Chemicals

Isopropyl-β-D-thiogalactopyranoside (IPTG), dithiothreitol (DTT), hemin, oxidized glutathione (GSSG), veratryl alcohol (VA), manganese(II) sulphate, Reactive Black 5 (RB5), 2,6-dimethoxyphenol (DMP), sodium tartrate and other chemicals were purchased from Sigma-Aldrich; urea and hydrogen peroxide were from Merck; and 2,2'-azino-bis(3-ethylbenzothiazoline-6-sulfonate) (ABTS) from Roche.

Design of VP Variants

VPi and VPi-br variants were designed in silico based on a comparative analysis of the mature P. eryngii VP (allelic variant VPL2; GenBank^™ AF007222) and P. ostreatus MnP4 (ID 1099081 in the P. ostreatus PC15 v2.0 genome sequence from the Joint Genome Institute, JGI, at http://genome.jgi.doe.gov/PleosPC15_2/PleosPC15_2.home.html). For this analysis: i) the amino acid sequence alignment of both enzymes was performed using the pairwise sequence alignment tools (Needle, Stretcher, Water and Matcher programs) available at the European Bioinformatics Institute (EMBL-EBI); and ii) the structural alignment of VPL2 (PDB: 2BOQ) and MnP4 (PDB: 4BM1) was carried out with PyMOL (http://pymol.org). From this analysis, the VPi coding sequence was prepared by replacing codons encoding eight amino acid residues in VPL2 with those present at homologous positions in MnP4. The substituted amino acids were Asp69 → Ser (TCC), Thr70 → Asp (GAC), Ser86 → Glu (GAG), Asp146 → Thr (ACC), Gln202 → Leu (CTC), His232 → Glu (GAG), Ser301 → Lys (AAG) and Gln239 → Arg (CGC). The introduction of the following additional mutations in VPi resulted in the VPi-br variant: Thr2 → Lys (AAG), Ala131 → Lys (AAG), Gln219 → Lys (AAA), Leu288 → Arg (CGT), Ala308 → Arg (CGC), Ala309 → Lys (AAG) and Ala314 → Arg (CGT). Both VPi and VPi-br sequences were synthesized by ATG:biosynthetics (Merzhausen, Germany) and cloned into the NdeI/BamHI restriction sites of the expression vector pFLAG1 (International Biotechnologies Inc., Cambridge, UK).

Other two VP variants were produced using the QuikChange^™ Site-Directed Mutagenesis kit (Stratagene, La Jolla, CA, USA). Each of them was obtained by mutagenic PCR using the expression vector pFLAG1 containing the VPi (pFLAG1-VPi) or the VPi-br (pFLAG1-VPi-br) coding sequences as template, and two primers consisting of a direct and a reverse oligonucleotide designed complementary to opposite strands of the same DNA region containing the desired mutations. To obtain VPi-ss (A49C/A61C), a first PCR was carried out using pFLAG1-VPi as template and primers for A49C mutation, 5’-CC CTT CGT TTG ACT TTC CAC GAT TGC ATC GGT TTC TCT CC-3' (only direct sequences are shown here and below, with the changed triplets underlined and the mutations introduced in bold). Then, a second PCR was made using pFLAG1-VPi containing the A49C mutation as template and primers for A61C mutation, 5’-GGC GGA GGA GGA TGT GAC GGT TCC ATC ATC GCG-3’. VPi-br-ss was obtained using pFLAG1-VPi-br as template and primers for A49C and A61C mutations in two consecutive PCR reactions.

PCR reactions were carried out in an Eppendorf Mastercycler Pro S using 10 ng of template DNA, 250 μM each dNTP, 125 ng of both direct and reverse primers, 2.5 units of Pfu Turbo DNA polymerase AD (Stratagene) and the manufacture´s reaction buffer. Reaction conditions were as follows: (i) a “hot start” of 95°C for 1 min; (ii) 18 cycles at 95°C for 50 s, 55°C for 50 s and 68°C for 10 min; and (iii) a final cycle at 68°C for 10 min. The plasmids obtained from the mutagenic PCR were transformed into Escherichia coli DH5α for plasmid propagation. At least one positive clone of each variant was completely sequenced using an ABI 3730 DNA Analyzer (Applied Biosystem), checked and used to transform E. coli W3110 for protein expression.

Enzyme Production, Activation and Purification

Native (i.e. wild-type recombinant enzyme) VP of P. eryngii (isoenzyme VPL2) and its mutated variants were expressed in E. coli W3110 after transformation with the corresponding expression vectors. Cells were grown in Terrific Broth medium [23] supplemented with 100 μg/ml of ampicillin at 37°C and 180 rpm until OD₅₀₀~1 (~3 h). Then protein expression was induced with 1 mM IPTG and cells were grown for a further 4 h. The apoenzyme accumulated in inclusion bodies and was recovered by solubilization in 50 mM Tris-HCl (pH 8.0) containing 8 M urea, 1 mM EDTA and 1 mM DTT for 1 h at 4°C. The subsequent in vitro folding of the solubilized protein was performed overnight at room temperature in a solution of 0.16 M urea, 20 μM hemin, 5 mM CaCl₂, 0.5 mM GSSG, 0.1 mM DTT and 0.1 mg/ml protein in 20 mM Tris-HCl at pH 9.5. The refolded enzyme was purified by Resource-Q chromatography using a 0–0.3 M NaCl gradient (2 ml/min flow rate, 20 min) in 10 mM sodium tartrate (pH 5.5) containing 1 mM CaCl₂. Finally the enzymes were dialyzed against 10 mM sodium tartrate (pH 5). The purified native VP and its variants showed an R_z value (A₄₀₇/A₂₈₀)~ 4 indicative of their high purity. Moreover, the UV-visible spectrum obtained in the 300–700 nm range was used to check the correct incorporation of heme into the enzyme [24]. Enzyme concentration was determined from the absorbance of the Soret band (ε₄₀₇ = 150 mM^-1 cm^-1) [12].

Kinetic Studies

Oxidation of Mn²⁺ was estimated by the formation of Mn³⁺-tartrate complex (ε₃₂₈ = 6500 M^-1 cm^-1) in 0.1 M sodium tartrate (pH 5) and that of VA by formation of veratraldehyde (ε₃₁₀ = 9300 M^-1 cm^-1) in the same buffer at pH 3 and 2.5. Oxidation of ABTS was followed by generation of its cation radical (ε₄₃₆ = 29300 M^-1 cm^-1) and that of RB5 by its disappearance (ε₅₉₈ = 30000 M^-1 cm^-1) both in 0.1 M tartrate buffer at pH 3.5 (RB5 also at pH 3). H₂O₂ concentration was determined using ε₂₄₀ = 39.4 M^-1 cm^-1 [25]. All enzymatic activities were measured as initial velocities taking linear increments (decreases for RB5) in the presence of 0.1 mM H₂O₂ using a Shimadzu UV-1800 spectrophotometer. Values and standard errors for apparent affinity constant (Michaelis constant, K_m) and maximal enzyme turnover (catalytic constant, k_cat) were obtained fitting the experimental measurements to the Michaelis-Menten model with SigmaPlot 12.0 software (Systat. Software Inc, California). Fitting of these constants to the normalized Michaelis-Menten equation υ = (k_cat/K_m)[S]/(1+[S]/K_m) yielded enzyme efficiency values (k_cat/K_m) with their standard errors.

Optimum pH Determination

The optimum pH for oxidation of different substrates was determined by measuring the enzymatic activity of the native VP and its mutated variants at saturating concentrations of VA (20 mM), RB5 (15 μM) and ABTS (7 mM) in 0.1 M Britton-Robinson (B&R) buffer [26], and Mn²⁺ (6 mM) in 0.1 M sodium tartrate over the pH range 2.5–5.5, using 0.1 mM H₂O₂ and 0.01–0.03 μM enzyme.

pH Stability Studies

The enzymes were incubated at pH 3, 3.5 and 7 in 0.1 M B&R buffer at 25°C for 120 h. Their residual activity was estimated after 1 min, 1 h, 4 h, 24 h and 120 h incubation by measuring the oxidation of ABTS (2 mM) using 0.1 mM H₂O₂ and 0.01 μM of enzyme in 0.1 M sodium tartrate (pH 3.5). The activity obtained with the sample incubated for 1 min at pH 5 was taken as reference (maximum activity). UV-visible spectra in the 300–700 nm range and absorbance at the Soret band (407 nm) were recorded during the incubation at pH 3, 3.5 and 7 by using an Agilent 8453 diode array spectrophotometer.

Thermal Stability Studies

The enzymes were incubated at different temperatures (30–80°C) in 10 mM sodium tartrate (pH 5) for 10 min using a Thermo Shaker TS-100 BIOSAN, and then chilled on ice for 5 min. Their residual activity was measured by ABTS (2 mM) oxidation using 0.1 mM H₂O₂ and 0.01μM enzyme in 0.1 M sodium tartrate (pH 3.5), at 25°C. The activity of the enzyme measured before incubation at each temperature was taken as reference (maximum activity) to calculate the percentage of residual activity. Data of residual activity for each temperature were fitted to a sigmoidal model and T₅₀ values (temperature at which the activity is half of the initial one after 10 minutes of incubation) were calculated from the fit.

Crystallization

Crystallization trials were carried out by the sitting drop vapor diffusion method, using 96-well MRC2 plates with 50 μl reservoir solution and the commercially available screenings from Emerald (Wizard classic crystallization screens I, II and III) and Jena Biosciences (JBScreen Classic Kits 1–10). Drops consisted of 0.2 μl of protein solution (10 mg/ml in 10 mM sodium tartrate buffer at pH 5.0) and 0.2 μl of reservoir solution. Crystallization was carried out at 22°C. Crystals suitable for X-ray data collection were obtained from these initial trials. Crystals of the mutant VPi were obtained in 0.1 M sodium MES buffer at pH 6.5, 25% PEG 4000 and 0.2 M MgCl₂; and cryoprotected with 20% glycerol. Crystals of the mutant VPi-br were obtained in 10% PEG 4000 and 20% 2-Propanol; and cryoprotected with Paratone-N oil. Crystals of the mutant VPi-ss were obtained in 0.1 M sodium citrate buffer at pH 5.4, 12% PEG 4000 and 0.2 M calcium acetate; and cryoprotected with Paratone-N oil.

Data Collection and Processing

Crystals were mounted in nylon loops and flash-frozen in liquid nitrogen in the mother liquor containing the cryoprotectant indicated above. All diffraction data were obtained at 100 K. X-ray diffraction intensities were collected at SOLEIL (Gyf-sur-Yvette, France) and ALBA (Barcelona, Spain) synchrotrons. Diffraction data were indexed, integrated, merged and scaled using the program XDS [27]. Data collection statistics are shown in Table 1.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Data collection and refinement statistics.

https://doi.org/10.1371/journal.pone.0140984.t001

The structures of the three mutants were solved by molecular replacement using the crystal structure of P. eryngii VPL (3FMU) as the search model and the program PHASER implemented in the PHENIX package [28].

The final models were obtained by consecutive rounds of refinement, performed with the PHENIX package; followed by manual model building, performed with Coot [29] using σ_A weighted 2Fo-Fc and Fo-Fc electron density maps. Solvent molecules were introduced in the structure automatically in the refinement as implemented in the PHENIX package and visually inspected. A total of 5% of reflections was used to calculate the R_free value throughout the refinement process. The structures were validated using MolProbity [30, 31]. Refinement and final model statistics are summarized in Table 1. The coordinates and structure factors have been deposited with the Protein Data Bank accession codes 5ABN, 5ABO and 5ABQ. All figures were produced with PyMOL.

Rational Design Strategy

P. eryngii VP (isoenzyme VPL2) and P. ostreatus MnP4 share a common structural scaffold (Fig 1A). Their crystal structures (PDB entries 2BOQ for VP, and 4BM1 for MnP4) superimpose with a root mean square deviation (rmsd) of only 0.75 Å between the C_α positions over 316 amino acid residues, covering 95% of the mature proteins. This high structural similarity between both proteins was the basis of our strategy aimed to improve the pH stability of VP, which consisted in identifying the stabilizing motifs putatively contributing to the high stability towards pH of MnP4, and their subsequent transfer into VP. First, the amino acid sequences of these two enzymes were aligned (a 63% sequence identity was found) (Fig 1B). They differ in 124 amino acids, 27 being charged residues in MnP4 and non-charged in VP (while VP only has 11 charged residues being neutral in MnP4). In order to identify those contributing to pH stability in MnP4, a comparative analysis of their position in the molecular structures and interaction with surrounding residues was performed by walking on the superimposed crystal structures of MnP4 and VP on the protein surface. The internal regions of the structures were not explored and subsequent changes at these locations were not performed in VP, with only one exception as described below. Our idea was to substitute only a few amino acid residues in the VP molecular structure minimizing the impact on the catalytic sites to maintain the activity. It is well known that a very subtle equilibrium between stability and activity exits and the improvement of one of these properties is often at the expense of the other.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Structural and amino acid sequence alignment of VP (isoenzyme VPL2) from P. eryngii and MnP4 from P. ostreatus.

(A) Superimposition of VP (PDB 2BOQ) (grey) and MnP4 (PDB 4BM1) (orange) crystal structures (shown as cartoons) highlighting the VP amino acid residues mutated in this work (shown together with the heme group as CPK-colored sticks, and labeled according to the color code described below); and (B) alignment of their amino acid sequences (labeled using the same color code) (vertical lines denote conserved residues, and colons and periods indicate conservative and semi-conservative substitutions, respectively). Residues explored in the structural comparative analysis of VP and MnP4 searching for putative stabilizing motifs are shown in bold in the amino acid sequence alignment. VP amino acids subsequently substituted with those of MnP4 to generate the VPi variant appear on red background; those substituted by basic residues present in MnP4, introduced into VPi to form the VPi-br variant, are shown on blue background; and alanines substituted by cysteines to form an extra disulfide bridge in VPi resulting in the VPi-ss variant are highlighted on green background.

https://doi.org/10.1371/journal.pone.0140984.g001

Four regions exposed to the solvent were identified in the MnP4 molecular structure (Fig 2, left column) as hotspots for rational design of a VP with improved stability. These regions exhibit extra ion pairs and hydrogen bond networks in MnP4, compared with VPL2 (Fig 2, middle column), which are responsible for strengthening helix-helix, loop-helix and intra-loop interactions. Based on these observations and considering that most ion pairs have a stabilizing role [32], a VP variant (VPi) containing eight substitutions (D69S/T70D/S86E/D146T/Q202L/H232E/Q239R/S301K) was engineered by introducing the residues involved in these interactions in the four targeted regions. After verifying its increased pH stability (studies described below), new putative stabilizing residues were searched in MnP4. A high number of basic residues with their side chains exposed to the solvent, most of them with no movement restrictions by interactions with surrounding amino acids, were identified in MnP4 (31 of a total of 34 present in the protein, including 20 lysines and 11 arginines). Seven of these exposed residues (4 lysines and 3 arginines) were introduced into VPi, and the VPi-br variant containing the 8 mutations present in VPi plus mutations T2K/A131K/Q219K/L288R/A308R/A309R/A314R was obtained.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Structural details of four solvent exposed regions (A, B, C and D) in MnP4 (left column), VP (middle column) and VPi variant (right column).

Residues mutated in VPi and their homologous in MnP4 and VP are highlighted in red color.

https://doi.org/10.1371/journal.pone.0140984.g002

A third approach to improve the stability of VP was the further structural stabilization of the distal Ca²⁺-binding site, responsible for maintaining the relative position of the distal histidine involved in enzyme activation by H₂O₂. For that, the VPi-ss variant was designed by adding a double mutation (A49C/A61C) to VPi. The two cysteines added to this variant should form an extra disulfide bond contributing to the structural stabilization of the loop containing two of the four amino acid residues that coordinate the distal Ca²⁺ ion. Finally, the VPi-br-ss variant was designed by combining all the mutations described above in a single VP molecule.

The four purified VP variants exhibited the characteristic UV-visible absorption spectrum of the native VP showing relative maxima at 407 nm (Soret band), and at 505 and 637 nm (charge transfer bands CT2 and CT1, respectively) (S1 Fig), which is indicative of an active peroxidase with a high-spin ferric heme [14]. These results proved the correct heme incorporation in the recombinant enzymes.

Effect of the Mutations on VP Catalytic Properties

Native VP and the designed variants were kinetically characterized at the three catalytic sites characteristic of this ligninolytic peroxidase (Mn²⁺ oxidation site, main heme access channel and catalytic Trp exposed to the solvent) [16] (Table 2). The optimum pH for oxidation of four different substrates was also determined (Fig 3), in both cases with the aim of identifying potential effects on the catalytic activity due to the mutations introduced. Three of the four variants exhibited a catalytic efficiency for Mn²⁺ oxidation similar to that of the native enzyme. VPi-br-ss was the most affected variant, with only a 40% decrease in efficiency, and all of them (including native VP) showed the same pH activity profile with the optimum at pH 4.5–5 (Fig 3A). With respect to the catalytic activity at the main heme access channel, the optimum pH (3.5) for ABTS oxidation did not experience any variation in the four variants (Fig 3B), although their catalytic efficiency suffered a 35–70% decrease at this pH. The activity of native VP (and that of VPi-br and VPi-br-ss) oxidizing ABTS was dramatically reduced at pH 3 (Fig 3B). By contrast, VPi and VPi-ss showed high activity levels with this substrate at this pH, and a 2.7 and 2.3-fold improved catalytic efficiency, respectively, compared with the native enzyme at its optimum pH (Table 2). Finally, the catalytic activity at the exposed Trp164 responsible for the oxidation of high redox potential substrates was characterized using VA (simple lignin model compound) and RB5 (recalcitrant diazo dye) as reducing substrates. VPi, VPi-ss and VPi-br shifted their optimum pH from 3 to 2.5 for VA oxidation (Fig 3C), and VPi and VPi-ss widened the optimum pH range with RB5 (between pH 3 and 3.5) (Fig 3D). In addition, with both substrates, the catalytic efficiency of VPi and VPi-ss at the lower pH values (pH 2.5 and 3 for VA and RB5 oxidation, respectively) was higher than that of the native enzyme at its optimum pH (Table 2). This effect was more significant for RB5 oxidation, mainly due to a ~8-fold increased affinity (K_m = 0.4 μM for these variants vs 3.4 μM for the native enzyme), and much less important for VA oxidation (k_cat /K_m increasing from 2.2 s^-1 mM^-1 to 3.4 and 3.6 s^-1 mM^-1 in VPi and VPi-ss, respectively).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Steady-state kinetic constants [k_cat (s^-1), K_m (μM), k_cat /K_m(s^-1 mM^-1)] of native VP and mutated variants for oxidation of Mn²⁺ (at the Mn-binding site), ABTS (at the main heme access channel), and VA and RB5 (at the catalytic Trp exposed to the solvent)^a.

https://doi.org/10.1371/journal.pone.0140984.t002

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Optimum pH for oxidation of Mn²⁺ (A), ABTS (B), VA (C) and RB5 (D) by native VP (●), VPi (○), VPi-br (■), VPi-ss (□) and VPi-br-ss (▲).

Reactions in 0.1 M B&R buffer for ABTS (7 mM), VA (20 mM) and RB5 (15 mM) oxidation; and 0.1 M sodium sodium tartrate for Mn²⁺ oxidation (pH 2.5–5.5), with 0.1 mM H₂O₂ at 25°C. Means and 95% confidence limits are shown.

https://doi.org/10.1371/journal.pone.0140984.g003

pH and Thermal Stability of VP Variants

The stability of native VP and its mutated variants was evaluated during incubation at pH 3, 3.5 and 7, both by measuring the residual activity (Fig 4) and by monitoring the evolution of the UV-visible spectra (Fig 5). The decrease of the Soret band at 407 nm, typical of a stable native VP at pH 5 [14], was followed as an indicative of the integrity of the heme environment (S2 Fig). The results revealed that VPi is significantly more stable than native VP at acidic and neutral pH. The 7-fold stability improvement observed after 1 h of incubation at pH 3 was very limited in time since both the native enzyme and the mutated variant were nearly completely inactivated after 4 h of incubation (Fig 4A and 4B). By contrast, the improvement at pH 3.5 and 7 was more extended in time. VPi retained 61% (at pH 3.5) and 55% (at pH 7) of the initial activity after 24 and 120 h, respectively, compared with the native enzyme which resulted almost completely inactivated under the same experimental conditions.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. pH stability of native VP (A) and mutated variants VPi (B), VPi-br (C), VPi-ss (D) and VPi-br-ss (E).

The enzymes were incubated in 0.1 M B&R buffer (at pH 3, 3.5 and 7) at 25°C, and their residual activity was measured after 1 (), 4 (□), 24 (■) and 120 h (), with 7 mM ABTS and 0.1 mM H₂O₂ in 0.1 M sodium tartrate pH 3.5, and referred to activity after 1 min incubation at pH 5. Means and 95% confidence limits are shown.

https://doi.org/10.1371/journal.pone.0140984.g004

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 5. Time course of the electronic absorption spectra of native VP (top) and VPi (bottom) at acidic and neutral pH.

UV-visible spectra of native VP and VPi after 0 (red line), 1 (green line) and 5 h (blue line) incubation at pH 3 (A and D), 3.5 (B and E) and 7 (C and F) in 0.1 M B&R buffer, at 25°C.

https://doi.org/10.1371/journal.pone.0140984.g005

This stability improvement, measured as a percentage of the initial activity, could be correlated with the increased stability of the heme environment as revealed from the analysis of the electronic absorption spectra obtained for native VP and VPi. Although the two enzymes suffered strong changes in their UV-visible spectra at pH 3 (Fig 5A and 5D), these changes were quantitatively less significant for VPi. A faster decrease in the intensity at 411 nm (Soret band) and 507 nm (CT2 band) was observed for native VP (Fig 5A), with the Soret maximum progressively blue-shifting (39 nm) to attain a maximum at 372 nm after 5 h. A shoulder at 545 nm (β band) and a maximum at 638 nm (CT1 band) were observed over the entire time of the experiment in both native VP and VPi. The spectrum described for the native VP after 5 h at pH 3 (exhibiting maxima at 372, 507, 545 and 638 nm) is compatible with that of a four-coordinate heme. During these 5 h, VPi experienced a 38% decrease in the maximum at 410 nm with a shoulder appearing near the 372 nm region (Fig 5D). This spectrum has intermediate characteristics between those of the native VP incubated at pH 3 and those previously described for the native VP at pH 4.5 (maxima at 409, 505 and 638 nm) [33] at which the enzyme has been described to be a stable high-spin heme protein. These spectral characteristics suggest that at least two VPi species are present under these conditions, one with the iron four-coordinated, and the other with the iron also coordinated by the proximal histidine. This indicates that VPi moves much more slowly than the native enzyme towards destabilization of the heme environment at pH 3.

On the other hand, only slight modifications were observed in the spectra of VPi both at pH 3.5 and 7 after 15 h of incubation (Fig 5E and 5F) unlike what happened with the native enzyme (Fig 5B and 5C). Native VP experienced a sharp drop in the Soret band at pH 3.5, following the same spectral evolution described for the enzyme incubated at pH 3, although with less dramatic changes (Fig 5B and S2B Fig). The spectral changes at pH 7 exhibited a different behavior. A slight drop and shift of the Soret band to 409 nm were observed concomitantly with both the progressive disappearance of the CT2 and CT1 bands (at 505 and 638 nm respectively), and the increasing of α and β bands at 565 and 530 nm at this pH (Fig 5C and S2C Fig) suggesting an hexacoordinated heme iron. Interestingly, the almost undetectable changes in the spectrum of VPi at pH 3.5 and 7 contrast with the loss of 39% (at pH 3.5) and 45% (at pH 7) of the initial activity after 24 h. All these results together suggest that the mutations introduced at residues exposed to the solvent in VPi ultimately contribute to increase the stability of the heme environment, necessary for activity of ligninolytic peroxidases. However, this does not seem to be enough to completely stabilize the enzyme.

VPi-br harboring seven additional basic residues exposed to the solvent, and VPi-ss containing two cysteines added to form an additional extra disulfide bond, further improved the stability of VPi at acidic pH (Fig 4C and 4D, respectively). Both variants retained 86% and 85% of the initial activity at pH 3.5 after 24 h of incubation compared to the 61% retained by VPi and far from the 1% of the native VP. Moreover, although they were not so stable at pH 3 as at pH 3.5, VPi-br and VPi-ss maintained 62% and 47% of the initial activity, respectively, at pH 3 after 1 h compared to the 38% showed by VPi and the 5% of native VP. The analysis of the UV-visble spectra of these two variants over time evidenced an evolution of the Soret band similar to that observed for VPi (S2 and S3 Figs).This is indicative of the integrity of the heme and its environment at acidic pH. Unlike what observed at low pH, VPi-ss only exhibited a slight increased stability after 1 h incubation at pH 7 compared with VPi (Fig 4D). Both VPi-br and VPi-ss not only did not show an additional improvement at longer times at neutral pH but, on the contrary, VPi-br significantly reversed the improvement attained in VPi (Fig 4C). The accumulation of all the mutations in the VPi-br-ss variant retained or slightly improved the stability of VPi-br and VPi-ss at pH 3 (43% of residual activity after 4 h) but worsen at pH 7 being even less stable than the native VP at this pH (Fig 4E).

Regarding thermal stability, VPi, VPi-br and VPi-ss exhibited T₅₀ values between 58°C-60°C (S4 Fig). These values are similar to that found for the native VP (61°C) and higher than that obtained for VPi-br-ss that showed a T₅₀ of 55°C (6°C lower than that of the native VP).

Crystal Structure Analysis of the VP Variants

Crystal structures of VPi, VPi-br and VPi-ss showing enhanced pH stability were solved (Table 1) and subsequently analyzed both to verify that the mutations introduced had generated the expected stabilizing motifs and to explain how these motifs contribute to increase the enzyme stability. The overall folding characterizing the native VP, as well as the key elements that maintain its molecular architecture (i.e. two structural Ca²⁺ ions and four disulfide bonds), and position of the heme group, appear conserved in the three mutated variants (Fig 6A shows these elements in VPi). The structural changes observed are in general restricted to the regions where mutations were introduced.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 6. Crystal structures of VPi, VPi-br and VPi-ss variants.

(A) Molecular structure of VPi (with 12 α-helices named from A to J, shown as cylinders) including general structural elements such as four disulfide bonds (cyan sticks) and two Ca²⁺ ions (green spheres); heme cofactor; the two catalytic histidines above and below the porphyrin plane; and mutated residues (all of them as CPK sticks) generating new H-bond and salt bridge interactions (yellow dashed lines) at four regions (named A to D) described in more detail in Fig 2. (B) Molecular structure of VPi-br, showing the same general elements described for VPi plus the seven solvent-exposed basic residues characterizing this variant (mutations described in VPi are also included in VPi-br but they have not been represented for simplifying purposes). (C) Structural detail of the VPi-ss variant showing the extra disulfide bond (formed by Cys49 and Cys61) that connects helices B and B'a (shown as cartoons); the amino acid residues (CPK sticks) and water molecules (w) coordinating the distal Ca²⁺ ion; and one of the four disulfide bonds naturally existing in native VP between cysteine residues 34 and 114 that connects helices B and D (also depicted as cartoon) (heme and axial histidines are also shown).

https://doi.org/10.1371/journal.pone.0140984.g006

VPi includes eight mutations distributed into four solvent-exposed regions (named A-D in Fig 6A). Region A (containing mutations D69S/T70D/S86E) is located at the heme distal side above the heme plane, whereas regions B, C and D (containing mutations D146T/Q239R, Q202L/H232E and S301K, respectively) are found at the proximal side below the heme plane. The three mutations introduced in region A fail to emulate the contacts identified in MnP4 (Fig 2A, left). However, compared with the native VP (Fig 2A, middle), they contribute to reinforce the interaction between helices B'_b and C by increasing the H-bond network in this area, as shown in the crystal structure (Fig 2A, right). Similarly, the two substitutions in region B strengthen the loop between helices H and I by interaction of the Arg239 guanidinium group with the Asp237 carboxylate (Fig 2B, right), mimicking that observed between Arg245 and Asp243 in MnP4 (Fig 2B, left). In addition, the two mutated residues in this region (Thr146 and Arg239) are able to retain the H-bond that connects the loop between helices H and I with the N-terminal end of helix E established between Asp146 and Gln239 in the native VP (Fig 2B, middle). Regarding the region C, the introduction of a glutamate at position 232 in helix H promotes the formation of a salt bridge between this amino acid and Arg227 (Fig 2C, right) emulating that observed between Glu238 and Arg233 in MnP4 (Fig 2C, left). This interaction, not existing in the native enzyme (Fig 2C, middle), reinforces an extensive H-bond network that anchors the helix H both to the C-terminal end of helix G and to Glu304 located at the C-terminal region of the protein consisting of 66 residues without clearly defined secondary structures (except for two 3-amino acids β strands and a single turn 3₁₀ helix). Finally, unlike what was described for the other regions, the S301K substitution included in region D (Fig 2D, right) do not have the expected effect. This should consist in formation of a new H-bond, as observed in MnP4 (Fig 2D, left). By contrast, the side-chain of Lys301 appears exposed to the solvent. Summarizing, three of the four protein regions containing mutations exhibit a modified distribution of hydrogen bonds and salt bridges that can be correlated with the improvement in pH stability previously described for the VPi variant.

VPi-br and VPi-ss crystal structures exhibited minor differences compared with VPi regarding the interactions involving the mutated residues in region A (VPi-br also regarding the interactions of residues in region C), and no significant changes were identified in the other regions described. The analysis of the VPi-br crystal structure also revealed that the seven solvent-exposed basic residues introduced in this variant present a configuration similar to that observed for the homologous residues in MnP4. Their side-chains appear interacting with water molecules through hydrogen bonds or do not present any interaction, and no contacts are observed with surrounding residues (Fig 6B). Finally, the VPi-ss structure confirmed that the two cysteines introduced at positions 49 and 61 form an extra (fifth) disulfide bond next to the distal Ca²⁺ binding site (Fig 6C). This region did not show any additional change compared with the native enzyme even though the two new cysteines occupy positions adjacent to Asp48, Gly60, Asp62 and Ser64 involved, together with two water molecules, in Ca²⁺ coordination.