The diheme cytochrome c₄ from Vibrio cholerae is a natural electron donor to the respiratory cbb₃ oxygen reductase

Abstract

The respiratory chain of Vibrio cholerae contains three bd-type quinol oxygen reductases as well as one cbb₃ oxygen reductase. The cbb₃ oxygen reductase has been previously isolated and characterized, however the natural mobile electron donor(s) which shuttles electrons between the bc₁ complex and the cbb₃ oxygen reductase is not known. The most likely candidates are the diheme cytochrome c₄ and mono-heme cytochrome c₅, which have been previously shown to be present in the periplasm of aerobically grown cultures of V. cholerae. Both cytochromes c₄ and c₅ from V. cholerae have been cloned and expressed heterologously in E. coli. It is shown that reduced cytochrome c₄ is a substrate for the purified cbb₃ oxygen reductase and can support steady state oxygen reductase activity of at least 300 e⁻¹/s. In contrast, reduced cytochrome c₅ is not a good substrate for the cbb₃ oxygen reductase. Surprisingly, the dependence of the oxygen reductase activity on the concentration of cytochrome c₄ does not exhibit saturation. Global spectroscopic analysis of the time course of the oxidation of cytochrome c₄ indicates that the apparent lack of saturation is due to the strong dependence of K_M and V_max on the concentration of oxidized cytochrome c₄. Whether this is an artifact of the in vitro assay or has physiological significance remains unknown. Cyclic voltammetry was used to determine that the midpoint potentials of the two hemes in cytochrome c₄ are 240 mV and 340 mV (vs SHE), similar to the electrochemical properties of other c₄-type cytochromes. Genomic analysis shows a strong correlation between the presence of a c₄-type cytochrome and a cbb₃ oxygen reductase within the β- and γ- proteobacterial clades, suggesting that cytochrome c₄ is the likely natural electron donor to the cbb₃ oxygen reductases within these organisms. These would include the β-proteobacteria Neisseria meningitidis and Neisseria gonnorhoeae, in which the cbb₃ oxygen reductases are the only terminal oxidases in their respiratory chains, and the γ- proteobacterium Pseudomonas stutzeri.

The genome of Vibrio cholerae encodes four respiratory oxygen reductases (1). There are three bd-type oxygen reductases which use ubiquinol as the natural electron donor (2), and one cbb₃-type heme-copper oxygen reductase (3), which is presumed to use a cytochrome c as its natural electron donor. The motivation of the current work is to identify the natural electron donor for the cbb₃-type heme-copper oxygen reductase.

The cbb₃-type oxygen reductases are members of the heme-copper superfamily which includes enzymes that perform oxygen reductase chemistry or nitric oxide reduction chemistry (4–7). The vast majority of currently identified oxygen reductases within the heme-copper superfamily are classified as being in one of three families, the A-, B- and C-families (6, 7). The cbb₃-type oxygen reductases (7–11) are members of the C-family and represent over 20% of the sequences of heme-copper oxygen reductases from the currently known genome sequences. In the past decade, cbb₃-type oxygen reductases have been purified and characterized from several bacterial species, including Rhodobacter sphaeroides (5, 12–17), Paracoccus denitrificans (18), Rhodobacter capsulatus (19–23), Vibrio cholerae (3, 24), Bradyrhizobium japonicum (25–31), Rhodothermus marinus (32), Pseudomonas stutzeri (33–37), Helicobacter pylori (38, 39), and Sulfurihydrogenibium azorense (10). It has been demonstrated by mass spectrometry that the cbb₃-type oxygen reductases have the signature histidine-tyrosine cross linked co-factor within their active site, but that the tyrosine comes from a different transmembrane span than for the A-family and B-family heme-copper oxygen reductases (15, 24).

The physiological role of the cbb₃-type oxygen reductases is generally defined by a high affinity for O₂, allowing these enzymes to function at lower oxygen concentrations. Therefore these enzymes are often expressed under microaerophilic growth conditions (9). These enzymes are of particular interest because they are present in a number of pathogenic species and are proposed to be important for virulence by facilitating growth under conditions of low oxygen (40). The cbb₃-type enzymes are the only oxygen reductases in Helicobacter pylori (41), Neisseria gonorrhoeae (42) and Neisseria meningitidis (43) and are, therefore, potential drug targets for these significant human pathogens.

To understand the physiological roles of the cbb₃-type oxygen reductases, it is important to know the natural electron donor(s). In most cases, the natural electron donor is not known. One exception is the cbb₃-type oxygen reductase from H. pylori, where the natural electron donor has been shown to be cytochrome c₅₅₃ (39). When the electron donors are not known, the purified enzymes are usually assayed using the artificial electron donor TMPD (reduced by ascorbate) as the electron donor. In a few instances, reduced horse heart cytochrome c can function as an electron donor (25), though often these assays are also performed in the presence of TMPD.

The proteomes of V. cholerae strains have previously been analyzed for the presence of c-type cytochrome binding domains by searching for the canonical sequence CXXCH (44). A total of 48 CXXCH motifs were identified in 31 proteins, of which 14 have N-terminal signal sequences, indicating these are cytochrome c's and are components of the bacterial envelope. Under aerobic growth conditions, six major c-type cytochromes are visualized by SDS-PAGE analysis using a stain for covalently bound heme (44). Two of these six cytochromes c have been identified as the CcoO and CcoP subunits of the cbb₃-type oxygen reductase (44). A third heme-containing band was tentatively identified as the YecK subunit of biotin sulfoxide reductase (44). It is likely that one of the heme-staining bands corresponds to the PetC gene, encoding the cytochrome c₁ component of the bc₁ complex, but this remains to be shown. The best candidates for the remaining two heme-staining bands in aerobically grown V. cholerae are the cycA and cycB gene products, corresponding to cytochrome c₄ and cytochrome c₅, respectively. Each of these cytochrome c's is predicted to have a cleaved N-terminal signal sequence and to be located within the periplasm. The predicted molecular weights of the mature cytochromes c₄ and c₅ (19.8 kDa and 11.5 kDa) also make these reasonable candidates for two of the heme-staining bands from aerobically grown V. cholerae (44). The V. cholerae genome also encodes two additional c-type cytochromes of unknown function that are predicted to be periplasmic (44). Their predicted molecular weights (9.1 kDa and 8.5 kDa) do not match the major heme-staining proteins reported for aerobically grown V. cholerae (44). It is likely, therefore, that cytochrome c₄ and/or cytochrome c₅, might be an electron donor to the cbb₃-type oxygen reductase and responsible for shuttling electrons between the bc₁ complex and the heme-copper oxygen reductase. For this reason, the genes encoding the di-heme cytochrome c₄ and mono-heme cytochrome c₅ were each cloned, expressed heterologously in E. coli and the proteins purified.

It is shown that cytochrome c₅ does not function as an electron donor to the purified cbb₃-type oxygen reductase from V. cholerae, whereas cytochrome c₄ can support oxygen reductase activity at a rate of at least 300 e⁻¹/s at 25°C.

Materials and Methods

Construction of expression plasmids

A heterologous E. coli expression system (45) was used to obtain recombinant cytochromes c₄ and c₅ from V. cholerae. Following the procedure successfully employed previously for the expression of cytochrome c₅₅₂ from Thermus thermophilus (46), the DNA fragments encoding the predicted sequence of the mature cytochromes c₄ and c₅ were fused to a DNA fragment encoding the signal peptide (MKISIYATLAALSLALPAGA) of cytochrome c₅₅₀ from Thiobacillus versutus (47). The genes of the mature cytochrome c's were obtained by PCR from genomic DNA. The oligonucleotides used were as follows: forward primer 5'-CAAGGCGCCCAGGCCCAAGGTAGTATCGAAG -3' and reverse primer 5'-AAGGATCCCTAGTGTAGGCCACCTAC -3' for cycA; forward primer 5'-CAAGGCGCCCAGGCTCTAACTGAAGCCGATA-3' and reverse primer 5'-AAGGATCCTTACAGGCCTGCGATCATA-3' for cycB. The primers were designed for introducing a 5' NarI site and a 3' BamHI site in order to create a full-length chimeric T. versutus / V. cholerae cycA or cycB gene and cloned into pET17b.

In addition, the 24^th residue of cytochrome c₅ was changed from Leu, to Gln to enhance the recognition by the E. coli signal peptidase to cleave the new cutting site. Both cytochromes c₄ and c₅ are predicted to have cleaved signal peptides and to be soluble, periplasmic proteins. The predicted cleavage sites are after Ala₃₆ and Ala₂₃ for cytochromes c₄ and c₅, respectively. This hybrid construct was then cloned into the commercial expression vector pET17b.

Cell growth and enzyme expression and purification

Conditions of cell growth and enzyme expression were similar to those previous reported (46). The recombinant expression plasmids were utilized in a strain that also contained plasmid pEC86, which expresses maturation genes for cytochrome c in E. coli (ccmABCDEFGH)(27). The two plasmids were co-transformed into E. coli BL21(DE3) competent cells which were then streaked on a Luria-Bertani (LB) agar plate containing 50 μg/ml ampicillin and 30 μg/ml chloramphenicol (27, 46). The cells were grown in 1 L of culture medium in a 2.8 L Fernbach flask to A₆₀₀ = 0.8–1, and then expression of the cytochrome c was induced with 1 mM IPTG. After 7 hours, the cells were harvested by centrifugation at 7,000 × g for 15 min. Generally, about 10 g of cell paste was collected from each liter of growth media. The cell pellet was a reddish-brown color, indicating the overexpression of cytochrome c.

Protein purification

Each of the recombinant cytochrome c's was purified by column chromatography, using procedures similar to those used for the purification of recombinant cytochrome c₅₅₂ from T. thermophilus (46). Briefly, the collected cell pellet was homogenized in 25 mM Tris-HCl (pH 7.5), 8 mM MgSO₄, DNase I, 0.1% Triton X-100 and protease inhibitor cocktail. The suspended cell mixture was passed through a microfluidizer three times at a pressure of 20,000 psi. The cell debris was spun down at 8000 rpm for 30 min at 4□°C. The supernatant was loaded onto a 5×20 cm CM-52 cellulose column (Whatman) and eluted with a gradient of 0 to 1 M NaCl in 25 mM Tris-HCl (pH 7.5) buffer. Fractions which were reddish-brown were collected and concentrated using concentrators (Amicon) with YM-10 membranes. The concentrated protein was loaded onto a 2.6×100 cm gel filtration column (Sephacryl S-100 High Resolution; GE Healthcare) equilibrated with 100 mM potassium phosphate buffer (pH 7.0) at flow rate 0.4 ml/min. The protein was collected and dialyzed overnight against 25 mM Tris-HCl and then loaded onto the buffer-equilibrated 1.6×20 cm DEAE-5PW column (Toso-HaaS). The column was eluted using a gradient of 0 to 1 M NaCl in 25 mM Tris-HCl (pH 7.5) buffer. The eluted protein was dialyzed and concentrated again as described above and then flash-frozen in liquid nitrogen and stored at −80 °C. The final yield of the purified cytochromes c₄ and c₅ were about 15 mg and 5 mg per liter of cell culture, respectively.

SDS-PAGE analysis

The purified c-type cytochromes were analyzed using SDS-PAGE. Protein was visualized using Coomassie Blue and heme staining (48) was used to identify proteins containing covalently attached heme c. Pre-cast 15% SDS-PAGE gels from ISC BioExpress were used. After electrophoreses, the gels were then incubated in 15 mL of 6.3 mM 3,3`,5,5`-tetramethylbenzidine (TMBZ from Sigma) and 35 mL of 0.25 M sodium acetate, pH 5.0, for 1 h. The gels were then stained for heme by adding H₂O₂ to a final concentration of 30 mM.

Spectrophotometric Measurements

Spectra of the isolated cytochromes were acquired with a Shimadzu UV−vis-2101PC spectrophotometer. The enzyme sample (3–17 μM) was examined in 25 mM Tris-HCl buffer at pH 7.5. The enzymes were oxidized with 2 μL of 1 mM Fe(CN)₆ and reduced with a small amount of solid dithionite, both obtained from Sigma. Spectra were scanned from 375 nm to 800 nm. The concentrations of the cytochrome c's were estimated from the spectra of the reduced cytochromes using extinction coefficients of 40 mM⁻¹cm⁻¹ at 553 nm for the di-heme cytochrome c₄ (49, 50) and 20 mM⁻¹cm⁻¹ at 555 nm for the mono-heme cytochrome c₅ (51).

Cyclic voltammetry

Reduction potentials were measured using protein film voltammetry (52, 53). The cytochrome c₄ was applied directly to a freshly polished pyrolytic graphite-edge (PGE) electrode surface and then placed into solution in a thermostated all-glass cell encased in a Faraday cage. Since oxygen does not interfere with measurements in the range of 100 mV–500 mV (vs. SHE), measurements were performed aerobically. Analogue-scan cyclic voltammetry was performed using a Bioanalytical Systems (West Lafayette, IN) CV-27 voltammograph with an in-house amplifier, and results were recorded via an in-house program. Data were analyzed using Fourier transformation and Origin 7.5 (OriginLab Corp., Northampton, MA). Solution pH value was controlled by using 10 mM HEPES buffer containing 2 M NaCl at pH 7.0.

Purification of the cbb₃-type oxidase from V. cholerae

The cbb₃-type oxidase from V. cholerae was purified as previously described (3).

Steady state kinetics using an oxygen electrode

Cytochrome c oxidase activity was measured polarographically at 25 °C using a YSI model 53 oxygen meter. The standard reaction mixture contained 1.8 mL of 50 mM sodium phosphate (pH 6.5), 50mM NaCl, 0.05% dodecylmaltoside, 10 mM sodium ascorbate and the cytochrome c to be tested. The cytochrome c₄ and cytochrome c₅ concentrations were varied in the range of 1 μM to 100 μM. The O₂ consumption reaction was initiated by the addition of the oxidase to a final concentration of 50 nM. The dependence of activity on ionic strength was measured by varying the concentration of NaCl from 10 mM to 150 mM using 20 μM of cytochrome c₄ reduced by 10mM sodium ascorbate in 10 mM sodium phosphate (pH 6.5) buffer, 0.05% dodecylmaltoside.

Steady state kinetics by stopped flow spectrophotometry

To prepare the pre-reduced cytochrome c₄, the sample was reduced with sodium dithionite and the excess dithionite was removed by gel filtration. The steady state kinetics between the cbb₃ oxidase from V. cholerae and the pre-reduced cytochrome c₄ was monitored at 25 °C with Applied Photophysics SX-17MV stopped-flow spectrometer as follow. One syringe of the stopped-flow apparatus was filled with 400 nM of the cbb₃ oxygen reductase in 50 mM sodium phosphate (pH 6.5), 150 mM NaCl, 0.05 % dodecylmaltoside, while the pre-reduced cytochrome c₄ in the same buffer was loaded into the other syringe. The concentration of cytochrome c₄ was varied from ~ 1 μM to 60 μM for different measurements. After mixing, the reaction was followed spectrometrically from 400 nm −700 nm for 50 s −1000 s using a photodiode array. Using a least square fitting method (54), the spectrum at each time point was deconvoluted into the component spectra of reduced and oxidized cytochrome c₄, which had been collected separately by mixing cytochrome c₄ with either sodium dithionite or potassium ferricyanide. The concentrations of the reduced and oxidized cytochrome c₄ at each time point were then used to calculate the reaction rate at each point. By using different starting concentrations of cytochrome c₄ in different measurements, the reaction rates at various combinations of reduced and oxidized cytochrome c₄ concentrations were obtained. The reaction rates were plotted against reduced cytochrome c₄ concentrations at a fixed oxidized cytochrome c₄ concentration, and the K_M and V_max for each plot were obtained by non-linear least square fitting to the standard Michaelis-Menten kinetics model. All data processing and analyses were performed with Mathematica.

Results

Expression and purification of recombinant cytochrome c₄ and cytochrome c₅

Transformants were found to express the cytochrome c's upon induction (46) and the recombinant proteins were isolated as described in the Methods section. The SDS-PAGE analysis of the recombinant cytochromes is shown in Figure 1. Staining with Coomassie Blue (Figure 1A), the purified cytochrome c₄ has a single band near 21~22kDa which contains two covalently bound heme (Figure 1B). The expected molecular weight of the mature cytochrome c₄ is 19,791 (187 amino acid residues) with 1,233 (two heme c moiety). Cytochrome c₅ has a single heme-staining component running near the expected molecular weight of 11,480, but there is a major protein contaminant that does not contain heme. Since cytochrome c₅ does not function as a reductant for the cbb₃-type oxidase (see below), no further effort was made to improve the purification.

Figure 1.

SDS-PAGE of the isolated recombinant cytochromes c₄ and c₅. (A) 15% SDS gel stained with Coomassie blue. (B) Identical gel stained for heme. The lanes from left to right contain the cytochrome c₅ (lane 1), cytochrome c₄ (lane 2) and molecular weight standards (lane 3).

UV-vis spectroscopy

Figure 2 shows the spectra of the reduced and oxidized forms of cytochromes c₄ and c₅. The spectra of each are consistent with previously published reports of homologues from other organisms (49–51). In the reduced form, the absorption spectrum of cytochrome c₄ has three maxima at 553nm, 523nm and 417nm, corresponding to the α-, β- and γ-(Soret) bands, respectively (Figure 2A). The maxima in the oxidized form of cytochrome c₄ are at 528nm and 409 nm. The spectra show two diagnostic features shared by all cytochrome c₄'s: (1) A split α-band with maxima at 553nm and 549nm. (2) A low ratio of the amplitude of the α-band compared to the β-band, $\frac{A_{553}}{A_{523}}=1.05$.

Figure 2.

Absorption spectra of fully reduced (solid line) and fully oxidased (broken line) V. cholerae (A) cytochrome c₄ (3μM) and (B) cytochrome c₅ (7μM). The insets show the α- and β-bands of the fully reduced cytochrome c₄ (10μM) and cytochrome c₅ (17μM). All the absorption spectra were detected under the condition of 25 mM Tris-HCl (pH=7.5), 25 °C.

Reduced cytochrome c₅ has maxima at 554.8nm, 524.6nm and 419nm, corresponding to the α-, β- and γ -(Soret) bands, respectively (Figure 2B). The α/β ratio $\frac{A_{555}}{A_{525}}=1.5$. The spectrum of the oxidized cytochrome c₅ has two peaks at 526nm and 413nm. These features are consistent with previous reports of cytochrome c₅'s (51).

Electrochemistry

Cytochrome c₄ was found to be amenable to direct electrochemical characterization by adhering the protein to a carbon electrode. The midpoint potentials of the two heme components of the recombinant cytochrome c₄ were measured by cyclic voltammetry (Figure 3). Four well defined redox peaks are observed and the $E_m^o$ values of the two heme groups were determined to be approximately 240 mV and 340 mV (vs SHE).

Figure 3.

Cyclic voltammetry of the recombinant cytochrome c₄ which was applied to a freshly polished pyrolytic PGE electrode. The solution contained 10 mM HEPES buffer and 2 M NaCl at pH 7.0. Values obtained (vs. SHE) were E_pc1 = 0.26 V, E_pc2 = 0.35 V, E_pa1 = 0.23 V, E_pa2 = 0.34 V and E_1/2 = (E_pc + E_pa)/2 = 0.24 V and 0.34 V.

Rates of oxygen reduction by the V. cholerae cbb₃-type oxidase using cytochromes c₄ and c5

Figure 4A compares the rates of oxygen reduction, measured with an oxygen electrode, by the V. cholerae cbb₃-type oxidase using ascorbate plus either i) horse heart cytochrome c, ii) recombinant cytochrome c₄ or iii) recombinant cytochrome c₅. These assays were performed in the absence of the mediator TMPD, which can itself function as a substrate for the cbb₃-type oxidase. The data (Figure 4A) show that neither the horse heart cytochrome c nor cytochrome c₅ is effective as a substrate, whereas cytochrome c₄ is capable of supporting turnover at rates over 300 e⁻¹/s at 25°C. This turnover is comparable to what is observed with 0.5 mM TMPD in the presence of ascorbate to maintain the TMPD reduced (not shown). The turnover using recombinant cytochrome c₅ is more than a factor of 10 less (7%) than that observed with cytochrome c₄, and horse heart cytochrome c is also a poor substrate (8%). Although the purity of recombinamt cytochrome c₅ was not high, it is unlikely that the contaminants are inhibitory. The oxidase activity with cytochrome c₄ is rapidly stopped upon the addition of 25 μM cyanide (Figure 4B), showing that the oxidase activity is catalyzed by the cbb₃-type oxygen reductase. The data in Figure 4A do not indicate saturation even at concentrations of cytochrome c₄ as high as 100 μM. Activity was compared using 50 nM of the oxidase and 20μM cytochrome c₄, varying the concentration of NaCl from 1 mM to 150 mM (Figure 4C). These data indicate that the apparent binding of cytochrome c₄ to the oxidase is not sensitive to ionic strength and that the lack of saturation of activity as a function of the concentration of cytochrome c₄ is not due to weak binding due to the selected ionic strength. It was determined, however, that under the conditions used for the assays with the oxygen electrode, the ascorbate does not maintain the cytochrome c₄ fully reduced.

Figure 4.

Steady state oxidase activity of the cbb₃ oxygen reductase from V. cholerae. (A) Dependence on the concentration of cytochrome c₄ (solid squares), cytochrome c₅ (open circles) and horse heart cytochrome c (solid triangles). The reaction mixture contained 50 mM sodium phosphate (pH 6.5), 50mM NaCl, 0.05% dodecylmaltoside, 10 mM sodium ascorbate, and the indicated amount of the cytochrome c, in a total volume of 1.8 ml. The O₂ consumption reaction was initiated by the addition of 50 nM of the oxidase. (B) Sensitivity of the oxidase activity to the addition of 25 μM cyanide. (C) Dependence of the oxidase activity on the concentration NaCl, measured with50 nM oxidase and 20 μM cytochrome c₄, in the presence of 10mM sodium ascorbate, 10 mM sodium phosphate (pH 6.5) buffer and 0.05% dodecylmaltoside.

Stopped flow spectrophotometry

To better understand the lack of saturation in the steady state kinetics measurements using the oxygen electrode, steady state activity was also measured by stopped flow spectrophotometry, mixing pre-reduced cytochrome c₄ with a solution containing the cbb₃-type oxidase. The oxidation of cytochrome c₄ was monitored by recording the change of the absorption spectrum from 400 nm to 700 nm using a photodiode array. The initial concentration of cytochrome c₄ was varied from 1 to 60 μM and spectra were obtained as a function of time for each initial condition. The spectra were deconvoluted to quantify the concentration of oxidized and reduced cytochrome c₄ in solution as a function of time. From this large matrix of data, the rate of oxidation of cytochrome c₄ could be determined for a wide range of concentrations of both the reduced and oxidized forms of cytochrome c₄. These data were plotted to show how the rate of oxidation varies as a function of reduced cytochrome c₄ in the presence of a specific concentration of oxidized cytochrome c₄ (Figure 5A–E). If the oxidized cytochrome c₄ is a competitive inhibitor for the reduced cytochrome c₄, one would expect that the data would show a constant V_max and increasing K_M as the concentration of oxidized cytochrome c₄ increases. However, this is not the case, and both V_max and K_M increase as the concentration of oxidized cytochrome c₄ increases (Figure 5F). One complication is that the spectral deconvolution with the least square method is not satisfactory in the case of cytochrome c₄ because cytochrome c₄ contains two hemes which have slightly different spectra. The assumption that either both are oxidized or both are reduced is not realistic. Attempts were not made to analyze the data with more complex models. Despite this, the current analysis indicates the following significant points.

• 1)Oxidized cytochrome c₄ has a strong effect on the activity between the cbb₃-type oxidase and reduced cytochrome c₄ The apparent K_M increases in the presence of relatively small amounts of oxidized cytochrome c₄ (Figure 5F). If the concentration of oxidized cytochrome c₄ is greater than 2.5 μM, the K_M of reduced cytochrome c₄ is too large to obtain a reasonably accurate fitting to the Michaelis-Menten model.
• 2)During the manipulation of the pre-reduced cytochrome c₄ to load the stopped flow syringe, a significant amount of autoxidation occurs. Hence, none of the data represent a starting condition in which cytochrome c₄ is actually fully reduced.
• 3)Because of the spectroscopic complications of having two different heme components in cytochrome c₄, data fits are not sufficient to extrapolate to obtain true values for either V_max or K_M. Phenomenologically, as the concentration of oxidized cytochrome c₄ increases, both the apparent V_max and K_M increase. The inhibitory influence of oxidized cytochrome c₄ and the increasing concentration of oxidized cytochrome c₄ that accumulates as the reaction proceeds, explains the lack of saturation under conditions used for the oxygen electrode assays (Figure 4A).

Figure 5.

The rates of oxidation of reduced cytochrome c₄ by the cbb₃ oxygen reductase from V. cholerae at different oxidized cytochrome c₄ concentrations determined by stopped-flow spectroscopy and least square fitting. The substrate [S] refers to reduced (not total) cytochrome c₄. The oxidized cytochrome c₄ concentrations were (A)1.5 μM; (B) 2.0 μM, (C) 2.5 μM; (D) 3.0 μM and (E) 7.5 μM. The curves show the non-linear least square fitting of the points to Michaelis-Menten kinetics. The K_M and V_max values for 1.5 μM, 2.0 μM and 2.5 μM oxidized cytochrome c₄ are plotted in (F).

Discussion

Although initially thought to be confined to proteobacteria (7, 11), the cbb₃ oxygen reductases are widely distributed among bacterial phyla (5, 10). Although there are exceptions, most archaea lack c-type cytochromes (55), and no example of a cbb₃-type oxidase in an archaea has been reported. Nearly all studies have focused on the role of the C-family (cbb₃) oxygen reductases in the aerobic respiratory chains of proteobacteria(8), exceptions being the cbb₃-type oxidases from Rhodothermus marinus (yet to be confirmed in the genome sequence) (32) and from Sulfurihydrogenibium azorense (10).

In most cases, the natural electron donors for these C-family oxygen reductases are not known. The current work has identified the di-heme cytochrome c₄ as a natural electron donor to the cbb₃-type oxygen reductase from the γ-proteobacterium, V. cholerae. Cytochrome c₄ is one of five cytochrome c's encoded in the genome of V. cholerae that are predicted to be soluble, periplasmic proteins (44). Cytochromes c₄ and c₅ appear to be among the major cytochrome c's expressed under aerobic growth conditions (44). Heterologous expression in E. coli was successful, generating soluble cytochromes c₄ and c₅. Cytochrome c₄ from V. cholerae has spectroscopic and electrochemical properties similar to the cytochrome c₄'s from other organisms. These include the high midpoint potentials (240 mV and 340 mV, vs SHE) of the two hemes, the split α-band and the low ratio of the peak heights of the α- and β-bands (50, 56, 57).

Cytochrome c₄ can clearly donate electrons to the cbb₃-type oxidase, supporting steady state reduction of O₂ at a rate of at least 300 e⁻¹/s. Rates attained using either cytochrome c₅ or horse heart cytochrome c were a 10- to 20-fold less under comparable conditions. The one caveat is that the steady state rate of cytochrome c₄ oxidase activity does not saturate and quite high concentrations (100 μM) are required to reach the turnover of 300 s⁻¹. The lack of saturation (Figure 4A) is unusual and not expected. By comparison, the steady state turnover of H. pylori cytochrome c₅₅₃ by the H. pylori cbb₃-type oxidase is 250 s⁻¹ with a K_m of 0.9 μM. The nearly linear dependence of oxygen reductase activity as a function of the concentration of cytochrome c₄ (Figure 4A) can be qualitatively explained being due to the influence of oxidized cytochrome c₄ (39). There is a significant concentration of oxidized cytochrome c₄ present at all times during the oxygen reductase assay and the presence of ascorbate even as high as 100 mM does not alleviate this problem. The data do show that the true V_max for the oxidation of cytochrome c₄ must be at least 300 s⁻¹, but the true K_m cannot be measured using this assay. In the steady state assay, the velocity is measured in the presence of 100 μM reduced cytochrome c₄ with about 25 μM oxidized c₄ also being present. The stopped flow experiments cannot be performed under these conditions, but the data show that the values of both V_max and K_M depend on the concentration of oxidized cytochrome c₄, and the stopped-flow data are consistent with the steady state kinetics. It is quite conceivable that the true K_m is in the range of 1 to 10 μM, but further work will be needed to determine this.

Since the oxidized and reduced forms of cytochrome c often have similar affinities for their redox partners, it is to be expected that oxidized cytochrome c₄ would competitively inhibit the activity between reduced cytochrome c₄ and the cbb₃-type oxidase. However, product inhibition can only explain the increase in the K_M but not the increase in V_max that is observed as the concentration of the product increases (Figure 5). Whether this kinetic behavior reflects an in vitro artifact or an unusual form of allosteric regulation is not clear. This unusual behavior might be related to the fact that cytochrome c₄ is a two-electron carrier. The cyclic voltammetry shows that the di-heme cytochrome c₄ exhibits two substantially different midpoint potentials, but the extent of cooperativity between the two hemes is not known. In principle, there could be two different binding sites on cytochrome c₄ for the cbb₃ oxygen reductase, each with a distinct K_M and V_max. In any event, further characterization of the oxidation kinetics of the diheme cytochrome c₄ is needed to address this question.

At this point, it is safe to conclude that cytochrome c₄ is definitely a substrate for the cbb₃-type oxidase whereas cytochrome c₅ is a very poor substrate, at best, under in vitro conditions.

The roles of cytochrome c₄ in other bacteria

Analysis of sequenced genomes demonstrates that the distribution of cytochrome c₄ is limited, with the majority of homologs being found within the β- and γ- proteobacterial clades. Cytochrome c₄ is also sporadically distributed within the α- proteobacterial clade as well as a few other bacterial phyla. Bacteria within the β- and γ-proteobacterial clades which contain a cbb₃ oxygen reductase are very likely to also contain a cytochrome c₄. Within these groups of bacteria, it is reasonable to assume that cytochrome c₄ is an electron donor to the cbb₃ oxygen reductase. It is noted that the cbb₃ oxygen reductases within the β- and γ- proteobacterial clades all contain the auxiliary subunit CcoP, a di-heme cytochrome c which is the likely electron acceptor from cytochrome c₄. A large number of the cbb₃ oxygen reductases outside the β- and γ- proteobacterial clades are missing the CcoP subunit and others have an alternative subunit, CcoR (10). Indeed, many organisms that are outside the β- and γ-proteobacterial clades which encode a cbb₃ oxygen reductase, do not contain cytochrome c₄ and, therefore, must be using a different electron donor to this enzyme.

Other γ-proteobacteria which contain cbb₃ oxygen reductases include Azotobacter vinlandii (58), and Pseudomonas stutzeri (33–37). Knock-out mutants of both cytochromes c₄ and c₅ in A. vinlandii support a role for both of these cytochromes in aerobic respiration, though there are no data concerning a specific role with the cbb₃ oxygen reductase. The cytochrome c₄ from P. stutzeri has been extensively characterized (49, 59–62), but its physiological role has not yet been defined.

Acidothiobacillus ferrooxidans is a γ-proteobacterium whose genome includes several cytochrome c₄'s but which does not contain a cbb₃-type oxidase (63–65). The cytochrome c₄'s are strongly implicated in the aerobic respiratory chain (66–68). A. ferrooxidans is an obligate chemolithotropic bacterium which grows at acidic pH (pH 1.5 to 4) and uses either ferrous iron or reduced sulfur compounds as an electron source. The cyc1 gene encodes a di-heme c₄-type cytochrome c₅₅₂ within an operon which includes an A-family oxygen reductase as well as the blue copper protein rusticyanin (69). These respiratory components comprise the electron transport chain between the ferrous iron oxidase and O₂. The cytochrome c₄ forms a complex with rusticyanin which is likely the immediate electron donor to the A-family oxygen reductase (68, 70). The structure of the Cyc1 cytochrome c₄ has been determined and the complex with rusticyanin has been modeled computationally (64, 70–72). Two additional genes encoding cytochrome c₄'s, cycA1 and cycA2, are in operons encoding two separate bc₁ complexes (66). It is suggested that when growing on reduced sulfur compounds, CycA2 shuttles electrons from the bc₁ complex encoded by the petII operon to one or several terminal reductases (66). When growing on Fe⁺², CycA1 is proposed to be part of an aerobic respiratory chain which runs in the reverse direction to generate NADH: CycA→bc₁ (petI)→Q→Complex I (66, 67).

The β-proteobacterial clade includes Neissseria meningitidis and Neisseria gonorrhoeae, which are each human pathogens and are notable also for the fact that the cbb₃ oxygen reductase is the only respiratory oxidase encoded in their genomes (42, 43). Knock-out mutations have implicated cytochrome c₄ along with cytochrome c_x (a homologue of cytochrome c₅₅₂ from T. thermophilus) as being an electron donor to the cbb₃ oxygen reductase in N. meningitidis (73). Cytochrome c₅ is required for electron transfer to the nitrite reductase in N. meningitidis (73). Knock-out mutations have also indicated a role of cytochrome c₄ along with cytochrome c₅ as substrates of the cbb₃ oxygen reductase in N. gonorrhoeae, but not cytochrome c₂ (42). Interestingly, the CcoP subunit of the cbb₃-type oxidase in N. gonorrhoeae has an extra cytochrome c-domain which is homologous to cytochrome c₅, and this domain is implicated in electron transfer to the nitrite reductase which is tethered to the outer membrane (74).

Outside the β- and γ- proteobacterial clades, there is limited information about the native electron donors for the cbb₃ oxygen reductase. One exception is H. pylori, in which the monoheme cytochrome c₅₅₃ (39) is the substrate for the cbb₃ oxygen reductase (38). H. pylori is an ε-proteobacterium. Closely related ε-proteobacteria which contain cbb₃ oxygen reductases and which might also use homologues of cytochrome c₅₅₃ as substrates are Campylobacter jejuni (41, 75) and Wolinella succinogenes (76, 77).

The cbb₃ oxygen reductases have been biochemically characterized from several α-proteobacteria: Rhodobacter sphaeroides (5, 12–17), Paracoccus denitrificans (18), Rhodobacter capsulatus (19–23) and Bradyrhizobium japonicum (25–28, 29, 30, 31), Rhodothermus marinus (32). Using knock-out mutants, it was concluded that both the A- and C-family oxygen reductases from R. sphaeroides can utilize either cytochrome c₂ or membrane-anchored cytochrome c_y (21).

Conclusion

Cytochrome c₄ has been shown to be the native electron donor for the C-family (cbb₃) oxygen reductase in V. cholerae. Within the γ- and β-protobacterial clades, there is a strong correlation for the co-existence of a C-family oxygen reductase and a cytochrome c₄, suggesting that in these organisms, it is likely that cytochrome c₄ is a natural electron donor. This does not exclude other electron donors from also being substrates within these organisms. Outside the γ-and β-protobacterial clades there is clear evidence that other cytochrome c's are substrates for C-family oxygen reductase, though in most cases, there are no data to identify these electron donors.

Abbreviations

The abbreviations used are

SDS-PAGE

sodium dodecyl sulfate-polyacrylamide gel electrophoresis

PEG

pyrolytic graphite-edge

TMPD

N,N,N',N'-tetramethyl-p-phenylenediamine

TMBZ

3,3`,5,5`-tetramethylbenzidine