Thermodynamic Characterization of a Triheme Cytochrome Family from Geobacter sulfurreducens Reveals Mechanistic and Functional Diversity

Abstract

A family of five periplasmic triheme cytochromes (PpcA-E) was identified in Geobacter sulfurreducens, where they play a crucial role by driving electron transfer from the cytoplasm to the cell exterior and assisting the reduction of extracellular acceptors. The thermodynamic characterization of PpcA using NMR and visible spectroscopies was previously achieved under experimental conditions identical to those used for the triheme cytochrome c₇ from Desulfuromonas acetoxidans. Under such conditions, attempts to obtain NMR data were complicated by the relatively fast intermolecular electron exchange. This work reports the detailed thermodynamic characterization of PpcB, PpcD, and PpcE under optimal experimental conditions. The thermodynamic characterization of PpcA was redone under these new conditions to allow a proper comparison of the redox properties with those of other members of this family. The heme reduction potentials of the four proteins are negative, differ from each other, and cover different functional ranges. These reduction potentials are strongly modulated by heme-heme interactions and by interactions with protonated groups (the redox-Bohr effect) establishing different cooperative networks for each protein, which indicates that they are designed to perform different functions in the cell. PpcA and PpcD appear to be optimized to interact with specific redox partners involving e⁻/H⁺ transfer via different mechanisms. Although no evidence of preferential electron transfer pathway or e⁻/H⁺ coupling was found for PpcB and PpcE, the difference in their working potential ranges suggests that they may also have different physiological redox partners. This is the first study, to our knowledge, to characterize homologous cytochromes from the same microorganism and provide evidence of their different mechanistic and functional properties. These findings provide an explanation for the coexistence of five periplasmic triheme cytochromes in G. sulfurreducens.

Introduction

The bacterium Geobacter sulfurreducens (Gs) is one of the most effective microorganisms for the bioremediation of radioactive and toxic metals in contaminated subsurface environments, and the conversion of organic compounds to electricity in microbial fuel cells (1,2). Gs is metabolically versatile, being able to couple the anaerobic oxidation of organic compounds, or molecular hydrogen, to the reduction of several terminal electron acceptors, such as Fe(III) and U(VI) (3–5).

Many of the terminal electron acceptors that can be used by Gs are insoluble and thus are unable to diffuse inside the cells. Therefore, reduction of these acceptors cannot occur in the periplasm and requires electron transfer across the outer membrane (6). Although numerous electron transfer proteins have been identified in Gs, the electron transfer pathways that allow this microorganism to obtain energy are still far from being understood (6). It is essential to obtain a detailed characterization of Gs electron transfer proteins before we can understand these electron transfer mechanisms and delineate their physiological roles. Elucidating the cellular strategies that allow energy production from the reduction of extracellular terminal acceptors is also of major interest for the design of improved biotechnological applications.

Soluble periplasmic c-type cytochromes are crucial for shuttling electrons from the cytoplasmic compartment to the outer membrane. A periplasmic triheme cytochrome (PpcA, encoded by gene GSU0612) was shown to be involved in Fe(III) reduction (7,8). In addition to its electron transfer capacity, previous studies also showed that PpcA is likely to contribute to the proton electrochemical gradient across the bacterial cytoplasmic membrane that drives ATP synthesis (9).

The genome of Gs has been sequenced (10) and its analysis has revealed the presence of four PpcA homologs, each containing three heme groups (11). The four PpcA homologs are small (∼10 kDa), soluble cytochromes designated as PpcB (GSU0364), PpcC (GSU0365), PpcD (GSU1024), and PpcE (GSU1760). They respectively share a 77%, 63%, 60%, and 56% amino acid sequence identity with PpcA (11). The x-ray structures of the five triheme cytochromes were recently determined (12). The tertiary structure of all of the proteins is similar, although local variations were observed in their structures. A representative structure of the triheme cytochrome family is shown in Fig. 1. Earlier knock-out studies of Gs cells suggested that, despite their structural similarity, these proteins may have functional specificity. In studies of Gs strains with genes encoding the above cytochromes deleted, it was shown that the cellular activities of Fe(III) reduction are affected differently (7). In fact, the knock-out mutant of PpcA was clearly impaired in its ability to reduce Fe(III)-oxides, a phenotype shared with PpcD, albeit to a smaller extent, whereas deletion of the other homologs did not impair Fe(III) reduction (7). Proteomics studies by Ding and co-workers (13) showed that all of these cytochromes are expressed in Gs cells. PpcA, PpcB, and PpcD are expressed in both Fe(III) oxides and Fe(III) citrate cultures, whereas PpcE was found only in the latter.

Figure 1.

Structure of the triheme cytochrome PpcE (12). Roman numerals indicate the hemes (black) in order of attachment to the CXXCH motif in the polypeptide chain (gray).

To investigate the molecular bases for these functional differences, we sought to achieve a detailed thermodynamic characterization of PpcA, PpcB, PpcD, and PpcE. This is the first time, to our knowledge, that a family of homologous soluble cytochromes from the same organism has been characterized in detail, allowing for their thermodynamic properties to be compared and their physiological role discussed.

Materials and Methods

All experimental details concerning preparation of the triheme cytochromes, NMR, and visible spectroscopy experiments are available in the Supporting Material.

Data analysis

The general theoretical framework that allows the thermodynamic properties of the redox centers in multiheme proteins to be studied in detail was previously described (14) and can be sketched as follows: In the particular case of a triheme cytochrome, three consecutive reversible steps of one-electron transfer convert the fully reduced state (stage 0, S₀) in the fully oxidized state (stage 3, S₃), and therefore four different redox stages can be defined. At each stage, microstates are grouped with the same number of oxidized hemes (Fig. 2). Additionally, within each microstate, the group responsible for the redox-Bohr effect may be protonated or deprotonated, leading to a total of 16 microstates (Fig. 2). The energies of the microstates may be expressed as the sums of terms for each center and the interactions between them (14). Thus, three reduction potentials and one pK_a plus six two-center interaction energies (three heme-heme and three redox-Bohr) are sufficient to characterize the system across the full range of pH values and solution potentials, independently of any structural model. The total reduced heme in each triheme cytochrome, or the fractional oxidation of any heme at any stage of oxidation, is therefore a function of 10 energy parameters. To obtain information about each microstate, it is necessary to monitor the stepwise oxidation of each individual heme, which for the particular case of heme groups displaying identical optical properties can be obtained with confidence by NMR spectroscopy (15,16). On the NMR timescale, under conditions of fast intramolecular electron exchange (between the different microstates within the same oxidation stage) and slow intermolecular electron exchange (between different oxidation stages), the individual heme signals can be discriminated in the different oxidation stages. Heme methyl resonances provide the most easily identifiable NMR signals of heme substituents because they also have some of the largest paramagnetic shifts, and thus are ideal for following the heme oxidation throughout the redox titration. Since the intrinsic paramagnetic shifts of the heme methyls are proportional to the oxidation fraction of that particular heme, m, they can be used to obtain the relative microscopic reduction potentials of the hemes in each stage of oxidation (15,16). For each pH, the paramagnetic shifts of one methyl group of each heme in oxidation stages 1–2 relative to the fully oxidized form give the averaged oxidation fraction for each heme in each stage (14). However, the fact that each oxidation step involves a single electron means that the relative shifts of any two methyl groups determine that of the third according to Eq. 1, and therefore there is redundancy when data for all three hemes are collected.

$$\sum _{m=1}^3\frac{\left({\delta }_{obs}^{m,S}-{\delta }_{obs}^{m,0}\right)}{\left({\delta }_{obs}^{m,3}-{\delta }_{obs}^{m,0}\right)}=S$$ (1)

Figure 2.

Electronic distribution scheme for a triheme cytochrome with a proton-linked equilibrium, showing the 16 possible microstates. The three inner circles represent the hemes, which can be reduced (black) or oxidized (white). The outer circles with solid and dashed lines represent the protonated and deprotonated microstates, respectively. The microstates are grouped according to the number of oxidized hemes in four oxidation stages (S_0–3) connected by one electron step. E_S (S = 1–3) is the macroscopic reduction potential, i.e., the solution potential at which the sum of the microstate populations in stage S equals the sum of the populations in stage S₋₁. P_0H and P₀ represent the reduced protonated and deprotonated microstates, respectively. P_ijkH and P_ijk indicate the protonated and deprotonated microstates, respectively, where i, j, and k represent the heme(s) that are oxidized in that particular microstate. As an example, the transitions between the fully reduced protonated and the corresponding protonated microstate in oxidation stage S₁ are indicated with the corresponding microscopic reduction potentials.

Therefore, NMR data only define the relative heme reduction potentials and heme redox interactions, which are calibrated with redox titrations monitored by visible spectroscopy (14,17). The data obtained from redox titrations followed by NMR and visible spectroscopy for PpcA, PpcB, PpcD, and PpcE were fitted simultaneously with a thermodynamic model previously described (9,14) and outlined above. The experimental uncertainty of the NMR data was evaluated from the line width of each NMR signal at half height, and the visible data points were given an uncertainty of 3% of the total optical signal.

Results

Assignment of heme signals in the reduced form

The heme proton resonances of PpcA and PpcB were previously assigned in the reduced state (18) by following the assignment strategy described for heme protons in multiheme ferrocytochromes (19). The same strategy was used in this work for specific and self-consistent assignments of the heme proton resonances of PpcD and PpcE (Table S1 and Fig. S1). Gs triheme cytochromes are structurally similar to tetraheme cytochromes c₃, with the exception that heme II and the corresponding fragment of the polypeptide chain are absent. Thus, to be consistent with the literature (9,20), the heme groups are numbered I, III, and IV (indicating the heme by order of attachment to the CXXCH motif in the polypeptide chain). As expected from the conserved heme disposition observed in all cytochromes (12), the same type of heme I and IV substituents, which are more exposed to the solvent (2¹CH₃, 18¹CH₃, and H20; Fig. 1), display similar chemical shifts. On the other hand, the chemical shifts of heme III substituents are more spread due to the contribution of the ring current shifts of neighboring porphyrin rings. In fact, several heme III resonances (namely, those of 20H^III, 2¹CH₃^III, 7¹CH₃^III, 18¹CH₃^III, 3¹H^III, and 8²CH₃^III) are shifted downfield. We also tested the heme proton assignments of PpcD and PpcE by comparing the observed heme proton chemical shifts with those calculated from their respective crystal structures. These shifts correlate very well, even for the protons subjected to the larger ring current effects (Fig. S1). The assignment of the heme protons was further confirmed by the observation of the interheme connectivities in the nuclear Overhauser enhancement spectroscopy (NOESY) NMR spectra for closest substituents (≤3.5 Å). The correlation between the measured heme proton chemical shifts for the four cytochromes is indicated in Fig. S2.

Thermodynamic characterization

The thermodynamic properties of PpcA were determined in an earlier study (9) using NMR data obtained at 500 MHz and relatively high ionic strength (500 mM) to slow down the intermolecular electron exchange between the different oxidation stages. These experimental conditions were identical to the ones used to characterize the triheme cytochrome isolated from Desulfuromonas acetoxidans, the only cytochrome c₇ reported in the literature at that time (21), and allowed a proper and detailed comparison of the redox properties of these two cytochromes (9). Despite these particular conditions, signal broadness was observed for the connectivities between the heme methyl signals in the different oxidation stages. As expected from the structural similarities among the triheme cytochromes, comparable problems were found for the PpcA homologs. To overcome these difficulties, two-dimensional exchange spectroscopy (2D-EXSY) NMR spectra were acquired at 800 MHz to favor a slow intermolecular exchange regimen on the NMR timescale. At this magnetic field, it was also possible to lower the ionic strength of the samples to match the physiological ionic strength recommended by the American Type Culture Collection for growth of G. sulfurreducens (22). Under these experimental conditions, well-resolved 2D-EXSY NMR spectra were obtained for all proteins, and discrete NMR signals connecting the different oxidation stages were observed for each heme methyl. The only exception was found for PpcC, which presents different conformations at intermediate stages of oxidation. This leads to splitting and excessive broadening of the NMR signals, preventing the thermodynamic characterization of this protein (23).

The heme methyl groups 12¹CH₃^I, 7¹CH₃^III, and 12¹CH₃^IV (PpcA and PpcB); 18¹CH₃^I, 7¹CH₃^III, and 18¹CH₃^IV (PpcD); and 18¹CH₃^I, 12¹CH₃^III, and 12¹CH₃^IV (PpcE) were used to monitor the stepwise oxidation of each heme at each pH. As an example, the stepwise oxidation of the PpcA, PpcB, PpcD, and PpcE heme groups followed by 2D-EXSY NMR is illustrated in Fig. S3. From these experiments, it is possible to obtain the chemical shifts of each methyl in all oxidation stages. The chemical shifts of the heme methyls in the different oxidation stages are listed in Table S2. Analysis of this table confirms that the extrinsic shifts for the methyls selected are not significant, since the sums of the oxidation fraction at each oxidation stage (Eq. 1) are close to integers, and therefore each methyl reflects the averaged oxidation state of its heme in each stage (14,16,24).

To determine the thermodynamic parameters of PpcA, PpcB, PpcD and PpcE, the pH dependence of the paramagnetic chemical shifts of each heme methyl, in the pH range 5.7–9.1, was fitted to the model described in the Materials and Methods section, together with the data from redox titrations followed by visible spectroscopy (Fig. 3). The thermodynamic parameters and the macroscopic pK_a values associated with the four stages of oxidation are indicated in Tables 1 and 2, respectively. The quality of the fittings obtained for the pH dependence of the paramagnetic chemical shifts and for the visible redox titrations clearly shows that the thermodynamic properties of all cytochromes are well described by the model used. In particular, the pH dependence of the paramagnetic shifts in the three oxidation stages (see data points in Fig. 3) are dominated by a single pK_a in each stage of oxidation, and thus there is no justification to consider more that one ionizable center within the framework of the thermodynamic model (14). Computational electrostatic models have also been proposed to analyze protonation and redox equilibria in proteins (for reviews, see Ullmann and Knapp (25) and Zheng and Gunner (26)).

Figure 3.

Fitting of the thermodynamic model to the experimental data for PpcA, PpcB, PpcD, and PpcE. The solid lines are the result of the simultaneous fitting of the NMR and visible data. The upper figures show the pH dependence of heme methyl chemical shifts at oxidation stages 1 (Δ), 2 (□), and 3 (○). The chemical shifts of the heme methyls in the fully reduced stage (stage 0) are not plotted, because they are unaffected by the pH. The last figure in each panel corresponds to the reduced fractions of the various cytochromes, determined by visible spectroscopy at pH 7 (○) and pH 8 (□). The open and solid symbols represent the data points in the reductive and oxidative titrations, respectively. The macroscopic reduction potentials (see Fig. 2) at pH 7 (right) and pH 8 (left) are indicated. The dashed line in the PpcE bottom panel represents a standard n = 1 Nernst curve.

Table 1.

Energy parameters (meV) for PpcA, PpcB, PpcD, and PpcE

Energy (meV)					Energy (meV)
PpcA	Heme I	Heme III	Heme IV	Redox-Bohr center	PpcB	Heme I	Heme III	Heme IV	Redox-Bohr center
Heme I	−154 (5)	27 (2)	16 (3)	−32 (4)	Heme I	−150 (3)	17 (2)	8 (2)	−16 (4)
Heme III		−138 (5)	41 (3)	−31 (4)	Heme III		−166 (3)	32 (2)	−9 (4)
Heme IV			−125 (5)	−58 (4)	Heme IV			−125 (3)	−38 (4)
Redox-Bohr center				495 (8)	Redox-Bohr center				426 (8)
PpcD	Heme I	Heme III	Heme IV	Redox-Bohr center	PpcE	Heme I	Heme III	Heme IV	Redox-Bohr center
Heme I	−156 (6)	46 (3)	3 (4)	−28 (6)	Heme I	−167 (4)	27 (3)	5 (3)	−12 (4)
Heme III		−139 (6)	14 (4)	−23 (6)	Heme III		−175 (4)	22 (3)	2 (4)
Heme IV			−149 (6)	−53 (6)	Heme IV			−116 (5)	−13 (4)
Redox-Bohr center				501 (8)	Redox-Bohr center				445 (10)

Diagonal terms (in boldface type) represent the oxidation energies of the three hemes (see Fig. 2) and the deprotonating energy of the redox-Bohr center (g_B) in the fully reduced and protonated proteins (14). Off-diagonal values are the redox (heme-heme) and redox-Bohr (heme-proton) interaction energies. The standard errors are given in parentheses. Following the nomenclature for the pairwise interacting centers model (14), the pK_a of the reduced proteins is given by g_BF/(2.3RT) and the pK_a of the oxidized ones is given by $\left(g_B+\sum _{i=1}^3{g}_{iB}\right)$ F/(2.3RT).

Table 2.

Macroscopic pK_a values of the redox-Bohr center for PpcA, PpcB, PpcD, and PpcE at each stage of oxidation

	pKa
Oxidation stage	PpcA	PpcB	PpcD	PpcE
0	8.6	7.4	8.7	7.7
1	8.0	7.1	8.1	7.6
2	7.2	6.8	7.4	7.5
3	6.5	6.3	6.9	7.4
ΔpKa	2.1	1.1	1.8	0.3

The values were calculated from the redox-Bohr center parameters indicated in Table 1.

When we compare the results previously reported for PpcA at 500 mM and 500 MHz (9) with those obtained under these new conditions, we see that there is no significant difference in the parameters. However, the errors associated with the parameters are smaller under the new conditions.

Discussion

In this work, we used NMR and visible spectroscopy techniques to structurally and thermodynamically probe a family of four triheme cytochromes (PpcA, PpcB, PpcD, and PpcE) from Gs. The first step was to assign the proton signals of each heme substituents in the NMR spectra (TableS1). This allowed us to compare the structural features of the heme core architectures of all cytochromes and to demonstrate for the first time, to our knowledge, their structural similarity in solution. The NMR spectra were also used as a starting point for the thermodynamic studies because they allowed us to follow the behavior of representative heme methyl signals through the oxidation of the proteins at different pH values. The experimental data thus obtained, together with data obtained from visible redox titrations, were fitted with the thermodynamic model that considers four interacting charged centers (three hemes and one protonatable center). This permitted us to determine the thermodynamic parameters for the four triheme cytochromes, and the specific mechanistic and functional properties of each member of this unusual family of Gs cytochromes.

Structural characterization of the heme core architecture in solution

Previous studies carried out on the reduced forms of PpcA and PpcB allowed assignment of the heme proton signals (18). The identification of the interheme NOE connectivities between protons of the closest heme groups showed that the structure of the PpcB heme core in solution is similar to that of the crystal structure, which was not the case for PpcA. In this work, we also assigned the heme proton signals of PpcD and PpcE (Table S1) following the strategy described for PpcA and PpcB. We confirmed the assignments by examining the interheme NOE connectivities measured from the 2D-NOESY NMR spectra acquired at different mixing times. The NOE connectivities expected to be observed on the basis of their crystal structure were observed in the 100 ms 2D-NOESY NMR spectra, indicating that their heme core structures in solution are similar. The good correlation obtained between the heme proton chemical shifts measured for PpcA and those of PpcB (root mean-square deviation (RMSD) = 0.11), PpcD (RMSD = 0.37), and PpcE (RMSD = 0.26) further confirms the similarity of the heme core solution structures of the four cytochromes (Fig. S2). Cytochromes PpcD and PpcE showed the largest deviations for some of the heme protons, which is not unexpected since these homologs have lower sequence identity with PpcA (60% and 56%, respectively). The best correlation was obtained for the heme IV protons, corroborating the finding that the structure of this heme and its surroundings is the most conserved feature among the four cytochromes (12).

Thermodynamic characterization

The analysis of the thermodynamic parameters of the four Gs cytochromes (Table 1) shows that, in the fully reduced and protonated form, the reduction potentials of the hemes are negative and different from each other. Although the heme core arrangement is well conserved among these cytochromes, as it is the tertiary structure of the proteins, local structural differences, such as charge and/or solvent exposure variation, were observed (12). These have been described as important factors in controlling the redox properties of the heme groups (27), and a simple correlation between the structures and the heme reduction potentials is not straightforward, particularly for multiheme proteins with interacting redox centers. The four cytochromes studied here differ in 41 out of ∼70 amino acids (Fig. S4), further complicating such a correlation. This was achieved only for PpcA and PpcB, which have a higher amino acid sequence identity (77%). In a previous study (18), we predicted the relative variation of the heme reduction potentials for these two homologs. In that work, however, only the qualitative variation of the heme oxidation profiles of PpcB and PpcA was studied at pH 6 and 8. Even in the absence of a detailed thermodynamic characterization of these proteins, as reported here, the higher oxidation fractions of PpcB heme III compared to those of PpcA allowed us to postulate that the lack of two positively charged lysines near heme III (Fig. S4) would destabilize its reduced form by electrostatic effects, favoring its oxidized form. That hypothesis is now confirmed in the work presented here. Indeed, the reduction potentials of hemes I and IV, which had no net charge variation in their vicinity between PpcA and PpcB, are very similar, whereas that of heme III is considerably lower in PpcB (Table 1).

The thermodynamic characterization of the four triheme cytochromes also showed that the redox interactions (between each pair of hemes) are positive, indicating that the oxidation of a particular heme renders the oxidation of its neighbors more difficult. The strongest redox interactions are observed between the closest pair of heme groups: I–III and III–IV. Although they were not identical, the distances between the heme iron atoms and the angles of the heme porphyrin ring planes showed that the heme core arrangement of the hemes is well conserved among these cytochromes (12). However, the magnitude of the redox interactions varies in each protein, particularly in PpcD and PpcE, which show higher values between hemes I–III. The structural comparison among this family of cytochromes revealed that PpcD and PpcE have one fewer residue before the heme III binding motif (12). The residue K49 appears to be conserved in the four proteins, but the structural comparisons suggest that this residue is equivalent to G48 in PpcA and PpcB. Thus, with one fewer residue at this site, PpcD and PpcE can form regular helices (Fig. 1), whereas the other two homologs have a hump in the α-helix in this position. Structural rearrangements that include subtle movements of charged groups and variations in local dielectric constants might explain these differences, making the correlation between the heme iron atoms' distances highly approximate (28). The same features have also been observed in tetraheme cytochromes c₃, where, despite a conserved heme core, the pairwise heme-heme interactions differ largely for the various proteins (29).

The redox-Bohr interactions (between the hemes and the redox-Bohr center) are negative, i.e., the oxidation of the hemes facilitates the deprotonation of the acid-base center and vice versa. The redox-Bohr interactions for PpcA and PpcD are higher than those of PpcB and PpcE (Table 1). The magnitude of the redox-Bohr effect at physiological pH can be easily inferred from the separation of the nonstandard n = 1 Nernst redox curves carried out at pH 7 and 8, which describe a 3e⁻/H⁺ process for each protein (Fig. 3). The different degree of the redox-Bohr effect is also reflected in the macroscopic pK_a values of the redox-Bohr center at each oxidation stage (Table 2). The larger redox-Bohr effect (ΔpK_a = pK_a^red − pK_a^ox) is observed for PpcA and PpcD, 2.1 and 1.8 pH units, respectively. The ΔpK_a value is lower for PpcB and much smaller for PpcE (1.1 and 0.3 pH units, respectively). For the cytochromes with a nonnegligible redox-Bohr effect (PpcA, PpcB, and PpcD), heme IV shows the highest redox-Bohr interaction, suggesting that in all proteins the redox-Bohr center is located in the vicinity of this heme. In fact, we previously identified propionate 13 of heme IV (P₁₃^IV) as the protonatable center that is responsible for the redox-Bohr effect in PpcA and PpcB (9,18), so it is conceivable that P₁₃^IV is also the redox-Bohr center in PpcD. The differences in the magnitude of the redox-Bohr effect observed for the cytochromes in this study may be a consequence of the combined contributions of fractional protonation of several acid-base groups (30).

Order of oxidation of the heme groups

The relative order of oxidation of the heme groups in the fully protonated and reduced proteins can be obtained from the values of the microscopic reduction potentials listed in Table 1: I-III-IV for PpcA, III-I-IV for PpcB and PpcE, and I-IV-III for PpcD. However, as mentioned above, the reduction potential of each heme is affected by the oxidation state of neighboring ones and by the pH. This is reflected in the individual oxidation profiles of the hemes shown in Fig. 4. These curves are substantially different from a pure Nernst curve, and the several crossovers clearly indicate that the electron affinity of each heme is tuned by the redox interactions. Thus, during the oxidation of each protein, the affinity of each redox center is modulated such that their apparent midpoint reduction potentials e_app (i.e., the point at which the oxidized and reduced fractions of each heme group are equally populated) are different in relation to those in the fully reduced protein (cf. Table 1 and Fig. 4). Consequently, the actual order of the midpoint reduction potentials at physiological pH is I-IV-III for PpcA, (III,I)-IV for PpcB and PpcE, and IV-I-III for PpcD. This shows how structural related proteins can specifically fine-tune their redox centers, a feature that cannot be envisaged from the macroscopic redox curves (see dashed lines in Fig. 4).

Figure 4.

Individual heme oxidation fractions (labeled with Roman numerals) for PpcA, PpcB, PpcD, and PpcE (solid lines). The dashed line indicates the global oxidation fraction of each protein. The curves were calculated as a function of the solution reduction potential at pH 7.5 using the parameters listed in Table 1. The midpoint reduction potentials (e_app) of the individual hemes are also indicated.

Relevant microstates in solution

The molar fractions of the 16 microstates of PpcA, PpcB, PpcD, and PpcE at physiological pH can be obtained from the thermodynamic parameters listed in Table 1 (Fig. 5). From the analysis of Fig. 5, it is clear that the dominant microstates are different in each protein. In the case of PpcA, oxidation stages 0 and 1 are dominated by the protonated forms P_0H and P_1H, respectively, while the redox-Bohr center is kept protonated. Stage 2 is dominated by the oxidation of heme IV and deprotonation of the redox-Bohr center (P₁₄), which remains deprotonated in stage 3 (P₁₃₄). Therefore, a route is defined for the electrons within PpcA: P_0H → P_1H → P₁₄ → P₁₃₄. In the case of PpcD, a different profile for electron transfer is observed. Oxidation stage 0 is dominated by the protonated form P_0H. However, the microstates of oxidation stage 1 are overcome by the P_0H curve, which earlier intercepts the P₁₄ curve. This microstate (P₁₄) dominates oxidation stage 2, whereas P₁₃₄ dominates stage 3. Thus, for this cytochrome, a different preferential route for electrons is established, favoring a proton-coupled 2e⁻ transfer step between oxidation stages 0 and 2: P_0H → P₁₄ → P₁₃₄. The small redox interaction between hemes I and IV, and the proximity of their reduction potentials in PpcD, in comparison with those of PpcA, favor a 2e⁻ transfer step in PpcD. Since the redox-Bohr parameters of both proteins are of similar magnitude, it can also be inferred that the redox-Bohr center does not control this pathway.

Figure 5.

Molar fractions of the 16 individual microstates (described in Fig. 2) of PpcA, PpcB, PpcD, and PpcE at pH 7.5. The curves were calculated as a function of the solution reduction potential using the parameters listed in Table 1. Solid and dashed lines indicate the protonated and deprotonated microstates, respectively. For clarity, only the relevant microstates are labeled.

In the case of PpcB and PpcE, several microstates are significantly populated in oxidation stages 1 and 2, and therefore no preferential pathway for electron transfer can be established.

These observations, together with the order of the midpoint reduction potentials, make it clear that heme III plays a key regulatory role in the functional mechanism of each cytochrome by controlling the microscopic redox states that the protein can access during the redox cycle and thus establishing preferential pathways for electron transfer. Indeed, the relative value of the reduction potential of heme III seems to be a crucial factor in allowing these proteins to couple electron transfer with deprotonation of the acid-base center. In fact, only PpcA and PpcD, for which heme III has the highest reduction potential, have preferential pathways for electron transfer with e⁻/H⁺ coupling being established.

Functional implications

Optimal growth of Gs occurs at pH 7–8 (K.P. Nevin, University of Massachusetts, personal communication, 2009) and is severely limited at pH 6 (31). A direct measurement of the periplasmic pH in the Gram-negative bacterium Escherichia coli showed that the periplasmic pH remains at or near the pH of the external medium (32). Thus, it is reasonable to assume that in the Gram-negative Gs, the periplasmic pH would also be closer to that of the external medium, with optimal growth conditions ranging from pH 7 to 8. Therefore, in this work we chose to use a pH value of 7.5 to discuss the functional consequences of the thermodynamic properties of the cytochromes. Additionally, cyclic voltammetry experiments on Gs biofilms have shown that the reduction potential of the Gs cells is −150 mV (33). Thus, as can be seen from Fig. 4 and Table 2, under these conditions of physiological redox potential (−150 mV) and pH (7.5), the cytochromes studied are not fully oxidized or fully reduced, and thus are functionally active (i.e., capable of receiving and donating electrons).

From the thermodynamic properties determined in this work, it is clear that, despite the high degree of sequence and structural homology, the Gs periplasmic triheme cytochromes behave quite differently. For PpcA and PpcD, dominant microstates emerge during protein oxidation (Fig. 5). In both cases, the oxidation progresses from particular protonated redox-microstates to particular deprotonated redox-microstates, showing how dominant microstates can confer the directionality of events (see Fig. 6). At the physiological redox potential and pH, PpcA can uptake a strongly reducing electron (−167 mV) and a weakly acidic proton (pK_a 8) from the donor associated with the cytoplasmic membrane. When it meets its physiological downstream redox partner, PpcA donates de-energized electrons (−109 mV) and a more acidic proton (pK_a 7.2), that is now sufficiently acidic to be released into the periplasm.

Figure 6.

Thermodynamic and mechanistic bases for energy transduction by PpcA (left) and PpcD (right). The functional pathway involving the significantly populated microscopic redox states in Fig. 5 is indicated by arrows in each scheme. The microstates are labeled as in Fig. 2. The pK_a values are those listed in Table 2, and the reduction potentials were calculated from the values indicated in Table 1. As a concrete example, the redox potential for the equilibrium involving the protonated microstates P_1H and P_14H for PpcA (−109 mV) is obtained by the sum of the energy terms corresponding to heme IV oxidation (−125 mV) and I-IV redox interaction (16 mV).

PpcD can also couple the transfer of electrons and protons, though by a different mechanism and covering a different redox potential range (see also Fig. 6). Indeed, the relevant microstates that are involved in the energy transduction in PpcA work in the range of −167 to −109 mV, whereas those of PpcD work in the range of −202 to −146 mV.

It is quite interesting that PpcA and PpcD can perform e⁻/H⁺ energy transduction in addition to their role in the electron transfer pathways that lead to the reduction of extracellular electron acceptors. This may represent additional mechanisms contributing to the H⁺ electrochemical potential gradient across the periplasmic membrane that drives ATP synthesis. The use of extracellular electron acceptors by Gs, in comparison with the use of fumarate, leads to a decrease in biomass production because of dissipation of the membrane potential by cytoplasm acidification (34). In contrast, the cytoplasmic protons produced from acetate oxidation are consumed in the cytoplasm when fumarate is the terminal electron acceptor. In metabolic modeling studies, Mahadevan et al. (34) showed that cellular growth in the presence of insoluble electron acceptors is possible only when additional e⁻/H⁺ coupling mechanisms (compared to those used in fumarate respiration) are present. The authors suggested that the most likely mechanism for generating additional membrane potential is the coupling of electron transfer to periplasmic cytochromes with proton translocation, so that additional membrane potential can be generated for ATP production. Thus, it can be proposed that PpcA and PpcD contribute to an additional transmembrane pH gradient by receiving weakly acidic protons (pK_a > 8.0) and electrons from donors associated with the cytoplasmic membrane. These protons will then be released in the more-acidic periplasmic space with a lower pK_a (<7.5) upon electron transfer to the acceptor.

On the other hand, PpcB and PpcE appear to be designed to perform different cellular functions, since no evidence for coupling of e⁻/H⁺ transfer was observed. Furthermore, the observation of different working potential ranges for these cytochromes suggests that they may have physiological redox partners distinct from those of PpcA and PpcD.

The distinct functional properties described here for the four Gs triheme cytochromes correlate with previous proteomics and knock-out mutant studies of Gs (7,13). Those studies showed that PpcA and PpcD play a significant role in the respiration of iron oxides. Their different detailed functional mechanisms with respect to reduction potentials and pK_a values suggests that they interact with distinct partners. These differences provide an excellent example of how structurally related proteins from the same microorganism can interact with different physiological partners, and establish a rationalization for the coexistence of five homologous periplasmic triheme cytochromes in Gs. This work provides the first step in unraveling the organization of the complex network of redox proteins found in the periplasmic space of the bacterium Gs.

Supporting Material

A detailed material and method section, four figures, and two tables are available at http://www.biophysj.org/biophysj/supplemental/S0006-3495(10)00481-9.