Introduction

The multicopper oxidases (MCOs) constitute a family of enzymes known to oxidize a variety of organic and/or inorganic substrates (Solomon et al. 2014; Santhanam et al. 2011; Pezzella et al. 2015; Dwivedi et al. 2011). They are widely distributed in fungi, bacteria, plants, and insects, involved in lignin formation in plant, pigment formation in fungi, copper homeostasis in bacteria, and iron metabolism in yeast (Nakamura and Go 2005; Giardina et al. 2010).

Laccases are one of the simplest members in the MCO family. Structurally, they exhibit a similar overall fold composed of three cupredoxin-like domains and a catalytic motif containing four copper atoms. Four copper atoms were classified into three types according to their spectroscopic features (Malmstrom 1982; Silva et al. 2012; Jones and Solomon 2015). The type 1 copper (T1 Cu) was characterized by an intense absorption band of the S(Cys) → Cu(II) charge transfer (CT) at around 600 nm, and a small parallel hyperfine coupling in electron paramagnetic resonance (EPR) (Jones and Solomon 2015). The type 2 copper (T2 Cu) exhibited no distinctive absorption features but a large parallel hyperfine coupling in EPR (Jones and Solomon 2015). Two type 3 coppers (T3 Cu) were anti-ferromagnetically coupled through a bridging hydroxide and consequently EPR is silent, but they exhibited a CT transition at around 330 nm (Chen et al. 2010). The T2 Cu and T3 Cu formed the trinuclear copper cluster (TNC).

The electron transfer (ET) in laccases includes intermolecular ET from the substrate to the T1 Cu ion, and intramolecular ET from the T1 Cu ion to the TNC via a conserved His-Cys-His triad (Wherland et al. 2014; Enguita et al. 2004). The oxidation of the substrate occurs at the T1 Cu site (Enguita et al. 2004). It is suggested that the solvent-exposed His residue, a first-sphere ligand of the T1 Cu atom, was at the substrate-binding site common in laccase, and it might be responsible for the electron transfer from the substrate to the T1 Cu ion (Matera et al. 2008; Polyakov et al. 2009; Kallio et al. 2009). The electrons were transferred at ~ 13 Å within the protein from the T1 Cu site to the TNC. At the TNC, the O2 molecule was bound, activated, and reduced to H2O (Jones and Solomon 2015; Wherland et al. 2014). The four-electron reduction process of O2 to two H2O molecules requires four protons and four electrons. Based on the sequence alignment and known crystal structures, two highly conserved carboxylate residues located close to the TNC were found to be involved in O2 reduction (Jones and Solomon 2015; Hugo et al. 2015). One Glu or Asp residue is situated at the dioxygen-entering channel. One Asp residue is situated at the water-leaving channel (Chen et al. 2010). Two acidic residues may facilitate the binding of O2, supply of protons, and cleavage of the O-O bond as indicated by studies performed on Fet3p-metallo oxidase from Saccharomyces cerevisiae, Tth-MCO from Thermus thermophilus HB27, CotA-laccase from Bacillus subtilis, and CueO-laccase from Escherichia coli (Santhanam et al. 2011; Hugo et al. 2015; Quintanar et al. 2005; Augustine et al. 2007; Silva et al. 2012; Chen et al. 2010; Bento et al. 2010; Kataoka et al. 2009; Ueki et al. 2006).

Our previous work demonstrated that T. thermophilus SG0.5JP17-16 laccase (lacTT) is thermoactive and chloride-tolerant, and it can degrade some synthetic dyes (Liu et al. 2015). Thus, the lacTT is a good candidate for the decolorization and detoxification of textile wastewaters. In this study, we built the 3D–structural model of lacTT and, in addition, analyzed the roles of two carboxylate residues E456 and D106 which are equivalent to those found in other MCOs (Hugo et al. 2015; Quintanar et al. 2005; Augustine et al. 2007; Kataoka et al. 2009; Ueki et al. 2006). For the first time, we identified another two key residues, E356 and D423 in lacTT, and explored their roles. The E356 residue may be important for the binding of the substrate and geometry of the T1 Cu site. The D423 residue may have similar function to that of the D106 residue. Furthermore, we developed the catalytic mechanism of laccase, which can provide a new strategy to modify the laccase for the industrial applications, for example, the treatment of synthetic dyestuff.

Materials and methods

Structural modeling and molecular docking

The laccase from Thermus thermophilus HB27 has 75% primary amino acid sequence identity to lacTT. The crystal structure of Thermus thermophilus HB27 laccase (lacHB27) was taken from the Protein Data Bank (PDB ID: 2XU9A) and used as the template to build the 3D model of lacTT. Modeling analysis was done by Discovery Studio Client 3.5 (DS 3.5, Accelrys, Inc., San Diego, CA, USA). The rationality of the 3D model for lacTT was examined via Ramachandran plot and Profile-3D scoring. The protein design procedure was used to build the mutants. Molecular docking was performed using an automated Dock Ligands subprogram of DS 3.5 to fit the substrate guaiacol into the model structure of lacTT; hydrogen bonds and bond lengths were analyzed.

Mutagenesis, expression, and purification

The lacTT gene was optimized for E. coli expression using Optimum Gene technology and synthesized by GenScript Biotech Co. (Nanjing, China). The accession number of the codon-optimized lacTT gene was BankIt2067778 (Seq1MG601742) in GenBank. Amplification of the laccase gene was performed by PCR using the primers TT-F/TT-R (Table S1) and the synthesized laccase gene as the template. Single amino acid substitutions were created using overlap extension PCR with the primers shown in Table S1. The wild-type and mutated genes (E356A, D106L, D106R, D106E, E456L, E456R, E456D, and D423A) were cloned into the Nco I and Hind III sites of plasmid pET-30a(+) (Novagen, Madison, WI) and expressed in E. coli BL21(DE3) (Novagen, Madison, WI).

E. coli BL21(DE3) containing recombinant plasmids were grown in Luria-Bertani (LB) medium supplemented with kanamycin (50 μg/mL) at 37 °C and 200 rpm. When the OD600 of the culture reached 0.6–1.0, 0.1 mM IPTG was added to induce protein expression, and the cells were grown for another 12–18 h, with shaking at 30 °C and 200 rpm. The cells were harvested by centrifugation, then resuspended in Buffer A (20 mM Tris-HCl, 500 mM NaCl, pH 7.5) containing 0.2 mM CuSO4, and sonicated on ice. The crude cell extracts were centrifuged at 8000×g for 30 min. The supernatant was collected, and the soluble wild-type (WT), E356A, D106L, D106R, D106E, E456L, E456R, E456D, and D423A lacTT proteins were purified by immobilized metal affinity chromatography (IMAC). The purification was performed with HisTrap™ FF crude column (GE Healthcare) using AKTA purifier FPLC system (GE Healthcare). The purity of the purified protein was checked by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE). Protein concentration was estimated using Bradford assay with bovine serum albumin (BSA) (Sigma-Aldrich) as the standard.

Enzymatic activity assay

A standard enzyme activity assay was performed in a 1.0-mL reaction mixture containing 50 mM sodium phosphate buffer (pH 6.0), 10 μM CuSO4 and 2.0 mM guaiacol (ε465 = 12,000 M−1 cm−1). The reaction started with the addition of enzymes at 90 °C. After incubation at 90 °C for 5 min, the mixture was immediately cooled in ice bath for 1 min to stop the reaction, and the absorbance was measured at 465 nm. One unit of enzyme was defined as the amount of enzyme that oxidizes 1 μmol guaiacol per minute.

Biochemical characterization

The enzymatic activities were recorded at 90 °C in Britton–Robinson buffer as the function of the pH values varying from 4.5–8.0, using guaiacol as substrate. The optimal pH values for lacTT and its mutants were determined (Fig. S1); no difference in the pH behavior was observed between mutants D106R, D106E, E456R, and wild-type lacTT, and the optimal pH value was 6.0. However, the optimal pH values of D106L, E356A, and D423A mutants moved slightly towards acidic value by 0.5 or 1 units. The optimal pH values of E456L and E456D mutants increased by 1.5 and 2 units, respectively. Figure S2 shows the enzymatic activity as a function of the temperature in the range of 60–95 °C. Optimal reaction temperatures of the wild-type enzyme and mutants were determined for 50 mM sodium phosphate buffer at pH 6. Except for the D106R mutant, which exhibits an optimal temperature of 80 °C, wild-type lacTT and other mutants display similar optimal temperature of 90 °C (Fig. S2). The thermostability of the enzymes was examined for 80 °C during 4 h. D106E and E456R mutants exhibit similar thermal stability to that of wild-type lacTT; the thermostability of the other mutants is reduced (Fig. S3). All assays were repeated for three times.

Kinetic studies

Kinetic assay was performed for wild-type lacTT and mutants in 50 mM sodium phosphate buffer (pH 6.0), supplemented with 10 μM CuSO4 and 0.125–4.0 mM guaiacol in air saturated conditions as described previously (Durao et al. 2006). The reaction started with the addition of enzymes at 90 °C. Kinetic parameters were obtained by fitting the data to the Lineweaver–Burk plot. All assays were repeated for three times.

UV-visible spectroscopy

UV-visible absorption spectra (300–800 nm) of the purified proteins were recorded by using a Hitachi U-3010 spectrophotometer at room temperature in 10 mM sodium phosphate buffer (pH 6.0). The concentration of the protein was 3 mg/ml.

CD spectroscopy

Far-UV circular dichroism (CD) spectra (190 to 260 nm) were recorded by using an Applied Photophysics Chirascan spectrometer at room temperature in 10 mM sodium phosphate (pH 6.0). The bandwidth in a 1-mm cell was 1 nm. The protein concentration was 0.3 mg/ml. Data were averaged, and the baselines were subtracted.

Results

Structural model of lacTT

So far, the structure of lacTT has not been available; therefore, we used homology modeling to build its three-dimensional structure. In order to select an appropriate template, we employed BLASTp to search PDB. It was found that lacHB27 (PDB entry: 2XU9A) has 75% sequence identical to that of lacTT, and the crystallographic structure of lacHB27 was resolved (Hugo et al. 2015). Based on homology modeling and energy minimization, the structural model of lacTT was generated, and it is presented in Fig. 1a. The quality of the model was evaluated by Ramachandran plot and Profile-3D. In Ramachandran plot, 93.4% of the amino acid residues were located in the optimum region, 3.4% were in the allowed region, and only 2.7% in the disallowed region (Fig. 1b). In Profile-3D graph, the compatibility score of the amino acid residues was 98.9% (Fig. 1c). Thus, the generated model can represent the structure of lacTT, and it will be employed in the following analyses.

Fig. 1
figure 1

Structural model of lacTT. a The modeled three-dimensional structure of lacTT comprised of three domains, showing α-helices in red, β-sheets in cyan, β-turns in green, loops in gray, T1 Cu in blue and trinuclear copper cluster (TNC): T2 Cu in green, and two T3 Cu in dark red. b Ramachandran plot of the simulated lacTT structure, showing the residues in the optimum region (blue), the allowed region (magenta), and the disallowed region (red triangle). c Profile-3D of the lacTT structure, indicating the compatibility score of the majority of the residues was greater than zero. d The geometry of the copper active sites: the copper ions (spheres) and their ligands

Full size image

Like other MCO structures (Hugo et al. 2015; Piontek et al. 2002; Enguita et al. 2003; Bello et al. 2012), the structural model of lacTT contains three cupredoxin-like domains and four copper atoms classified into two active sites, a mononuclear T1 Cu center situated in domain 3 and a TNC located at the interface between domains 1 and 3 (Fig. 1a). The active Cu sites of lacTT share primary coordination sphere characteristics of other bacterial MCOs. The T1 Cu atom is beset in a tetrahedral geometry formed by His397, His455, Cys450, and Met460 residues (Fig. 1d), where the substrate was oxidized. In UV-visible spectrum (Fig. 2a), a strong absorption band appeared at around 608 nm due to the charge transfer transition between Scys450 and T1 Cu2+ (Solomon 2006). There is a methionine-rich motif on the top of the substrate-binding pocket. The TNC is comprised of a T2 Cu and two T3 Cu atoms (denoted as T3α Cu and T3β Cu). The T3α Cu coordinated by His137, His402, and His449 residues is ~ 5 Å away from the T3β Cu having His97, His135, and His451 ligands (Fig. 1d). At about 330 nm of the UV-visible spectrum, a broad shoulder absorption band was recorded, and it is due to the charge transfer transition between the T3 Cu ions and a hydroxyl bridged between two T3 Cu ions (Cole et al. 1990). The T2 Cu atom is braced by His95 and His400 residues; no absorption band was observed for the T2 Cu atom.

Fig. 2
figure 2

UV-visible absorption spectra of the lacTT proteins were measured at room temperature in 10 mM sodium phosphate buffer, pH 6.0. The protein concentration was 3 mg/ml. a WT (black) and E356A (red). b WT (black) and E456D (blue), and E456R (green) and D456L (red) mutants. c WT (black) and D106E (blue), and D106R (green) and D106L (red) mutants. d WT (black) and D423A (blue)

Full size image

The effect of a second-sphere Glu residue at the T1 Cu site on the oxidation of the substrate

It is well known that in MCOs, the substrate is oxidized at the T1 Cu site, and the intermolecular ET from the substrate to the T1 Cu ion is controlled most likely by the ligands surround the T1 Cu ion (Kataoka et al. 2007; Marcus and Sutin 1985). The geometry formed by the primary coordination ligands was tuned by the second-sphere residues via H-bonding networks (Span et al. 2017). In lacTT, the second-sphere residues comprise E356, which is located at the T1 Cu site and is H-bonded to the His397 ligand of the T1 Cu atom, Glu456, D106, and D423. The later three residues are positioned close to the TNC (Fig. 3).

Fig. 3
figure 3

a Four second-sphere residues distributed in the lacTT. The color scheme is same as that in Fig. 1. The E356 residue (red) is located at the T1 Cu site and in the region of the substrate binding. At the top of the T1 Cu site, there is a methionine-rich loop, which may be a switch of controlling the substrate entering. The E456 residue (yellow) is between the T1 Cu site and the TNC, at the O2-entering channel. The D423 residue (magenta) is at the T2-T3β Cu atom side of the TNC. The residue D106 (magenta) is set to the T2-T3α Cu atom side of the TNC, at the exit of the reduced water. b The zoom-in of the T1 Cu center highlighting the locations of the residues E356 and E456. c The zoom-in of the TNC center highlighting the locations of the residues E456, D423, and D106

Full size image

In order to further understand the role of the E356 residue, we mutated Glu356 to Ala356. CD spectra of the WT laccase and the E356A mutant are similar, as shown in Fig. 4a, suggesting that the global structure of the enzyme is not affected by site-directed substitution of the E356 residue. The distance between copper ions in the E356A mutant is identical to that in the wild-type enzyme as reported in Table S2. Thus, the replacement of the E356 residue by alanine does not affect the relative position of the active Cu sites. We characterized the biochemical properties of the E356A mutant. The optimal pH value of the E356A mutant is the same as that of WT enzyme, which is approximately 6.0 (Fig. S1A). However, the kinetic study suggests that the catalytic efficiency (K cat /K m ) of the E356A enzyme decreased ~ 13-fold as compared to that of WT laccase, as it can be observed in Table 1. The comparison of UV-visible spectra for WT laccase and the E356A variant reveals that the absorption intensity is similar at 330 nm, but the absorption intensity at 608 nm for the E356A mutant is 60% of the wild-type enzyme (Fig. 2a and Table S3), reflecting the lower Cu loading at the T1 Cu site in the mutant. Thus, the efficiency of the E356A mutant should be adjusted to eightfold lower than that of the wild-type laccase. From the structural model of lacTT, the side chain Oε1 atom of the E356 residue is 4.7 Å away from the side chain Nδ atom of His397. They can form the H-bond interaction, while the E356 residue is negative charged at the optimal pH 6.0, the carboxyl group of the E356 residue may H-bond to the side chain Nε atom of His397 (Fig. 5a). These interactions are required to retain the geometry of the T1 Cu site and to enhance the nucleophilicity of the His ligand. The substitution of the E356 residue by an apolar alanine eliminated these interactions. The side chain of the A356 residue pointed to the opposite direction of the H397 ligand (Fig. 5b). The kinetic assay showed that the K cat value of the E356A mutant decreased ~ 4-fold, while the K m value increased ~ 4-fold relative to that of WT enzyme (Table 1). We compared the substrate-binding pockets in WT enzyme (Fig. 6a) and the E356A mutant (Fig. 6b). The pocket in the E356A mutant became deeper, suggesting that the E356 residue is related to the binding of the substrate, for example, probably responsible for holding the substrate at the correct position to form the optimal interactions between the substrate and the catalytic residues from the enzyme. Therefore, we inferred that the E356 residues, serving as a secondary coordination sphere residue, affected the catalysis of the enzyme mainly by orientating the histidine ligands of the T1 Cu atom and/or the binding of the substrate. We tried to further clarify the details of the interactions between the substrate and the enzyme via the docking study. Unfortunately, the substrate could not be docked into the substrate-binding pocket, as shown in Fig. S4, most probably due to the closed conformation of the methionine-rich loop blocking the entering of the substrates.

Fig. 4
figure 4

CD spectra of the lacTT proteins were measured at room temperature in 10 mM sodium phosphate buffer, pH 6.0. The protein concentration was 3 mg/ml. a WT (black) and E356A (red). b WT (black) and E456D (blue), E456R (green), and D456L (red) mutants. c WT (black) and D106E (blue), and D106R (green) and D106L (red) mutants. d WT (black) and D423A (blue)

Full size image
Table 1 Apparent steady-state kinetic constants measured at the saturating concentrations of O2 for guaiacol by the lacTT proteins
Full size table
Fig. 5
figure 5

Hydrogen-bonding networks between T1 Cu ligands and the E356 residue. The T1 Cu atom is braced in a tetrahedral geometry formed by the residues M460, C450, H455, and H397. a The E356 residue is within hydrogen-bonding distance with the H397 residue, a ligand of the T1 Cu atom. b The A356 residue cannot form the H-bond interaction with the H397 residue

Full size image
Fig. 6
figure 6

Surface presentation showed the shape of the substrate-binding pockets for the wild-type lacTT (a) and the E356A mutant (b). The Glu356 residue in the substrate-binding pocket is in red. The substitution of the E356 residue by alanine resulted in a deeper pocket

Full size image

The electric transfer from T1 Cu to TNC coordinated by a Glu residue at the TNC

The electron was abstracted at the T1 Cu site, then, rapidly transferred to the TNC via the Cys-His pathway. At the TNC, the dioxygen obtained the electrons and was reduced to the water (Jones and Solomon 2015). The bioinformatics analysis based on the sequence identity and structure revealed that the residues Glu456, D106, and D423 in lacTT are highly conserved (Figs. 3 and S5), and would control the intramolecular electron transfer from T1 Cu to the TNC, the reduction of the dioxygen, and/or the release of the water.

The G456 residue was located between the T1 Cu site and T3α Cu atom. Its carboxyl side chain points towards the O2-entering channel, within the H-bond distance with the histidine ligands of the T3α Cu atom. The distances from the side chain Oε atoms of the Glu456 residue to the Nδ1 atom of His449 and the Nδ1 atom of His137 are 4.79 and 3.46 Å, respectively (Fig. 7a); consequently, this residue is important for the geometry of the TNC. The E456 residue in lacTT is equivalent to Glu451 in lacHB27, Glu498 in the CotA-laccase, Glu506 in the CueO-laccase, and Glu487 in Fet3p-metallo oxidase (Hugo et al. 2015; Chen et al. 2010; Bento et al. 2010; Quintanar et al. 2005; Augustine et al. 2007; Kataoka et al. 2009; Ueki et al. 2006). Previous studies show that the acidic residue at this position is involved in the proton transfer procedure during the dioxygen reduction to the water (Hugo et al. 2015; Chen et al. 2010; Bento et al. 2010; Quintanar et al. 2005; Augustine et al. 2007; Kataoka et al. 2009; Ueki et al. 2006). To examine its ability to shuttle the proton, the E456 residue was replaced by the leucine, arginine, and aspartic acid, respectively, by site-directed mutagenesis. The catalytic activities of WT lacTT and its mutants were determined by using the guaiacol as the reducing substrate at their optimum pH value. Compared with the wild-type enzyme, the E456L, E456R, and E456D mutants exhibited lower catalytic efficiencies (kcat/Km), shown in Table 1. The catalytic efficiency of the E456D mutant decreased eightfold. The below UV-visible spectral data revealed that 61% of population of the T1 Cu site in the E456D mutant were filled with the copper atoms (Table S3); the efficiency of the E456D mutant should be adjusted to fivefold lower than that of the wild-type laccase. The aspartic acid had the ability to shuttle the proton, but the transfer efficiency became lower due to the shorter side chain than that of glutamic acid. In E456L and E456R mutants, the apolar leucine and the basic arginine residues at the reaction pH value could not channel the proton for the reduction of the dioxygen, the E456L and E456R mutants showed 30 to 40 times lower efficiency in relation to the wild type.

Fig. 7
figure 7

Hydrogen-bonding networks formed between the amino acid residue at the position of 456 and the histidine ligands of the copper centers. a In the wild-type lacTT. b In the E456D mutant. c In the E456R mutant. d In the E456L mutant

Full size image

In the UV-visible spectrum, the wild-type lacTT presented two absorption bands. One maximum absorption band appeared at 608 nm, and the other at 330 nm. The E456D mutant showed two lower intensity peaks; the maximum absorption intensity at 608 nm was 61% of the wild-type enzyme (Fig. 2b and Table S3). After treatment with potassium bichromate, this absorption intensity did not change; ruling out that the reduction of the T1 Cu ion resulted in the decrease of the peak intensity for the E456D mutant. Thus, some of T1 Cu sites in the E456D mutant lack copper. The E456L and E456R mutants show higher peak intensity at 608 nm compared to the wild-type enzyme, reflecting that the occupancy of T1 Cu sites in two mutants is slightly higher. CD spectra of three variants were similar to that of the wild-type laccase, as shown in Fig. 4b. The structure of each mutant can superimpose well that of the wild-type laccase (Fig. S6C). These data clearly indicate that their overall structures are essentially the same. However, the differences are also observed for the copper centers and their surrounds; the distances between three coppers at the TNC in three mutants increase, as shown in Table S2. The E456R mutant was most affected, showing the longest distance between any two coppers. These findings reflect a loose structure at the TNC center in terms of the variants, which might be related to a slower electron transfer rate, consistent with the lower K cat value for variants.

In the wild-type lacTT, the E456 residue is situated next to the T3α Cu atom and its carboxyl side chain points towards the T3α Cu (Fig. 3). The side chain Oε2 atom of Glu456 was 2.98 Å away from the Nε atom of His449 and 3.74 Å from the Nε atom of His137 (Fig. 7a). In the E456D mutant, aspartic acid residue has a shorter side chain, and the orientation of the side chain is altered, which increases the distances from the side chain Oδ atom of Asp456 to His449 and His137 ligands of the T3α Cu atom to 3.92 and 4.43 Å, respectively (Fig. 7b), resulting in the orientation change of the His ligands and/or the decrease of the interactions between the His ligands and the D456 residue. Furthermore, these effects are transferred to His135, His451, and Cys450 via β-sheets, and to His455 via α-helix (ligands of T3β and T1 Cu atoms). It is suggested that, in E456D mutant, the coordination of all Cu centers can be affected. These results are consistent with the intensity alteration of the absorption peaks in the UV-visible spectra and the change of the distances between any two coppers in the mutant described above. The alteration of interacting networks and the ligand orientation of the Cu atoms are also observed in the E456R and E456L mutants. In the E456R mutant, the longer, positively charged side chain of the R456 residue is extended to H-bond distance with the His135 and His451 ligands of the T3β Cu atom (Fig. 7c). In the E456L mutant, these interactions could not be formed, and the orientation of the L456 residue was altered (Fig. 7d), the catalytic behavior of the enzyme was changed. By combining the structural, UV-visible and CD spectral data, it was found out that the secondary coordination sphere residue E456, apart from channeling the proton for the reduction of the O2 molecule (Chen et al. 2010; Hugo et al. 2015; Kataoka et al. 2009; Bento et al. 2010), formed hydrogen-bonding networks with His ligands of the copper centers to maintain the geometry of the copper centers.

The reduction of the dioxygen and/or the release of the reduced water coordinated by two Asp residues at the TNC

Previous extensive studies have revealed that the reduction of O2 to 2H2O occurred at the TNC and involved two two-electron transfers (Augustine et al. 2010). The dioxygen was converted to the peroxide intermediate (PI) by first two-electron transfer, and this step was rate-limiting, followed by fast second two-electron reduction from PI to the native intermediate (NI) (Lee et al. 2002; Palmer et al. 2002; Yoon and Solomon 2007). In the decay of the PI to the NI, one proton was required to cleave the peroxide −O–O− bond (Palmer et al. 2002) and shuttled by a Glu residue nearby the TNC center (Chen et al. 2010; Bento et al. 2010). On the other hand, an Asp residue at the TNC, such as Asp116 in the CotA-laccase, Asp112 in the CueO-laccase, Asp106 in lacHB27, and Asp94 in Fet3p-metallo oxidase, was also important for the reduction of dioxygen to water (Silva et al. 2012; Hugo et al. 2015; Quintanar et al. 2005; Kataoka et al. 2009). This residue did not provide the proton but coordinated an OH− ion and an H2O molecule connecting the T2 Cu atom (Silva et al. 2012; Augustine et al. 2007; Bento et al. 2010).

In lacTT, the Asp106 residue was at the T3α-T2 edge, its carboxylate group, within the H-bond distances with the His402 ligand of the T3α Cu atom (3.85 Å) and the His95 ligand of the T2 Cu atom (2.71 Å), points towards the water-leaving channel (Figs. 3 and 8a). The bioinformatics studies demonstrate that this residue is equivalent to the Asp residue in other MCOs (Fig. S5) (Silva et al. 2012; Hugo et al. 2015; Quintanar et al. 2005; Kataoka et al. 2009). To get insight into the roles of the Asp106 residue, we mutated the Asp106 to leucine, arginine, and glutamic acid. The kinetic assay revealed that three mutants exhibited similar catalytic efficiency, ~ 4-fold decrease as compared with that of the wild-type laccase (Tables 1 and S3). CD spectra of the wild-type enzyme and three mutants show no significant difference (Fig. 4c), reflecting the similarity of their global structures. This was indicated by the structural superimposition of mutants with WT enzyme (Fig. S6B). However, there are some small differences at the Cu centers and in the orientations of the Cu ligands; Table S2 shows that in three mutants, the distance between each two Cu atoms at the TNC becomes remarkably longer as compared with that of the wild-type enzyme. Furthermore, in the D106E mutant, one carboxyl O atom of the side chain of E106 still points to the water-leaving channel, but the other orientates to the interior of the protein (Fig. 8b). Although the substitution of the D106 residue by glutamic acid retained the positive charge at this position, the increase of side chain length caused the change of interacting networks at the TNC, and the increases of the COO−E106-NδH95 distance from 2.71 to 4.26 Å (Fig. 8b) and the COO−E106-NδH402 distance from 3.85 to 4.06 Å. These changes may contribute to the decrease in the catalytic efficiency of the D106E mutant. In D106L mutant, the side chain of the apolar leucine residue points to the water-leaving channel and His ligands of the Cu center, but there are no H-bond interactions between them (Fig. 8c). In the case of the D106R mutant, negatively charged side chain of the arginine residue oriented to the opposite direction of the Asp residue in WT enzyme. The histidine ligands of the T3α and T2 Cu atoms cannot form H-bond interactions with the R106 residue any more (Fig. 8d). These findings suggest that the D106 residue is involved in the catalytic reaction by hydrogen-bonding networks to the histidine ligands of the Cu centers.

Fig. 8
figure 8

Hydrogen-bonding networks formed between the amino acid residue at the position of 106/423 and the histidine ligands of the copper centers. a In the wild-type lacTT. b In the D106E mutant. c In the D106L mutant. d In the D106R mutant. e In the D423A mutant

Full size image

In UV-visible spectra, compared with the WT enzyme, the D106E, D106R, and D106L mutants show lower intensity peak at 330 nm, suggesting the Cu occupancy of the Cu sites at the TNC is lower, consistent with the loose structure of the TNC demonstrated by the longer distance between two Cu atoms in mutants (Table S2). For the peak at 608 nm, the higher intensity was observed for the D106E and D106L mutants, while the D106R enzyme shows a lower intensity peak, reflecting different occupancy at the T1 Cu site in three mutants.

The Asp423 and Asp106 residues are symmetrically distributed at both sides of the T2 Cu atom (Fig. 8a). The D423 residue is situated at the T2-T3β side and at the water-leaving channel. The carboxyl side chain of the D423 residue also points towards the water-leaving channel (Fig. 3), within hydrogen-bonding distances with the His97 ligand of T3β Cu and the His400 ligand of T2 Cu. Thus, both Asp residues may play the key roles in the catalytic activity of lacTT. To address the importance of the D423 residue in interacting networks with His ligands, the D423 residue was replaced by an apolar alanine. In this case, the catalytic efficiency of the D423A mutant was 46% of the WT enzyme. Considering that the Cu loading at the T1 Cu site in the D423A mutant was 70% of the WT enzyme, reflected by the lower intensity of the peak at 608 nm in the UV-visible spectrum of the D423A mutant (Fig. 2d and Table S3), thus, the catalytic efficiency of the D423A mutant was adjusted to ~ 65% of the WT enzyme. Again, the A423 residue could not form H-bond interactions with His400 bound to the T2 Cu atom and His97 bound to the T3β Cu atom (Fig. 8a, e). In the D423A mutant, the alteration in the orientation of the His ligands may damage the catalytic behavior of the enzyme. It is interesting that the lack of the interactions in the D423A mutant did not lead to the alterations of the distance between the Cu atoms at the TNC (Table S2) and the overall structure of the enzyme (Figs. 4d and S6D).

Discussion

Laccase within the multicopper oxidase family appears to share similar structural characteristics, including absolutely conserved histidine ligands of the active Cu centers which form the primary copper coordination sphere (Hugo et al. 2015; Quintanar et al. 2005; Augustine et al. 2007; Kataoka et al. 2009; Ueki et al. 2006). The N-terminal amino group of histidine ligates the Cu. Its orientation and/or position is directly related to the catalytic activity of the enzyme. Noncovalent secondary sphere residues around the primary coordinating ligands affect the catalysis of the enzyme by steric, electrostatic, and/or hydrogen-bonding interactions. Hydrogen-bonding network formed by secondary sphere residues was found to be most common in O2-activating enzymes (Chen et al. 2010; Span et al. 2017). They play an important role in stabilizing coordination geometries, adjusting ligand donor strengths, and mediating proton transfer reactions (Span et al. 2017; Holm et al. 1996; Makris et al. 2007).

In this study, the structure of lacTT was predicted by using homology modeling. It exhibits an overall folding similar to that of other multiple copper oxidases, comprising of three domains and four copper atoms. Four second-sphere residues, E356, E456, D106, and D423, were identified, and their roles were analyzed and discussed. It was suggested that the E356 residue maintains the geometry of the T1 Cu site by interacting with His ligands and may be related to the binding of the substrate revealed by site-directed mutagenesis of the E356 residue.

The E456 residue was situated at the entry channel of dioxygen. It is highly conserved in MCOs. The substitution of the E456 residue by aspartic acid, arginine, or leucine was used to examine its role. The replacement of this residue did not lead to change of the overall structures (Figs. 4 and S6C). But the E456D mutant shows a fivefold lower efficiency of the wild-type enzyme (Table 1), consistent with the ability of the Asp residue to shuttle the proton. The efficiency of the E456D mutant decreased due to the alteration of interacting networks at the Cu centers (Fig. 7). Similar behavior was observed for the CotA laccase (Chen et al. 2010). However, the difference in the catalytic behavior was also found between CotA and lacTT enzymes. For the CotA laccase, equivalent Glu498 residue was mutated to leucine or threonine, no enzymatic activity was noted (Chen et al. 2010). In contrast, for lacTT, the E456R and E456L mutants could not have the ability to shuttle the proton, but they still showed 3 and 4% activity of the wild-type enzyme, respectively. One explanation is that there was an alternative proton transfer pathway in lacTT although the transfer efficiency was very low. Or the solvent entered the TNC, acting as a proton donor, retrieved the activity to some extent. On the other hand, structural data of lacTT indicate that there are hydrogen-bonding networks established between the E456 residue and His ligands of the T3α and T3β Cu atoms (Fig. 7a). The replacement of this residue leads to a loose structure at the TNC, increasing the distance between the Cu atoms (Table S2), which can decrease the electron transfer rate in the protein. It is suggested that the E456 residue can be required to maintain the geometry of TNC. Two functions of the E456 residue may be explained by two conformations—one conformation is the protonated state, which can provide the proton to the peroxide bond, and the other one is the carboxylate state, which can maintain the geometry of the TNC by forming hydrogen-bonding networks with His ligands of the cupper ions. At the optimal reaction condition of pH 6.0, the Glu456 residue can be present in the carboxylate state considering the pK a value of 3.4 in solution, while it can also be abnormally protonated as described in CotA laccase depending on the oxidation and TNC loading state (Bento et al. 2010). Two states of the Glu456 residue may be actualized by conformational dynamics. X-ray crystallographic studies show that equivalent Glu451 residue in lacHB27 is present in both conformations (Hugo et al. 2015).

In the catalytic cycle of MCOs, two intermediates (PI and NI) were established experimentally. In the peroxy intermediate, three copper ions at the TNC coordinate the O-O bond, while the T2 Cu ion had a coordinating hydroxide (Yoon and Solomon 2007) (Fig. 9a). In the native intermediate, the O-O bond was cleaved and the oxygen atoms presented as a μ3-oxo with three copper atoms at the TNC and a μ2-OH bridging two T3 Cu ions (Yoon et al. 2007) (Fig. 9b). In both intermediates, the D106 and D423 residues in lacTT could H-bond to His ligands of T2 Cu and T3 Cu (Fig. 9a, b). These hydrogen-bonding networks can increase the nucleophilicity of the N-terminal amino group of histidine ligands and facilitate the stability of oxyl intermediates at the TNC (Fig. 9a, b) (Span et al. 2017; Yoon and Solomon 2007; Yoon et al. 2007). It is worth noticing that when the D106 and D423 residues are mutated to an apolar residue, the activities of two mutants display significant differences. The catalytic efficiency of the D106L mutant was 2.5-fold lower than that of the D423A mutant (Tables 1 and S3) considering the Cu loading at the T1 Cu site in two mutants (Fig. 2c, d). This can be explained by another role of the D106 residue. The D106 residue may mediate the release of the reduced water. Structural data show that the D423 residue is located at the inner of the water-leaving channel at the β-sheet in domain 1. The D106 residue is situated at the exit of the water-leaving channel at the solvent-exposed loop in domain 3 (Figs. 3 and 9c). The flexibility of this loop can provide for the opportunity of the D106 residue switching between two functions. The mutation of the D106 residue to leucine did not only damage the hydrogen-bonding networks but also lost the ability to carry the water. This role of the D106 residue was also observed in other MCOs (Hugo et al. 2015; Quintanar et al. 2005; Augustine et al. 2007; Kataoka et al. 2009; Ueki et al. 2006). X-ray studies indicate that the Asp106 residue in lacHB27 does not take part in both the provision of the proton and the cleavage of the O-O bond, but coordinated a water molecule and a hydroxyl group (Hugo et al. 2015).

Fig. 9
figure 9

The schematic diagram of two intermediates, peroxide intermediate (a) and native intermediate (b), in the catalytic cycle of the laccase. In two intermediates, the D423 residue formed the hydrogen-bonding interactions with H97 of T3β Cu and H400 of T2 Cu; the D106 residue H-bound with H402 of T3α Cu and H95 of T2 Cu, and formed a network with a water and a hydroxyl connecting with the T2 Cu atom. c Two acidic residues at the TNC were symmetrically distributed at both sides of the T2 Cu atom, and their side chains pointed towards the water-leaving channel. The D423 residue was at a β-sheet and at the inner of the channel, while the D106 residue was located at the exit of the channel, and at a flexible loop

Full size image

In summary, two acidic residues, Glu456 and Asp106 in lacTT, were examined, and they exhibit a role similar to that of equivalent acidic residues in other MCOs as described in previous work (Hugo et al. 2015; Quintanar et al. 2005; Augustine et al. 2007; Kataoka et al. 2009; Ueki et al. 2006). In addition, we, for the first time, identified two key second-sphere residues, E356 and Asp423, and discussed their roles. Four second-sphere residues affect the activity of the enzyme, mainly via hydrogen-bonding networks.