Kinetic analysis and structural studies of a high‐efficiency laccase from Cerrena sp. RSD1

A high‐efficiency laccase, DLac, was isolated from Cerrena sp. RSD1. The kinetic studies indicate that DLac is a diffusion‐limited enzyme. The crystal structure of DLac was determined to atomic resolution, and its overall structure shares high homology to monomeric laccases, but displays unique substrate‐binding loops from those in other laccases. The substrate‐binding residues with small side chain and the short substrate‐binding loop IV broaden the substrate‐binding cavity and may facilitate large substrate diffusion. Unlike highly glycosylated fungal laccases, the less‐glycosylated DLac contains one highly conserved glycosylation site at N432 and an unique glycosylation site at N468. The N‐glycans stabilize the substrate‐binding loops and the protein structure, and the first N‐acetylglucosamine is crucial for the catalytic efficiency. Additionally, a fivefold increase in protein yield is achieved via the submerged culture method for industrial applications. Database The atomic coordinates of the structure of DLac from Cerrena sp. RSD1 and structural factors have been deposited in the RCSB Protein Data Bank (PDB ID: 5Z1X).

A high-efficiency laccase, DLac, was isolated from Cerrena sp. RSD1. The kinetic studies indicate that DLac is a diffusion-limited enzyme. The crystal structure of DLac was determined to atomic resolution, and its overall structure shares high homology to monomeric laccases, but displays unique substrate-binding loops from those in other laccases. The substrate-binding residues with small side chain and the short substrate-binding loop IV broaden the substrate-binding cavity and may facilitate large substrate diffusion. Unlike highly glycosylated fungal laccases, the less-glycosylated DLac contains one highly conserved glycosylation site at N432 and an unique glycosylation site at N468. The N-glycans stabilize the substratebinding loops and the protein structure, and the first N-acetylglucosamine is crucial for the catalytic efficiency. Additionally, a fivefold increase in protein yield is achieved via the submerged culture method for industrial applications.

Database
The atomic coordinates of the structure of DLac from Cerrena sp. RSD1 and structural factors have been deposited in the RCSB Protein Data Bank (PDB ID: 5Z1X).
Laccases are important biocatalysts for a number of applications, including delignification, bioremediation, biosensor, and analytical applications [1], and have been studied continually and extensively. H 2 O 2 -independent laccase is more eco-friendly than other lignin-modifying enzymes because it only requires O 2 and produces water as the by-product of catalysis [2,3]. Laccase can be found in bacteria, insects, fungi, and plants, and fungal laccases show higher redox potentials than other laccases [4][5][6][7]. At least sixty fungal strains have been reported to have laccase activities, including ascomycetes, deuteromycetes, and basidiomycetes [8]. In whiterot basidiomycetes, laccases are used to oxidize phenolic compounds, leading to the degradation of lignin [9]. 2,2 0 -Azino-bis(3-ethylbenz-thiazoline-6-sulfonic acid) (ABTS) is a common substrate of laccases as well as a mediator that can extend its substrate range from phenolic compounds to nonphenolic substrates [10].
However, the difficulty of the large-scale production of fungal laccases hinders their industrial usage because of the long generation times of fungi and multiple enzyme isoforms, as well as complicated purification steps. Even though the heterologous expression is a promising technique, the expression of fungal laccases in Escherichia coli has not been reported [11]. The protein yield and laccase activity obtained from the expression of fungal laccase in Pichia pastoris are considerably low compared to the native forms due to differences in the glycosylation patterns [12,13].
Fungal laccases typically have 3 to 10 glycosylation sites, and 10 to 50% of their molecular weight is attributed to glycosylation [1,14,15]. The deglycosylation of laccase in Pycnoporus sanguineus affects its enzyme kinetics, and the loop-glycan interaction is crucial for laccase function in Lentinus sp. [12,16]. However, structural data are limited due to low resolution and the high flexibility of glycan molecules. Previous studies have indicated that the loss of laccase activity in Rigidoporus lignosus and Pleurotus ostreatus is associated with the protein conformation rather than with the coordination of metal site geometry at the active site [17]. A portion of the structure of Rigidoporus lignosus laccase is sensitive to high temperatures, which leads to the loss of enzyme activity [18]. Many studies have focused on the improvement of the enzyme activity of laccase via protein engineering approaches, such as random mutation [19], site-directed mutation [20], and directed evolution [21]. Most of the mutations on the surface and in loop regions have been suggested to increase enzyme activity [21,22]. The active site and the copper coordination geometry are highly conserved among laccases. Therefore, the residues on the protein surfaces and in loop regions must play a crucial role in improving the enzyme activity. Unfortunately, these prior studies lacked either kinetic data or structural evidence to justify how these residues affect the laccase activity.
The main focus of this study was to discover a highly active laccase and to understand what structural features governing the enzyme kinetics. A unique laccase called DLac was isolated from Cerrena sp. RSD1. It displays high catalytic efficiency in the diffusionlimited range compared to all previously studied enzymes. A productive submerge culture system was established for DLac production, purification, and crystallization. The crystal structure was determined to atomic resolution to investigate the correlations between the structural properties and the catalytic efficiency. The observed glycosylation also implicates its important role in the stabilization of substrate-binding loops (SBLs) and domain 3 (D3).

Results and Discussion
Identification of DLac from Cerrena sp. RSD1 As high-efficiency fungal laccases are required for industrial applications, Taiwanese indigenous fungi were isolated from rice straw composts from Shan-hua in Taiwan. The dominant microbes in the degrading compost, which revealed lignocellulosic enzyme activity for rice straw degradation, are well adapted to growth under low pH and high temperature conditions. To screen for fungi producing high-efficiency laccases, we have developed a high-throughput screening method (Fig. 1A) based on the color change on fungi-cultured plates containing guaiacol or Remazol Brilliant Blue R (RBBR) [23,24]. A fungal strain exhibiting the strongest laccase activity was isolated, and the laccase secreted from a submerged culture of the Cerrena sp. RSD1 was designated DLac. Phylogenetic analysis using 18S rDNA sequence and internal transcribed spacer (ITS) sequences indicated that the Cerrena sp. RSD1 belongs to the genus Cerrena (Fig. 2). The crude filtrate of Cerrena sp. RSD1 and purified DLac were analyzed by SDS/PAGE (Fig. 3A). The purified DLac protein band observed on the gel was subsequently isolated for N-terminal amino acid sequencing. The N-terminal sequence 'AVGPVTDI' and highly conserved copper-binding regions were used to design the degenerate primers for DLac gene sequencing. The full-length DLac contains 2168 bp with a 1551-bp coding sequence encoding 516 amino acids, including a signal peptide of 21 amino acids (Fig. 4). The mature protein has 495 amino acids with a calculated molecular mass of 52.9 kDa. NCBI BLAST analysis indicated that the gene with highest homology to DLac is the laccase 7 precursor of Cerrena sp. HYB07 (GenBank accession no. KF317949). The cDNA and amino acid sequences shared 98.6 and 100% identity, respectively [25].

Production, purification, and characterization of DLac
Here, we used a submerged culture to produce DLac and its production was fivefold enhanced in an 8-day culture by adding 2,5-dimethylaniline as an inducer ( Fig. 1B-C). The utilization of 2,5-dimethylaniline as an inducer has also been observed in the production of other basidiomycete laccases [26,27]. The crude filtrate was collected, and the DLac protein was further purified using anion-exchange and size-exclusion chromatography. We then characterized the chemical properties of DLac using ABTS, 2,6-dimethoxyphenol (2,6-DMP) and guaiacol as substrates. The optimal pH value for ABTS is pH 3.0, and pH 4.0 is the optimal pH reaction value for DMP and guaiacol (Fig. 5A). Both ABTS and guaiacol have optimal temperatures of 65°C, and that for DMP is 50°C (Fig. 5B). The kinetic parameters of DLac were also determined using ABTS, DMP, and guaiacol as substrates. DLac has the highest turnover rate, substrate affinity, and catalytic efficiency for ABTS, with a k cat of 52 515 s À1 , a K M of 36 lM, and a k cat /K M of 1.5 9 10 9 s À1 ÁM À1 (Table 1). The catalytic efficiency was higher than 10 8 s À1 ÁM À1 , suggesting that DLac is a diffusion-limited enzyme [28]. We further compared the kinetic parameters of DLac with those of other fungal laccases, which have high oxidation efficiency of ABTS. The k cat of DLac is higher than those of other fungal laccases with the exceptions of POXA3a and POXA3a from Pleurotus ostreatus, while its K M value is average.
As ABTS is the mediator for laccase-mediator system applications, we have chosen this substrate to determine the thermostability of DLac. The T 50 and T m values of DLac are 63.3°C and 62.9°C that is close to its optimal temperature 65°C, respectively (Fig. 5C). The heat-induced conformational change increases the enzyme activity of DLac, similar to a combination of SDS and heat-induced bacterial laccase from Azospirillum lipoferum after catalytic activation [29]. Interestingly, two thermostable basidiomycete laccases, laccase Lac-4.8 from Physisporinus rivulosus and laccase I from basidiomycete PM1, showed thermal activation after preheating [23,30,31], suggesting that thermally induced transitions in laccases are relevant to enzyme activation. In contrast, previous studies have indicated that the optimal temperature for several enzymes is much lower than their T m [32][33][34].
To study the kinetic comparison between DLac and other laccases, we further purified three commercial laccases, Trametes versicolor laccase (TvLac), Agaricus bisporus laccase (AbLac), and recombinant Myceliophthora thermophila laccase (MtLac) (Fig. 3B-D), and compared their kinetics parameters for ABTS oxidation (Table 1). Different from AbLac, the catalytic efficiencies of DLac, TvLac, and MtLac were higher than 10 8 s À1 ÁM À1 , suggesting that these laccases are also considered to be diffusion-limited enzymes. Our results further indicate that, when the temperature increases from 30 to 65°C, both the k cat and K M of DLac, TvLac, and MtLac increase ( Table 1). This is consistent with previous findings that the substrate diffusion rate is the factor determining the catalytic efficiency of diffusion-limited enzymes [28]. Notably, using ABTS as the substrate, fungal laccases show catalytic efficiencies in a range from lM À1 Ás À1 to nM À1 Ás À1  [35,36]. The produced DLac with a high enzyme activity for ABTS oxidation displays high potential for industrial applications.

Overall structure
To investigate the relationship between the structure and the activity of DLac, the crystal structure of DLac was determined to a resolution of 1.38 A (Fig. 6A). The DLac crystals belong to space group P2 1 2 1 2 1 , with two molecules in an asymmetric unit. The current structure was refined to an R work of 12.6% and an R free of 15.0% ( Table 2). The overall structure of DLac is similar to the crystal structure of TvLac and consists of thirty bstrands, five a-helices, and five 3 10  (arrow) and basidiomycete species were compared with 18S sequence (A) and ITS sequence (B) data from GenBank using BLASTn search. The numbers in parentheses are accession numbers of the sequences. The sequences were aligned using Clustal W, and phylogenetic evolutionary analyses were conducted using MEGA6.06 by the neighbor joining method.
are also linked through a disulfide bridge between C85 and C484 (Fig. 6A). Three potential N-glycosylation sites, N432, N468, and N474, were predicted by the NetNGlyc 1.0 Server [25]. However, only two Nglycosylation sites, N432 and N468 in D3, were observed in the crystal structure of DLac (Fig. 6A). Both chains A and B observed in the crystal structure display identical glycosylation patterns.

Active site and functional loops
The active site of DLac consists of four copper atoms. A mononuclear T1 copper (T1Cu) is localized between D2 and D3, and a T2/T3 trinuclear cluster, including one T2 copper and two T3 coppers (T3 and T3 0 ), is located between D1 and D3 (Fig. 6A). The T1Cu is coordinated by H393, C449, and H454, and the T2/T3 trinuclear cluster is bound by eight highly conserved histidine residues in laccases (Fig. 6C). The T2 copper is coordinated by H64 and H396 and an axial water molecule, while each T3 copper is coordinated to three histidines (H111, H398, and H448 for the T3 copper and H66, H109, and H450 for the T3 0 copper) and one mutual oxygen molecule. The substrate-binding cavity of DLac was determined by superimposition with the laccase soaked with ABTS from Bacillus subtilis (Fig. 6D). Therefore, the SBLs of DLac were identified as loop I: 159-165, loop II: 264-270, loop III: 332-337, and loop IV: 386-393.

Comparative analysis of the sequence and structure among laccases
The protein sequence and crystal structure of DLac were compared with the known structures of basidiomycete laccases, including TvLac and laccases from Trametes trogii [37], Coriolopsis gallica [38], Lentinus sp. [12], Trametes hirsute [35], and Cerrena maxima [39] (Figs 7-9). The phylogenetic tree using the primary amino acid sequence of DLac locates it far from other basidiomycete laccases (Fig. 7A). To determine the precise differences between DLac and basidiomycete laccases, a structural superimposition was performed to calculate the root-mean-square deviation (RMSD) between the structures. Besides DLac, the structures of basidiomycete laccases are highly similar with a RMSD of 0.223~0.471 A for 481~499 Ca atoms. However, DLac is distinguished from other laccases with an RMSD of 0.491~0.553 A for 469~480 Ca atoms (Fig. 7B). To further identify the unique feature in the crystal structure of DLac, it was compared in detail to that of TvLac, which shows the highest sequence identity and the lowest RMSD value relative to DLac. Sequence alignment between DLac and TvLac indicates that different fragments appear in the N-terminal, C-terminal, and dispersed loops (Fig. 6B). The observation of conservation of the active site geometry but with differences in the loop combinations is consistent with previous studies [7,14]. Among the loops, we highlight the SBLs as targets, which directly affect the substrate diffusion into active site. All SBLs are different in either length or amino acid composition (Fig. 6B). Both DLac and TvLac show similar K M values, while a difference occurs in k cat (Table 1). We thus suggest that the difference in SBLs between DLac and TvLac may play an important role in determining the kinetic properties.

Substrate-binding loops of DLac
We further compared the SBL regions of DLac to other basidiomycete laccases (Fig. 8A). According to length and similarity of sequence, we classified SLB into three patterns. The heterodimeric laccases, POXA3a and POXA3b, from Pleurotus ostreatus which displays  highest activity [36] were also included by sequence alignment result of their large subunit that is identical between isoforms [40]. The SBL IV in DLac and Pleurotus ostreatus laccases are shorter than those in others, with deletion of two and five amino acids, respectively. Additionally, the SBL IV of DLac is rich in glycine in comparison with other laccases, providing more flexibility at the active site, similar to a previous study on HIV-1 integrase, which showed higher enzyme activity after glycine-to-alanine substitutions on the active site loop [41]. To identify the functional residues involved in substrate binding, the crystal structure of DLac was superimposed with that of laccases from Trametes trogii and Bacillus subtilis complexed with their substrates, methylbenzoic acid (4MA, Fig. 8B) and ABTS (Fig. 8C), respectively. The nonpolar residues V162, L265, F332, F337, L389, and G390 in the SBLs interact with the 4MA by hydrophobic contact. L389 on loop IV of DLac supplies a larger side chain than the corresponding proline residue in other laccases. Additional five residues, N264, V388, G391, P392, and H393 of DLac, are involved in ABTS-binding. Among these residues, the side chains of V162, L265, and G391 of DLac are smaller than the corresponding residues from other laccases. Previous studies on TvLac suggested that nonpolar amino acids with smaller side chains give more space in the binding cavity for the entry of larger substrates [42,43]. The contact between large substrate, ABTS, and SBL IV is much closer than that of small substrate 4MA. DLac and Pleurotus ostreatus laccases containing substrate-binding residue 162V, and shorter loop IV show higher k cat than that of other laccases (Fig. 8D). In contrast to Pleurotus ostreatus laccases, the V388, L389, and G390 on loop IV of DLac may contribute to increase ABTS-binding affinity. Moreover, the kinetic study of DLac indicated the substrate affinity of small substrates (DMP and guaiacol) is lower than that of large substrate (ABTS) ( Table 1). This property is consistent with prior studies of Pleurotus ostreatus laccases [36]. In view of the fact that the turnover rate of diffusion-limited enzymes depends on the diffusion rate of substrates, we suggest that the smaller side chains of V162, L389, and G341 and the shorter loop IV of DLac generates a larger cavity for  Although the SBLs were similar, the glycosylation plays another important role in determining the kinetic property, for example, isoforms from Pleurotus ostreatus that display identical SBLs but show difference between glycosylation leads to twofold difference in k cat [36,40]. Here, the glycosylation site of DLac is another distinctive feature from other laccases. Two asparagine residues (N432 and N468) were glycosylated with a disaccharide and a monosaccharide, respectively (Fig. 6A). The first N-acetylglucosamine (N-GlcNAc) on N432 has hydrophobic interaction with the side chain of A385, which is contiguous to the SBL IV, mainly stabilizing loop IV (Fig. 9A). The second GlcNAc on N432 interacts with the side chain of N327 by hydrogen bonding, which is extended by SBL III, indirectly stabilizing loop III. Therefore, the glycans attached to N432 result in stabilizing the substrate-binding pocket. The function of the two-first GlcNAcs on N432 in DLac is consistent with the study of Lentinus sp. laccase [12]. The mono-GlcNAc on N468 interacts with N361, T373, and V467 via hydrogen bonds and hydrophobic interactions (Fig. 9B). All these residues that interact with the mono-GlcNAc on N468 are localized in D3, resulting in stabilization of the protein structure.
We further compared N-glycosylation sites at position 54, 217, 432, and 468 between DLac and basidiomycete laccases (Fig. 9C). N432 is highly conserved in all basidiomycete laccases [12], while N468 is found only in DLac. Two N-glycosylation sites, N54 and N217, commonly observed in basidiomycete laccases, were replaced by aspartic acids at D54 and D217 in DLac. D54 in DLac, which indicates that glycosylation at N54 is not important for enzyme activity, was also reported as nonessential for enzyme activity using sitedirected mutagenesis [12]. Furthermore, the previous study also indicated the glycosylation at N217 in Lentinus sp. laccase plays an important role in stabilizing D2-D3 loops [12]. However, no glycosylation site at D217 was found in DLac. We thus superimposed the D2-D3 loops between DLac and Lentinus sp. laccase (Fig. 9D). The result indicated the one amino acid shorter and threonine-rich region, TTTQTT, in DLac was corresponding to ATAPPSA in Lentinus sp. laccase. The residue of T291 in DLac forms hydrogen bond with D214 and shares hydrophobic contact with Y257. Both D214 and Y257 are conserved between DLac and Lentinus sp. laccase. Additionally, the glycosylation level of DLac, which is relatively lower than other laccases (Fig. 9C), is estimated from the difference between the 55.0 kDa obtained from SDS/PAGE analysis and its calculated size of 52.9 kDa. All the Nglycans detecting from the crystal structure of DLac functionally stabilize loop interactions in the substratebinding site and in D3, which differs from other basidiomycete laccases.

Effect of glycosylation on the enzyme kinetics of DLac
To understand the effect of glycosylation on the kinetic efficiency, we deglycosylated DLac using   treatment with endoglycosidase H (Endo H) and peptide-N-glycosidase F (PNGase F). The Endo H treatment, which leaves only one remaining GlcNAc on the glycosylation sites, had no significant effect on k cat / K M . In contrast, the catalytic efficiency decreased by 41% when all the glycan was cleaved by PNGase F (Fig. 10A). This result is consistent with the study of deglycosylation and site-directed mutagenesis in Lentinus sp. laccase [12]. Both Endo H and PNGase F deglycosylation led to similar decreases in the k cat of DLac (Fig. 10B). The substrate affinity (45% decrease in K M ) was increased by Endo H treatment, while it did not change with PNGase F treatment (Fig. 10C). We conclude that the presence of the first GlcNAc residue in DLac is essential for its high catalytic efficiency. Clarification of the functionality of glycans in DLac is helpful in choosing the proper heterologous expression system.
In conclusion, the high-efficiency laccase, DLac, has been isolated from Cerrena sp. RSD1, and the protein yield was fivefold improved using a submerged culture in the presence of the inducer, 2,5-dimethylaniline. This diffusion limited DLac consists of the small hydrophobic residues in SBL I, II, and IV and a short SBL IV, which provides a broad cavity for substrate-binding, hence enhancing the diffusion rate of large substrates. Structural evidence also indicates its N-glycans stabilize the SBLs III and IV and the D3. The maintenance of the first GlcNAc is essential for enzyme efficiency. This study presents the large substrate-binding cavity, and the low glycosylation level of DLac contributes to its high efficiency to oxidize bulky substrate, ABTS, and these structural features can be used for protein engineering.

Enzyme activity and kinetic studies
The standard laccase activity was determined by spectrophotometric measurements using the Synergy TM Mx Monochromator-Based Multi-Mode Microplate Reader (BioTek). The standard reaction mixture contains 198 lL of 0.5 mM ABTS (e 420 = 3.60 9 10 4 M À1 Ácm À1 ) in pH 3.0 McIlvaine's buffer and 2 lL of the enzyme solution. The enzyme assay was performed at 30 or 65°C. One unit of enzyme activity was defined as 1 lmole of substrate transformed to product per minute [44]. The kinetic study of DLac was performed using 1.25~800 lM ABTS at pH 3.0, 3.13~2000 lM DMP (e 477 = 1.48 9 10 4 M À1 Ácm À1 ) at pH 4.0, and 18.78~3000 lM guaiacol (e 465 = 1.2 9 10 4 M À1 Ácm À1 ) at pH 4.0. The initial rates were acquired from the linear portion of the experimental curve, and the kinetic parameters were determined using the Michaelis-Menten model in SigmaPlot 10.0.

Submerged cultivation and laccase production
The Cerrena sp. RSD1 was cultured on 3.9% potato dextrose agar (PDA, Becton, Dickinson and Company) plates. To screen for the fungal strains exhibiting the highest laccase activity, we used PDA plates containing 0.04% guaiacol (Sigma#G5502) or RBBR (Sigma#R8001). For laccase production, twelve 4-day-old mycelial plugs (8 mm in diameter) were inoculated into 6.0 g of autoclaved rice straw submerged in 60 mL of minimal medium (10.0 gÁL À1 glucose and 4.4 gÁL À1 ammonium tartrate, pH 5.5) and 0.4 mM CuSO 4 . The fungal cultures were incubated in the bottles agitated at 100 r.p.m. and 30°C for 16 days. On the fourth day, 20 mM 2,5-dimethylaniline (Sigma#102253) was added as the inducer. The enzyme production was monitored by measuring the laccase activity in the supernatant.

Cloning of the DLac sequence
Genomic DNA was extracted using the WizardÒ Genomic DNA purification kit (Promega); total RNA was isolated with the TRIzolÒ Reagent (AmbionÒ). Total cDNA was labeled with an anchor sequence by RT-PCR using Super-ScriptÒ III Reverse Transcriptase (Novagen) and the RT-polyT primer. To clone the laccase gene from the Cerrena sp. RSD1, the cDNA of DLac was specifically amplified using the degenerated primer NP2. Other degenerate primers Cu1F, Cu3R, and Cu4R, which were designed from    conserved copper-binding regions 1, 3, and 4 in fungal laccases [45], were used for DLac sequencing. To obtain the signal peptide sequence, 5 0 -RACE was performed with SMART II TM A oligonucleotide, the GSP1 primer and 5 0 PCR Primer II A. The ORF and the full gene of DLac were amplified using primers DLacF and DLacR from the total cDNA or genomic DNA. All the PCRs in this study were amplified using TAKARA Ex TaqÒ DNA polymerase (TaKaRa Bio, Inc.) and performed in a T3000 Thermocycler (Biometra, Goettingen, Germany). The primers used for the PCR amplification are listed in Table S1.

Purification of laccases and deglycosylation analysis
Rice straw was submerged in 20 mM Tris/HCl buffer (pH 8.0) at 4°C overnight to wash out the enzyme absorbed on the rice straw. The supernatant was filtrated through INSIDE C eRAM minifilter MSKTB060010070-0.14 l (TAMI) and concentrated using a Xampler Ultrafiltration Cartridge (UFP-10-E-4MA, GE Healthcare) with the QuixStand System. After concentration, DLac was purified with a HiTrap Q column and a HiPrep 16/60 Sephacryl S-100 high-resolution column. The purity of the protein was verified by SDS/PAGE. Three commercial laccases, including TvLac (#51639 and #38429), AbLac (#40452), and MtLac (#SAE0050), were purchased from Sigma-Aldrich. All these commercial laccases were dialyzed in 20 mM Tris/ HCl (pH 8.0) and purified by a HiTrap Q column and Superdex 75 10/300 GL column. All the laccases were stored in 20 mM Tris/HCl with 150 mM NaCl (pH 8.0). For deglycosylation assays, purified DLac was treated with either Endo H (New England BioLabs) or Remove-iT TM PNGase F (New England BioLabs) for 24 h at 37°C according to the manufacturer's instructions. The residual activity was measured by standard enzymatic assay.

Effects of pH and temperature on the enzyme activity
The pH and temperature effects on the enzyme activity were evaluated by preparing the substrate suspensions in McIlvaine's buffer in the pH range of 2.2-8.0 at 30°C and in the temperature in the range of 30-65°C at pH 3.0 for ABTS and at pH 4.0 for 2,6-DMP and guaiacol. The T 50 value was assessed in a gradient thermocycler (Biometra T  The D2-D3 loops of DLac (green) and Lentinus sp. laccase (orange) were superimposed. The residues D214 (yellow), Y257 (yellow), and T291 (green) in DLac are represented in a stick model. The protein sequence from DLac aligned with Lentinus sp. laccase. Residues in g-helices, b-strands, and loops are shown as red cylinders, blue arrows, and black lines, respectively. The loops involved in substrate binding are marked with green lines, and the linkage between domains is marked with gray lines and labeled the D2-D3. The residues at 217 and residues at 214, 257, and 291 are marked with red and green numbers on the amino acid sequence, respectively. The TTTQTT region in DLac and its corresponding region in Lentinus sp. laccase are colored in green and orange background, respectively.
Gradient Thermocycler), corresponding to the temperature at which the enzyme retains 50% of its activity after a 10min incubation. Subsequently, 176 lL aliquots were used in triplicate assays at temperatures between 25 and 82°C. After a 10-min incubation, samples were immediately chilled in ice water for 10 min. Thereafter, 2 lL of each sample (the protein concentration was 5 lgÁmL À1 ) was used to measure the enzyme activity.
Near-UV circular dichroism (CD) spectropolarimetry CD experiments were performed on a Jasco-815 spectropolarimeter equipped with a six-position Peltier temperature control system (PTC-4245). The sample was prepared with 40 lM DLac in 4 mM Tris/HCl and 30 mM NaCl buffer (pH 8.0). The CD spectra were measured between 260 and 320 nm in a 1.0 mm light path length quartz cuvette with a step size of 1.0 nm. The signals at 280 nm were recorded during heating from 25 to 95°C at 1°C intervals and at the rate of 1°C per minute.

Crystallization and data collection
Crystals of the DLac were grown by mixing 1 lL of protein (15 mgÁmL À1 ) with 1 lL of reservoir solution using the sitting-drop vapor diffusion method at 20°C. The DLac crystals were obtained in a reservoir solution of 1.4 M ammonium sulfate, pH 7.6. All the crystals were flash-frozen with 20-25% glycerol as a cryoprotectant, and the diffraction patterns were recorded at cryogenic temperatures. The diffraction data were collected at a wavelength of 1.000 A using the synchrotron BL15A beamline of NSRRC in Taiwan, with a MX300HE CCD detector. The diffraction data were processed and scaled using the program HKL2000 [46].

Structure determination and refinement
The DLac crystal structure was determined by molecular replacement using the program MOLREP of the CCP4 program suite [47], and the crystal structure of TvLac [48] (PDB ID: 1GYC) was used as a search model. Throughout the refinement, 5% of randomly selected data were set aside for cross-validation with the R free values. Manual modifications of the models were performed using the program Coot [49]. Difference Fourier (F o -F c ) maps were calculated to locate the solvent molecules. All the crystal structures were refined using Refmac5 [50], including the individual isotropic B-factor refinement. The data collection and final model statistics are shown in Table 2. The molecular figures were produced using UCSF Chimera [51]. The atomic coordinates and structure factors of the DLac crystal structure have been deposited in the PDB ID: 5Z1X.