INTRODUCTION

Top Abstract Introduction Procedures Results & Discussion References

The photosynthetic bacterium Rhodospirillum rubrum is able to utilize CO as its sole energy source when grown anaerobically in the dark (1). The key enzyme in this metabolism is the CO-induced, carbon monoxide dehydrogenase (CODH),¹ which is the product of the cooS gene (2, 3). CODH catalyzes the reversible oxidation of CO to CO₂ and the reduction of a special ferredoxin that is the cooF gene product (3, 4). CODH has only two metal clusters, which have been designated as centers B and C (5). The latter contains a nickel atom bridged by an unknown ligand to an iron atom in one of the two [Fe₄S₄]^2+/1+ clusters found in CODH (5, 6). The second cubane, which is not linked to nickel, is the sole component of center B. CO oxidation occurs at center C, and center B is proposed to mediate the transfer of electrons to CooF or artificial electron acceptors such as methyl viologen (4, 7, 8). When purified from R. rubrum grown on nickel-depleted medium, CODH lacks both nickel and the ability to oxidize CO, yet retains both [Fe₄S₄]^2+/1+ clusters (5, 9). Nickel-deficient CODH can be fully activated by treatment with Ni²⁺.

CODH from R. rubrum is closely related to another class of anaerobic CO-oxidizing enzymes that, unlike CODH, also catalyze the synthesis of acetyl coenzyme A from CO, coenzyme A, and a methyl donor. The CO-oxidizing acetyl-CoA synthase (CODH/ACS), found in acetogenic and methanogenic bacteria, possesses center B and center C analogs as well as a third and unidentified redox center, which has been invoked in recently proposed mechanisms for the enzymatic oxidation of CO (25, 28). CODH/ACS also contains a second NiFeS cluster (center A) that serves as the site of acetyl-CoA synthesis (for a recent review of CODH/ACS, see Ragsdale and Kumar (10)). Unlike the monomeric CODH from R. rubrum, the CODH/ACS enzymes are composed of multiple subunits among which the CODH and ACS activities are divided. In the Clostridium thermoaceticum and Methanothrix soehngenii enzymes, the CODH activity appears to be localized in the  (CmbB) and  (CdhA) subunits, respectively (3, 11). This is consistent with the significant sequence similarity of cmbB and cdhA genes to the cooS gene, which encodes the single subunit of the R. rubrum CODH (3).

Although the structures of neither R. rubrum CODH nor the CODH/ACS enzymes are known, analysis of the deduced amino acid sequences of cooS, cmbB (24), and cdhA (21) as well as two related genes (22, 23) from genome sequences allows the identification of conserved motifs and amino acid residues. In the absence of structural information, site-specific substitution of conserved amino acid residues is most likely to provide insight into the immediate environment of the metal clusters of CODH. Of the conserved residues, histidine 265 of the R. rubrum CODH was chosen for substitution as it is the only histidine conserved in all known CooS-like gene products. The selection of histidine was also guided by previous extended x-ray absorption fine structure studies that suggest one or more nitrogen-donor ligands to the nickel center of R. rubrum CODH (6).

In this study, the effect of substitution of valine for histidine 265 resulted in an intact CODH enzyme (designated as H265V) which partially lacked nickel and had a dramatically lower activity. The ability of nickel-deficient H265V CODH to incorporate nickel in vitro is shown here, and the catalytic activity and binding of CO and cyanide to nickel-treated H265V CODH has been investigated. The nickel-containing form of the enzyme does not exhibit the unique C_red1 and C_red2 EPR signals attributed to reduced forms of center C that were characterized in previous spectroscopic studies of CODH and CODH/ACS enzymes (see Hu et al. (5) and references therein). From these studies, we have developed the hypothesis that the ligand-mediated, electronic coupling between nickel and iron in center C has been disrupted or altered by the amino acid substitution. This does not adversely affect substrate or inhibitor binding; however, the modification greatly diminishes the ability of the enzyme to oxidize bound substrate.

	EXPERIMENTAL PROCEDURES

Top Abstract Introduction Procedures Results & Discussion References

Site-directed Mutagenesis of cooS-- The conversion of CODH His-265 to Val used the restriction site elimination method (12) with some modifications (13). Standard media, antibiotic usage, and mating protocols were employed (14). The pUC19-derived template plasmid (pCO11, in Escherichia coli strain UQ1155) bears a 2.2-kb SalI fragment that encompasses all of cooS, the gene encoding CODH. This template was mutagenized in vitro using a selection primer that converted an AflIII site to a BglII site in the vector, with a mutagenic primer that converted the His-265 codon (CAT) to a Val (GTT), and simultaneously created a HpaI restriction cleavage site (GTTAAC). The mutated cooS region was excised as a 1.2-kb NcoI-HindIII fragment, then cloned into similarly cut pCO12 (E. coli strain UQ1161), which bears coo DNA extending 2.0 kb upstream of the NcoI site and 1.6 kb downstream of the HindIII site. The entire coo region was subsequently introduced (as a PvuII fragment) into the mobilizable vector pUX19 cut with HincII. This construct, in E. coli S17-1 (UQ324), was mated to R. rubrum UR485 (cooS::aacC1), and the desired Km^rGm^rNx^r recombinant was isolated. The merodiploid recombinant was resolved by growth for several generations in SMN liquid culture supplemented with Nx, followed by isolation of a Km^sGm^s derivative, indicative of homologous recombination. The construct was verified by Southern analysis of chromosomal DNA, and the sequence of one strand of the NcoI-HindIII fragment cloned into pCO12 was verified (data not shown).

Cell Growth and Enzyme Purification-- Nickel-supplemented cell cultures of R. rubrum, strain UR499, were grown in medium (15) supplemented with 0.05 mM NiCl₂ and 3 g/liter each of yeast extract and casamino acids. Nickel-depleted cultures were grown according to previously published methods (9). Enzyme purification was performed according to the method of Bonam and Ludden (15). All buffers used in the purification of nickel-deficient H265V from nickel-depleted cultures were passed over a metal-chelating column of Bio-Rad Chelex-100 cation exchange resin, and all buffer solutions contained 1 mM EDTA.

Assays-- Protein concentrations were determined by the bicinchoninic acid colorimetric method using bovine serum albumin (grade A, Sigma) as a standard (16). The CO oxidation activity of H265V was measured by the CO-dependent, methyl viologen reduction assay (7, 9). One unit of activity equals 1 µmol of CO oxidized per min. The catalytic activity of the reverse reaction, CO₂ reduction, was measured according the method described by Ensign (17). For the reverse reaction, 1 unit of activity equals 1 µmol of CO produced per min. SDS-polyacrylamide gel electrophoresis and immunoblot analyses were performed according to published methods (18).

Sample Preparations-- All samples were prepared in an anaerobic glove box (Vacuum/Atmospheres Dri-Lab glove box model HE-493) with an N₂ atmosphere containing less than 1 ppm O₂. The buffer used in each experiment was 100 mM MOPS that was adjusted to pH 7.3 with NaOH unless otherwise stated. Trace metal contaminants in buffers were removed by passing buffer solutions through a column of Bio-Rad Chelex-100.

Assay of Purified H265V in the Presence of Divalent Metals-- The ability of divalent metal cations to activate nickel-deficient H265V was tested by adding solutions of CoCl₂, ZnCl₂, MnCl₂, NiCl₂, or Fe(II) to solutions of nickel-deficient enzyme containing equimolar amounts of methyl viologen and sodium dithionite or titanium(III) citrate (7, 19). Activity was measured versus time following the addition of metal salt. The K_m for nickel activation was determined according to published methods (19).

Preparation of Nickel-treated H265V-- Nickel-treated samples of H265V, freed from exogenous nickel, were prepared according to published methods (7).

Metal Analysis-- Cobalt, iron, manganese, nickel, and zinc were measured by inductively coupled plasma mass spectrometry at the University of Wisconsin Soil and Plant Analysis Laboratory. Iron and nickel were additionally measured using a Perkin-Elmer 3030 atomic absorption spectrometer. Enzyme samples were desalted on a column of Sephadex G-25 equilibrated in 25 mM MOPS that was adjusted to pH 7.3 with KOH. A second column of Bio-Rad Chelex-100 anion exchange resin, equilibrated in the same buffer, was employed as a final preparation step to remove trace quantities of adventitiously bound metals.

Cyanide Binding and Inhibition-- The forward, second-order rate constant for cyanide inhibition and the dissociation constants for CO and cyanide were determined according to previously published methods (8).

Oxidation and Reduction of H265V-- Oxidized H265V was prepared by adding a slight excess of thionine (E°' = 56 mV) or indigo carmine (E°' = 125 mV) to enzyme samples desalted on Sephadex G-25 to first remove excess dithionite. Indigo carmine-oxidized, nickel-deficient samples of H265V prepared for EPR and UV-visible spectroscopy were passed through columns of Bio-Rad AG 1X8 and Chelex-100 cation-exchange resin to remove adventitiously bound dye (7) and metals, respectively. Enzyme samples were fully reduced by adding 0.1 M sodium dithionite to a final concentration of 2 mM. Fully reduced, dithionite-free H265V CODH was oxidatively titrated with a 2.4 mM standardized thionine solution and analyzed by EPR spectroscopy.

Time-dependent Reduction of H265V by CO-- The CO-dependent reduction of the [Fe₄S₄]^2+/1+ clusters was followed over time by measuring the loss in absorbance at 420 nm of solutions containing oxidized H265V that were incubated under a gas phase of 100% CO (7). Extinction coefficients previously reported for oxidized and reduced wild-type CODH are 34.3 and 20.1 mM¹ cm¹, respectively, at 420 nm (7).

EPR and UV-visible Spectroscopy-- EPR spectra were recorded using a Varian E-15 spectrometer equipped with an Oxford Instruments ER910A cryostat in the laboratory of Prof. George Reed in the Enzyme Institute, University of Wisconsin-Madison. UV-visible spectra of samples anaerobically sealed in quartz cuvettes were recorded using a Shimadzu UV-1601PC spectrophotometer.

	RESULTS AND DISCUSSION

Top Abstract Introduction Procedures Results & Discussion References

Purification of H265V CODH-- In R. rubrum strain UR499, the H265V CODH was predominantly membrane-associated and accumulated to levels equivalent to those observed for CODH in wild-type strain UR2. These conclusions were based upon estimates of the amount of CODH protein in extracts of strain UR499 that were analyzed by immunological methods using anti-CODH antibody (data not shown). Thus it appeared that H265V CODH accumulated to normal levels and associated with its in vivo electron acceptor, CooF, which has been shown to serve as the membrane anchor for CODH. H265V CODH behaved like the wild-type enzyme in all steps of the purification except the heat treatment. Only 20% of H265V CODH was recovered following heating for 7 and 5 min at 60 and 80 °C, respectively, compared with a 60% yield reported for wild-type CODH (15). This suggested that H265V CODH might be less heat stable than the wild-type enzyme.

Metal Content of H265V CODH-- The metal content of H265V CODH isolated from UR499 cultures grown in nickel-supplemented or nickel-depleted media were compared with the metal content of CODH from wild-type cultures grown similarly (Table I). Also included in the comparison is H265V CODH (from nickel-supplemented cells) that was treated with Ni²⁺ following purification. All forms of H265V contain 8 iron atoms, consistent with the presence of an intact center B and Fe₄S₄ component of center C. Thus, the substitution of valine for histidine 265 has no effect on the ability of the enzyme to ligand a full compliment of iron; UV-visible spectra of H265V CODH confirms this conclusion (see below).

Treatment of Nickel-deficient H265V CODH with Divalent Metal Cations-- As mentioned above, the activity of as-purified H265V CODH increased from 0.7 to 2.4 units/mg following treatment with Ni²⁺. Likewise, the activity of nickel-deficient H265V CODH increased from 0.01 to 2.2 units/mg when treated with Ni²⁺ over a 3-h period (Fig. 1). Nickel-deficient H265V CODH was also treated with Co²⁺, Zn²⁺, Fe²⁺, or Mn²⁺. Of these, only Co²⁺ gave a form of the enzyme having measurable activity, which was approximately 20% of that observed with Ni²⁺ (Fig. 1; data for Zn²⁺, Fe²⁺, and Mn²⁺ not shown). This is in contrast to the 1000-fold greater increase in activity produced by Ni²⁺ versus Co²⁺ treatments of the nickel-deficient wild-type enzyme (7). The activity of the cobalt-containing wild-type CODH, judged by the rate at which the Fe₄S₄ are reduced by CO (7), is similar to the activity levels measured in this work for the nickel and cobalt-treated forms of H265V CODH. It seems interesting, therefore, that the amino acid substitution appears to greatly impair the activity of the nickel but not the cobalt-containing form of CODH. Because the K_m for CO for nickel-treated H265V CODH is similar to that of the wild-type enzyme (Table II; discussed below), low activity does not reflect an inability to bind CO; rather, it suggests an inability to catalyze the oxidation of CO that is bound to center C of H265V CODH.

View larger version (13K): [in this window] [in a new window]   Fig. 1.   Time-dependent increase in activity following treatment of nickel-deficient H265V CODH (0.01 nickel atom per monomer) with Ni2+ or Co2+. At t = 0, a solution of CoCl2 (), NiCl2 (), or buffer (, control) was added to buffer solutions containing enzyme, methyl viologen, sodium dithionite and titanium(III) citrate (7). At the indicated time points, 15-µl aliquots were removed and assayed for activity by the CO-dependent, methyl viologen reduction assay. Mn2+, Zn2+, and Fe2+ were also tested; however, treatments of nickel-deficient H265V CODH with these metal cations did not produce measurable levels of activity (data not shown). Concentrations were as follows: CODH, 3.1 µM; metal salts, 5 mM; methyl viologen, 0.4 mM; sodium dithionite, 0.1 mM; and titanium(III) citrate, 0.3 mM.

UV-visible Spectra of H265V CODH-- UV-visible absorption spectra of oxidized and reduced H265V CODH were found to be identical to those of wild-type CODH (data not shown), demonstrating that histidine 265 is not required for assembly of the Fe₄S₄ clusters. The molar extinction coefficients at 420 nm (₄₂₀) for oxidized and reduced H265V CODH were 36.2 and 20.1 mmol¹ cm¹, respectively. These values are essentially identical to those of the wild-type enzyme (7). Nickel does not contribute to the UV-visible spectrum of the enzyme (7).

View larger version (9K): [in this window] [in a new window]   Fig. 2.   Time-dependent reduction of Fe4S4 by CO. The reduction of H265V CODH in a solution of CO-saturated buffer was determined by measuring the magnitude of absorbance loss at 420 nm. Fully oxidized, 3-ml samples of A, nickel-deficient H265V CODH (0.01 nickel atom per monomer), and B, nickel-treated H265V CODH (2.4 nickel atoms per monomer) were placed under 100% N2 in double-stoppered cuvettes. After measuring the initial absorbance at 420 nm (100% oxidation), the cuvettes were evacuated and flushed with 100% CO and mixed for three minutes before resuming the time-course readings at 420 nm. The absorbance of fully-reduced (i.e. 0% oxidation) samples was measured following the addition of sodium dithionite. CODH concentrations were A, 3.9 µM, and B, 3.6 µM.

Activity of H265V CODH-- As noted above, nickel-containing H265V CODH has at least 1000-fold lower catalytic activity compared with the wild-type enzyme. Both the forward (CO oxidation) and reverse (CO₂ reduction) activities were decreased by similar factors (Table I) although the K_m values for CO and methyl viologen for H265V CODH were not significantly different from those of the wild-type enzyme (Table II). With the exception of V_max, the kinetic parameters of H265V CODH were very similar to those of the wild-type enzyme. Therefore, we conclude that H265V CODH binds substrates with normal affinities but somehow lacks the ability to catalyze the oxidation of CO. The defect in activity does not appear to be in the ability to transfer electrons from center C to center B (or the reverse). If this were the case, one would expect that, upon treatment of the oxidized enzyme with CO, the Fe₄S₄ component of center C would become rapidly reduced followed by a very slow transfer of electrons to center B. This would be observed as an initially rapid decrease in A₄₂₀ due to reduction of one Fe₄S₄ center (center C), followed by a very slow decrease in A₄₂₀, reflecting the reduction of the second Fe₄S₄ center (center B). Based upon the data shown in Fig. 2, it appears that the defect occurs prior to or at the point of electron transfer to the Fe₄S₄ clusters of H265V CODH.

pH Dependence of CO Oxidation Activity by H265V CODH-- The mechanism of CO oxidation may involve the attack of a hydroxyl on bound CO, and it was suggested to us by Dr. S. Ensign that the pH profiles of the wild-type CODH and H265V CODH might, therefore, differ dramatically. The pH profile of H265V CODH activity showed a nearly linear increase from 1.3 units/mg at pH 6.5 to 4.2 units/mg at pH 9.0, and the activity level reached a plateau of approximately 4.6 units/mg from pH 9.5 to 11 (data not shown). This is nearly identical to the effect of pH on the activity of wild-type CODH, and the two pH profiles are nearly superimposable (data not shown). The defect in CODH activity would not appear to be an inability to generate a hydroxyl group for attack on metal-bound CO, otherwise, a much greater increase in activity at higher pH would be expected for the H265V CODH.

Inhibition of H265V CODH by Cyanide-- Cyanide is a slow, tight-binding, competitive inhibitor of wild-type CODH (8). Analysis of the inhibition of nickel-treated H265V CODH by cyanide in the absence or presence of CO yielded results similar to those previously reported for the wild-type enzyme (Table II). Both wild-type and H265V forms of the enzyme bind substrate and inhibitor with similar affinity; therefore, H265V CODH does not appear to suffer from any loss of binding affinity for substrate or substrate analog.

EPR Spectroscopy of H265V CODH-- The EPR spectrum of dithionite-reduced H265V CODH exhibits several overlapping S = 1/2 signals (Fig. 3). The most prominent has g values of 2.04, 1.92, and 1.88 (g_av = 1.94) and is characteristic of the [Fe₄S₄]⁺ form of center B (B_red) of wild-type CODH. Another signal having g values of 2.07, 1.92 and 1.88 (g_av = 1.96) is evident (Fig. 3A), and this signal appeared to lose intensity with increasing nickel content (Fig. 3, B and C) and, therefore, was presumed to be due to a reduced version of center C in H265V CODH that lacked nickel. This signal did not shift significantly following the addition of cyanide (data not shown), which is what one would anticipate considering that cyanide is not expected to bind to the active site lacking nickel (8). Note that cyanide strongly perturbs the C_red1 (g_av = 1.86) EPR signal (20, 26).

View larger version (22K): [in this window] [in a new window]   Fig. 3.   EPR spectra of H265V CODH reduced with dithionite. Ni-deficient H265V (A), as-purified H265V (B), Ni-treated H265V (C), and buffer plus wild-type CODH (D, control). EPR conditions were microwave power, 5 mW; frequency, 9.24 GHz; modulation amplitude, 12.5 G; temperature, 10 K.

View larger version (13K): [in this window] [in a new window]   Fig. 4.   EPR spectra of H265V CODH containing 2.4 mol of nickel per mol of enzyme (A) reduced with CO and (B) reduced with CO followed by KCN addition. EPR conditions were microwave power, 5 mW; frequency, 9.24 GHz; modulation amplitude, 12.5 G; temperature, 10 K.

Summary-- When purified from nickel-grown cultures, H265V CODH has a low nickel content that appears to result from a diminished capacity to incorporate nickel in vivo. When treated with either Co²⁺ or Ni²⁺ in vitro, the activity level of nickel-deficient H265V CODH increases, yet remains more than 1000-fold less active than does wild-type CODH. The kinetic data and slow rate of CO-dependent reduction of Fe₄S₄ clusters implicate a rate-limiting step that occurs prior to or at the point of electron transfer to the cubanes. The affinity of H265V CODH for substrates, however, is not diminished as a result of the substitution. The complete absence of g_av = 1.86 (C_red1) and g_av = 1.87 (C_red2) EPR signals suggest that the unique FCII subsite is not formed in the presence of nickel, perhaps as a consequence of the alteration of the ligand-mediated, electronic interaction between nickel and iron components of the center C of H265V CODH.