each arsenate-resistant, and the purified substituted His-88 ArsC enzymes had wild type catalytic activity.
Thus His-88 is unrelated to catalysis. Cells bearing four different arsC genes with mutations in the codon for His-8 were each arsenate-sensitive. Three of the gene products were purified and shown to be catalytically inactive.
The kinetics of DEPC modification indicated that both histidine residues were modified. Through the use of altered proteins having only His-8 or His-88, it was shown that His-88 reacted rapidly with DEPC. His-8 reacted more slowly but with complete loss of activity. The activity of the DEPC-modified enzyme could be restored by treatment with neutral hydroxylamine, and spectroscopic determination of N-carbethoxyhistidine residues demonstrated regeneration of His-8 by treatment with hydroxylamine.
The substrate arsenate protected His-8 from DEPC modification, suggesting that it is in or near the active site. Since the source of reducing equivalents for arsenate reduction are glutathione and the thiol carrier glutaredoxin (6), it is reasonable that a cysteine residue would be involved in catalysis, and Cys-12 has previously been identified as an active site residue (7). What is the relationship between His-8 and Cys-12? The reductase has a pH optimum of about 6.5, considerably lower than the pK a of 8.3 for free cysteine. Other proteins with active site cysteine thiolates have acidic pH optimum, and in such proteins there is a basic residue near the cysteine that forms a stable ion pair with it (8 -10). Therefore the possibility of His-8 and Cys-12 forming a thiolate-imidazolium charge pair was investigated. The pK a of Cys-12 was found to be approximately 6.5, the same as the pH optimum of the reductase reaction. When another basic residue, arginine, was substituted for His-8, the pK a remained the same. Although catalytic activity is not maintained in the H8R enzyme, charge pairing is. When a neutral residue, glycine, was substituted for His-8, the pK a of Cys-12 increased to 8.3. These results are consistent with His-8 and Cys-12 forming a charge pair that lowers the pK a of the cysteine thiolate. This increases the nucleophilicity of Cys-12 at the pH optimum of the enzyme.