The electron transfer (ET) properties of two types of high-valent hemes were studied within the same protein matrix; the bis-FeIV state of MauG and the Compound I state of Y294H MauG. The latter is formed as a consequence of mutation of the tyrosine which forms the distal axial ligand of the six-coordinate heme that allows it to stabilize FeIV in the absence of an external ligand. The rates of the ET reaction of each high-valent species with the type I copper protein, amicyanin, were determined at different temperatures and analysed by ET theory. The reaction with bis-FeIV wild-type (WT) MauG exhibited a reorganization energy (λ) that was 0.39 eV greater than that for the reaction of Compound I Y295H MauG. It is concluded that the delocalization of charge over the two hemes in the bis-FeIV state is responsible for the larger λ, relative to the Compound I state in which the FeV equivalent is isolated on one heme. Although the increase in λ decreases the rate of ET, the delocalization of charge decreases the ET distance to its natural substrate protein, thus increasing the ET rate. This describes how proteins can balance different ET properties of complex redox cofactors to optimize each system for its particular ET or catalytic reaction.
High-valent hemes in proteins participate in a variety of important biological processes. They are used in the oxidation of substrates, as in oxygenases including cytochrome P450 enzymes [1,2] and as intermediates in the breakdown of peroxides in peroxidases . The hemes are usually present as Compound I or Compound ES, in which the ferryl heme iron is FeIV=O with a cation radical present on the porphyrin ring or a nearby tryptophan or tyrosine residue, respectively . The bis-FeIV state of MauG is an alternative strategy by which to stabilize a high-valent heme iron which utilizes two hemes . In MauG from Paracoccus denitrificans this high-valent state has one heme present as FeIV=O with an axial ligand provided by His35 and the other present as FeIV with two axial ligands provided by His205 and Tyr294 and no exogenous ligand (Figure 1) [6,7]. Ultrafast electron transfer (ET) between the hemes in the bis-FeIV state is mediated by hopping through Trp93, which lies between the hemes. This leads to charge resonance (CR) stabilization of this high-valent heme species . Trp93 is reversibly oxidized to a radical species during this process. It is believed that as a consequence of the CR stabilization, the high-valent state actually comprises an ensemble of resonance structures with the true bis-FeIV state as the dominant species [8,9]. The function of MauG is unique among other enzymes that use high-valent hemes as reaction intermediates. This di-c-type heme enzyme catalyses the posttranslational modification of a precursor of methylamine dehydrogenase (preMADH) to generate the protein-derived tryptophan tryptophylquinone (TTQ) cofactor . TTQ is generated by the six-electron oxidation by MauG of two tryptophan residues in preMADH. The oxidizing intermediate is the bis-FeIV species. In contrast with the mechanism of heme-dependent oxygenases, the hemes of MauG make no direct contact with the substrate. Instead preMADH binds to the surface of MauG with the residues on preMADH that are oxidized at a distance of 40 and 19 Å (1 Å=0.1 nm), respectively, from the two heme irons . Due to the long distance, the catalytic reactions require long range ET from the substrate to the hemes.
Hemes are also involved in many non-catalytic physiological ET reactions. Nearly all of these involve the FeII and FeIII redox states as the electron donor and acceptor. The long range ET reaction from preMADH to the bis-FeIV hemes of MauG, which is required as a part of the catalytic mechanism, has been studied and the ET parameters for this reaction have been determined . Similar long range ET studies have not been performed with other types of high-valent hemes. In the present study we present a direct comparison of the ET properties of the hemes in the bis-FeIV and Compound I redox states within the same protein framework of MauG.
The bis-FeIV redox state of WT MauG and the Compound I redox state of Y294H MauG are illustrated
Mutation of the Tyr294 axial ligand of the six-coordinate heme of MauG resulted in a heme with histidine–histidine axial ligation rather than tyrosine–histidine ligation . The crystal structure of Y294H MauG showed that apart from the ligand substitution, this mutation did not significantly perturb the structure of the diheme region of MauG. Whereas addition of H2O2 to MauG results in formation of the bis-FeIV state, addition of H2O2 to Y294H MauG resulted in formation of a Compound I species because the six-coordinate heme with histidine–histidine ligation could not stabilize FeIV. The high-valent Compound I species in Y294H MauG was unable to oxidize preMADH and catalyse TTQ biosynthesis. It was concluded that the inability of Compound I in Y294H MauG to oxidize preMADH was because the ET distance from the substrate to this single heme was too great (∼40 Å), whereas the delocalization of the oxidizing equivalent over both hemes in the bis-FeIV state of MauG reduced the ET distance to a more manageable distance (∼19 Å). Thus, it appears that a major function of the six-coordinate heme of MauG in TTQ biosynthesis is to shorten the ET distance, which exponentially increases the ET rate. Unfortunately, because of the inability of Y294H MauG to react with preMADH, this meant that it was not possible to compare the ET properties of the different high-valent redox states of wild-type (WT) and Y294 MauG.
Recently an inter-protein ET reaction between the cupredoxin amicyanin and MauG was described . Amicyanin is a small beta-barrel protein from P. denitrificans with a type 1 copper centre that consists of a single copper ion coordinated by His53, His95, Cys92 and Met98 . The ET reaction that was studied was that from CuI amicyanin to bis-FeIV MauG. A docking model of the MauG–amicyanin complex placed the copper of amicyanin 16.7 Å from the porphyrin ring and 17.9 Å from the heme iron of the five-coordinate heme of MauG . This distance matched that estimated from analysis of the temperature-dependence of the rate of the ET reaction (kET). The shorter ET distance to the five-coordinate heme in MauG–amicyanin complex than in the preMADH–MauG complex suggested that it might be possible to study an ET reaction of the Compound I redox state of Y294H MauG in complex with amicyanin. In the present study the ET reaction between Y294H MauG and amicyanin is analysed and compared with the reaction with WT MauG. This allows a direct comparison of the ET properties associated with the ET from the bis-FeIV and Compound I redox states. This provides a unique opportunity to study an inter-protein ET reaction involving a Compound I redox centre in the same protein matrix as the bis-FeIV redox centre, and it allows for direct comparison of the ET parameters, i.e. reorganization energy (λ) and electronic coupling (HAB) associated with ET from these two different types of high-valent heme redox states.
MATERIALS AND METHODS
Determination of kET
The rates of the ET reactions from reduced amicyanin to Compound I Y294H MauG were determined using an On-Line Instruments (OLIS) RSM1000 stopped-flow rapid scanning spectrophotometer. Each reaction was performed in 10 mM potassium phosphate buffer, pH 7.5, at the indicated temperature. One syringe contained the limiting reactant, 2 μM Y294H MauG, and the second syringe contained various concentrations of reduced amicyanin. Amicyanin was reduced by addition of an equimolar amount of sodium dithionite . The high-valent species of Y294H MauG was generated by addition of equimolar H2O2 . After rapidly mixing of the reactants, the absorbance spectrum was monitored over the range from 365 to 435 nm to observe the conversion of Compound I Y294H MauG to diferric Y294H MauG. Kinetic data were reduced by factor analysis using the singular-value decomposition (SVD) algorithm and then globally fit using the fitting routines of OLIS GlobalWorks software. In each of the reactions that were performed, the observed rate constant (kobs) was best fit to a single-exponential relaxation. The rate constants that were obtained at different concentrations of amicyanin were analysed by eqn (1) to determine the limiting first-order rate constant (kET) for the reaction and the Kd for complex formation. The errors which are listed in the text are standard errors of the fits.
Analysis of kET by ET theory
Data for the temperature-dependence of kET were analysed using ET theory (eqn 2) . The terms in the equation are temperature (T), free energy (ΔG°), electronic coupling (HAB), the reorganization energy (λ), Planck's constant (h) and the gas constant (R). ΔG° is calculated by eqn (3), where n is equal to the number of electrons transferred, F is Faraday's constant and ΔEm is the difference in the oxidation–reduction midpoint potential value of the electron donor and acceptor sites. Data were also analysed by eqn (4) where Ea is the activation energy and A is the pre-exponential factor.
In silico docking of amicyanin and Y294MauG
Determination of kET for the reaction of Y294H MauG with reduced amicyanin
The formation of the high-valent species in Y294H MauG is accompanied by a decrease in intensity of the Soret peak, which occurs immediately after addition of H2O2. Subsequent addition of reduced amicyanin results in an increase in intensity of the Soret peak and return of the absorbance spectrum to that of the diferric state (Figure 2A). The kinetics of this reaction fit well to a single exponential transition. The kinetic plot derived from the global fit of the SVD-reduced three-dimensional data set is shown in Figure 2(B). Analysis of the time course of the reaction at a single wavelength (404 nm) also fit well to a single exponential (Figure 2C). The reaction of the high-valent state of Y294H MauG with various concentrations of reduced amicyanin at pH 7.5 at 25°C exhibited saturation behaviour (Figure 2D). The fit of this data to eqn (1) yielded a limiting first-order rate constant (kET) of 33±4 s−1 and a Kd of 300±128 μM. For comparison, the reaction of bis-FeIV WT MauG with reduced amicyanin exhibited a kET of 22 s−1 and a Kd of 165 μM . Similar monophasic kinetics was observed for the reaction of bis-FeIV WT MauG with reduced amicyanin. It should be noted that the reduction in either high-valent MauG species, bis-FeIV or Compound I, requires two electrons whereas the oxidation of CuI amicyanin requires one electron. Thus, two molecules of CuI amicyanin must be oxidized to observe the complete reaction. In each reaction, the change in absorbance fit best to a single exponential. The monophasic kinetics means that the reaction of the first CuI amicyanin is followed by dissociation of the complex and reaction with a second CuI amicyanin. If the dissociation/association steps are very rapid relative to kET and excess CuI amicyanin is present, rebinding of the CuII amicyanin after the single turnover is not a factor and monophasic kinetics is observed.
The reduction of Compound I Y294H MauG by reduced amicyanin
Temperature-dependence of kET for the reaction of Y294H MauG with reduced amicyanin
The kET for the ET reaction from reduced amicyanin to the Compound I Y294H MauG was determined at temperatures from 10°C to 27°C (Figure 3A). The high-valent state of Y294H MauG was unstable at higher temperatures, and so this limited the range that could be studied. The Em value for the Compound I/diferric couple of Y294H MauG is unknown. Em values have been determined for the Compound I/FeIII couple in soybean ascorbate peroxidase , Arthromyces ramosus peroxidase , yeast cytochrome c peroxidase , and horseradish peroxidase type A2  and type C [23–26]. These Em values range from 717 to 954 mV, with an average of 887 mV. As such, for the analysis of the temperature-dependence of the ET rates by eqn (2) the Em value of 887 mV was used for Y294H MauG. That value and the known Em value of amicyanin at pH 7.5 of 265 mV  were used to determine the ΔG° for the ET reaction from the ΔEm. The fit of the data to eqn (2) yielded values of λ=1.95±0.08 eV and HAB=0.03±0.01 cm−1. For comparison, the reaction of bis-FeIV WT MauG with reduced amicyanin exhibited values of λ=2.34±0.16 eV and HAB=0.6±0.1 cm−1 . Thus, the mutation decreased the magnitudes of both λ and HAB. These data were also analysed by eqn (3) (Figure 3B). This yielded values of Ea of 8.2±0.3 kcal/mol for the reaction with WT MauG and 4.6±0.3 kcal/mol for the reaction with Y294H MauG.
The temperature-dependence of kET from reduced amicyanin to bis-FeIV WT MauG and Compound I Y294H MauG
Docking model of the complex of Y294H MauG and amicyanin
Docking model of the amicyanin–Y294H MauG complex
The value of λ reflects the amount of energy needed to optimize the system for ET, or in other words the energy required to bring the reactant and product states to the state in which the ET event occurs. The value of λ is composed of an inner sphere λin, which derives from optimization bond lengths and bond angles of the inner shell of atoms of the redox centre for the ET event, and an outer sphere λout, which reflects reorientation of solvent molecules that is associated with the ET event. In small molecules, these two parameters are fairly well distinguishable. However, in protein ET reactions, the line between these two terms can become blurred as electron density is distributed asymmetrically about the redox cofactor or metal-ligating residues. Computational methods have been used to predict λ for protein ET reactions which tend to focus on the contributions of the protein medium [28,29]. λ values have also been determined for ET reactions between native redox cofactors and redox-active tags such as Ruthenium complexes [30,31]. ET reactions between the copper of azurin a disulfide radical anion were used to determine λ . The proteins used in the present study are among the few for which λ values have been experimentally determined for the reactions with their natural redox partners: the reactions of amicyanin with MADH  and cytochrome c-551i , and the reactions between MauG and preMADH and MADH . λ values have also been determined for some other ET reactions between quinoprotein dehydrogenases redox protein partners . However, in general there have been relatively few studies in which λ values have been experimentally determined for ET reactions between two protein-bound redox cofactors.
The structures of WT and Y294H MauG suggest that there should be little difference in the value of λout during ET into the high-valent hemes since the only solvent accessible portion of the diheme site is the FeIV=O of the five-coordinate heme [6,36]. The six-coordinate heme is shielded from solvent by the protein. Therefore, the 0.39 eV decrease in the experimentally determined λ that was caused by the Y294H mutation is most likely a consequence of a decrease in λin. It was previously suggested that the relatively large λ values associated the ET reactions of WT bis-FeIV MauG with preMADH  and amicyanin  could be attributed to the fact that λin associated with the bis-FeIV state could require alterations in lengths and angles of multiple bonds within the two hemes and intervening Trp93. In Y294H MauG, the ET reaction associated with the high-valent Compound I redox state would only require such changes in bond lengths and angles of the five-coordinate heme, as there is no change in the redox state of the six-coordinate heme. This would explain the decrease in λ that was caused by the Y294H mutation.
The Y294H mutation also caused a decrease in HAB. Comparison of the docking models of the complexes of amicyanin with WT MauG and Y294MauG does not provide a clear explanation for this. HAB is dependent upon the ET distance and the nature of the intervening medium. ET through amino acid bonds in proteins will exhibit a larger HAB than for ET through empty space [37,38]. For interprotein ET reactions, a through-space segment of the ET pathway is required to get from one protein to the other. As such, the nature of the protein/protein interface in the ET protein complex can be a critical determinant of the overall HAB. An increase of less than an angstrom in the length of the through-space jump can significantly decrease the overall HAB . Such distances cannot be reliably ascertained from docking models. It should be noted that although the structures of the diheme sites of WT amicyanin and Y294H amicyanin are indistinguishable, there were some other changes in structure of the protein. In order for His294 to replace Tyr294 as a heme axial ligand, residues 291–314 in the C-terminal region of the Y294H MauG are displaced relative to WT MauG . Although this region is not at the interface with amicyanin, it does leave open the possibility that the mutation may have also caused subtle changes in structure at the protein surface that affect the interface. The fact that the Y294H mutation increased the Kd for complex formation with amicyanin from 165 to 265 μM supports the notion that subtle changes have occurred at the protein/protein interface. These could account for the decrease in HAB.
Analysis of the temperature-dependence of kET by eqn (4) (Figure 3B) also clearly illustrates the effect of the Y294H mutation on λ and HAB as both the slope and y-intercept of the lines are altered. The change in slope reflects the decrease in Ea for the ET reaction. Since this is an ET reaction, the term in eqn (2) which corresponds to Ea in eqn (4) is (∆G°+λ)2/4λ. Thus, using the experimentally determined values of λ for the two reactions, it is calculated that the ΔEa caused by the mutation of −3.6 kcal/mol corresponds to −0.156 eV. This describes the overall effect on the reaction of the 0.39 eV decrease in λ that was caused by the mutation. The change in the y-intercept caused by the mutation reflects the change in HAB, which is present in the pre-exponential term in eqn (2).
In addition to providing some insights into the factors which contribute to the magnitude of λ for ET reactions involving hemes, as well as other cofactors, these results provide some insight into why Nature uses alternative mechanisms by which to stabilize and use utilize high-valent FeV equivalent heme species. These results clearly indicate that Compound I is a more efficient electron acceptor in redox reactions than bis-FeIV by virtue of the lower λ associated with the ET reaction. Single-heme enzymes which function as oxygenases utilize the Compound I species to catalyse enzymatic reactions requiring a strong oxidant in which the substrate is in close proximity to the heme. In this case, it makes sense to not delocalize the oxidizing potential but to keep it centred on the heme. The results presented herein further suggest that this also minimizes the energy barrier for the short range ET that is part of the catalytic mechanism in the active site. In contrast, the diheme redox centre of MauG catalyses a reaction at the surface of the protein that is distant from the heme site. The ability to delocalize the oxidizing potential over two hemes which are themselves separated by several angstroms significantly reduces the distance for the long range ET required for catalysis. Although this delocalization results in a significant increase in the λ associated with the reaction, that deficiency is more than made up for by the decrease in ET distance by approximately 20 Å. The importance of this consideration is evidenced by that fact that Y294H MauG, which forms the Compound I state instead of the bis-FeIV, was unable to oxidize the natural substrate preMADH over the longer distance .
Brian Dow designed and performed experiments, analysed data and wrote the manuscript. Victor Davidson designed experiments, analysed data and wrote the manuscript.
We thank Yu Tang for providing technical assistance.
This work was supported by the National Institute of General Medical Sciences of the National Institutes of Health [grant number R37GM41574] to V.L.D.
- bis-FeIV MauG
redox state of MauG with one heme as FeIV=O and the other as FeIV
oxidation–reduction midpoint potential
the biosynthetic precursor protein of MADH with incompletely synthesized TTQ