Elucidating the Catalytic Mechanism of Sulfite Oxidizing Enzymes using Structural, Spectroscopic and Kinetic Analyses

Abstract

Sulfite oxidizing enzymes (SOEs) are molybdenum cofactor dependent enzymes that are found in plants, animals and bacteria. Sulfite oxidase (SO) is found in animals and plants, while sulfite dehydrogenase (SDH) is found in bacteria. In animals, SO catalyzes the oxidation of toxic sulfite to sulfate as the final step in the catabolism of the sulfur-containing amino acids, methionine and cysteine. In humans, sulfite oxidase deficiency is an inherited recessive disorder that produces severe neonatal neurological problems that lead to early death. Plant SO (PSO) also plays an important role in sulfite detoxification, but in addition serves as an intermediate enzyme in the assimilatory reduction of sulfate. In vertebrates the proposed catalytic mechanism of SO involves two intramolecular one-electron transfer (IET) steps from the molybdenum cofactor to the iron of the integral b-type heme. A similar mechanism is proposed for SDH, involving its molybdenum cofactor and c-type heme. However, PSO, which lacks an integral heme cofactor, uses molecular oxygen as its electron acceptor. Here we review recent results for SOEs from kinetic measurements, computational studies, electron paramagnetic resonance (EPR) spectroscopy, electrochemical measurements, and site-directed mutagenesis on active site residues of SOEs and of the flexible polypepetide tether that connects the heme and molybdenum domains of human SO. Rapid-kinetic studies of PSO are also discussed.

In animals, sulfite oxidase (SO) is a molybdenum cofactor dependent enzyme, that catalyzes the oxidation of sulfite to sulfate, using ferricytochrome c (cyt c_ox) as the physiological electron acceptor (eq 1) (1-4):

$${{\mathrm{SO}}_3}^{2\text{-}}+{\mathrm{H}}_2\mathrm{O}\to {{\mathrm{SO}}_4}^{2\text{-}}+2{\mathrm{H}}^{+}+2{\mathrm{e}}^{\text{-}}$$ (1)

This reaction is biologically essential, serving as the final step in the catabolism of sulfur containing amino acids, cysteine and methionine. SO also functions in detoxifying exogenously supplied sulfite and sulfur dioxide.

The SO family is comprised of plant assimilatory nitrate reductases (NR) and sulfite oxidizing enzymes (SOEs) found in animals, plants and bacteria. The SOEs include plant and animal SO and bacterial sulfite dehydrogenase (SDH). In animals, SO is located in the mitochondrial intermembrane space (5, 6). Plant SO (PSO) is localized in the peroxisomes and does not react with cyt c, but instead, PSO uses molecular oxygen as its terminal electron acceptor (7, 8). SDH is located in the periplasm (9), and the enzyme from Starkeya novella is the most thoroughly characterized example.

SO is physiologically vital in human metabolism, and severe neurological damage and early childhood death result from deficiency of SO activity due to an inborn metabolic defect. The symptoms of inherited sulfite oxidase deficiency include dislocation of the ocular lenses, attenuated growth of the brain and mental retardation (10, 11). Other symptoms that have been reported in patients with SO deficiency include poor feeding, hypoactivity, dyspnea, diffuse edema and seizures (12). The disease results either from molybdenum cofactor deficiency, which arises from a defect in its biosynthesis, or from certain point mutations within the SO enzyme (isolated SO deficiency) (13-15).

Prior to 2010 there had been no successful treatment for SO deficiency. Very recently the case of an infant female patient, who was diagnosed with molybdenum cofactor deficiency at 6 days old, has been reported (16). The patient showed a homozygous G175R (GGG-to-AGG) change in exon 10 of the MOCS1 gene, a mutation which disrupts the production of cyclic pyranopterin monophosphate (cPMP) a key step in the biosynthesis of the molybdenum cofactor (16-18). After substitution therapy with purified cyclic pyranopterin monophosphate (cPMP) (13), starting at 36 days old, all urinary metabolites of sulfite oxidase (sulfite, S-sulfocysteine, and thiosulfate) and xanthine oxidase deficiency (xanthine and uric acid) returned to almost normal readings and remained constant for more than one year. Thus, knowledge of the biosynthetic pathway of the molybdenum cofactor has made possible the first clinical treatment of SO deficiency. It is possible that an increased understanding of the catalytic mechanism of SOEs will result in additional treatment protocols for these disorders.

The development of recombinant human sulfite oxidase (HSO) has made it possible to prepare a library of mutations to study isolated SO deficiency, especially the effects of point mutations on the kinetic and spectroscopic properties of purified HSO (19-23). In a previous review Feng and coworkers have discussed the effects of several mutations on the intramolecular electron transfer (IET) kinetics of SO (24). For example, three mutations that have been associated with SO deficiency in patients (HSO^R160Q (21), HSO^G473D and HSO^A208D (20) are near the active site of SO. Laser flash photolysis measurements of their IET rates showed that the k_et values were decreased by three orders of magnitude relative to wild type. Thus, IET of these variants became rate-limiting and equal to the overall turnover rate (k_cat) determined from steady-state kinetic studies (24). Feng also discussed several future research directions concerning SO that still needed to be addressed. These topics included: 1) studies of electron transfer, using SDH as a model; 2) the importance of the flexible tether connecting the heme and molybdenum domains in docking the two domains for electron transfer; 3) pulsed EPR experiments to elucidate the nature of the sulfite/sulfate-bound and chloride-bound forms of SO; 4) computational modeling and simulations to provide further understanding of chloride-bound SO and to promote greater understanding of the role of specific amino acids in IET and SO catalysis. This review addresses these questions using molecular biology, rapid kinetics (laser flash photolysis and stopped-flow), advanced EPR spectroscopy, protein crystallography, and computational modeling.

Structural Motif of Sulfite Oxidizing Enzymes

Crystal structures have been solved for SO from Arabidopsis thaliana (25), chicken liver (26) and SDH from S. novella (27) (Figure 1). In addition, structures of recombinant wild type chicken SO proteins with and without sulfate-bound and recombinant R138Q chicken SO have been reported (28). Recently, structures of mutant forms of SDH have been solved and will be discussed in this review (29-31).

Figure 1.

Top row: Ribbon diagrams of the crystal structures of SOEs: 1A CSO (26), 1B PSO (32), 1C SDH (29); blue represents the molybdenum and C-terminal domain; red represents the heme domain. Bottom row: Selected amino acid residues near the molybdenum active site of the corresponding SOEs of the top row.

Active Site

The molybdenum active site is buried within a positively charged pocket, and has tyrosine, histidine and arginine residues that are conserved throughout SOEs and plant assimilatory nitrate reductases and have direct interactions with sulfate in CSO (26). The molybdenum active sites exhibit similar square pyramidal geometry around the molybdenum atom (Figure 1). There is an axial terminal oxo group, and the equatorial plane has three sulfur atoms (two from the dithiolene moity of the pyranopterin cofactor and one from a cysteine residue) and a water/hydroxo ligand.

Chicken Sulfite Oxidase (CSO)

CSO is the only intact eukaryotic crystal structure solved to date and possesses both a heme and a molybdenum domain. It is a 110 kDa homodimer and each subunit has three domains, an N-terminal b₅-type heme domain (Figure 1A, colored red), a central molybdenum binding domain, and a C-terminal interface domain (Figure 1A, colored blue) (26). The distance between the heme and molybdenum centers, within this structure is ~ 32 Å. However, the b₅-type heme and the molybdenum domains are connected by a flexible polypeptide tether that varies in length, depending upon the SO species. CSO has a tether that contains 13 amino acids, whereas the tether in HSO is 14 amino acids in length (based upon sequence alignment, Figure 2). The sequence of the tether of CSO is very different from that of HSO.

Figure 2.

Sequence alignment of the flexible tether regions of CSO and HSO. Amino acids highlighted in red are conserved between the two species, while those in blue are similar. Mutations discussed in this work are indicated in the HSO sequence: proline residues mutated to alanines are in bold type, and deleted residues are underlined.

The crystal structures reported for two forms of recombinant CSO (wt rCSO and rCSO^R138Q) contain only the C-terminal dimerization domain and the catalytic core domain, which houses the molybdenum cofactor center (28). Neither of these structures exhibit a heme domain. Superposition of the structures of wt rCSO and native CSO show that they have nearly identical structures. Both the wt rCSO and rCSO^R138Q SOEs were crystallized in the resting state (without sulfate) and with sulfate bound. wt rCSO contains two bound sulfates per monomer, whereas rCSO^R138Q contains only one bound sulfate per monomer. The active site of SO has three positively charged residues, R138, R190 and R450, which are presumed to help guide the negatively charged substrate into the active site of the enzyme. The active site pocket of the sulfate-bound and resting state of the wt rCSO and rCSO^R138Q forms, showed significant changes in the position of R450, indicating that this residue undergoes a conformational change upon substrate binding.

Plant Sulfite Oxidase

The crystal structure of PSO reveals that this enzyme is a homodimer, with an overall fold similar to CSO. PSO has two domains: an N-terminal Moco and a C-terminal dimerization domain (32), with no additional heme domain (Figure 1B). The active site is positively charged and has the conserved tyr, his, arg motif found in CSO (Figure 1E). The purpose of SO in plants is to detoxify excess sulfite in the cell, preventing sulfitolysis and also to serve as an intermediate enzyme in the assimilatory reduction of sulfate (33).

Sulfite Dehydrogenase

Sulfite dehydrogenase is a heterodimer, containing both a molybdenum-binding subunit (SorA) and a smaller heme c containing subunit (SorB) (29). The domains of the SorA subunit have the same overall fold as the Moco and dimerization domains of both CSO and PSO (Figure 1C). Unlike the eukaryotic CSO, SDH does not possess a flexible polypeptide tether. Also, the heme and molybdenum centers in bacterial SDH are within a closer proximity (~ 16 Å), than the distance found in CSO. The SorB subunit and the heme b domain of chicken liver SO have a similar overall shape. The crystal structures of the SDH^Y236F (30), SDH^R55M and SDH^H57A (31) mutants of SDH have also been solved. Combining these results with those for wt SDH has enabled potential electron transfer pathways between the heme and molybdenum redox centers to be postulated and evaluated by steady-state kinetics (31) and laser flash photolysis measurements (34).

The structure of SDH^R55M

Within SDH^wt, the side chain of R55 occupies a space that is close to the substrate binding site, forming hydrogen bonds with the equatorial oxo ligand of the molybdenum cofactor, Gln33 OE1 and a water molecule. It also forms a salt bridge with the propionate-6 of the heme moiety of the cytochrome subunit, which effectively locks the propionate group into the appropriate position for one of the proposed electron transfer pathways (29). However, within the SDH^R55M crystal structure, the side chain of M55 does not occupy the same space as does R55 in the wt structure. M55 bends away, packing into a small cavity between the side chains of L121 and Q33. The space previously occupied by R55 in SDH^wt is empty (Figure 3A) (31).

Figure 3.

Top: superposition of the active site structures of SDH^wt and SDH^R55M, showing that the M55 side chain (red) in SDH^R55M does not occupy the same space as R55 (blue) in SDH^wt. M55 bends away, packing into a small cavity between the side chains of L121 and Q33, and the space previously occupied by R55 in SDH^wt is empty (31). Bottom: structure of the active site of SDH^H57A, showing the two disordered conformations of the R55 side chain (31).

The structure of SDH^H57A

Kappler and co-workers found that the SDH^H57A structure confirms that the substitution of H57 with an alanine identifies an additional water molecule that sits within close proximity to the imidazole ring in SDH^wt (31). The water molecule is able to form hydrogen bonds with the N5 and O4 of the molybdopterin cofactor. Like SDH^Y236F, SDH^H57A appears to have increased the mobility of the R55 side chain and the interacting propionate-6 side chain of the heme (30, 31). Kappler and co-workers found that both Y236 and H57 are necessary to stabilize R55 in a position for optimal hydrogen bonding to the heme 6–propionate (Figure 3B).

Mechanism of the Sulfite Oxidizing Enzymes

The catalytic mechanism of animal sulfite oxidase, originally proposed by Hille (Figure 4), (35-38) has become widely accepted. The proposed mechanism of animal SO starts with the fully oxidized resting Mo(VI)/Fe(III) state. Two-electron oxidation of sulfite to sulfate occurs at the molybdenum center, which is reduced to Mo(IV). The first intramolecular electron transfer (IET) step occurs when the reduced Mo(IV) transfers one electron to the oxidized Fe(III) b₅ heme. The resultant Mo(V)/Fe(II) species contains a diamagnetic Fe(II) center and a paramagnetic 4d¹ Mo(V) center which is detectable by electron paramagnetic resonance (EPR) spectroscopy (39-42) (see following section). The Mo(V)Fe(II) species transfers an electron to the physiological electron acceptor (cyt c)_ox, producing the Mo(V)/Fe(III) state. A second IET step generates the Mo(VI)/Fe(II) state, and reduction of a second molecule of cyt c regenerates the fully oxidized Mo(VI)/Fe(III) state of the enzyme. The portion of the catalytic mechanism, in which the oxidized enzyme is reduced by two electrons and sulfate is produced, is referred to as the reductive half reaction. The oxidative half reaction thus involves two one-electron transfers of electrons to cyt c to yield oxidized SO and two molecules of reduced cyt c.

Figure 4.

Proposed oxidation state changes occurring at the Mo and Fe centers of animal SO during the catalytic oxidation of sulfite and the concomitant reduction of (cyt c)_ox.

EPR Spectroscopy

The utility of EPR spectroscopy for investigating the Mo(V) sites of SOEs formed in the reductive half reaction of Figure 4 has been recognized since the early 1980s, with the identification of three distinct CW EPR forms of vertebrate SO: low pH (lpH); high pH (hpH); and phosphate inhibited (P_i) (43-45). The coordination structures of the Mo(V) sites of SOEs obtained by pulsed EPR spectroscopy have been discussed previously (46), and the applications of pulsed EPR methods to SOEs have also been extensively reviewed (47). Therefore, only a few more recent EPR structural results for SOEs are presented here.

Bray and co-workers originally proposed that chloride and phosphate were directly coordinated to Mo(V) in the lpH and P_i forms, respectively, but no ^35,37Cl (I = 3/2) or ³¹P (I = 1/2) hyperfine interactions (hfi) were directly detected for these nuclei by CW EPR (42-45, 48). In 1996, electron spin echo envelope modulation (ESEEM) studies between 8-18 GHz detected ³¹P hfi for the P_i form of vertebrate SO and demonstrated that it has an equatorial monodentate phosphate group (42). However, until recently, the role of Cl⁻ in lpH SO remained elusive. A multifrequency CW EPR experiment using HSO and ³⁵Cl⁻- and ³⁷Cl⁻-enriched solutions showed very small spectral differences that were interpreted as indicating the presence of an axial Cl⁻ ligand trans to the terminal oxo group, but the limited resolution of the CW EPR spectra precluded reliable determination of the hfi and nuclear quadrupole interaction (nqi) parameters required to define the structure of lpH SO (49). Very recently, the magnetic resonance parameters of ^35,37Cl (I = 3/2) in lpH SO were finally measured by ESEEM methods, and density functional theory (DFT) calculations favored a structure in which the Cl⁻ anion is hydrogen bonded to the equatorial OH ligand and surrounded by several proton donors from nearby water molecules and charged amino acid side chains (50).

At some stage of the catalytic cycle of Figure 4, the product, SO₄²⁻, has to be released from the active site. This release might occur from either the Mo(IV) or Mo(V) states of the reductive half reaction (Scheme 1). In 1982 Bray et al. reported that the presence of excess sulfite leads to another EPR-active Mo(V) species at low pH, which they attributed to a sulfite complex of the Mo center (51). Experimental evidence for a Mo(V) species with a sulfur-containing ligand was obtained only recently from pulsed EPR investigations of samples reduced with sulfite labeled with ³³S (I=3/2) at low pH, which allowed, for the first time, direct detection of coupling to ³³S. This coupling was initially observed in PSO from A. thaliana and was assigned to a Mo(V)–SO₄ complex (52). It was suggested that a conformational change of some of the nearby amino acid residues blocked hydrolysis of the sulfate ligand, and a structure similar to the P_i form was proposed (42), with a monodentate sulfate ligand (52). Subsequently, numerous examples of blocked forms have been observed for mutants of HSO and of SDH (22, 53). More recently, it was shown that the blocked form is also produced in chloride-depleted samples of wt HSO. Addition of excess chloride to the blocked forms of SOEs converts them, at least partially, to the familiar lpH form (54).

Scheme 1.

Possible Reaction Pathways for the Catalytic Oxidation of Sulfite by SDH or SO^a

^aFor the sake of simplicity, the enzyme and substrate are depicted as Mo^VIO and SO₃^2-, respectively. The pathways differ in the sequence of steps which transform the enzyme-substrate complex, Mo^IV(OSO₃^2-), into product (sulfate) and the reoxidized Mo(VI) state of the enzyme. The pathway colored red is the one commonly proposed, in which product release precedes reoxidation of the enzyme. For the pathway colored black, the enzyme-substate complex is oxidized by two electrons prior to product release (34). See the text for additional discussion.

The hydrolysis of the product (sulfate) with H₂¹⁷O in the reductive half reaction (Figure 4) provides a convenient method for introducing ¹⁷O (I = 5/2) into the Mo(V) center of SOEs (46). It has long been known that the lpH and hpH forms show distinctly different ¹⁷O hfi (55), but it was not possible to investigate the details of their hfi and nqi parameters prior to the development of high resolution variable frequency pulsed EPR spectrometers (56) and sophisticated DFT programs to relate experimental results to electronic and molecular structure (57, 58). At high pH, SOEs exhibit two types of ¹⁷O atoms in pulsed EPR experiments. One has a large hfi of ~30-40 MHz and can be associated with the equatorially coordinated O-containing ligand. The second has a much smaller hfi (~ 5-6 MHz). For hpH SO, this latter 17O signal was assigned to the axial oxo-Mo(V) group from comparison to the results for model compounds and from extensive DFT calculations (39, 59). At low pH (~6) the blocked forms of SOEs also exhibit two types of ¹⁷O signals (22). The larger hfi (~18 MHz) is again associated with the equatorial O-containing ligand, but the origin of the smaller ¹⁷O hfi (5-6 MHz) is less clear because its nqi value (~ 4-5 MHz) is substantially larger than those for known axial oxo ligands in model compounds (1.5 MHz) (59). For studies of blocked forms of SOEs in H₂¹⁷O, it is also important to note that the oxygen atoms of sulfite exchange very rapidly (60). Consequently, all of the O atoms of a coordinated sulfite (reactant) or sulfate (product) anion will be labeled with ¹⁷O in the reductive half reaction of Figure 4. Recent extensive DFT calculations suggest that for the blocked form the smaller ¹⁷O hfi arises from remote O atoms on the sulfite (or sulfate) group bound to Mo(V) (61). The DFT calculations for the sulfite-coordinated structure originally proposed by Bray et al. (51) give somewhat better agreement with the experimental nqi parameters for this weakly coupled ¹⁷O and for the central ³³S atom compared to the sulfate-coordinated structure proposed previously by us (52). Figure 5 summarizes the current view of the structures of the lpH, hpH and blocked forms of SOEs.

Figure 5.

Proposed structures of the molybdenum active sites of SOEs from pulsed EPR experiments and DFT calculations for the lpH (50), blocked (61), and hpH forms (58). The blocked form shown on the left has bound reactant (sulfite); the form shown on the right has bound product (sulfate). See text and (61) for details.

The availability of more than 4 million genome sequences from databases and more than 40,000 protein structures from the protein data bank (PDB) has made it possible to analyze protein folds of SOE's in all domains of life using bioinformatics (62, 63). After multiple sequence alignments, Dahl and Kappler determined a phylogenetic relationship with three distinguishable groups of SOEs (63). Group 1 includes the pathogen enzymes; group 2 contains the classic SOEs and nitrate reductases; and group 3 contains enzymes from Archaea, phototrophic and soil bacteria. Chicken, human, and plant SO, along with SDH are all included in group 2. This group contains a dimerization domain and a molybdenum domain (26, 29, 32).

Similar bioinformatic analyses performed by Weiner and co-workers identified an E. coli protein, YedY as a part of the SOEs family (62, 63). The crystal structure of YedY has been solved and has revealed that this protein has only the molydopterin binding SO fold (64). Yed Y is categorized in group 3. Its heme counterpart is a separate monoheme cytchrome b protein known as YedZ (64). To date, very little is known about the function of YedY. Sequence comparison of plant, chicken, nitrate reductase, S. novella SDH and YedY reveal that there are fifteen residues that are conserved throughout these five structures (64). One of these conserved residues is the molybdenum coordinating cysteine residue (64).

Intramolecular Electron Transfer (IET) Kinetic Studies in Sulfite Oxidizing Enzymes

Intramolecular electron transfer (IET) is essential for the catalyzed reaction of SO. In animal SO and SDH, IET occurs between the heme and the molybdenum centers of the enzyme (Figure 4). The IET process can be studied using laser flash photolysis techniques. Laser flash photolysis experiments can be performed anaerobically in solutions containing 5-deazariboflavin (dRF) and freshly prepared semicarbazide or EDTA (AH₂) as a sacrificial reductant. It has been previously shown that dRF radicals are generated with a laser pulse and will then rapidly reduce the Fe(III) heme center of animal SO and S. novella SDH by one electron (Figures 4 and 6). This process corresponds to the reverse of the second physiologically essential IET step that is boxed in Figure 4. The initial rapid increase in absorbance at 555 nm is due to the reduction of the heme from the Fe(III) to the Fe(II) form (Eq. (4)) by the dRF radical. This process is followed by a slower absorption decay that is due to the net IET from the Fe(II) to Mo(VI), i.e., the intramolecular re-oxidation of Fe(II) (forming the Fe(III)/Mo(V) species (Eq. (5)). The laser flash photolysis technique follows the IET process in the reverse direction of the enzymatic turnover.

Figure 6.

Kinetic transient obtained at 555 nm upon photoexcitation of a solution containing wild type human SO, dRF, and 0.5 mM fresh semicarbazide hydrochloride (pH 7.4). The portion of the figure outlined by the orange box points to the initial heme reduction by dRFH·; this process is pseudo first order, and its rate depends on protein concentration. The dark blue box points to heme reoxidation due to the subsequent IET between the Mo and Fe centers; this process is independent of protein concentration, consistent with its intraprotein nature. The red solid line indicates a single-exponential fit to the IET phase. K_eq=b/a. Copyright Elsevier publishing; reproduced from Feng, C., Tollin, G., and Enemark, J. H. (2007) Sulfite Oxidizing Enzymes, Biochim. Biophys. Acta 1774, 527-539, with permission.

The methodologies used to obtain rate and equilibrium constants for IET in SO have been described previously (65) and are summarized below. The laser flash photolysis apparatus system has been extensively described (66) as has been the basic photochemical process by which 5-deazariboflavin semiquinone (dRFH^•) is generated by reaction between triplet state dRF and the sacrificial reductant and used to reduce redox-active proteins (eqs. 2-5) (67-69).

$$\mathrm{dRF}\stackrel{\mathrm{h}\nu }{\to }{}^1\mathrm{dRF}\to {}^3\mathrm{dRF}$$ (2)

$${}^3\mathrm{dRF}+{\mathrm{AH}}_2\to {\mathrm{dRFH}}^{•}+{\mathrm{AH}}^{•}$$ (3)

$${\mathrm{dRFH}}^{•}+\mathrm{Mo}\left(\mathrm{VI}\right)∕\mathrm{Fe}\left(\mathrm{III}\right)\to \mathrm{dRF}+\mathrm{Mo}\left(\mathrm{VI}\right)\mathrm{Fe}\left(\mathrm{II}\right)$$ (4)

$$\mathrm{Mo}\left(\mathrm{VI}\right)\mathrm{Fe}\left(\mathrm{II}\right)\underset{k_{\mathrm{r}}}{\overset{k_{\mathrm{f}}}{\rightleftharpoons }}\mathrm{Mo}\left(\mathrm{V}\right)∕\mathrm{Fe}\left(\mathrm{III}\right)$$ (5)

The IET rate constant can be calculated by fitting the heme reoxidation curve with an exponential function (eq. 6) where the IET rate constant is the sum of the forward and the reverse electron transfer rate constants (k_f and k_r, respectively, in eq. 7).

$$\frac{\mathrm{d}{A}_{5555}}{\mathrm{d}t}=a+b\text{exp}\left(\text{-}{k}_{\mathrm{et}}t\right)$$ (6)

$$k_{\mathrm{et}}=k_{\mathrm{f}}+k_{\mathrm{r}}$$ (7)

The equilibrium constant can then be calculated using the parameters a and b, which are determined from the kinetic traces. A₀ is the absorbance extrapolated to t = 0, assuming that the photochemically-induced reduction of SO is instantaneous.

$$a={\mathrm{A}}_{\mathrm{o}}\frac{k_{\mathrm{r}}}{k_{\mathrm{et}}}={\mathrm{A}}_{\mathrm{o}}\frac{k_{\mathrm{r}}}{k_{\mathrm{f}}+k_{\mathrm{r}}}$$ (8)

$$b={\mathrm{A}}_{\mathrm{o}}\frac{k_{\mathrm{f}}}{k_{\mathrm{et}}}={\mathrm{A}}_{\mathrm{o}}\frac{k_{\mathrm{f}}}{k_{\mathrm{f}}+k_{\mathrm{r}}}$$ (9)

$$K_{\mathrm{eq}}=\frac{k_{\mathrm{f}}}{k_{\mathrm{r}}}=\frac{b}{a}$$ (10)

The forward and reverse rate constants (k_f and k_r, respectively) of IET can then be calculated from the equilibrium constant (eq. 10), thereby providing quantitative information about interdomain electron transfer in the enzyme. Note that k_f in these flash photolysis experiments is actually the reverse of the net physiological IET direction.

Tether Linking the Heme and Molybdenum Domains of SO: Past and Recent Studies

Until recently, little has been known about the role of the tether within animal SO (Figure 2), and this linkage is not present in the plant and bacterial SOE's. The initial work on this tether was performed by Feng and co-workers and has been discussed previously (24, 70). They proposed that interdomain motion that decreases the Mo to Fe distance is essential for rapid IET and that the flexible tether linking the two domains of SO facilitates this motion. Consistent with this, laser photolysis revealed that the IET rate constant in CSO decreased with increasing solvent viscosity. Steady-state kinetics and EPR measurements on SO were also consistent with this hypothesis (70).

Kawatsu and Beratan have used a computational model to explore the mechanisms of IET in proteins such as SO, in which domains containing two cofactors are linked by a flexible tether (71). They found that the constrained conformational flexibility of the tether introduces an entropic component to the effective donor–acceptor interaction potential that produces a kinetic regime intermediate between “unimolecular” and “bimolecular”. Their calculations predicted that for SO the tether length may control the transition between the electron tunneling and diffusion-limited regimes. Recently, Pushie and George used molecular dynamics calculations to investigate the large-scale domain motions of CSO (72). They observed motion of the N-terminal heme domain into an orientation similar to that postulated for rapid electron transfer. The simulations also probed the dynamics of specific active site residues to further understand the structural and thermodynamic details of SO (72). Very recently, Utesch and Mroginski investigated the interdomain motion of CSO by combining steered molecular dynamics with all-atom molecular dynamics (73). They obtained a stable 3D “docked” structure with a relatively short Mo···Fe distance (~19 Å), similar to that proposed previously from the rapid IET rates measured experimentally (24, 65). The best theoretically calculated IET pathways for the “docked” model involved 1-3 interfacial water molecules and gave IET rates that were within an order of magnitude of the experimental values.

Prior to 2010 there were no experimental data from site-specific mutants to investigate the role of conformational changes of the tether that had been proposed from the effects of viscosity on IET (70) and from theoretical calculations (71). Recently, the proposed role(s) of the tether in facilitating conformational change and reactivity have been explored by altering both the length and flexibility of the tether in HSO by site-specific mutagenesis, and the reactivities of the resulting variants were studied using laser flash photolysis and steady-state kinetics assays (79).

In order to change the flexibility of the tether, two conserved proline residues were mutated to alanines, P105 (adjacent the heme domain) and P111 (at the center of the tether) (Figure 2). The double mutant, P105A/P111A has also been made. While P105 is conserved within CSO and HSO, P111 is not. However, P111 is conserved within other mammalian forms. Figure 7 illustrates the dependence of k_et on pH and mutation. The data show that the P105A mutation decreases k_et by approximately 3-fold at pH 7.4. In contrast, the P111A mutation has only a minimal effect on k_et. The IET rate constants for the P105A/P111A double mutant are similar to those for P105A, suggesting that the P105A change is the primary cause of the decrease in IET that is observed in this mutant.

Figure 7.

IET Rate Constants for the Proline to Alanine Tether Mutants (79).

The equilibrium constants (K_eq) for the P111A mutant show a pH dependence with a trend that is similar to wt, i.e. K_eq increases with increasing pH. However, this trend is not seen with P105A. The kinetic data indicate that the P111A mutation in the middle of the tether has little effect upon the IET kinetics. However, the much larger effects on the IET kinetics for the P105A mutation suggest that this conserved proline, which is adjacent to the heme domain, promotes a tether conformation that results in optimal IET.

In addition to these proline to alanine mutations, several non-conserved residues (K108, V109, A110, T112 and V113, Figure 2) have been deleted from the tether in order to determine the effects of tether length on the IET kinetics of HSO. Deletions of three (ΔK108V109A110 (ΔKVA)), four (ΔK108V109A110T112 (ΔKVAT)), and five (ΔK108V109A110T112V113 (ΔKVATV)) non-conserved amino acids from the tether result in steadily decreasing IET rate constants, from 467 ± 19 s^-1 in wt HSO down to 5.59 ± 0.03 s^-1 in ΔKVATV HSO at pH 7.4 (Figure 8). Interestingly, the largest decrease occurs on deletion of the fifth residue (ΔKVATV). Additionally, all three deletion mutants showed decreasing IET rates with increasing pH from 6.8 to 7.6. K_eq was significantly larger for ΔKVATV HSO (0.91 ± 0.03) in comparison to the wt HSO value (0.46 ± 0.02) and the other tether deletion mutants, but all the enzymes have a maximal K_eq around pH 7.4. To further understand the cause of this large increase in K_eq for ΔKVATV, spectroelectrochemistry studies were performed on wt and ΔKVATV HSO (Table 1), which will be discussed in the spectroelectrochemistry section of this review.

Figure 8.

IET Rate Constants for Tether Deletion Mutants (79).

Table 1.

Electrochemical Midpoint Potentials of SOEs cofactors^a

	HSO	CSO	SDH	P105A HSO	ΔKVATV HSO
pH	7.5	7.5	7.1	7.5	7.5
Heme Cofactor (Fe (III/II))	+ 62 ± 2c (79)	~+ 69d (78)	+ 177d (82)	+ 56 ± 3c (79)	+ 42 ± 2c (79)
Molybdenum Cofactor (VI/V)	+ 41 ± 1b (79)	n.d.	+ 264d (82)	+ 36 ± 1b (79)	+ 41 ± 1b (79)

^aAll potentials are in mV vs SHE.

^bCalculated from ΔE^o_{Reaction (Heme-Mo)} and the experimental heme potential.

^cDetermined by spectroelectrochemistry, see Figure 9.

^dDetermined by Protein Film Voltammetry

In parentheses are the cited references.

The steady-state kinetics revealed that the largest effect observed was on the Michaelis-Menten constants (K_m) for P111A and P105A/P111A, which were approximately five-fold higher than for wt (data not shown). The increase in K_m caused a 30% decrease in the catalytic efficiency (k_cat/K_m) of the enzyme. From these results it was concluded that the P111A mutation may affect the binding of sulfite to the enzyme. In addition, the steady-state data revealed that the k_cat values for each of these mutations are much smaller than k_et. Therefore, the IET process is not rate-limiting in catalysis, which is also true of the wt enzyme.

The steady-state kinetic parameters for ΔKVA and ΔKVAT HSO, are slightly higher in comparison to the wt HSO. However, the turnover number for ΔKVATV HSO (10.6 ± 0.3 s^-1) is approximately one-third that of wt (26.9 ± 0.5) at pH 8.0. Unlike the proline to alanine mutations, IET for the ΔKVATV deletion mutant became rate limiting in catalysis. Thus, decreasing the length of the tether apparently influences the docking of the heme domain onto the molybdenum domain, thereby changing the distance and relative orientation of the two redox centers. This effect is the probable cause for the decrease in k_et.

Very recently the tether sequence of CSO (Figure 2) was introduced into HSO. The expectation was that this chimeric HSO protein would show IET kinetics that were faster than wt HSO, and more similar to those of CSO. Surprisingly, however, IET for the chimeric HSO was actually slower than wt HSO (74). This result suggests that the composition of the non-conserved tether sequence of animal SOs may have become optimized for individual species. The kinetic results for all of the tether mutants point to a critical role for the tether in facilitating the IET reaction between the heme and molybdenum centers.

Mechanistic considerations based upon conserved arginine residues in the active site of SOE's

A new analysis of possible mechanistic pathways of sulfite oxidizing enzymes has emerged from laser flash photolysis and EPR studies on wt and R55 variants of SDH (34). Three variants of SDH (R55M, R55Q and R55K) were made to explore the role of R55 in the pathway for IET in SDH. The SDH^R55Q variant had the same Arg to Gln substitution that is observed in R160Q in humans, which is a fatal mutation, causing sulfite oxidase deficiency. HSO^R160Q shows a substantial decrease in IET, which is almost 400 fold lower than HSO^wt (122 ± 5 s^-1) (21). For SDH^R55Q the laser flash photolysis data revealed that k_et was 3-fold larger than for SDHwt. Both K_eq and k_et were mildly pH dependent within SDH^R55Q. At pH 5.1, there was a 2-fold increase in K_eq for this variant (34).

The IET rate constant for SDH^R55K was unchanged with varying pH, which is a similar trend to SDH^wt. However, the IET rate could not be determined at pH 6.0. IET in SDH^R55K is much slower (15 ± 0.7 s^-1) when compared to SDH^wt. Between pH 5.1 (0.59 ± 0.02) and 5.5 (0.34 ± 0.03) the SDH^R55K variant showed a decrease in K_eq. These data are much different than those for HSO^R160K, which showed a decrease in IET.

In the SDH^R55M variant the hydrogen bonding interactions within the active site are disrupted. The IET rate constant for SDH^R55M (217 ± 25 s^-1) was 2-fold larger than SDH^wt (122 ± 5 s^-1) at pH 6, and increased 4-fold (535 ± 44) by pH 5.1. While the IET rate constant for SDH^R55M was pH-dependent; K_eq was nearly pH independent. These behaviors are the opposite of those observed for SDH^wt. Therefore, both the forward and the reverse IET rate constants for SDH^R55M increase with decreasing pH, raising the possibility that SDH^R55M has a more efficient tunneling pathway for IET (21, 34).

The structure of SDH^R55M shows that there is a sulfate anion bound within the active site, which is different from SDH^wt, SDH^Y236F (30) and SDH^H57A (31). The pulsed EPR studies of SDH^R55M show only a hpH–type signal, similar to SDH^wt enzyme, with no pH dependence (53). SDH^R55Q exhibits the blocked form, with bound sulfate at low pH, which was verified by ³³S couplings observed upon reduction with ³³S-labeled sulfite (53). This blocked feature has also been observed in the fatal mutant, HSO^R160Q (22).

SDH^R55M and SDH^R55Q are the first SOE active site variants to display an increase in the IET rate constant. This effect is the opposite of that observed in HSO^R160Q (21). These results indicate that the conserved arginine within the active site of SOEs is not a required part of the electron transfer pathway between the heme and molybdenum domains. However, the result is consistent with the positively charged arginine influencing the docking between the two domains in animal SO. Additional kinetic studies of SDH catalysis have shown that the R55 residue is important for effective substrate binding, and that KMsulfite,app for SDHR55M is 2-3 orders of magnitude larger than that for SDH^wt (31). This result is also consistent with that data for SDH^R55K, which showed a K_Msulfite,app factor of ~10 times larger than that for SDH^wt (31).

Based upon the kinetic data and the pulsed EPR studies of SDH^wt and variants, and similar studies on HSO^R160Q, alternative mechanistic possibilities for the catalytic cycle of SOEs must be considered. Scheme 1, proposed by Emesh et al., summarizes alternative possibilities for the catalytic cycle of SOEs (34). The sequence colored in red is the prevailing view (1, 24, 75), in which product is released (k₂) from the enzyme-substrate complex. This is followed by two sequential one-electron oxidations of the Mo center (k₄ and k₇). This mechanism was discussed previously above (see Figure 4).

A second alternative (colored in blue) would be for the enzyme-substrate complex to be oxidized by one electron (k₃ in Scheme 1). This is followed by product dissociation (k₅) and then another oxidation by one electron (k₇). Emesh et al. suggested that this pathway is operative in SDH^wt and the R55 variants because IET rates are significantly faster than catalytic turnover (34). Direct evidence for the one-electron oxidized enzyme-substrate complex, i.e., the Mo^V(OSO₃^2-) species of Scheme 1, has been obtained by pulsed EPR studies of HSO^R160Q that has been reduced by sulfite labeled with ³³S (I=3/2) (22) and other blocked forms of SOEs (31, 52, 53).

A third possibility shown in black in Scheme 1 is for the enzyme-substrate complex to be oxidized by two electrons back to the resting state Mo(VI) (k₃ and k₆) before release of sulfate (k⁸). The IET and steady-state kinetics of the SDH^Y236F variant are consistent with this pathway. This variant is catalytically competent even though it does not exhibit IET by laser flash photolysis (steps k₇ and k_-7) (30).

Kinetics of Plant SO

Hille and co-workers have recently reported in-depth rapid kinetics and steady-state studies of PSO, using stopped-flow methods. The data demonstrated that superoxide is produced during the course of the reaction between the reduced enzyme and molecular oxygen and then decays to hydrogen peroxide (76). The rate constant was linearly dependent on the concentration of molecular oxygen, and the second-order rate constant was k_ox = 8.7 ± 0.5 × 10⁴ M^-1s^-1. They also found that the reaction, in the presence of cyt c was biphasic, with a fast and a slow phase. This result indicated that the first equivalent of superoxide was generated from the rapid one-electron oxidation of the fully reduced Mo(IV) state to Mo(V). This reaction was followed by the slower oxidation of the Mo(V) to the Mo(VI) state.

Electrochemistry

Studies of Animal SO

As discussed above, SO is a multi-domain enzyme, containing heme and molybdenum redox centers. An extensive micro-coulometric study of native CSO by Spence et al. investigated the reduction potentials of the Fe(III/II), Mo(VI/V) and Mo(V/IV) couples as a function of pH and anions in the medium (77). The Fe(III/II) couple was nearly independent of pH, whereas the Mo(VI/V) couple showed a pH dependence of ~60 mV per pH unit, characteristic of an accompanying proton transfer. More recently, Elliott et al. investigated native CSO by protein film voltammetry (PFV) (78). A catalytic current was observed in the “potential domain” of CSO upon addition of sulfite. From the limiting current, they obtained a turnover number, k_cat of 2-4 s^-1, which is significantly slower than the reported turnover numbers of SO in solution (100 s^-1) (38). The PFV data suggested that a large fraction of the SO molecules immobilized on the electrode surface do not engage in catalysis, but undergo non-catalytic electron transfer. The observed behavior is independent of the electrode and does not seem to be an artifact of the surface-protein interaction. The midpoints of the catalytic waves (E_mid) in the PFV experiments show very small pH dependence, which is only consistent with the wave arising from action at the Fe(III/II) center (77) and identifies it as the central distributive site for catalytic ET. The redox potential at pH 7.5 is approximately +65 mV vs SHE, which is similar to the heme potential of HSO^wt determined from spectroelectrochemistry (see discussion below) (79).

The reduction potential of the b₅ heme of wt and mutant HSO has recently been measured using spectroelectrochemistry, following procedures described previously for heme proteins (77, 79-81). Since the ratio of oxidized to reduced forms of each redox species is directly related to the absorbances of the optical spectra via Beer's law (assuming the extinction coefficients of the reduced and oxidized species are different at the chosen wavelength), the change in absorbance with respect to applied potential can be fit to the Nernst equation (eq. 11) using the nonlinear-least-squares fitting algorithm in the software Origin©.

$$E_{app}=E^0+2.303(\mathrm{RT}∕\mathrm{nF}){\text{log}}_{10}(\left[\mathrm{Ox}\right]∕\left[\mathrm{Red}\right])$$ (11)

where E_app is the applied potential, E⁰ is the midpoint potential determined from these data, and [Ox] and [Red] are, respectively, the concentrations of the Fe(III) and Fe(II) states of the b₅ heme of HSO.

Based upon kinetic results by Johnson-Winters et al. (79) the laser flash photolysis data for ΔKVATV resulted in K_eq values that were approximately 0.91, compared to 0.46 for the wt. To confirm whether this was a result of the change in the midpoint potential between the heme and molybdenum, spectroelectrochemistry studies were performed to determine the Fe(III/II) potentials for wt and ΔKVATV HSO. The corresponding midpoint potentials vs. SHE are +62 ± 2 mV for wt HSO and +44 ± 2 mV for ΔKVATV HSO (Figure 9, Table 1). Thus, the Fe (III/II) potential for ΔKVATV HSO is 18 ± 3 mV more negative than wt, supporting the laser flash photolysis kinetic data and favoring the Fe(III)/Mo(V) species in eq. 5. Using the K_eq value from the laser flash photolysis data and the heme potential, the potential of the molybdenum center can be calculated, as well. The calculated potentials of the Mo(VI)/Mo(V) couple for ΔKVATV and wt HSO are +41 ± 2 and +42 ± 2 mV, respectively (Table 1). Therefore, it was concluded that the change in K_eq for ΔKVATV was primarily due to a change in the potential of the heme. Perhaps the deletion of five amino acids from the tether changes the exposure of the heme to solvent, thereby causing a shift in the potential and in K_eq.

Figure 9.

Spectroelectrochemical titration of the b₅ heme of wild-type HSO at pH 7.5 and 27°C (79). The inset shows the fit of the data to eq. 11 at 413 nm (78); black = wt; red = ΔKVATV mutation.

Spectroelectrochemistry has also been used to measure the Fe(III/II) potential of the P105A mutant which shows much slower rates of IET compared to wt, but similar K_eq values (Figure 7). Table 1 shows that the Fe(III/II) potential for P105A is essentially identical to wt, as is the calculated Mo(VI/V) potential. Thus, the P105A mutation does not change the thermodynamic driving force for IET (as seen from K_eq) or the individual potentials of the two redox centers. Therefore, the 2-4 fold decrease in k_et for P105A mutant compared to wt (Figure 7) presumably reflects differences in their IET pathways.

Studies of SDH

From PFV studies of SDH^wt on a pyrolytic graphite working electrode, McEwan and coworkers identified the heme and molybdenum potentials as: E_m,8 (Fe(III/II)) +177 mV; E_m,8 (Mo(VI/V)) +211 mV; and E_m,8 (Mo(V/IV)) -118 mV vs SHE (82). The Michaelis-Menten constant (K_m) for SDH^wt obtained upon saturation of the catalytic current with sulfite was similar to previous steady-state assay values (26 μM) (13). Also, using square wave voltammetry, they were able to show that the SDH^wt potentials of the heme and molybdenum exhibit a pH dependence.

Attempts to extend the PFV measurements to the SDH^R55 variants, however, were not successful (34). Also, it was not feasible to directly determine the Mo(VI/V) potentials of the SDH^R55 variants by EPR titrations, because of the large amount of protein required for such experiments. However, the Mo(VI/V) potentials of SDH^wt can be calculated by using K_eq (1.57 at pH 5.2) from the laser flash photolysis experiments and the Fe(III/II) potential determined from spectroelectrochemistry, as described above for HSO. For SDH^wt this procedure gave a ΔE° of 12 mV and a calculated Mo(VI/V) potential of 196 mV (34). This calculated Mo(VI/V) potential is very different from the +381 mV reported for the direct experimental measurement for SDH^wt at pH 5.2 by PFV (82). It is unknown why these potential values vary by so much. However, it could be due to the fact that the technique of PFV requires adsorption of the protein on an electrode surface and immobilization within a surfactant film. This is very different than spectroelectrochemical determination of the potential in solution (34).

Conclusions

As noted above, a review of SOEs in 2007 by Feng (24) and coworkers concluded with a list of topics to be addressed in future research in this area. Substantial progress has already been made on several of these topics by the combined approaches of molecular biology, rapid kinetics, advanced spectroscopy, protein crystallography, and computational modeling. Scheme 2 summarizes the current status of research on SOEs and some of the remaining challenges. Specifically, recent advances include:

1. Flash photolysis studies of variants of HSO and of bacterial SDH have provided new insights concerning the amino acids involved in direct IET through the protein medium. In humans, the HSO^R160Q mutation is fatal, and the rate of IET in the purified protein is dramatically reduced relative to HSO^wt. However, the analogous SDH^R55Q mutation of the conserved active site arginine in bacterial SDH does not decrease the rate of IET. In SDH the heme and Mo domains are locked in place by strong interactions between the two subunits, whereas in HSO the heme and Mo domains are linked by a flexible tether and domain reorientation is required in order to accomplish IET. These combined flash photolysis results suggest that in HSO the primary role of the positively charged R160 residue is to aid in docking the heme domain with the Mo domain for efficient IET and catalysis (34). The relative unimportance of R55 in SDH implies that this conserved active site arginine is not directly on the pathway for IET. Calculations of the IET pathway(s) for SDH should provide further insight concerning this hypothesis.

2. Substantial progress has been made in elucidating the effect of the tether linking the heme and Mo domains of HSO on the rates of IET and steady-state kinetics. We have shown that both the length and composition of the tether strongly affect the rate of IET (79). Theoretical calculations of the effects of interdomain motion on IET are beginning to appear (71), and the first molecular mechanics calculations on the time evolution of the structure of CSO have recently been described (72, 73). The increasing availability of GPU-based computer clusters should make molecular mechanics calculations of longer time trajectories of interdomain motions in SO more computationally tractable in the future (83). This is critical since IET occurs on the millisecond time scale, which is presently computationally inaccessible.

3. Pulsed EPR spectroscopy of “difficult” nuclei (¹⁷O, ³³S, ^35,37Cl) with I > ½ has been combined with DFT calculations to elucidate the structures of the Mo(V) states of SOEs and to gain insight concerning their catalytic mechanisms. The long-standing question of the structure of the lpH form of SO has been resolved from ^35,37Cl pulsed EPR and DFT calculations; the chloride ion is in the second coordination sphere and hydrogen-bonded to the equatorial Mo-OH group and surrounding water molecules and proton donors from amino acid side chains (50). Observation of ³³S couplings in SOEs reduced by ³³S-labeled sulfite at low pH in chloride-depleted solutions has shown that the blocked form is a common Mo(V) species in SOEs. Addition of excess chloride converts the blocked form to the lpH form in most cases (54). DFT calculations on models for ³³S- and ¹⁷O-labeled blocked SO suggest that a structure with O-bound sulfite is most likely. The weakly coupled ¹⁷O atom detected in the blocked form is assigned to remote oxygen atoms of the sulfite (sulfate) ligand (61). Detection of the nqi parameters for strongly coupled equatorial ¹⁷O ligands remains difficult, but recent advances in W-band ESEEM methods hold promise for future advances (84).

4. X-ray crystal structures of several variants of SDH (30, 31) have appeared, and structures of the Mo domain of recombinant CSO and CSO^R138Q have been described (28). However, to date the only structure of an intact form of vertebrate SO is the original report of native CSO at 1.9 Å resolution (26). The crystallization and structure determination of recombinant HSO, both wt and mutants, remain as challenges for future research on SO.

Scheme 2.

Overview of the current state of research on sulfite oxidizing enzymes (SOEs). From bioinformatic analyses, the diverse sulfite oxidase family of proteins that contains the same molybdenum cofactor center (1) can be classified into at least three groups (63, 85). This review has focused on the SOEs. Animal sulfite oxidases possess two redox active domains and present fundamental biophysical problems relating to intramolecular electron transfer (IET), the relationship of IET rates to steady-state kinetics, the overall conformation of the protein, and the molecular dynamics of the motion of the two domains relative to one another. The effects of extensive mutations of the tether connecting the heme and molybdenum domains of human SO have been discussed here. However, as yet there are no X-ray structural results of the intact protein for any of these variants. The detailed structures of the molybdenum centers of these variants have been determined from analysis of high resolution pulsed EPR spectra of their Mo(V) states as a function of pH, anions in the media, and mutations of nearby amino acid residues. A long term hope is that these studies will contribute to the continuing development of molecular medicine to treat sulfite oxidase deficiency (13, 16).

Abbreviations

SO

sulfite oxidase

SOE

sulfite oxidizing enzymes

dRF

5-deazariboflavin

DFT

density functional theory

EPR

electron paramagnetic resonance

SDH

sulfite dehydrogenase

NR

nitrate reductase

CSO

chicken sulfite oxidase

HSO

human sulfite oxidase

PSO

plant sulfite oxidase

cyt c

cytochrome c

IET

intramolecular electron transfer

k_et

electron transfer rate constant

K_eq

equilibrium constant for intramolecular electron transfer

k_f and kr

microscopic rate constants for the forward and reverse directions, respectively, of intramolecular electron transfer

Moco

molybdopterin cofactor

wt

wild type

cPMP

cyclic pyranopterin monophosphate

nqi

nuclear quadrupole interaction

hfi

hyperfine interaction

lpH

low pH

hpH

high pH

P_i

phosphate inhibited