|Home | About | Journals | Submit | Contact Us | Français|
Cytochrome c oxidase (CcO) catalyzes the reduction of molecular oxygen to water using ferrocytochrome c (cyt c2+) as the electron donor. In this study, the oxidation of horse cyt c2+ by CcO from Rhodobacter sphaeroides, was monitored using stopped-flow spectrophotometry. A novel analytic procedure was applied in which the spectra were deconvoluted into the reduced and oxidized forms of cyt c by a least-squares fitting method, yielding the reaction rates at various concentrations of cyt c2+ and cyt c3+. This allowed an analysis of the effects of cyt c3+ on the steady-state kinetics between CcO and cyt c2+. The results show that cyt c3+ exhibits product inhibition by two mechanisms: competition with cyt c2+ at the catalytic site and, in addition, an interaction at a second site which further modulates the reaction of cyt c2+ at the catalytic site. These results are generally consistent with previous reports, indicating the reliability of the new procedure. We also find that a 6xHis-tag at the C-terminus of the subunit II of CcO affects the binding of cyt c at both sites. The approach presented here should be generally useful in spectrophotometric studies of complex enzyme kinetics.
Cytochrome c oxidase (CcO) is the terminal enzyme in the respiratory electron transport chains of mitochondria and many bacteria[1–3]. It catalyzes the electron transfer from ferrocytochrome c (cyt c2+) to molecular oxygen and contributes to the proton motive force across the membrane through proton pumping coupled to the electron transfer process.
Although enzyme kinetics studies were initially carried out with bovine or horse CcO, there have been numerous studies on respiratory oxygen reductases from prokaryotes, such as Rhodobacter sphaeroides and Paracoccus denitrificans, within the last two decades because these bacterial complexes are highly homologous to the mitochondrial enzymes and can be readily mutated for functional analyses[4–6]. Crystal structures of CcO from several organisms including Bos taurus (bovine), Paracoccus denitrificans and R. sphaeroides are now available. Although no crystal structure of CcO complexed with its substrate cyt c has been reported, the cyt c binding domain of CcO has been mapped to several residues located in the periplasmic domain of subunit II through biochemical and mutagenesis studies [10–14]. Structural models for the CcO-cyt c complexes have also been analyzed by computational methods[15, 16]. At present, it is generally accepted that the cyt c binding site located in subunit II of CcO is made up of an outer surface of negatively charged residues and a central non-polar area, featuring the interaction modes similar to those observed in reaction centers, cytochrome bc1 and cytochrome c peroxidase. Multiple reports have also shown that a conserved tryptophan residue located at the center of the interaction domain in CcO subunit II mediates electron transfer from cyt c to the CuA center of subunit II[12–14]. The negatively charged residues and the conserved tryptophan residue at the binding site are illustrated in Fig. 1.
Despite the general consensus on the site of cyt c2+ oxidation on CcO, there is still no universally agreed model on the steady-state kinetics of CcO with regards to cyt c2+[17–19]. It has long been known that the activity of CcO is highly influenced by ionic strength, due to electrostatic interactions involved in binding with its substrate, cyt c[10, 11, 13, 20]. The high affinity for cyt c at low ionic strength leads to lower enzyme activity as the slow dissociation of cyt c becomes rate-limiting. Since the reaction involves the formation of a CcO-cyt c complex, classical Michaelis-Menten kinetics has generally been used to describe the steady-state enzyme activity. Kinetics studies carried out at low ionic strength consistently yield non-hyperbolic Michaelis-Menten kinetics [10, 21]. Residues important for the interaction between CcO with cyt c2+ and for interprotein electron transfer have been identified through mutagenesis studies of bacterial enzymes [10, 11, 13], and it is clear that there is only one catalytic site in CcO for the oxidation of cyt c2+. The most satisfactory kinetic model proposed includes, in addition to the catalytic site for the oxidation of of cyt c2+, a second, nearby allosteric binding site for cyt c. There is negative cooperativity in the binding affinities of cyt c between the two sites . More recently, it has been argued, based on studies with the cytochrome caa3 complex from Bacillus subtilus, that the non-hyperbolic steady-state kinetic behavior may originate from intrinsic properties of the enzyme intermediates, whose populations change with ionic strength and cyt c concentration [18, 23]. Given the enormous amount of data reported from studies carried out with different versions of cyt c and diverse enzymes, not only from the mammalian sources but also from prokaryotes, it is challenging to derive a model that is able to explain all the findings.
Two fundamentally different experimental methods have been widely used to monitor the steady-state turnover at various cyt c2+ concentrations. In the first method, the rate of oxygen consumption is measured polarographically with a Clark-type oxygen electrode [13, 21, 23, 24]. This approach requires the addition of artificial electron donors such as ascorbate and TMPD (N,N,N′,N′-tetramethyl-p-phenylenediamine) to maintain cyt c in the reaction mixture in the fully reduced state. The second approach directly monitors the oxidation of cyt c2+ spectrophotometrically by following the absorbance at its characteristic 550 nm wavelength [10, 11, 22, 23, 25, 26]. This procedure involves pre-reduction of cyt c with dithionite or ascorbate, and then removal of the excess reductants by gel-filtration prior to the measurement. Regardless of the assays employed to measure the activity of CcO, the steady-state kinetic analyses of the reaction rates at varying cyt c2+ concentrations have been predominantly performed with simple linear regression methods such as the Eadie-Hofstee method, which are known to be unreliable due to distorted error distributions [27, 28].
In this report, the steady-state kinetics between the aa3-type CcO from R. sphaeroides and horse cyt c are described. The commercially available horse cyt c was used instead of the physiological substrate (cyt c2 from R. sphaeroides) because it has been routinely used for activity assays of prokaryotic CcOs [10, 13, 14]. The horse cyt c and the R. sphaeroides cyt c2 have very similar structures and have been shown to use the same binding site of CcO for electron transfer . The oxidation of cyt c2+ was followed with a stopped-flow spectrophotometer, and least-squares deconvolution was applied to obtain the concentrations of both cyt c2+ and cyt c3+ during the entire time course of the reaction. This allows one to explicitly model the possible role of cyt c3+ as an inhibitor and to analyze its effect on the steady-state kinetics between CcO and cyt c2+.
Horse cyt c was purchased from Sigma-Aldrich and used without further purification. TMPD, sodium ascorbate and other chemicals used were of laboratory grade or better.
The CcO enzyme with a 6xHis-tag at the C-terminus of subunit I (HT-I CcO) was expressed from R. sphaeroides and purified from the membranes solubilized with dodecyl maltoside (DDM) using Ni-NTA affinity chromatography as described in a recent report . The CcO enzyme with a 6xHis-tag at the C-terminus of subunit II (HT-II CcO) was expressed from the pCH37 plasmid in R. sphaeroides strain YZ200 and purified with the same procedure as HT-I CcO. The YZ200 strain and pCH37 plasmid were kindly provided by Dr. Shelagh Ferguson-Miller at Michigan State University [30, 31]. The concentration of the enzyme was estimated from the reduced-minus-oxidized difference spectrum with an extinction coefficient of 24 mM−1cm−1 for 605-minus-630 nm .
The oxidation of horse cyt c2+ by CcO at 25 °C was followed with an Applied Photophysics SX-17MV stopped-flow spectrophotometer as described in a recent report . Before the measurement, horse cyt c was reduced with sodium dithionite and the excess reductant was removed by Sephadex G-25 gel-filtration chromatography. The reaction buffer was either 5 mM or 50 mM potassium phosphate (pH 7.5) and contained 0.05 % DDM and 1 mM EDTA. About 400–1600 spectra were recorded with a photodiode array detector for a duration of 4–40 s after a known quantity of CcO (5–10 nM) was mixed with varying amounts of cyt c2+ ranging from 0.75 – 40 μM. The fully oxidized and fully reduced spectra of cyt c were collected separately with the same stopped-flow instrument by mixing a known quantity (~2–10 μM) of cyt c with a very small amount of CcO (less than 1 nM) or excess sodium dithionite and used as standard spectra in the least-squares deconvolution of each recorded spectrum (described in Section 2.4). This analysis yields the concentrations of both cyt c3+ and cyt c2+ at each time at which a spectrum was recorded. In this way, the change in the concentrations of both cyt c3+ and cyt c2+ are followed as CcO oxidizes cyt c2+ (Fig. 2). The reaction rate at any time step can be calculated from the slope of either the cyt c2+ or cyt c3+ trace and each rate is associated with specific concentrations of cyt c3+ and cyt c2+. All the calculations were performed with Mathematica (Wolfram Research).
The spectra recorded by the stopped-flow spectrophotometer during the oxidation of cyt c2+ by CcO were fitted to the standard spectra of fully oxidized and fully reduced cyt c with a least-squares regression method . The formula can best be illustrated in a matrix equation (Equation I). Matrix S contains i standard spectra, each spectrum composed of absorbance values at j wavelengths and matrix A represents the spectrum being analyzed as a linear combination of contributions of each of the different species whose individual spectra are included in the matrix S.
Solving Equation II yields the concentrations of individual species (matrix C) in the recorded spectrum (matrix A). Some examples of the least-squares fitting of the recorded spectra performed in this study are displayed in Fig. 3. Mathematica was used to carry out the least-squares fitting procedure.
After the spectrum at each time point was deconvoluted into the component spectra of reduced and oxidized cyt c using the least-squares fitting method described in the Section 2.4 (Fig. 3), the reaction rate at each time point was calculated from the slope of either the cyt c2+ or cyt c3+ trace (Fig. 2). Each rate obtained is associated with the specific amounts of cyt c2+ and cyt c3+ at that particular time point. By using different starting concentrations of cyt c in different measurements, the reaction rates at various combinations of reduced and oxidized cyt c concentrations were obtained. For example, for two different stopped-flow spectrophotometric experiments with the respective starting cyt c2+ concentrations of 2 and 4 μM, after exactly 1 μM of cyt c2+ has been oxidized by the enzyme, the two experiments will generate the reaction rates corresponding to 1 and 3 μM of cyt c2+ respectively although they both have the same 1 μM of cyt c3+ at their respective time points. For each cyt c3+ concentration ranging from ~ 0.5 μM – 12 μM, the rates of oxidation of cyt c2+ by CcO at different cyt c2+ concentrations was fitted to classical Michaelis-Menten kinetics with the non-linear regression method. The Eadie-Hofstee plots shown in Fig. 4 are strictly for visualization purposes. The apparent KM and Vmax values determined from the non-linear regression method were used in further analyses of the effects of cyt c3+ on the steady-state kinetics between the CcO and cyt c2+. These calculations were performed with OriginPro 8 software (OriginLab Corporation).
Fig. 4 illustrates the Eadie-Hofstee plots showing the relation between the turnover of CcO and [cyt c2+] at each value of [cyt c3+]. There is no indication of biphasic kinetics even at the low ionic strength of 5 mM potassium phosphate. The apparent KM and Vmax values determined from non-linear regression to the Michaelis-Menten kinetics model are plotted against the concentration of cyt c3+ in Fig. 5. If cyt c3+ competes with cyt c2+ at the active site, we would expect that only the apparent KM would be affected by cyt c3+, according to Equation III, while Vmax would remain unchanged . At low ionic strength of 5 mM potassium phosphate, the competitive inhibition by cyt c3+ is observed for [cyt c3+] > 3 μM (Fig. 5a). For [cyt c3+] < 3 μM, the apparent Vmax is a function of [cyt c3+] and the apparent KM is more sensitive to the change in [cyt c3+]. This result can best be interpreted as cyt c3+ binding at a second allosteric site, leading to the weaker binding and faster dissociation of cyt c bound at the active site. At concentrations of [cyt c3+] > 3 μM, the allosteric site appears to be saturated and cyt c3+ simply behaves as a competitive inhibitor, acting at the catalytic site. From the first phase ([cyt c3+] < 3 μM), the actual KM for cyt c2+ can be estimated as ~ 0.2 μM and the KD for cyt c3+ at the allosteric site is ~ 1.5 μM. The KI for cyt c3+ at the catalytic site cannot be calculated with Equation III from the first phase, since this does not represent pure competitive inhibition. However, the KI for cyt c3+ can be estimated to be comparable to the KM for cyt c2+. With cyt c3+ bound at the second site, the KM for cyt c2+ and KI for cyt c3+ at the catalytic site can be determined from the second phase ([cyt c3+] < 3 μM) as 3.3 μM and 6.4 μM respectively (Table 1).
At high ionic strength of 50 mM potassium phosphate, both the apparent KM and Vmax values increase with [cyt c3+] up to 10 μM of cyt c3+, the highest concentration analyzed in this study (Fig. 5b). It appears that the second site is not saturated even at [cyt c3+] = 10 μM, reflecting the weaker binding at high ionic strength. The actual KM for cyt c2+ in the absence of cyt c3+ is also raised to ~ 10.7 μM, which is in good agreement with the value obtained from the polarographic method (unpublished data). The relatively large errors in the calculated kinetic parameters especially for [cyt c3+] > 5 μM are due to the limited range of [cyt c2+] utilized relative to the high apparent KM values in the corresponding region.
At the low ionic strength of 5 mM potassium phosphate, the apparent KM increases while the apparent Vmax is unchanged up to ~ 6 μM cyt c3+ (Fig. 5c). It seems that the His-tag at the C-terminus of the subunit II eliminates the allosteric effect of cyt c3+ bound at the second site when [cyt c3+] < 6 μM. Under these conditions, the KM for cyt c2+ and KI for cyt c3+ are determined to be 2.4 μM and 4.7 μM, respectively. Thus, the His-tag also lowers the affinity at the active site for by about an order of magnitude, which is consistent with the reported weaker interaction between the R. sphaeroides cyt c2 and the HT-II CcO . The effect of cyt c3+ is no longer simple competitive inhibition when [cyt c3+] is higher than 6 μM, as shown by the slow decline of the apparent Vmax. The data suggest that at concentrations of cyt c3+ > 6 μM, it does bind to the allosteric binding site on HT-II CcO and, when bound, the cyt c3+ reduces the rate of electron transfer between cyt c2+ (at the catalytic site) and CcO.
Compared to the polarographic approach, the spectrophotometric approach in general has the advantage for the investigation of the steady-state kinetics between CcO and cyt c because the reaction mixture includes only CcO and cyt c, and there are no complications arising from the presence of additional reductants such as TMPD. In all previous reports of assays using the spectrophotometric approach, only the initial reaction rates were used, and it was assumed that cyt c was fully reduced at time zero, just prior to being mixed with the enzyme [10, 11, 22]. The reaction rate was then determined from the change in the absorbance of cyt c at its characteristic wavelength of 550 nm. In the present work, we apply least-squares spectral fitting to the recorded spectra to determine the concentrations of both cyt c2+ and cyt c3+ at each time step with high accuracy (Fig. 2). Although, in principle, it is sufficient to monitor the conversion of cyt c2+ to cyt c3+ from the absorbance values at one or two characteristic wavelengths, the current method allows one to determine the absolute concentrations of each species in solution with no additional assumptions. In addition, the least-squares spectral fitting method can readily trace the reaction progress of more than two chemical species and has great potential for applications in more complex systems.
With the knowledge of the concentrations of both the substrate and the product, one can explicitly treat cyt c3+ as an inhibitor and investigate the possibility of product inhibition in the steady-state kinetics. In addition, the data analysis is no longer restricted to the use of initial rates as long as the steady-state assumption is valid, i.e. the concentration of cyt c2+ is substantially larger than that of CcO. This approach allows one to systematically explore the effects of cyt c3+ on the reaction between CcO and cyt c2+.
We find that cyt c3+ influences the rate of oxidation of cyt c2+ by (at least) two distinct mechanisms. First, cyt c3+ acts as a competitive inhibitor, presumably by competing for binding to the catalytic site on CcO with a similar but slightly weaker affinity. This is consistent with previous results obtained with the CcO enzymes [10, 26, 35–37], thus confirming the reliability of the new approach. The binding affinities were also found to be highly influenced by the ionic strength, as expected from the involvement of electrostatic interactions between the two proteins. It should be noted that competitive product inhibition by cyt c3+ is not unique to the CcO enzymes, but has also been reported in other complexes involving cyt c [38, 39].
The KM values of the high affinity phase between R. sphaeroides CcO and horse cyt c at the low ionic strength, determined by the polarographic approach, although consistent with the dissociation constants from binding studies [13, 21], are about one order of magnitude lower than the values reported both in the current work as well as values reported from previous measurements using the spectrophotometric approach. Polarographic assays previously performed in our laboratory indicated that the high affinity phase is highly dependent on the TMPD concentration (unpublished work) and is likely due to the reaction between TMPD and the cyt c-CcO complex. Although the reaction between TMPD and the cyt c-CcO complex at low ionic strength has been characterized [25, 40, 41], and the resulting possible complications in the analysis of steady-state kinetics have been discussed [13, 17, 22], a full quantitative analysis has not been achieved. This is in part due to the difficulty in estimating the cyt c-CcO population and its reaction rate with TMPD in the presence of other factors involved in the polarographic assay.
In previous spectrophotometric studies, cyt c was assumed to be fully reduced at the beginning of the reaction and the effects of a small percentage of cyt c3+ were ignored [10, 22, 35]. In the current work, this assumption is avoided, and the concentration of cyt c3+ is explicitly determined. Although we found that cyt c was mostly reduced at the beginning of each measurement, even a small percentage of cyt c3+ can have significant effects on the turnover, especially at low ionic strength where the binding becomes very tight and even seemingly small effects can be disproportionately represented in the Eadie-Hofstee plots . Therefore, even though Eadie-Hofstee plots are shown in Fig. 4, linear transformations of the Michaelis-Menten model are not used in deriving the kinetic parameters in this study. Our results indicate that the steady-state kinetics between R. sphaeroides CcO and horse cyt c2+ has only a single phase even at low ionic strength (Fig. 4).
The second mechanism by which cyt c3+ influences the rate of oxidation of cyt c2+ is by binding at a nearby allosteric site, leading to higher maximum turnover. Our results suggest that the interactions at the second allosteric site must also involve charged residues because the affinity is substantially lower at high ionic strength. It is not surprising that the introduction of the His-tag to a location very close to the catalytic site hinders the binding of cyt c (Fig. 1). The His-tag possibly induces a modification of the configuration of cyt c3+ at the allosteric site, resulting in slightly less efficient electron transfer from cyt c2+ at the active site. The calculated buried surface area at the interaction domains between P. denitrificans CcO and cyt c552 is about twice as large as the average protein-protein binding area expected for a dynamic complex  and is substantially larger than the contact areas observed in the crystal structures of the reaction centers-cyt c2 complex  and of the cytochrome bc1-cyt c complex. Therefore, it is very reasonable to postulate that two cyt c molecules can be accommodated at the unusually large binding domain of CcO, leading to the negative cooperativity in the binding.
The kinetic model based on the current findings, although different from some previous kinetic models [17, 18], is very similar to the most accepted model for the steady state kinetics of the mitochondrial CcO with cyt c2+, in which it is proposed that cyt c2+ can bind to a second, allosteric site, resulting in weakening the affinity at the active site for cyt c2+ . In the current work, using the R. sphaeroides enzyme, the proposed second site has a preference for the oxidized form of cyt c. This difference may result from small structural differences between the mitochondrial and bacterial CcO enzyme, i.e., cyt c2+ binds to the second site more effectively in the mammalian enzyme, but not in the bacterial version. Despite the similar overall structure between the reduced and oxidized forms of cyt c, subtle changes such as a reorientation of a heme propionate groups and alternations in hydrogen bond patterns have been detected upon reduction of cyt c . In fact, differential binding to the oxidized and reduced forms of cyt c has been reported for the bacterial photosynthetic reaction centers [39, 45, 46] and for the cytochrome bc1 complex . It should be noted that the effect of the proposed binding of cyt c3+ to the allosteric site on the apparent Vmax is an increase of only about 20%, from ~ 400 to 500 electrons/s at low ionic strength and from ~ 800 to 1050 electrons/s at high ionic strength (Fig. 5). Therefore, a slightly weaker binding by cyt c2+ at the allosteric site probably does not alter the turnover enough to significantly affect the single kinetic curve obtained at each cyt c3+ concentration (Fig. 5).
Although electrostatic interactions are crucial for the substrate binding at CcO, these are most critical for long-range recognition between the two molecules. Short range hydrophobic interactions, on the other hand, play a more direct role in achieving the configurations required for electron transfer as indicated by structural and mutagenesis studies [10, 13, 43, 48, 49]. In non-prototypical CcO such as cytochrome ba3 from Thermus thermophilus  and cbb3 oxidase from Vibrio cholera , few electrostatic residues are believed to be involved in the interactions with their respective cyt c2+ substrates. Recent sequence comparisons of the mammalian CcO enzymes have revealed rapid evolutionary changes of the charged residues at the binding site into the neutral ones in anthropoids, and it was speculated that subsequent changes in the enzyme’s kinetics could be part of the evolution process to achieve enlarged brains and longer life spans in primates . The improved spectrophotometric approach described in this study can be very useful in further investigations of the steady-state kinetics of different CcO enzymes or any other enzymatic systems involving two or more spectrally different species.
We thank Dr. Shelagh Ferguson-Miller for providing us with pCH37 vector and YZ200 strain.
This work was supported by the National Institute of Health Grant HL016101 to R. B. G.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.