Unless otherwise stated, all reagents were of analytical or higher grade and were purchased from Sigma-Aldrich, Merck and Fluka. Solutions were prepared in bi-distilled water or Milli-Q water where indicated.

The DNA sequence of cytochrome c ₅₅₂ from W. succinogenes (WS0700) was obtained from the GenBank (http://www.ncbi.nlm.nih.gov/genbank) to design the primers. All the analyzed protein sequences were obtained from the Protein database at NCBI (http://www.ncbi.nlm.nih.gov/protein). Multi-sequence alignments were performed using the Clustal Omega (Sievers et al., 2011). Prediction of a signal peptide was performed using SignalP—6.0 (Teufel et al., 2022). A structural model of the globular domain of Ws cytochrome c ₅₅₂ was generated using AlphaFold colab (Jumper et al., 2021), which has a per-residue confidence greater than 90, indicating a high degree of confidence in the coordinates of the polypeptide chain (Supplementary Figure S2, in the Supplementary Material). The heme group was manually added, according to the following procedure. The coordinates of the heme group were those of M. nauticus cytochrome c ₅₅₂ after superposition of the model structure with this structure (global RMDS of 1.8 Å). The superposition of the two structures (Supplementary Figure S9, in the Supplementary Material) showed that the atoms coordinating the heme iron (Nε2 of His14 and Sδ of Met58), as well as the sulfur atoms of the two cysteine residues (Cys10 and Cys13) forming the thioether bond with the heme, are superimposed. In this way, the thioether bond was formed with the vinyl groups at the β-pyrrole positions two and four of the heme, and this model structure was used to generate the figures shown in the manuscript. In this structure, the distance between the coordinating atoms and the heme iron was similar to other structures (2.0 Å and 2.2 Å for the N-Fe and S-Fe, respectively), and the angles of the thioether bond were 100° and 109° (which compares well with 105° and 108° in the M. nauticus cytochrome c ₅₅₂ structure).

The gene encoding W. succinogenes cytochrome c ₅₅₂ (WS0700), hereafter referred to as Ws cytochrome c ₅₅₂, was amplified by PCR using a set of primers (forward: CAT​GCC​ATG​GCT​GAT​GGT​GCA​ACC​CTC​TA, and reverse: CTG​GCT​CGA​GTT​ACT​TGA​GCG​TGG​AGA​TGT​AT) with a NcoI and a XhoI restriction site at the 5′ and 3′-ends, respectively. The amplified fragment encoding the mature cytochrome c ₅₅₂, excluding its signal peptide (residue 1 to 17, Supplementary Figure S1, in the Supplementary Material), was cloned into a pET22b (+) plasmid, hereafter designated as pET22-Wscytc. The plasmid confers ampicillin resistance and adds an N-terminal signal peptide (pelB) to direct the protein to the periplasm. This cloning strategy introduced an additional Met residue, at the N-terminus after cleavage of the signal peptide, giving the mature protein 82 residues.

Escherichia coli DH5α (Invitrogen) was used for plasmid propagation. Ws cytochrome c ₅₅₂ was produced in E. coli C41 (DE3) (Merck) co-transformed with pET22-Wscytc and pEC86 (which harbors the ccm genes for production of the machinery for c-type heme biosynthesis and maturation (Arslan et al., 1998), and confers chloramphenicol resistance). Four to five colonies of the co-transformed E. coli C41 (DE3) were used to inoculate 50 mL of Luria-Bertani (LB) medium (10 g tryptone, 10 g NaCl and 5 g yeast extract, per liter) supplemented with 100 μg/mL ampicillin and 30 μg/mL chloramphenicol and grown overnight at 37°C, 210 rpm. Fresh 2xYT medium (16 g tryptone, 5 g NaCl and 10 g yeast extract, per liter), supplemented with the same antibiotics, was inoculated with 1% of the pre-inoculum. Cultures were incubated under orbital shaking at 37°C, 210 rpm until an OD_{600 nm} of 0.6 was reached. At this point, gene expression was induced with 0.25 mM IPTG for 18 h at 30°C, 120 rpm. Cells were harvested at 8000 × g, six°C, 15 min, and resuspended in 50 mM Tris-HCl, pH 7.6 containing protease inhibitors (cOmplete™, Mini, EDTA-free, Protease Inhibitor Cocktail Tablets, Roche).

The periplasmic fraction was obtained by four freeze-thaw cycles and separated from spheroplasts and cell debris by centrifugation at 39,000 g, 6°C, 45 min. Purification was performed in a single chromatographic step, using a cationic exchange chromatography. The periplasmic fraction was diluted 10× with cold milli-Q water and loaded onto a CM52 (Cytiva) cation-exchange chromatographic column (12.5 cm x 3 cm ∅, 90 mL), equilibrated with 5 mM sodium phosphate buffer, pH 7. Protein not adsorbed on the matrix was eluted with 5 mM sodium phosphate buffer, pH 7, and then a linear gradient between 0 and 500 mM NaCl was applied. The Ws cytochrome c ₅₅₂ eluted with 125 mM NaCl, and the fractions with A_408.5nm/A_280nm above 4.7 were considered pure, combined, and concentrated over a 3 kDa MWCO membrane using a Vivaspin 15 centrifugal device (Sartorius). The final Ws cytochrome c ₅₅₂ fraction was buffer exchanged to 5 mM Na-phosphate pH 7.0 using a desalting PD-10 column (Cytiva). A 15% SDS-PAGE and 10% PAGE stained for protein (Coomassie blue) and heme content (Goodhew et al., 1986) was also used throughout the purification to verify the protein purity. Ws cytochrome c ₅₅₂ was stored in small aliquots at—80°C until further use. Protein and heme content were estimated in the same solution for which the CD spectra in the far-UV region was obtained: protein was quantified using the extinction coefficient at 205 nm determined for CD (see below) and heme content by the pyridine hemochrome assay (Berry and Trumpower, 1987). Molecular mass was determined by LC-ESI-MS (Bruker) using a μPACTM analytical column from PharmaFluidics. The Ws cytochrome c ₅₅₂ from the same purification batch was used for all the spectroscopic and biophysical analyses: the protein was desalted in the buffer with different pH values (see below) and then used in the different experiments.

The UV-visible spectra were recorded on a Shimadzu UV-1800 spectrophotometer, connected to a computer, using a 1 cm path quartz cuvette. The concentration of the protein was estimated using the reported molar extinction coefficient of 80.1 mM⁻¹cm^–1 at pH 7.6, 100 mM Na-phosphate buffer (Moura et al., 1988). Ws cytochrome c ₅₅₂ in the reduced state was obtained by reduction with a solution of sodium dithionite, with a final concentration of 5 mM.

Circular dichroism spectra were acquired in an Applied Photophysics ChirascanTM qCD spectrometer (Leatherhead, Surrey, United Kingdom). Far-UV (190–260 nm) spectra were acquired in a sample of 8.5–10.6 μM W cytochrome c ₅₅₂, in 5 mM Na-phosphate buffer, pH 5.0, 6.0, 7.0, 8.0, and 9.0, using a 1 mm path length cuvette with a total volume of 300 μL. In the visible region (260–800 nm) the spectra were recorded for a 35 μM W cytochrome c ₅₅₂ sample, in 5 mM Na-phosphate buffer, pH 7.0, using a 10 mm path length cuvette with a total volume of 2.3 mL. The concentration of the samples was determined using the absorbance at 205 nm and the extinction coefficient of 290,700 mM⁻¹ cm⁻¹, based on the amino acid content (Anthis and Clore, 2013), or the absorbance at 408 nm and the extinction coefficient mentioned before. CD data were reported in mean residue ellipticity ([θ]_MRE) as a function of wavelength. The [θ]_MRE is the CD raw data (in deg) corrected for the Ws cytochrome c ₅₅₂ concentration of the solution using Eq. 1: θ MRE = θ ⁡ deg ⁡ × MRW 10 × L × C (1) where MRW (mean residue weight) is the molecular mass of Ws cytochrome c ₅₅₂ divided by the number of peptide bonds (81), C is the concentration of the protein in the sample in g•mL⁻¹, and L is the pathlength of the cell in cm.

The CD spectra between 190 and 260 nm are an average of three spectral acquisitions at 25°C, with a bandwidth and step-size of 1 nm, and acquired with a time per point of 3 s. The far-UV data analysis to determine the secondary structure content was performed using the BeStSel server (https://bestsel.elte.hu/) (Micsonai et al., 2021; Micsonai et al., 2022). The CD spectra between 260 and 800 nm are an average of three spectral acquisitions at 25°C, with a bandwidth and step-size of 1 nm, and acquired with a time per point of 3 s. The temperature-dependent CD spectra were acquired between 10°C and 94°C, with a stepped ramp mode of 1 s per point, an increase of 2°C for each measurement, and a stabilization period of 1 min between each point.

The unfolding process was analyzed using the Gibbs–Helmholtz method - to fit the change in ellipticity at a single wavelength as a function of temperature, considering a two-state transition from a folded native state to an unfolded state, and assuming that the heat capacity of the folded and the unfolded states are equal, ΔCp = 0 (Greenfield, 2004). For the Gibbs–Helmholtz method, the experimental data were fitted using Eqs. 2–5 and using the SOLVER add-in program of Microsoft Excel. The molar ellipticity at any given temperature (T), (θ)_T, is given by Eq. 2: ϴ T = α × ϴ F − ϴ U + ϴ U , , (2) in which the fraction folded at any given temperature, α, is given by Eq. 3: α = K 1 + K (3) and the folding constant, K, at any given temperature is given by Eq. 4: K = ⁡ exp − Δ G R × T (4) and ΔG is given by Eq. 5: Δ G = Δ H × 1 − T T M (5) in which T_M is the temperature at which the fraction folded, α, is 0.5.

For the differential scanning calorimetry (DSC) experiments, the NanoDSC instrument (TA Instruments) was loaded with degassed buffers (baselines and reference cell) and protein solution (83–93 μM Ws cytochrome c ₅₅₂ in the sample cell). Each protein sample was passed through a desalting PD-10 column (Cytiva) equilibrated in the appropriate buffer (5 mM Na-phosphate buffer at pH 5.0, 6.0, 7.0, 8.0, and 9.0) and then diluted to the desired concentration in the same buffer. The temperature was raised from 10 to 120°C at a scan rate of 1°C/min, and 6 bar. The thermograms were analyzed with TA instruments NanoAnalyze software using a non-two-state model to fit the data and obtain the melting temperature (T _M ), the calorimetric (ΔH) and van’t Hoff (ΔH _v ) enthalpies. The corresponding baseline was subtracted from each sample scan.

Cyclic voltammetry assays were performed in a single compartment cell with a three-electrode configuration. A cysteamine-modified gold disk (ϕ = 2 mm) was used as the working electrode and a platinum foil and a saturated calomel electrode (SCE) were used as the secondary and reference electrodes, respectively. The gold electrode was modified by immersion in a 2 mM cysteamine solution for 30 min at room temperature (RT), followed by washing by immersion in deionized water. The protein (5 μL of approx. 550 μM) was then placed on the modified gold electrode and a cellulose membrane (3.5 kDa cut-off) was then applied for protein entrapment and studied in a thin-layer regime. The used electrolyte was a mixed 20 mM phosphate/acetate buffer/0.1 M NaCl, at different pH values or only 20 mM phosphate/0.1 M NaCl, pH 7. Assays were performed at RT in a strictly anaerobic environment inside an anaerobic chamber (c_O2<0.1 ppm). Different potential windows and scan rates were tried. All the potentials were converted and presented in the normal hydrogen electrode (NHE) reference scale. Controls were obtained using the same procedure and conditions without the presence of the protein.

Analysis of the primary sequence of the mature cytochrome c ₅₅₂ from W. succinogenes (81 residues) reveals the presence of an N-terminal heme-binding motif and a conserved methionine residue, about 40 residues away toward the C-terminus, as a distal axial ligand of the heme iron. These features suggest that this c-type cytochrome belongs to class I, subdivision b (Pettigrew and Moore, 1987).

In Figure 1 is shown the sequence alignment of the primary sequence of the mature cytochrome c ₅₅₂ from W. succinogenes with two c-type cytochromes that are electron donors of nitrous oxide reductase from M. nauticus (Dell’Acqua et al., 2008; Dell’acqua et al., 2011), cytochrome c ₅₅₂, and from Paracoccus denitrificans, cytochrome c ₅₅₀ (Pearson et al., 2003; Dell’acqua et al., 2011), and the cytochrome c ₅₅₃ from Nitratidesulfovibrio vulgaris Hildenborough (formerly known as Desulfovibrio vulgaris Hildenborough), which has a low reduction potential (vide infra) compared to other class I c-type cytochromes (Pettigrew and Moore, 1987). The sequence identity between these cytochromes is relatively low (34%, 35%, and 23% with the c-type cytochromes from M. nauticus, Nitratidesulfovibrio vulgaris Hildenborough, and P. denitrificans, respectively), but they all share the conserved heme-binding motif and the position of the axial methionine toward the C-terminus. The alignment shows that cytochrome c ₅₅₀ from P. denitrificans is longer, with an inserted region between residue 78 and 110 containing the coordinating methionine. The high sequence homology at the two extremities could indicate a transposition.

The structural model of this protein was obtained using AlphaFold colab (Supplementary Figure S2, in the Supplementary Material), and the heme group was added as described in Materials and Methods. This structure shows that it is composed of 4 α-helices and that the heme group is more solvent exposed than in the case of P. denitrificans cytochrome c ₅₅₀ (Figure 2D; Supplementary Figure S9 in the Supplementary Material), since the loops covering the heme and an extra helix are missing (Figure 2D, RMSd of 3.3 Å and TM-score of 0.55; Supplementary Table S1, in the Supplementary Material), as expected from the analysis of the sequence alignment (Figure 1). Comparison of its structure with that of the small c-type cytochrome from M. nauticus and N. vulgaris Hildenborough (RMSd of 1.8 and 1.0 Å, and TM-score of 0.86 and 0.89, respectively, Supplementary Table S1, in the Supplementary Material), shows that their structure and heme accessibility are similar (Figure 2; Supplementary Figure S9, in the Supplementary Material), although in the case of M. nauticus cytochrome c ₅₅₂ the heme is less exposed to the solvent because it is a dimer (Supplementary Figure S11, in the Supplementary Material). The superposition of the four structures highlights that the long loop in P. denitrificans cytochrome c ₅₅₀ is a α-helix in the other structures, and that there are also two short β-sheets that are absent in the small cytochromes. An important feature is that, although the structures have differences, the heme coordinating residues are well aligned in the structure (Supplementary Figures S9 and S10; Supplementary Table S2 in the Supplementary Material), although this is not shown in the sequence alignment (Figure 1). This structural difference supports the classification of Ws cytochrome c ₅₅₂ and the other two small cytochromes as belonging to subdivision Ib, whereas P. denitrificans cytochrome c ₅₅₀ belongs to subdivision Ia (Pettigrew and Moore, 1987).

The DNA fragment amplified by PCR and inserted into the pET22b expression vector encoded only the globular region of Ws cytochrome c ₅₅₂, starting at A18, and due to cloning an additional Met residue was added at the N-terminus (Supplementary Figure S1, in the Supplementary Material). This recombinant protein has 82 residues, after removal of the PelB signal peptide, and an expected molecular mass of 9342.7 Da (8728.2 Da plus 614.5 Da, which is the sum of the molecular mass of the polypeptide chain and one heme group).

The Ws cytochrome c ₅₅₂ was isolated from the periplasm of E. coli in a single chromatographic step, consisting of a cationic exchange chromatography, since the pI of this protein is 8.6. The cytochrome c was considered pure according to its SDS-PAGE and PAGE profile (Supplementary Figure S3, in the Supplementary Material) with a single band, and this fraction had a purity ratio [A_552nm–A_570nm (reduced form)/A_280nm (oxidized form)] of 1.20. This procedure has an average yield of 46 mg of pure Ws cytochrome c ₅₅₂ per liter of growth medium. The heme/protein ratio of the purified cytochrome c, as determined by heme and protein concentration (see Materials and Methods), was 0.89 ± 0.02, confirming the presence of 1 c-type heme bound to the polypeptide chain, as expected. The molecular mass of the purified Ws cytochrome c ₅₅₂ was determined by LC-ESI-MS to be 9342.53 ± 0.04 Da, confirming the correct processing of the signal peptide and the presence of the c-type heme.

The UV-visible spectrum of the heterologous cytochrome c shows similar spectral features to the same protein isolated from W. succinogenes (Moura et al., 1988) (Figure 3A), namely, the Soret band at 409 nm, and absorption bands with maxima at 620 nm and 695 nm, indicating the presence of a high-spin state, and a methionyl residue (Met59, numbering according to the mature sequence), respectively. The extinction coefficient estimated based on the heme content is identical to that reported, 80.1 mM⁻¹cm⁻¹ at pH 7.6 (Moura et al., 1988).

After reduction, the Soret band shifts to 417 nm, with the appearance of the β and α bands at 523 nm and 552 nm, respectively (Figure 3A). The maximum absorption of the α-band at 552 nm, designates this cytochrome as cytochrome c ₅₅₂, as has been for others (Hon-nami and Oshima, 1977; Saraiva et al., 1994; Samyn et al., 1998).

As previously observed, the high-spin state decreases with increasing pH, as the ratio A_621nm/A_695nm decreases (from pH 5.0 to pH 7.0 the ratio is around 6.0, decreasing to 4.7 and 4.5 at pH 8.0 and 9.0, respectively) (Table 1), indicating a pKa around 7.5 (Figure 5B), as previously reported (inflection point at pH 7.3 for the absorbance at 620 nm) (Moura et al., 1988).

The folding state of a protein and its secondary structure content can be determined by analyzing its circular dichroism spectra in the far-UV region (190–260 nm). The CD spectrum of the as-isolated Ws cytochrome c ₅₅₂ shows a positive peak at 192 nm, and two negative peaks at 208 and 222 nm, with a ratio θ_222nm/θ_208nm of 0.8. These features indicate the presence of a folded protein composed mainly of α-helix (Figure 3B). This spectrum is similar to that of cytochrome c ₅₅₂ from Hydrogenobacter thermophilus and cytochrome c ₅₅₃ from N. vulgaris Hildenborough, cytochromes c composed mainly of α-helix (Wittung-Stafshede, 1999; Oikawa et al., 2005). The secondary structure content was estimated by analyzing the far-UV spectrum at 25°C, pH 7.0 using the BeStSel algorithm (Micsonai et al., 2021; Micsonai et al., 2022) and indicates the presence of 27% α-helices, with 10% of β-sheets (Table 1). These values are in agreement with the analysis of the coordinates of the structural model of this protein, which predicts a protein mainly α-helix with a small percentage of structure as β-sheets (40% α-helices, with 7% of β-sheets).

The near-UV and visible CD spectra of the oxidized Ws cytochrome c ₅₅₂ were collected in a pH range of 5–9 (Supplementary Figure S6, in the Supplementary Material). In the near-UV region, the presence of two positive bands at 282 nm and 290 nm is observed, which can be attributed to the single tryptophan residue present in the sequence, W33, which is located near the heme propionates in the predicted structure (Supplementary Figure S12, in the Supplementary Material). These bands do not change over the pH range studied. The visible region of the CD spectrum of the oxidized Ws cytochrome c ₅₅₂ at pH 7 has a broad negative peak between 320 nm and 400 nm with a maximum at 370 nm (Figure 3C), which has also been observed in other c-type cytochromes (Vinogradov and Zand, 1968). A negative well-defined S-shaped band is observed with a negative minimum at 407 nm and a negative maximum at 416 nm corresponding to the Soret Cotton effect. This Cotton effect on the Soret band of CD spectra has usually been associated with transitions of the heme with nearby aromatic side chains (Hsu and Woody, 1971; Pielak et al., 1986) and is a fingerprint of heme integrity. The visible CD spectra do not change significantly with the pH (Supplementary Figure S6, in the Supplementary Material, suggesting that the tertiary structure of the protein, and in particular the heme pocked is not being affected by changes in the pH. The fact that the Soret band is negative can also be explained by the parallel propionate groups on the porphyrin ring (Nagai et al., 2015), which again remains unchanged with pH for Ws cytochrome c ₅₅₂. Upon reduction, the broad band with a maximum at 370 nm becomes a negative minimum and the Soret CD spectrum shows a well-defined sharp band with an inflection point at 422 nm, and a negative peak at c. a. 417 nm, followed by a positive peak at 433 with a shoulder at 440 nm. There is also a positive peak at 556 nm, which corresponds to the peak of the α-band. The changes observed in the visible CD spectra between the oxidized and the reduced forms indicate that the heme environment changes upon reduction. This has also been observed for other c-type cytochromes (Vinogradov and Zand, 1968).

The influence of pH on the thermostability of cytochrome c ₅₅₂ was studied by CD in the far-UV region and differential scanning calorimetry (DSC) (Figure 4), which provide complementary information on the unfolding equilibrium. The CD in the far-UV region monitors the change in secondary structure content, while DSC provides information about the overall unfolding process of the protein.

The far-UV CD spectra at different pH values, acquired at 25°C have similar spectral features to those described at pH 7.0: the peaks have the same maxima and minima (no shifts were observed) (Supplementary Figure S5, in the Supplementary Material). However, the ratio of θ _222nm /θ _208nm decreases at higher and lower pH values, indicating that there is a change in the alpha-helix content. Analysis of these spectra using the BeStSel algorithm confirms this small change in the secondary structure content (Table 1), with a decrease in alpha-helix content at pH 5.0, 6.0 and 9.0.

The thermostability studied by CD in the far-UV region shows that this protein is highly stable up to 60°C, independent of pH (Figure 4A). The T_M could only be estimated for pH 8.0 and 9.0, because at lower pH values no stabilization of the ellipticity up to 94°C was observed. At these two pH values, the estimated T_M was 73°C, but the process is not reversible (Supplementary Figure S7, in the Supplementary Material).

The thermogram obtained by DSC shows a single sharp endothermic peak, which was fitted to a one-peak transition model, from which a transition temperature, and a van’t Hoff enthalpy of unfolding, ΔH_vH, was estimated (Table 1). The shape of the thermogram shows that there is a sharp decrease, indicating that the protein aggregates in solution (Supplementary Figure S4, in the Supplementary Material), which is confirmed by the small ratio ΔH_cal/ΔH_vH (Table 1). This phenomenon occurs for all the pH values studied. The unfolding temperature shows the same trend as that observed for CD, indicating that this cytochrome c ₅₅₂ is more stable at lower pH values.

The results show that Ws cytochrome c ₅₅₂ has a quasi-reversible electrochemical behavior, as expected considering similar protein systems (Santos et al., 2015; Teixeira et al., 2019). The cyclic voltammogram at pH 7.0 shows a redox process, with broad anodic and cathodic peaks (Figure 5A). The formal reduction potential calculated from the average of E_pa and E_pc is +97 ± 5 mV vs. NHE. This experimentally estimated formal reduction potential is in agreement with the previously reported value for this protein of +105 ± 15 mV, obtained by potentiometry (Moura et al., 1988). The relative difference can be explained by the different techniques used.

The study of the redox behavior at different pH values, shows that at pH values lower than pH 7.0, the potential waves are relatively better defined, and a single redox process is observed (Supplementary Figure S8, in the Supplementary Material). Above pH 7.0, a broadening of the waves (both cathodic and anodic) occurs, but it is not possible to distinguish the processes. This behavior, if only pH 7.0 was considered, could be related to some dispersion of the heterogeneous electron transfer rate constants, due to multiple electron transfer pathways. However, the global redox behavior at different pH values points to the coexistence of two slightly different protein forms regarding the heme environment, which is consistent with the hypothesis of having an oxygenated species (OH⁻) or a water molecule binding/leaving at the distal axial position. The pH dependence of the redox potential considering only one acid/base group model is fitted with a pK _a values for the oxidized and reduced states: pK _ox and a pK _red of 6.9 and 7.3, respectively, with a calculated theoretical reduction potential of 0.122 V (pH 0) (Figure 5B).

W. succinogenes cytochrome c ₅₅₂ is an 81-residue c-type cytochrome that has been successfully heterologously produced in E. coli. It is an interesting protein because it is one of the shortest c-type cytochromes isolated to date and has unusual spectroscopic and redox properties, namely, a low reduction potential (+97 mV, pH 7.0) and the coexistence of a high and a low spin form at room temperature. Other small cytochromes have a similar reduction potential, but the high/low spin coexistence has not been reported (Table 2). Comparing its model structure, presented here, for the first time, it is clear that due to the shorter polypeptide chain, the heme is more exposed to the solvent, and not in the usual hydrophobic environment with only the exposed heme edge to the solvent (Figure 1) as in the case of P. denitrificans cytochrome c ₅₅₀. This solvent exposure may explain its lower reduction potential. Indeed, cytochrome c ₅₅₃ from N. vulgaris Hildenborough also has a low reduction potential, which has been attributed to the greater exposure of the heme to the solvent (Blackledge et al., 1996). An exception is the cytochrome c ₅₅₂ from M. nauticus, which has a much higher reduction potential despite being 88 residues long and having a similar structure (Table 2). However, this cytochrome c ₅₅₂ is a dimer in which the propionate groups are shielded from the solvent because they are located at the dimer interface (Supplementary Figure S11, in the Supplementary Material).

The effect of pH on the redox properties of W. succinogenes cytochrome c ₅₅₂ differs from that reported for N. vulgaris cytochrome c ₅₅₃, both in the shape of the cyclic voltammograms and in the estimated pKa. Similar to other class I c-type cytochromes, N. vulgaris cytochrome c ₅₅₃ shows no effect on the reduction potential up to pH 10.0 (Verhagen et al., 1994), after which there is a marked decrease in the reduction potential, which has been attributed to the loss of the methionine as a distal axial ligand with a pK _a of 10.9 (Verhagen et al., 1994). The pH dependence of Ws cytochrome c ₅₅₂ was studied only up to pH 9.0 to avoid its destabilization close to its pI (of 8.6) and shows a close pK _ox/pK _red, of 6.9/7.3, a lower value. Furthermore, the voltammograms indicate the presence of two forms, with similar reduction potential. Curiously, M. nauticus cytochrome c ₅₅₂ has a similar pKa observed in the redox potential determined by potentiometry (Saraiva et al., 1994), but not by cyclic voltammetry (data not shown) (Coutinho, 2013).

The observed band at 620 nm decreases in intensity with the increasing of the pH, while the band at 695 nm (due to the axially coordinated methionine) increases in intensity. This can be explained by the fact that the methionine ligand becomes strongly bound to the heme iron, and thus the heme becomes low spin, while at low pHs the methionine can be partially replaced by a low-field ligand, possibly OH⁻, which can be protonated to water. This could explain the observed low-spin EPR signal (Moura et al., 1988), since it can have a temperature-dependent high/low-spin equilibrium (Barreiro et al., 2023a). Another possible explanation for the observed equilibrium could be the protonation/deprotonation of the propionates. In any case both hypotheses do not promote structural changes in the heme pocket, since the near-UV/visible CD spectra do not change with pH (Supplementary Figure S6, in the Supplementary Material).

The thermostability of W. succinogenes cytochrome c ₅₅₂ has been studied for the first time by CD and DSC. The T_M values estimated by the two complementary techniques are similar, but no other thermodynamic parameters can be estimated because the unfolding process is irreversible. This c-type cytochrome is rather stable, and its stability decreases close to its pI. The effect of pH on the thermostability of some c-type cytochromes has been studied at very low pH values (Kuroda et al., 1992). However, there are no reports on the effect of pH on the thermostability of c-type cytochromes in the range studied here, so the destabilization at high pH, observed for Ws cytochrome c ₅₅₂, may be due to a change in the hydrogen bonds that stabilize the structure of this protein.

The unfolding temperature observed at pH 7.0 is relatively higher than that observed for other proteins isolated from mesophilic bacteria (Nobrega et al., 2017), although some exceptions have been reported (Barreiro et al., 2023b). However, the T_M is relatively low (90°C, at pH 5.0) compared to that of cytochrome c ₅₅₂ from thermophilic organisms: Hydrogenophilus thermoluteolus and H. thermophilus, which have a T_M of 108°C and 121°C, respectively, at pH 5.0 (Oikawa et al., 2005; Nakamura et al., 2006). These two c-type cytochromes are as small as Ws cytochrome c ₅₅₂, containing about 80 residues. Interestingly, however, although composed of only 4 α-helices, there is a loop protecting the 6-propionate group of the porphyrin ring (Supplementary Figure S12, in the Supplementary Material), which is absent in W. succinogenes cytochrome c ₅₅₂.

Furthermore, Pseudomonas aeruginosa cytochrome c ₅₅₁, being an 82-residue protein from a mesophilic organism, has a T_M of 85°C, at pH 7.0 (Sanbongi et al., 1989), which compares well with the unfolding temperature of W. succinogenes cytochrome c ₅₅₂, at pH 7.0 (80°C). The structure of this cytochrome c ₅₅₁ is similar to the previously mentioned highly thermostable c-type cytochromes, with the 4 α-helices and the loop protecting the 6-propionate, so perhaps the exposed heme does not affect the thermostability of the protein as much. In fact, some residues have been proposed to be the main cause of thermostability between H. thermophilus cytochrome c ₅₅₂ and P. aeruginosa cytochrome c ₅₅₁ (Oikawa et al., 2005), suggesting that the stabilization may be due to side-chain interactions.