A variant of high biotechnological interest (called 2-1B) was obtained by directed evolution of the Pleurotus eryngii VP (versatile peroxidase) expressed in Saccharomyces cerevisiae [García-Ruiz, González-Pérez, Ruiz-Dueñas, Martínez and Alcalde (2012) Biochem. J. 441, 487–498]. 2-1B shows seven mutations in the mature protein that resulted in improved functional expression, activity and thermostability, along with a remarkable stronger alkaline stability (it retains 60% of the initial activity after 120 h of incubation at pH 9 compared with complete inactivation of the native enzyme after only 1 h). The latter is highly demanded for biorefinery applications. In the present study we investigate the structural basis behind the enhanced alkaline stabilization of this evolved enzyme. In order to do this, several VP variants containing one or several of the mutations present in 2-1B were expressed in Escherichia coli, and their alkaline stability and biochemical properties were determined. In addition, the crystal structures of 2-1B and one of the intermediate variants were solved and carefully analysed, and molecular dynamics simulations were carried out. We concluded that the introduction of three basic residues in VP (Lys-37, Arg-39 and Arg-330) led to new connections between haem and helix B (where the distal histidine residue is located), and formation of new electrostatic interactions, that avoided the hexa-co-ordination of the haem iron. These new structural determinants stabilized the haem and its environment, helping to maintain the structural enzyme integrity (with penta-co-ordinated haem iron) under alkaline conditions. Moreover, the reinforcement of the solvent-exposed area around Gln-305 in the proximal side, prompted by the Q202L mutation, further enhanced the stability.

INTRODUCTION

Lignin is the main reservoir of aromatic compounds in living organisms and a valuable feedstock for the sustainable production of chemicals and materials [1,2]. Owing to its aromatic and heterogeneous nature, lignin is highly resistant to chemical and biological degradation [3]. White-rot fungi produce ligninolytic peroxidases involved in lignin degradation, and they are the only organisms in Nature able to extensively mineralize lignin [4,5]. Ligninolytic peroxidases include three families: LiPs (lignin peroxidases) (EC 1.11.1.14), which oxidize high-redox-potential substrates through an exposed tryptophan radical formed by electron transfer to the haem [6]; MnPs (manganese peroxidases) (EC 1.11.1.13), which oxidize Mn2+ to Mn3+ at a specific manganese-binding site near one of the haem propionates [7,8]; and VPs (versatile peroxidases) (EC 1.11.1.16), which combine the catalytic properties of LiPs and MnPs (due to the simultaneous presence of the exposed tryptophan residue and the manganese-binding site, mentioned above) at the same time that they oxidize the typical substrates of GPs (generic peroxidases) (EC 1.11.1.7) at the main haem access channel [9,10].

The above enzymes have evolved to overcome the recalcitrant nature of lignin taking advantage of their high redox potential, unspecific oxidation mechanism and other catalytic properties. Their use in industrial processes that require the oxidation of lignin and other phenolic and non-phenolic aromatic compounds and dyes is therefore very promising [11]. In this way, ligninolytic peroxidases could be used in the deconstruction of the lignocellulosic biomass and production of biofuels, materials and chemicals in lignocellulose biorefineries, in bioremediation processes for the treatment of recalcitrant dye wastes, or in bleaching applications in paper pulp manufacture [2,12,13]. However, there are some drawbacks that prevent the industrial application of these enzymes as they are produced in Nature. Among them, it is possible to mention their relative low stability towards some conditions of pH, temperature or hydrogen peroxide concentration, in which industrial processes are often carried out (including the alkaline conditions often used in lignocellulose processing) [2].

2-1B is an enzyme variant obtained by directed evolution of the VP from Pleurotus eryngii in Saccharomyces cerevisiae [14]. During the evolutionary process, seven mutations were introduced in the sequence of mature VP. As a consequence, 2-1B showed enhanced expression levels, increased thermostability and alkaline stability, and improved oxidation of some substrates. In the present study, five VP variants containing one or several of the mutations present in 2-1B were designed, heterologously expressed, and their biochemical properties carefully evaluated, at the same time that the crystal structures of two of the most relevant variants were solved and used in molecular simulations. The final objective was to understand how the substitutions introduced by directed molecular evolution modulate the alkaline stability and catalytic properties of the enzyme, with the aim of using this information in future peroxidase engineering work.

MATERIALS AND METHODS

Chemicals

IPTG, DTT, EDTA, haemin, oxidized glutathione, VA (veratryl alcohol), manganese(II) sulfate, RB5 (Reactive Black 5), DMP (2,6-dimethoxyphenol), sodium tartrate and other chemicals were purchased from Sigma–Aldrich; urea and hydrogen peroxide were from Merck; and ABTS [2,2′-azinobis-(3-ethylbenzothiazoline-6-sulfonic acid)] was from Roche.

Directed mutagenesis

All VP variants except 2-1B (Table 1) were produced using the QuikChange™ Site-Directed Mutagenesis kit (Stratagene). Each mutation was introduced by PCR using the expression plasmid pFLAG1 (International Biotechnologies) containing the CDS (coding DNA sequence) of P. eryngii VPL (allelic variant VPL2; GenBank® accession number AF007222), named pFLAG1-VPL2, or the same plasmid containing already mutated CDSs as templates and two primers, a direct and a reverse primer designed complementary to opposite strands of the same DNA region containing the desired mutation. PCRs were carried out in an Eppendorf Mastercycler Pro S using 10 ng of template DNA, 250 μM of each dNTP, 125 ng of both direct and reverse primers, 2.5 units of Pfu Turbo AD polymerase (Stratagene) and the manufacturer'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. Clones harbouring mutations were transformed into Escherichia coli DH5α. One positive clone of each variant was selected, sequenced (PerkinElmer ABI Prism 377) and checked to confirm that the desired mutations had been properly introduced.

Table 1
Seven mutations present in 2-1B and their combination in the five intermediate variants analysed
VariantMutation
2-1B E37K/H39R/V160A/T184M/Q202L/D213A/G330R 
H39R H39R 
EG E37K/G330R 
EHG E37K/H39R/G330R 
EHGQ E37K/H39R/Q202L/G330R 
EHGT E37K/H39R/T184M/G330R 
VariantMutation
2-1B E37K/H39R/V160A/T184M/Q202L/D213A/G330R 
H39R H39R 
EG E37K/G330R 
EHG E37K/H39R/G330R 
EHGQ E37K/H39R/Q202L/G330R 
EHGT E37K/H39R/T184M/G330R 

Heterologous expression

Wild-type recombinant (hereinafter referred to as native) VP and its directed variants were expressed in E. coli W3110 after transformation with the corresponding plasmids. The gene encoding the 2-1B variant was obtained by digesting the plasmid used for S. cerevisiae expression [14], cloned into pFLAG1 (resulting in the pFLAG-VPL2-2-1B construct) and transformed into E. coli W3110 for expression.

Cells were grown in Terrific broth [15] at 37°C until reaching a D500 of ∼1 (∼3 h). Then protein expression was induced with 1 mM IPTG and cells were grown for a further 4 h. The apoenzyme was produced as inclusion bodies and was recovered in a 50 mM Tris/HCl (pH 8.0) solution containing 8 M urea, 1 mM EDTA and 1 mM DTT. The subsequent in vitro folding of the solubilized apoenzyme was carried out in a solution of 0.16 M urea, 20 μM haemin, 5 mM CaCl2, 0.1 mM DTT, 0.5 mM oxidized glutathione and 0.1 mg/ml protein concentration in 20 mM Tris/HCl buffer, at pH 9.5, at room temperature overnight. The refolded enzyme was purified by Resource-Q chromatography using a 0–0.3 M NaCl gradient (2 ml/min for 20 min) in 10 mM sodium tartrate (pH 5.5) containing 1 mM CaCl2. Finally, the purified enzyme was dialysed against 10 mM sodium tartrate (pH 5). The proteins showed Reinheitszahl values (Rz, A407/A280) of ∼4 confirming their high purity. Their UV–visible spectra in the 300–700 nm range confirmed that they were correctly folded [16]. Enzyme concentrations were determined from the Soret absorbance (ε407=150 mM−1 · cm−1) [17].

pH stability studies

Native VP and its directed variants (1 μM) were incubated at different pH values (from 2 to 9) in 0.1 M B&R (Britton–Robinson) buffer (0.1 M phosphoric acid, 0.1 M boric acid and 0.1 M acetic acid; pH adjusted with NaOH) [18] at 4°C for 120 h. At different times (1 min, 1 h, 4 h, 24 h and 120 h), aliquots were taken and their residual activity was evaluated 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 from samples incubated for 1 min at pH 5 was taken as reference (maximum activity) to calculate the percentage of residual activity. Native VP and those variants showing improved alkaline stability (2-1B, EHG and EHGQ) (8 μM) were incubated at pH 5 or pH 8 in 0.1 M B&R buffer for 120 h at 4°C, and their electronic absorption spectra (in the 300–800 nm range) were recorded using an Agilent 8453 diode array UV–visible spectrophotometer.

Kinetic constants

Oxidation of Mn2+ (Mn3+–tartrate complex ε238=6500 M−1 · cm−1) was estimated at pH 5; that of VA (veratraldehyde ε310=9300 M−1 · cm−1) at pH 3; and those of RB5 (ε598=30000 M−1 · cm−1), ABTS (cation radical ε436 29300 M−1 · cm−1) and DMP (coerulignone dimeric product ε469=55000 M−1 · cm−1) at pH 3.5. All enzymatic activities were measured as initial velocities taking linear increments of absorbance due to the appearance of the reaction product (decreases in the case of RB5). Reactions were performed in 0.1 M tartrate buffer, at 25°C, in the presence of 0.1 mM H2O2 using a variable-wavelength Shimadzu UV-1800 spectrophotometer.

Steady-state kinetic constants were calculated from the estimated oxidation of increasing substrate concentrations until enzyme saturation was observed. Values and standard errors for affinity constant (Michaelis constant, Km) and maximal enzyme turnover (catalytic constant, kcat) were obtained fitting the experimental measurements to the Michaelis–Menten model using SigmaPlot 12.0 software. Fitting of these constants to the normalized equation v=(kcat/Km)[S]/(1+[S]/Km), where [S] is the substrate concentration, yielded the catalytic efficiency values (kcat/Km) with their corresponding standard errors.

Crystallization, data collection and crystal structure determination

Crystallization trials were carried out by the sitting-drop vapour-diffusion method, using 96-well MRC2 plates with 50 μl of 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 of 2-1B were obtained in 0.1 M Tris/HCl (pH 7.8), 5% PEG3350, 1.25 M (NH4)2SO4 and 7.5% glycerol; and cryoprotected raising the amount of glycerol up to 30%. Crystals of EHG were obtained in 0.1 M sodium acetate (pH 4.5), 1.26 M (NH4)2SO4 and 0.2 M NaCl; and cryoprotected with 0.8 M Li2SO4.

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 the X06DA beamline at the Swiss Light Source (Villigen, Switzerland) using a wavelength of 1.0000 Å (1 Å=0.1 nm). Diffraction data were indexed, integrated, merged and scaled using the program XDS [19]. Data collection statistics are shown in Table 2. 

Table 2
Crystallographic data collection and refinement statistics of VP EHG and VP 2-1B

Values in parentheses correspond to the highest resolution layer.

EHG2-1B
Data collection 
 Space group P2121P212121 
 Cell constants a=55.2, b=104.3, c=76.7 Å a=55.2, b=106.3, c=107.9 Å 
 Resolution range (Å) 50.00–1.70 (1.89–1.70) 50.00–2.00 (2.12–2.00) 
 Number of total reflections 646657 577925 
 Number of unique reflections 49391 43670 
Rmerge (%) 19.2 (145.6) 32.5 (171.0) 
 Completeness (%) 99.6 (97.6) 99.2 (96.9) 
 <I/σ(I)> 11.9 (1.9) 8.7 (1.5) 
 CC(1/2) 99.7 (70.7) 99.2 (63.2) 
 Solvent content (%)/Matthews coefficient 55.3/2.75 37.5/1.97 
 Subunits per asymmetric unit 
 Wilson B factor (Å225.0 27.0 
Refinement 
 Resolution range 50.0–1.70 Å 50.0–2.00 Å 
 Working reflections 49358 43615 
Rwork/Rfree 19.7/22.8% 18.6/23.6% 
 Protein atoms (non-hydrogen) 2423 4764 
 Haem group 
 Ca2+ 
 Water molecules 362 266 
 SO4 ions 
 Mean B factors (Å2
  Protein atoms (non H) 20.4 20.7 
  Heme group 16.2 16.2 
  Ca2+ 14.5 14.8 
  Water molecules 29.1 24.2 
  SO4 ions 43.3 25.0 
 Deviations from ideality 
  RMSD bond lengths 0.006 Å 0.008 Å 
  RMSD angles 0.886° 0.928° 
 Ramachandran plot statistics 
  Preferred (%) 98.17 98.44 
  Allowed (%) 1.83 1.41 
  Outliers (%) 0.00 0.16 
  PDB code 5FNE 5FNB 
EHG2-1B
Data collection 
 Space group P2121P212121 
 Cell constants a=55.2, b=104.3, c=76.7 Å a=55.2, b=106.3, c=107.9 Å 
 Resolution range (Å) 50.00–1.70 (1.89–1.70) 50.00–2.00 (2.12–2.00) 
 Number of total reflections 646657 577925 
 Number of unique reflections 49391 43670 
Rmerge (%) 19.2 (145.6) 32.5 (171.0) 
 Completeness (%) 99.6 (97.6) 99.2 (96.9) 
 <I/σ(I)> 11.9 (1.9) 8.7 (1.5) 
 CC(1/2) 99.7 (70.7) 99.2 (63.2) 
 Solvent content (%)/Matthews coefficient 55.3/2.75 37.5/1.97 
 Subunits per asymmetric unit 
 Wilson B factor (Å225.0 27.0 
Refinement 
 Resolution range 50.0–1.70 Å 50.0–2.00 Å 
 Working reflections 49358 43615 
Rwork/Rfree 19.7/22.8% 18.6/23.6% 
 Protein atoms (non-hydrogen) 2423 4764 
 Haem group 
 Ca2+ 
 Water molecules 362 266 
 SO4 ions 
 Mean B factors (Å2
  Protein atoms (non H) 20.4 20.7 
  Heme group 16.2 16.2 
  Ca2+ 14.5 14.8 
  Water molecules 29.1 24.2 
  SO4 ions 43.3 25.0 
 Deviations from ideality 
  RMSD bond lengths 0.006 Å 0.008 Å 
  RMSD angles 0.886° 0.928° 
 Ramachandran plot statistics 
  Preferred (%) 98.17 98.44 
  Allowed (%) 1.83 1.41 
  Outliers (%) 0.00 0.16 
  PDB code 5FNE 5FNB 

The structures of the two VP variants were solved by molecular replacement using the crystal structure of P. eryngii VPL (PDB code 3FMU) as the search model and the program PHASER implemented in the PHENIX package [20]. The final models were obtained by consecutive rounds of refinement, performed with the PHENIX package; followed by manual model building, performed with Coot [21] using σA weighted 2FoFc and FoFc 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 Rfree value throughout the refinement process. The structures were validated using MolProbity [22,23]. Refinement and final model statistics are summarized in Table 2. The co-ordinates and structural factors have been deposited with the PDB. All Figures were produced with PyMOL (http://www.pymol.org/).

Molecular dynamics

Structural and dynamic differences between native VP and its 2-1B variant were studied by MD simulations. Hydrogen addition, ligand optimization, completion of the two terminal residues and protonation state adjustment of ionizable residues (glutamate, histidine and aspartate) in both crystal structures were performed with the protein preparation wizard [24] at pH 8. MD simulations were then produced using the Desmond software and its Maestro interface from Schrodinger. Enzymes were solvated in an orthorhombic box with at least 10 Å from the protein surface, followed by NaCl addition for an ionic force of 0.15 M. The default relaxation protocol in Desmond was performed followed by a 20 ns trajectory production with the OPLS-2005 [25] force field at 300K. The NPT ensemble with Nose–Hoover thermostat [26] and a Martyna–Tobias–Klein barostat [27] were employed.

RESULTS

pH stability of the different variants

VP 2-1B (containing E37K/H39R/V160A/T184M/Q202L/D213A/G330R mutations) expressed in E. coli showed improved alkaline stability (Figure 1D) compared with native VP (Figure 1A). After 24 and 120 h of incubation at pH 8, it retained 82% and 78% of residual activity respectively, whereas native VP was fully inactivated. Moreover, after 1 h of incubation at pH 9, native VP retained only 5% of the initial activity, whereas 2-1B maintained 36% of activity. The differences in stability between both enzymes at neutral and most acidic pH conditions (in the pH 3.5–7 range) were minimal, and 2-1B was less stable at the lowest pH value (pH 3) assayed.

pH stability of native VP, and EHG, EHGQ and 2-1B mutated variants

Figure 1
pH stability of native VP, and EHG, EHGQ and 2-1B mutated variants

Residual activities of VP (A), and the EHG (B), EHGQ (C) and 2-1B (D) variants, after 1 h (dark grey bars), 4 h (white bars), 24 h (black bars) and 120 h (light grey bars) of incubation in 0.1 M B&R buffer at different pH values (3–9) and 4°C. Residual activity measured as ABTS (2 mM) oxidation in 0.1 M sodium tartrate (pH 3.5) using 0.01 μM enzyme and 0.1 mM H2O2. Results are means±95% confidence limits for replicate assays. See Table 1 for the mutations present in each of the variants.

Figure 1
pH stability of native VP, and EHG, EHGQ and 2-1B mutated variants

Residual activities of VP (A), and the EHG (B), EHGQ (C) and 2-1B (D) variants, after 1 h (dark grey bars), 4 h (white bars), 24 h (black bars) and 120 h (light grey bars) of incubation in 0.1 M B&R buffer at different pH values (3–9) and 4°C. Residual activity measured as ABTS (2 mM) oxidation in 0.1 M sodium tartrate (pH 3.5) using 0.01 μM enzyme and 0.1 mM H2O2. Results are means±95% confidence limits for replicate assays. See Table 1 for the mutations present in each of the variants.

To analyse the effect of each of the individual mutations in 2-1B, and the effect of combining them, several single and multiple VP variants were designed (Table 1), expressed in E. coli and characterized. The H39R variant showed high instability at acidic and alkaline pH values compared with native VP (Figure 2A). Native VP maintained ∼90% of the initial activity in the pH 4–7 range after 120 h of incubation, whereas H39R only showed 41%, 68% and 21% activity at pH 4, 5 and 6 respectively (and no activity at pH 7). The following VP variant analysed, EG containing mutations E37K/G330R, exhibited similar pH stability to native VP and no improvement in the alkaline range was observed (Figure 2B).

pH stability of H39R, EG and EHGT VP variants

Figure 2
pH stability of H39R, EG and EHGT VP variants

Residual activities of H39R (A), EG (B) and EHGT (C) after incubation in 0.1 M B&R buffer at different pH values (3–9) and 4°C for 1 h (dark grey bars), 4 h (white bars), 24 h (black bars) and 120 h (light grey bars). Residual activity measured as ABTS (2 mM) oxidation in 0.1 M sodium tartrate (pH 3.5) using 0.01 μM enzyme and 0.1 mM H2O2. Results are means±95% confidence limits for replicate assays. See Table 1 for the mutations present in each of the variants.

Figure 2
pH stability of H39R, EG and EHGT VP variants

Residual activities of H39R (A), EG (B) and EHGT (C) after incubation in 0.1 M B&R buffer at different pH values (3–9) and 4°C for 1 h (dark grey bars), 4 h (white bars), 24 h (black bars) and 120 h (light grey bars). Residual activity measured as ABTS (2 mM) oxidation in 0.1 M sodium tartrate (pH 3.5) using 0.01 μM enzyme and 0.1 mM H2O2. Results are means±95% confidence limits for replicate assays. See Table 1 for the mutations present in each of the variants.

Surprisingly, the EHG variant, which combined the mutations of the two variants analysed previously (E37K/H39R/G330R), displayed an increase in the alkaline stability compared with native VP (Figure 1B). This triple variant maintained 55% of the initial activity after 120 h of incubation at pH 8, whereas native VP not only showed no activity under these conditions, but was already inactivated after 24 h of incubation (Figure 1A). In a similar way, after 1 h of incubation at pH 9, EHG exhibited 46% of the initial activity, whereas native VP only retained 5%. Moreover, this variant maintained a similar stability to that of native VP at acidic and neutral pH values. However, the values of pH stability observed for EHG were not as high as those obtained for 2-1B, which retained 78% of activity after 120 h of incubation at pH 8 (Figure 1D).

The next VP variant analysed was EHGQ (E37K/H39R/Q202L/G330R), which harboured the three mutations of EHG plus the Q202L mutation, also found in 2-1B. EHGQ showed 67% of residual activity after 120 h of incubation at pH 8 (Figure 1C), a value higher than that of EHG (55% of activity). EHGQ also retained a low percentage of the initial activity (24%) after 1 h of incubation at pH 9 compared with native VP, which was completely inactivated. Therefore, it was determined that the Q202L substitution also contributed to the enhancement of the pH stability in 2-1B. As described previously for EHG, its stability at neutral and acidic pH was similar to that of the native VP.

The last variant tested, EHGT (E37K/H39R/T184M/G330R), harboured the three mutations introduced in EHG plus the T184M mutation, also found in 2-1B. After 120 h at pH 8, EHGT displayed 30% of the initial activity (Figure 2C) compared with 55% of activity retained by EHG. In consequence, we concluded that the T184M substitution did not contribute to the alkaline stability of 2-1B.

pH effect on the UV–visible spectra

The electronic absorption spectra of native VP and its variants were recorded after incubation at alkaline pH 8, and at pH 5 at which all of them are stable, to check the eventual modification of the haem interactions with its protein environment. As expected, all of the spectra at pH 5 exhibited a Soret maximum at 407 nm, maxima at 505 nm and 637 nm corresponding to ligand-to-metal charge transfer bands (CT2 and CT1 respectively), and the β band at 541 nm (Figure 3, black lines), which correspond to the typical spectrum of the ferric enzyme with high-spin penta-co-ordinated haem iron.

Electronic absorption spectra of native VP and variants after treatments at pH 5 and pH 8

Figure 3
Electronic absorption spectra of native VP and variants after treatments at pH 5 and pH 8

UV–visible spectra of native VP (A), and the EHG (B), EHGQ (C) and 2-1B (D) variants incubated for 120 h in 0.1 M B&R buffer at pH 5 (black) and pH 8 (grey) at 4°C. Amplified (×5) 450–800 nm regions are shown. Main maxima are indicated.

Figure 3
Electronic absorption spectra of native VP and variants after treatments at pH 5 and pH 8

UV–visible spectra of native VP (A), and the EHG (B), EHGQ (C) and 2-1B (D) variants incubated for 120 h in 0.1 M B&R buffer at pH 5 (black) and pH 8 (grey) at 4°C. Amplified (×5) 450–800 nm regions are shown. Main maxima are indicated.

However, the spectrum of native VP after incubation at pH 8 suffered important changes (Figure 3A, grey lines). The Soret maximum was red-shifted to 413 nm and showed less intensity, the CT bands at 505 nm and 637 nm disappeared and a new maximum rose at 538 nm (β band) together with a shoulder at 559 nm (α band). In contrast, 2-1B mostly preserved the spectrum when incubated at pH 8 (Figure 3D, grey lines), maintaining the Soret maximum at 407 nm, as well as the CT1, CT2 and β bands at 637, 505 and 541 nm respectively. In the same way, the EHG and EHGQ variants also preserved the ferric high-spin haem spectrum with all maxima unaffected (Figures 3B and 3C, grey lines, respectively).

Catalytic properties

Different substrates are oxidized at specific VP sites: (i) high-redox-potential VA and RB5 are oxidized at the catalytic tryptophan residue; (ii) low-redox-potential ABTS and DMP are oxidized both at the catalytic tryptophan residue (high-efficiency site) and at the haem channel in direct contact with the haem edge (low-efficiency site); and (iii) Mn2+ is oxidized at the Mn2+-binding site, in direct contact with a haem propionate. The steady-state kinetic constants for the oxidation of the above five substrates by native VP and four variants were measured (Table 3).

Table 3
Kinetic constants, kcat (s−1), Km (μM) and kcat/Km (s−1 · mM−1), for oxidation of VA, RB5, Mn2+, ABTS and DMP by native VP, 2-1B and three intermediate variants (see Table 1)

Reactions were carried out at 25°C in 0.1 M sodium tartrate (pH 3 for VA, pH 5 for Mn2+ and pH 3.5 for RB5, ABTS and DMP). Results are means±95% confidence limits of replicate assays. ABTS and DMP oxidation showed biphasic kinetics enabling calculation of two sets of constants (for high- and low-efficiency sites). ND, not determined because the kinetic curve of the high-efficiency site overlapped with that of the low-efficiency site.

VPEHGEHGQEHGT2-1B
VA kcat 5.8±0.1 5.2±0.1 4.4±0.04 5.9±0.1 3.8±0.1 
 Km 2600±190 2110±150 1780±60 5910±190 2710±270 
 kcat/Km 2.2±0.1 2.5±0.1 2.5±0.1 1.0±0.1 1.4±0.1 
RB5 kcat 5.5±0.3 7.8±0.4 6.7±0.3 5.1±0.1 6.7±0.3 
 Km 3.4±0.3 4.3±0.5 3.5±0.3 2.4±0.1 7.6±0.6 
 kcat/Km 1.6±0.1 1.8±0.1 1.1±0.1 2.2±0.1 0.8±0.1 
ABTS (high efficiency) kcat 8.1±0.2 8.3±0.3 7.2±0.3 39±7 ND 
 Km 3.0±0.2 2.6±0.2 2.3±0.2 13±3 ND 
 kcat/Km 2700±140 3100±180 3200±150 3050±160 ND 
ABTS (low efficiency) kcat 208±6 89±2 90±2 117±2 234±8 
 Km 1020±70 400±40 280±20 61±3 98±9 
 kcat/Km 204±10 219±15 320±20 1920±70 2370±150 
DMP (high efficiency) kcat 7.1±0.1 2.9±0.1 4.1±0.1 6.3±0.3 2.9±0.4 
 Km 38±4 17±2 16±3 11±2 27±7 
 kcat/Km 186±16 168±18 250±38 559±80 112±18 
DMP (low efficiency) kcat 30±1 40±4 56±4 46±2 91±2 
 Km 10500±400 34500±6050 34100±5290 3070±420 7470±380 
 kcat/Km 2.8±0.1 1.1±0.1 1.6±0.1 15±2 12±1 
Mn2+ kcat 211±4 208±2 188±2 160±4 143±3 
 Km 130±11 7020±190 6790±200 5590±371 5710±350 
 kcat/Km 1640±130 30±1 28±1 29±1 25±1 
VPEHGEHGQEHGT2-1B
VA kcat 5.8±0.1 5.2±0.1 4.4±0.04 5.9±0.1 3.8±0.1 
 Km 2600±190 2110±150 1780±60 5910±190 2710±270 
 kcat/Km 2.2±0.1 2.5±0.1 2.5±0.1 1.0±0.1 1.4±0.1 
RB5 kcat 5.5±0.3 7.8±0.4 6.7±0.3 5.1±0.1 6.7±0.3 
 Km 3.4±0.3 4.3±0.5 3.5±0.3 2.4±0.1 7.6±0.6 
 kcat/Km 1.6±0.1 1.8±0.1 1.1±0.1 2.2±0.1 0.8±0.1 
ABTS (high efficiency) kcat 8.1±0.2 8.3±0.3 7.2±0.3 39±7 ND 
 Km 3.0±0.2 2.6±0.2 2.3±0.2 13±3 ND 
 kcat/Km 2700±140 3100±180 3200±150 3050±160 ND 
ABTS (low efficiency) kcat 208±6 89±2 90±2 117±2 234±8 
 Km 1020±70 400±40 280±20 61±3 98±9 
 kcat/Km 204±10 219±15 320±20 1920±70 2370±150 
DMP (high efficiency) kcat 7.1±0.1 2.9±0.1 4.1±0.1 6.3±0.3 2.9±0.4 
 Km 38±4 17±2 16±3 11±2 27±7 
 kcat/Km 186±16 168±18 250±38 559±80 112±18 
DMP (low efficiency) kcat 30±1 40±4 56±4 46±2 91±2 
 Km 10500±400 34500±6050 34100±5290 3070±420 7470±380 
 kcat/Km 2.8±0.1 1.1±0.1 1.6±0.1 15±2 12±1 
Mn2+ kcat 211±4 208±2 188±2 160±4 143±3 
 Km 130±11 7020±190 6790±200 5590±371 5710±350 
 kcat/Km 1640±130 30±1 28±1 29±1 25±1 

In general, the oxidation of substrates at the catalytic tryptophan residue was not affected by the substitutions introduced in 2-1B, since the kinetic constants for the oxidation of VA, RB5, and ABTS and DMP (in the high-efficiency site) were mostly unaltered in this and the other variants. The only exception was EHGT which exhibited a 3-fold higher catalytic efficiency (559 s−1 · mM−1) for DMP oxidation, compared with that of native VP (186 s−1 · mM−1), due to a decrease in the affinity constant (Km).

Mn2+ oxidation was impaired in 2-1B and the rest of analysed variants since their efficiency values were decreased ∼60-fold, changing from 1640 s−1 · mM−1 in native VP to 25–30 s−1 · mM−1 in the variants. The decline in oxidation of Mn2+ was mainly due to loss of substrate affinity since the Km values for Mn2+ were substantially (40–50-fold) higher than for the native VP enzyme (Table 3).

Interestingly, the oxidation of ABTS and DMP at the low-efficiency site was improved in 2-1B and EHGT. The efficiency for the oxidation of ABTS showed a 9–12-fold increase in both variants, mainly due to a reduction in the Km values. In a similar way, the efficiency in the oxidation of DMP exhibited a 5-fold increase. In EHGT, this improvement was mainly due to the decrease in the Km. However, in 2-1B, a decrease in the Km value combined with an increase in the kcat was observed.

In contrast with the results obtained for EHGT, EHG showed no improvement in ABTS and DMP oxidation. Since the only difference between these two variants is the T184M mutation present in EGHT, it was concluded that this substitution is involved in the improved oxidation of low-redox-potential substrates by EGHT (and probably by 2-1B).

Analysis of crystal structures

The crystal structures of 2-1B and EHG were obtained, analysed and compared with that of native VP (PDB code 2BOQ) [34] to explain, from a structural point of view, both the enhanced alkaline stability and the modified catalytic properties of these variants. The simultaneous presence of H39R, E37K and G330R mutations was a determinant for the higher alkaline stability found in 2-1B. Residues 37 and 39 are partially solvent-exposed, close to the haem (at 7.1 and 3.3 Å respectively in the native VP structure) (Figure 4A) and close to the Mn2+-binding site (formed by Glu-36, Glu-40 and Asp-175). They belong to helix B, which is located directly above the haem and harbours the distal histidine (His-47) and the distal arginine (Arg-43) that are involved in catalysis. Moreover, the distal histidine is involved in the hexa-co-ordination of the haem iron observed at alkaline pH. The new Arg-39 in EHG and 2-1B (Figures 4C, 4D and 4E) showed a different orientation of its side chain relative to the position of His-39 in native VP (Figure 4B). The new orientation enabled the formation of a new hydrogen bond with the distal propionate of the haem, with this interaction being absent from native VP. Instead, in VP, His-39 formed two hydrogen bonds with Cys-34 and Pro-190, which were lost in the improved variants.

E37K and H39R mutations in 2-1B and EHG crystal structures

Figure 4
E37K and H39R mutations in 2-1B and EHG crystal structures

Crystal structure of 2-1B (A) with details of the region in front of the haem propionates in native VP (B) and the EHG (C) and 2-1B (D) variants showing the haem, the mutated residues at positions 37 and 39, and other surrounding residues [in CPK (Corey–Pauling–Koltun) colours]; and a stereo view of this region in EHG (E) showing the (2FoFc) electron density map, contoured at 1.0 σ, of the same amino acid residues depicted in (C). 2-1B showed the same electron density (not shown). Hydrogen bonds are shown as green dashed lines, Ca2+ ions are shown as green spheres and secondary structure is shown as a coloured cartoon.

Figure 4
E37K and H39R mutations in 2-1B and EHG crystal structures

Crystal structure of 2-1B (A) with details of the region in front of the haem propionates in native VP (B) and the EHG (C) and 2-1B (D) variants showing the haem, the mutated residues at positions 37 and 39, and other surrounding residues [in CPK (Corey–Pauling–Koltun) colours]; and a stereo view of this region in EHG (E) showing the (2FoFc) electron density map, contoured at 1.0 σ, of the same amino acid residues depicted in (C). 2-1B showed the same electron density (not shown). Hydrogen bonds are shown as green dashed lines, Ca2+ ions are shown as green spheres and secondary structure is shown as a coloured cartoon.

The E37K substitution changed an acidic residue for a basic one. The new lysine residue forms electrostatic interactions with Glu-40 in 2-1B (Figure 4D) and with Glu-36 and Glu-83 in EHG (Figure 4C). In contrast, the side chain of Glu-37 in VP does not make this type of interaction (Figure 4B). On the other hand, Glu-37 is close (<7 Å) to several acidic residues (such as Glu-40 and Glu-36 on helix B, Glu-83 on helix C, Glu-26 on helix A, and Asp-30 and Asp-175 on two loops). Therefore the introduction of the new Lys-37 in 2-1B and EHG variants contributed to reduce the putative destabilizing acid–acid interactions in the region where this mutation is located.

G330R was also involved in the pH-stability improvement of 2-1B. This is the penultimate residue in the VP C-terminal tail, and its position could not be determined in the crystals of native VP and its mutated variants owing to the high mobility of this area. For this reason, we could not verify in which way this mutation contributes to enhance the stability. Nonetheless, it is interesting to note that residue 330 is close to residues 37 and 39, all of them around the Mn2+-binding site.

Q202L, at the bottom of the molecule (Figure 5A) where the mutations V160A, T184M and D213A are also located (Figure 5D), was the last mutation identified as stabilizing VP under alkaline conditions. The introduction of Leu-202 in 2-1B led to the disappearance of two hydrogen bonds present in native VP, one with Ala-235 and other with Glu-304 (Figures 5B and 5C). As a consequence, the position of the side chain of some surrounding residues changed in the 2-1B variant, especially that of Gln-305 forming three new interactions with Gln-245, Cys-307 and Thr-310 in 2-1B. These interactions reinforced and joined two solvent-exposed regions of the protein: (i) the helix I; and (ii) the loop between a 310 helix (from Leu-300 to Asp-302) and the C-terminal tail. None of these interactions appeared in native VP.

Q202L mutation in 2-1B crystal structure

Figure 5
Q202L mutation in 2-1B crystal structure

Crystal structure of 2-1B (A) with a detail of the region around residue 202 in native VP (B) and the 2-1B (C) variant; and a stereo view of this region in 2-1B (D) showing the (2Fo−Fc) electron density map, contoured at 1.0 σ, of the mutated residues at positions 160, 184, 202 and 213. Analysed residues are shown as CPK (Corey–Pauling–Koltun) sticks, hydrogen bonds are shown as green dashed lines, Ca2+ ions are shown as green spheres and secondary structure is shown as a coloured cartoon.

Figure 5
Q202L mutation in 2-1B crystal structure

Crystal structure of 2-1B (A) with a detail of the region around residue 202 in native VP (B) and the 2-1B (C) variant; and a stereo view of this region in 2-1B (D) showing the (2Fo−Fc) electron density map, contoured at 1.0 σ, of the mutated residues at positions 160, 184, 202 and 213. Analysed residues are shown as CPK (Corey–Pauling–Koltun) sticks, hydrogen bonds are shown as green dashed lines, Ca2+ ions are shown as green spheres and secondary structure is shown as a coloured cartoon.

Regarding the structural Ca2+ ions, described as essential for pH stability, both of them appeared in the 2-1B and EHG structures co-ordinated in the same way as in the native VP.

Finally, the T184M mutation was analysed in the 2-1B structure since it was involved in the enhanced oxidation of ABTS and DMP at the low-efficiency site. Residue 184 is solvent-exposed and located below the Mn2+-binding site. No significant differences were observed between native VP and 2-1B in either the position of the surrounding residues or the interactions between them.

Molecular dynamics

Taking advantage of the solved crystal structures, the dynamics of native VP and 2-1B in solution were studied by MD simulations at pH 8. Since alkaline inactivation has been related to the formation of a hexa-co-ordinated iron complex with the distal histidine residue, we monitored the evolution of this distance aiming at possible correlations with the stability of the enzyme. Figure 6(A) shows the iron to His-47 distance along a 20 ns MD for both enzymes. Although this distance is similar in both crystals, the addition of thermal fluctuations at room temperature reveals a different behaviour, with a higher distance (∼1 Å difference on average) for the 2-1B variant. Such distance elongation is the result of a new hydrogen bond between Arg-43 and His-47 (Figure 6C), which is not formed in native VP because of the greater distance between these two residues (Figures 6B and 6D). Overall, inspection of the simulated trajectory indicates a slight shift (along with a higher mobility) in the Arg-43 position, facilitating its interaction with His-47. This deviation towards the histidine residue is mainly due to the electrostatic repulsion produced by the introduction of three new positive residues in the propionate region (E37K, H39R and G330R). In addition, the H39R mutant introduces a stronger interaction with Glu-40, weakening the contact of this glutamic acid with Arg-43. In native VP, Arg-43 interacts mostly with Glu-40, keeping it away from His-47 and allowing a closer distance of the distal histidine with the haem iron, facilitating the inactivation of the enzyme.

Molecular dynamics simulations

Figure 6
Molecular dynamics simulations

Distance profiles between the haem iron and the distal His-47 (A) and between the distal His-47 and Arg-43 (B); and representative snapshots, with the average distances, extracted from the MD simulations of 2-1B (C) and native VP (D). All distances are in Å.

Figure 6
Molecular dynamics simulations

Distance profiles between the haem iron and the distal His-47 (A) and between the distal His-47 and Arg-43 (B); and representative snapshots, with the average distances, extracted from the MD simulations of 2-1B (C) and native VP (D). All distances are in Å.

DISCUSSION

The 2-1B variant of P. eryngii VP contains seven mutations, introduced during the evolution process in S. cerevisiae, that conferred, among other properties, significantly improved alkaline stability, thermostability and enhanced activity on some low-redox-potential substrates [14]. In the present study, we identified some of the structural basis of the alkaline stabilization of this VP variant, by analysing the stability at basic pH and catalytic properties of a series of (single and multiple) variants expressed in E. coli.

First, we confirmed that 2-1B expressed in E. coli retains the improved properties in terms of alkaline stability and activity at acid pH on low-redox-potential compounds observed when expressed in S. cerevisiae. Although the mutations present in 2-1B also contributed to improve the enzyme thermostability when expressed in this eukaryotic host, no differences between 2-1B and native VP were observed when both enzymes were produced in E. coli (results not shown). These results suggest that additional factors contribute to the thermostability improvement in the evolved variant secreted by S. cerevisiae. Unlike what was observed for thermostability, 2-1B produced in E. coli showed a significantly increased alkaline stability (retaining 80% and 20% activity after 120 h at pH 8 and pH 9 respectively) compared with native VP (which was completely inactivated), although this improvement was not as high as what was found when the variant was expressed in S. cerevisiae (100% and 60% activity at pH 8 and pH 9 respectively) [14]. These differences in both thermostability and alkaline stability are likely to be due to the lack of glycosylation of enzymes expressed in E. coli, as reported for temperature stability of Phanerochaete chrysosporium ligninolytic peroxidases [28]. In general, glycosylation has been described to give robustness and protection against enzyme inactivation [29,30] and could therefore improve the thermostability of yeast-expressed 2-1B and reinforce the alkaline-stabilizing effect of the mutations introduced in this evolved variant.

Explaining the alkaline stability

During alkaline inactivation, ligninolytic peroxidases lose the structural Ca2+ ions responsible for maintaining the architecture of the haem environment. This loss leads to a relaxation of the protein structure and hexa-co-ordination of the haem involving the distal histidine residue [31]. These changes are reflected in the electronic absorption spectrum, which is indicative of the high- or low-spin states of the haem iron and its co-ordination state [32]. The P. eryngii native VP was inactivated in 24 h upon alkaline incubation (pH 8), showing the spectrum of a low-spin hexa-co-ordinated haem iron. This is characterized by the red-shifted Soret band and the appearance of the α and β bands. Similar changes have been reported for Bjerkandera adusta VP [33] and other ligninolytic peroxidases, such as LiP [31] and MnP [32] from P. chrysosporium. In contrast, the EHG variant showed improved alkaline stability and conserved the UV–visible spectrum (including the Soret maximum and the CT1 and CT2 bands) unchanged during incubation at pH 8. This spectrum is indicative of a peroxidase with a high-spin ferric state of the haem iron, which is completely functional [34,35]. From these data, it can be inferred that the three new basic residues present in EHG (Lys-37, Arg-39 and Arg-330) reinforced the haem environment enabling the maintenance of penta-co-ordinated haem iron at pH 8.

The stability improvement achieved with the simultaneous introduction of these basic residues could be rationalized by comparative structural analysis and MD simulations of the EHG, 2-1B and native VP crystal structures. The new Arg-39 appeared to be hydrogen-bonded to the haem-distal propionate in the molecular structure of the two variants. The connections of the haem with the apoprotein have been described as being essential for stability [36,37] and engineering new connections with the haem has been proposed as a strategy to increase stability [38]. The haem in ligninolytic peroxidases is non-covalently bound and shows no polar interactions with the apoprotein, with the exception of those that the propionates make. So, the new connection made by Arg-39 seems to be important for the stability achieved, as it anchors the haem to the protein (helix B) and contributes to fix the relative position of His-47 with respect to the iron ion.

On the other hand, the new Lys-37 forms new electrostatic interactions. These interactions have been positively correlated with stability in different families of proteins [3942]. Those of Lys-37 with Glu-36 and Glu-83 in EHG and with Glu-40 in 2-1B reinforced the structure of helix B, located directly above the haem and harbouring distal histidine (His-47) and arginine (Arg-43) residues, which are essential in the reaction of the enzyme with H2O2 [43] and in the stabilization of the oxidized forms of the haem iron [44,45]. Moreover, as in the case of H39R mutation, fixing the helix B position reduces the possibilities of His-47 to approach the haem and generate the inactive form with hexa-co-ordinated iron. Helix B also contains Asp-48, one of the residues co-ordinating the distal calcium, which is required for pH stability. In addition, a high number of acidic residues were observed around residue 37. At alkaline pH 8, these residues would be deprotonated. In consequence, a high negative charge would characterize this region. The introduction of a basic residue (Lys-37) would reduce the negative charge avoiding detrimental interactions and reinforcing the structure of VP at alkaline pH. At acidic pH, however, the acidic residues will be protonated and no accumulation of negative charge would be produced, explaining the lack of acidic stability improvement of EHG and 2-1B.

The G330R mutation also contributed to the improved stability found in 2-1B and EHG. Although the role played by Arg-330 could not be studied in the crystal structures, it is likely that this residue is located in the same area as residues 37 and 39 and forms new interactions stabilizing the surrounding region containing a high number of acidic residues.

It was remarkable that the stabilizing effect produced by Lys-37, Arg-39 and Arg-330 was only detected when they appeared together, since the H39R variant was unstable and the E37K/G330R variant showed a similar stability to that of the native VP. The sum of the new interactions formed by these residues results in the stabilization of VP, but, individually, they make the protein unstable due to the vulnerability of the region where they are (near the haem and the Mn2+-binding site), being a clear effect of positive epistasis unveiled by directed evolution [14]. According to the structural analysis described above, this effect is caused by the three basic residues simultaneously contributing to stabilize the relative position of the distal histidine. The MD simulations performed under alkaline conditions were quite enlightening in this regard. They revealed that the electrostatic repulsion, formed as a consequence of the introduction of Lys-37, Arg-39 and Arg-330, along with other interactions in which these and other surrounding residues are involved, is able to maintain the distal histidine for longer in a position far enough from the porphyrin to keep the haem iron in a penta-co-ordinate active state.

The EHGQ variant showed higher stability than EHG and almost achieved the alkaline stability found in 2-1B, so it was concluded that: (i) the Q202L mutation also contributed to the enzyme stabilization; and (ii) the four mutations introduced in EHGQ are responsible for the most part of the enhancement towards alkaline stability found in 2-1B. From the analysis of the crystal structure of 2-1B, it was observed that the Q202L substitution promoted the reorganization of Gln-305 and the formation of new interactions in the area which could explain the improvement observed due to this mutation. Although residue 202 is in the core of the protein, the residues affected by the new interactions are solvent-exposed. The solvent-exposed regions are very sensitive to pH changes, so the reinforcement of these areas could be determinantal to strengthening the structure and avoiding the alkaline inactivation at pH 8. Residue 202 would be an example of a residue located far from the haem and the rest of catalytic sites that is able to increase the stability of the enzyme.

Explaining the modified catalytic properties

Regarding the catalytic properties of 2-1B, it was found that this variant is more efficient than the native enzyme oxidizing low-redox-potential substrates, mainly due to an increased affinity (a decreased Km) in the low-efficiency site. The same behaviour was reported previously for 2-1B expressed in S. cerevisiae [14]. With the data in hand, this fact was attributed to the presence of the T184M mutation. It is possible to speculate that the substitution introduced could help with a better substrate accommodation at the main haem-access channel. However, the absence of structural differences in the environment of the residue at position 184 after comparative structural analysis of the native enzyme and 2-1B has not allowed us to give a rational explanation for the effect produced by this mutation on the catalytic activity.

The simultaneous worsening in manganese oxidation is explained by the introduction of Lys-37 and Arg-39 [14]. These two residues could compromise the Mn2+ binding by forming electrostatic interactions with Glu-36 and Glu-40 (two of the three residues involved in manganese-co-ordination together with Asp-175) in EHG and 2-1B respectively. Moreover, it is also probable that the new interaction established between the haem internal propionate and the apoprotein in 2-1B and EHG affects the Mn2+ binding and oxidation, considering that this propionate is involved in these two events [46]. Interestingly, the kinetic constants for Mn2+ oxidation in 2-1B and the rest of the variants were similar to those obtained for VP E36D and E40D variants [46], which suffered a shortening of the lateral chain of Glu-36 and Glu-40 (replacement by aspartic acid). These substitutions led to a worsening in the Mn2+ oxidation reflected in a high increase (25-fold) of the Km and a consequent decrease in the efficiency. In the same way, 2-1B and derived variants showed Km values 43–54-fold higher. It is likely that electrostatic interactions with Lys-37 in EHG and 2-1B were limiting the mobility of Glu-36 and Glu-40, leading to an impaired activity towards manganese in a similar way to that which happened with the E36D and E40D variants. Finally, substrate oxidation at the catalytic tryptophan residue was not impaired by the mutations introduced in 2-1B given that the oxidation of high-redox compounds such as VA and RB5 by this variant was similar to that found in native VP. This is indicative of the maintenance of the high redox potential of the enzyme, determined by the position of the proximal histidine residue and by the strength of the interaction between this residue and the haem iron [47,48].

Conclusions

We were able to provide a structural–functional basis for the improved alkaline stability of the 2-1B variant by designing and characterizing several (individual and multiple) intermediate variants with the same mutations found in 2-1B. The E. coli expression system used not only facilitated variant generation by site-directed mutagenesis, but also enabled crystallization of the non-glycosylated forms of two of them. In this way, the introduction of three basic residues in VP (Lys-37, Arg-39 and Arg-330) resulted in new interactions at the haem environment that avoided the hexa-co-ordination of the haem iron, which leads to enzyme destabilization at pH 8. Moreover, the reinforcement of the solvent-exposed area around Gln-305, in the proximal side, further enhanced the stability. Interestingly, the biochemical characterization of the intermediate variants revealed discrete stability improvements, and the combination of most of them was required to stabilize VP. This stabilization could be obtained by random combination of mutations using directed molecular evolution, and would be impossible by a rational approach since the basis for the synergistic improvements observed could not be predicted a priori. Once VP has been stabilized at alkaline pH, the next step to exploit this new property in industrial applications should be to make the enzyme active at this pH (the optimum of 2-1B and native VP is between pH 3 and 5 for the different substrates that they are able to oxidize [14,17], and no activity at alkaline pH has been observed). Directed evolution approaches are currently being explored to attain this objective with very promising results.

AUTHOR CONTRIBUTION

Verónica Sáez-Jiménez performed most of the experimental biochemical work. Sandra Acebes and Victor Guallar contributed to the MD analyses. Francisco Medrano and Antonio Romero crystallized VP variants. All authors participated in the interpretation and discussion of results. Verónica Sáez-Jiménez, Angel Martínez and Francisco Ruiz-Dueñas contributed data integration and writing.

FUNDING

This work was supported by HIPOP [grant number BIO2011-26694], NOESIS [grant number BIO2014-56388-R] and DEWRY [grant number BIO2013-43407-R] projects of the Spanish Ministry of Economy and Competitiveness (MINECO), and by the PEROXICATS [grant number KBBE-2010-4-265397] and INDOX [grant number KBBE-2013-7-613549] European projects. V.S.-J. and F.J.R.-D. are grateful for the financial support of an FPI fellowship and a Ramón y Cajal contract of the Spanish MINECO, respectively.

Abbreviations

     
  • ABTS

    2,2′-azinobis-(3-ethylbenzothiazoline-6-sulfonic acid)

  •  
  • B&R

    Britton–Robinson

  •  
  • CDS

    coding DNA sequence

  •  
  • CT

    charge transfer

  •  
  • DMP

    2,6-dimethoxyphenol

  •  
  • GP

    generic peroxidase

  •  
  • LiP

    lignin peroxidase

  •  
  • MnP

    manganese peroxidase

  •  
  • RB5

    Reactive Black 5

  •  
  • VA

    veratryl alcohol

  •  
  • VP

    versatile peroxidase

References

References
1
Ragauskas
 
A.J.
Beckham
 
G.T.
Biddy
 
M.J.
Chandra
 
R.
Chen
 
F.
Davis
 
M.F.
Davison
 
B.H.
Dixon
 
R.A.
Gilna
 
P.
Keller
 
M.
, et al 
Lignin valorization: improving lignin processing in the biorefinery
Science
2014
, vol. 
344
 pg. 
1246843
 
[PubMed]
2
Martínez
 
A.T.
Ruiz-Dueñas
 
F.J.
Martínez
 
M.J.
del Río
 
J.C.
Gutiérrez
 
A.
 
Enzymatic delignification of plant cell wall: from nature to mill
Curr. Opin. Biotechnol.
2009
, vol. 
20
 (pg. 
348
-
357
)
[PubMed]
3
Ruiz-Dueñas
 
F.J.
Martínez
 
A.T.
 
Microbial degradation of lignin: how a bulky recalcitrant polymer is efficiently recycled in nature and how we can take advantage of this
Microb. Biotechnol.
2009
, vol. 
2
 (pg. 
164
-
177
)
[PubMed]
4
Martínez
 
A.T.
Speranza
 
M.
Ruiz-Dueñas
 
F.J.
Ferreira
 
P.
Camarero
 
S.
Guillén
 
F.
Martínez
 
M.J.
Gutiérrez
 
A.
del Río
 
J.C.
 
Biodegradation of lignocellulosics: microbiological, chemical and enzymatic aspects of fungal attack to lignin
Int. Microbiol.
2005
, vol. 
8
 (pg. 
195
-
204
)
[PubMed]
5
Floudas
 
D.
Binder
 
M.
Riley
 
R.
Barry
 
K.
Blanchette
 
R.A.
Henrissat
 
B.
Martínez
 
A.T.
Otillar
 
R.
Spatafora
 
J.W.
Yadav
 
J.S.
, et al 
The Paleozoic origin of enzymatic lignin decomposition reconstructed from 31 fungal genomes
Science
2012
, vol. 
336
 (pg. 
1715
-
1719
)
[PubMed]
6
Hammel
 
K.E.
Cullen
 
D.
 
Role of fungal peroxidases in biological ligninolysis
Curr. Opin. Plant Biol.
2008
, vol. 
11
 (pg. 
349
-
355
)
[PubMed]
7
Gold
 
M.H.
Youngs
 
H.L.
Gelpke
 
M.D.
 
Manganese peroxidase
Met. Ions Biol. Syst.
2000
, vol. 
37
 (pg. 
559
-
586
)
[PubMed]
8
Fernández-Fueyo
 
E.
Acebes
 
S.
Ruiz-Dueñas
 
F.J.
Martínez
 
M.J.
Romero
 
A.
Medrano
 
F.J.
Guallar
 
V.
Martínez
 
A.T.
 
Structural implications of the C-terminal tail in the catalytic and stability properties of manganese peroxidases from ligninolytic fungi
Acta Crystallogr. D Biol. Crystallogr.
2014
, vol. 
70
 (pg. 
3253
-
3265
)
[PubMed]
9
Ruiz-Dueñas
 
F.J.
Morales
 
M.
García
 
E.
Miki
 
Y.
Martínez
 
M.J.
Martínez
 
A.T.
 
Substrate oxidation sites in versatile peroxidase and other basidiomycete peroxidases
J. Exp. Bot.
2009
, vol. 
60
 (pg. 
441
-
452
)
[PubMed]
10
Morales
 
M.
Mate
 
M.J.
Romero
 
A.
Martínez
 
M.J.
Martínez
 
A.T.
Ruiz-Dueñas
 
F.J.
 
Two oxidation sites for low redox-potential substrates: a directed mutagenesis, kinetic and crystallographic study on Pleurotus eryngii versatile peroxidase
J. Biol. Chem.
2012
, vol. 
287
 (pg. 
41053
-
41067
)
[PubMed]
11
Martínez
 
A.T.
Ruiz-Dueñas
 
F.J.
Gutiérrez
 
A.
del Río
 
J.C.
Alcalde
 
M.
Liers
 
C.
Ullrich
 
R.
Hofrichter
 
M.
Scheibner
 
K.
Kalum
 
L.
, et al 
Search, engineering and applications of new oxidative biocatalysts
Biofuels Bioprod. Biorefining
2014
, vol. 
8
 (pg. 
819
-
835
)
12
Ragauskas
 
A.J.
Williams
 
C.K.
Davison
 
B.H.
Britovsek
 
G.
Cairney
 
J.
Eckert
 
C.A.
Frederick
 
W.J.
Hallett
 
J.P.
Leak
 
D.J.
Liotta
 
C.L.
, et al 
The path forward for biofuels and biomaterials
Science
2006
, vol. 
311
 (pg. 
484
-
489
)
[PubMed]
13
Husain
 
Q.
 
Peroxidase mediated decolorization and remediation of wastewater containing industrial dyes: a review
Rev. Environ. Sci. Bio.
2010
, vol. 
9
 (pg. 
117
-
140
)
14
García-Ruiz
 
E.
González-Pérez
 
D.
Ruiz-Dueñas
 
F.J.
Martínez
 
A.T.
Alcalde
 
M.
 
Directed evolution of a temperature-, peroxide- and alkaline pH-tolerant versatile peroxidase
Biochem. J.
2012
, vol. 
441
 (pg. 
487
-
498
)
[PubMed]
15
Sambrook
 
J.
Russell
 
D.W.
 
Molecular Cloning
2001
Cold Spring Harbor
Cold Spring Habor Laboratory Press
16
Pérez-Boada
 
M.
Doyle
 
W.A.
Ruiz-Dueñas
 
F.J.
Martínez
 
M.J.
Martínez
 
A.T.
Smith
 
A.T.
 
Expression of Pleurotus eryngii versatile peroxidase in Escherichia coli and optimisation of in vitro folding
Enzyme Microb. Technol.
2002
, vol. 
30
 (pg. 
518
-
524
)
17
Ruiz-Dueñas
 
F.J.
Martínez
 
M.J.
Martínez
 
A.T.
 
Molecular characterization of a novel peroxidase isolated from the ligninolytic fungus Pleurotus eryngii
Mol. Microbiol.
1999
, vol. 
31
 (pg. 
223
-
236
)
[PubMed]
18
Britton
 
H.T.S.
Robinson
 
R.A.
 
Universal buffer soluttons and the association constant of veronal
J. Chem. Soc.
1931
(pg. 
1456
-
1462
)
19
Kabsch
 
W.
 
XDS
Acta Crystallogr. Sect. D Biol. Crystallogr.
2010
, vol. 
66
 (pg. 
125
-
132
)
20
Adams
 
P.D.
Afonine
 
P.V.
Bunkoczi
 
G.
Chen
 
V.B.
Davis
 
I.W.
Echols
 
N.
Headd
 
J.J.
Hung
 
L.W.
Kapral
 
G.J.
Grosse-Kunstleve
 
R.W.
, et al 
PHENIX: a comprehensive Python-based system for macromolecular structure solution
Acta Crystallogr. Sect. D Biol. Crystallogr.
2010
, vol. 
66
 (pg. 
213
-
221
)
21
Emsley
 
P.
Lohkamp
 
B.
Scott
 
W.G.
Cowtan
 
K.
 
Features and development of Coot
Acta Crystallogr. Sect. D Biol. Crystallogr.
2010
, vol. 
66
 (pg. 
486
-
501
)
22
Chen
 
V.B.
Arendall
 
W.B.
Headd
 
J.J.
Keedy
 
D.A.
Immormino
 
R.M.
Kapral
 
G.J.
Murray
 
L.W.
Richardson
 
J.S.
Richardson
 
D.C.
 
MolProbity: all-atom structure validation for macromolecular crystallography
Acta Crystallogr. Sect. D Biol. Crystallogr.
2010
, vol. 
66
 (pg. 
12
-
21
)
23
Davis
 
I.W.
Leaver-Fay
 
A.
Chen
 
V.B.
Block
 
J.N.
Kapral
 
G.J.
Wang
 
X.
Murray
 
L.W.
Arendall
 
W.B.
Snoeyink
 
J.
Richardson
 
J.S.
Richardson
 
D.C.
 
MolProbity: all-atom contacts and structure validation for proteins and nucleic acids
Nucleic Acids Res.
2007
, vol. 
35
 (pg. 
W375
-
W385
)
[PubMed]
24
Sastry
 
G.M.
Adzhigirey
 
M.
Day
 
T.
Annabhimoju
 
R.
Sherman
 
W.
 
Protein and ligand preparation: parameters, protocols, and influence on virtual screening enrichments
J. Comput. Aided Mol. Des.
2013
, vol. 
27
 (pg. 
221
-
234
)
[PubMed]
25
Kaminski
 
G.A.
Friesner
 
R.A.
Tirado-Rives
 
J.
Jorgensen
 
W.L.
 
Evaluation and reparametrization of the OPLS-AA force field for proteins via comparison with accurate quantum chemical calculations on peptides
J. Phys. Chem. B
2001
, vol. 
105
 (pg. 
6474
-
6487
)
26
Nosé
 
S.
 
A unified formulation of the constant temperature molecular dynamics methods
J. Chem. Phys.
1984
, vol. 
81
 (pg. 
511
-
519
)
27
Martyna
 
G.J.
Tobias
 
D.J.
Klein
 
M.L.
 
Constant pressure molecular dynamics algorithms
J. Chem. Phys.
1994
, vol. 
101
 (pg. 
4177
-
4189
)
28
Nie
 
G.J.
Reading
 
N.S.
Aust
 
S.D.
 
Relative stability of recombinant versus native peroxidases from Phanerochaete chrysosporium
Arch. Biochem. Biophys.
1999
, vol. 
365
 (pg. 
328
-
334
)
[PubMed]
29
Olden
 
K.
Bernard
 
B.A.
Humphries
 
M.J.
Yeo
 
T.K.
Yeo
 
K.T.
White
 
S.L.
Newton
 
S.A.
Bauer
 
H.C.
Parent
 
J.B.
 
Function of glycoprotein glycans
Trends Biochem. Sci.
1985
, vol. 
10
 (pg. 
78
-
82
)
30
Tams
 
J.W.
Welinder
 
K.G.
 
Glycosylation and thermodynamic versus kinetic stability of horseradish peroxidase
FEBS Lett.
1998
, vol. 
421
 (pg. 
234
-
236
)
[PubMed]
31
George
 
S.J.
Kvaratskhelia
 
M.
Dilworth
 
M.J.
Thorneley
 
R.N.F.
 
Reversible alkaline inactivation of lignin peroxidase involves the release of both the distal and proximal site calcium ions and bishistidine co-ordination of the haem
Biochem. J.
1999
, vol. 
344
 (pg. 
237
-
244
)
[PubMed]
32
Youngs
 
H.L.
Moënne-Loccoz
 
P.
Loehr
 
T.M.
Gold
 
M.H.
 
Formation of a bis(histidyl) heme iron complex in manganese peroxidase at high pH and restoration of the native enzyme structure by calcium
Biochemistry
2000
, vol. 
39
 (pg. 
9994
-
10000
)
[PubMed]
33
Verdín
 
J.
Pogni
 
R.
Baeza
 
A.
Baratto
 
M.C.
Basosi
 
R.
Vázquez-Duhalt
 
R.
 
Mechanism of versatile peroxidase inactivation by Ca2+ depletion
Biophys. Chem.
2006
, vol. 
121
 (pg. 
163
-
170
)
[PubMed]
34
Pérez-Boada
 
M.
Ruiz-Dueñas
 
F.J.
Pogni
 
R.
Basosi
 
R.
Choinowski
 
T.
Martínez
 
M.J.
Piontek
 
K.
Martínez
 
A.T.
 
Versatile peroxidase oxidation of high redox potential aromatic compounds: site-directed mutagenesis, spectroscopic and crystallographic investigations of three long-range electron transfer pathways
J. Mol. Biol.
2005
, vol. 
354
 (pg. 
385
-
402
)
[PubMed]
35
Santoni
 
E.
Jakopitsch
 
C.
Obinger
 
C.
Smulevich
 
G.
 
Comparison between catalase-peroxidase and cytochrome c peroxidase: the role of the hydrogen-bond networks for protein stability and catalysis
Biochemistry
2004
, vol. 
43
 (pg. 
5792
-
5802
)
[PubMed]
36
Henriksen
 
A.
Mirza
 
O.
Indiani
 
C.
Teilum
 
K.
Smulevich
 
G.
Welinder
 
K.G.
Gajhede
 
M.
 
Structure of soybean seed coat peroxidase: a plant peroxidase with unusual stability and haem-apoprotein interactions
Protein Sci.
2001
, vol. 
10
 (pg. 
108
-
115
)
[PubMed]
37
Mirza
 
O.
Henriksen
 
A.
Ostergaard
 
L.
Welinder
 
K.G.
Gajhede
 
M.
 
Arabidopsis thaliana peroxidase N: structure of a novel neutral peroxidase
Acta Crystallogr. Sect. D Biol. Crystallogr.
2000
, vol. 
56
 (pg. 
372
-
375
)
38
Kamal
 
J.K.A.
Behere
 
D.V.
 
Kinetic stabilities of soybean and horseradish peroxidases
Biochem. Eng. J.
2008
, vol. 
38
 (pg. 
110
-
114
)
39
Vogt
 
G.
Woell
 
S.
Argos
 
P.
 
Protein thermal stability, hydrogen bonds, and ion pairs
J. Mol. Biol.
1997
, vol. 
269
 (pg. 
631
-
643
)
[PubMed]
40
Kumar
 
S.
Tsai
 
C.J.
Nussinov
 
R.
 
Factors enhancing protein thermostability
Protein Eng.
2000
, vol. 
13
 (pg. 
179
-
191
)
[PubMed]
41
Vinther
 
J.M.
Kristensen
 
S.M.
Led
 
J.J.
 
Enhanced stability of a protein with increasing temperature
J. Am. Chem. Soc.
2011
, vol. 
133
 (pg. 
271
-
278
)
[PubMed]
42
Gromiha
 
M.M.
Pathak
 
M.C.
Saraboji
 
K.
Ortlund
 
E.A.
Gaucher
 
E.A.
 
Hydrophobic environment is a key factor for the stability of thermophilic proteins
Proteins
2013
, vol. 
81
 (pg. 
715
-
721
)
[PubMed]
43
Hiner
 
A.N.P.
Raven
 
E.L.
Thorneley
 
R.N.F.
García-Canovas
 
F.
Rodríguez-López
 
J.N.
 
Mechanisms of compound I formation in heme peroxidases
J. Inorg. Biochem.
2002
, vol. 
91
 (pg. 
27
-
34
)
[PubMed]
44
Erman
 
J.E.
Vitello
 
L.B.
Miller
 
M.A.
Shaw
 
A.
Brown
 
K.A.
Kraut
 
J.
 
Histidine 52 is a critical residue for rapid formation of cytochrome c peroxidase compound I
Biochemistry
1993
, vol. 
32
 (pg. 
9798
-
9806
)
[PubMed]
45
Vitello
 
L.B.
Erman
 
J.E.
Miller
 
M.A.
Wang
 
J.
Kraut
 
J.
 
Effect of Arginine-48 replacement on the reaction between cytochrome c peroxidase and hydrogen peroxide
Biochemistry
1993
, vol. 
32
 (pg. 
9807
-
9818
)
[PubMed]
46
Ruiz-Dueñas
 
F.J.
Morales
 
M.
Pérez-Boada
 
M.
Choinowski
 
T.
Martínez
 
M.J.
Piontek
 
K.
Martínez
 
A.T.
 
Manganese oxidation site in Pleurotus eryngii versatile peroxidase: a site-directed mutagenesis, kinetic and crystallographic study
Biochemistry
2007
, vol. 
46
 (pg. 
66
-
77
)
[PubMed]
47
Banci
 
L.
Camarero
 
S.
Martínez
 
A.T.
Martínez
 
M.J.
Pérez-Boada
 
M.
Pierattelli
 
R.
Ruiz-Dueñas
 
F.J.
 
NMR study of Mn(II) binding by the new versatile peroxidase from the white-rot fungus Pleurotus eryngii
J. Biol. Inorg. Chem.
2003
, vol. 
8
 (pg. 
751
-
760
)
[PubMed]
48
Banci
 
L.
Bertini
 
I.
Turano
 
P.
Tien
 
M.
Kirk
 
T.K.
 
Proton NMR investigation into the basis for the relatively high redox potential of lignin peroxidase
Proc. Natl. Acad. Sci. U.S.A.
1991
, vol. 
88
 (pg. 
6956
-
6960
)
[PubMed]

Author notes

The co-ordinates and structure factors for the 2-1B and EHG variants of the Pleurotus eryngii VP have been deposited in the PDB under codes 5FNB and 5FNE respectively.