Engineering Bifunctional Laccase-Xylanase Chimeras for Improved Catalytic Performance*

Background: Rational design methods can be used to create chimeric enzymes with novel catalytic combinations. Results: Bifunctional enzymes combining xylanase and laccase activities showed enhanced catalytic activity and stability. Conclusion: Formation of an inter-domain interface alters enzyme conformation that enhances catalytic performance of the chimera. Significance: Deeper understanding of structural principles of protein fusion can improve the design of novel catalysts. Two bifunctional enzymes exhibiting combined xylanase and laccase activities were designed, constructed, and characterized by biochemical and biophysical methods. The Bacillus subtilis cotA and xynA genes were used as templates for gene fusion, and the xynA coding sequence was inserted into a surface loop of the cotA. A second chimera was built replacing the wild-type xynA gene by a thermostable variant (xynAG3) previously obtained by in vitro molecular evolution. Kinetic measurements demonstrated that the pH and temperature optima of the catalytic domains in the chimeras were altered by less than 0.5 pH units and 5 °C, respectively, when compared with the parental enzymes. In contrast, the catalytic efficiency (kcat/Km) of the laccase activity in both chimeras was 2-fold higher than for the parental laccase. Molecular dynamics simulations of the CotA-XynA chimera indicated that the two domains are in close contact, which was confirmed by the low resolution structure obtained by small angle x-ray scattering. The simulation also indicates that the formation of the inter-domain interface causes the dislocation of the loop comprising residues Leu-558 to Lys-573 in the laccase domain, resulting in a more accessible active site and exposing the type I Cu2+ ion to the solvent. These structural changes are consistent with the results from UV-visible electronic and EPR spectroscopy experiments of the type I copper between the native and chimeric enzymes and are likely to contribute to the observed increase in catalytic turnover number.

The plant cell wall is a complex structure in which mixed linked glycans and lignin are closely associated in a compound structure that forms a protective sheath around the cellulose fibers (1,2). The degradation of this complex structure requires an enzymatic mixture that includes ligninases and hemicellulases, and consequently laccases and xylanases have attracted considerable interest and have been extensively studied. Laccases (benzenediol:oxygen oxidoreductase; EC 1.10.3.2) are important components of the lignolytic complex responsible for lignin decomposition (3) and are widely distributed among fungi, higher plants, and bacteria (4). The CotA laccase is a multicopper oxidase present in the endospore coat of Bacillus subtilis (5) and is composed of three greek key ␤-barrel cupredoxin domains and harbors a total of four copper atoms, which are classified on the basis of their spectroscopic properties as follows: one type 1 (T1) copper, one type 2 (T2), and two type 3 (T3) copper ions (6). The T1 copper ion gives the protein a blue color that is typical of multicopper proteins and is a mononuclear catalytic center involved in substrate oxidation. The T2 and T3 copper ions form a trinuclear center, responsible for O 2 reduction and the release of H 2 O (7,8). Laccases are exceptionally versatile enzymes, catalyzing reactions with different and frequently recalcitrant substrates, which lends these enzymes considerable biotechnological potential (9,10).
Xylan is the second most abundant polysaccharide found in nature, representing a large proportion of the hemicelluloses present in wood and graminaceous plant tissues (11). Xylanases (endo-1,4-␤-xylanase, EC 3.2.1.8) are important enzymes for hemicellulose degradation, hydrolyzing the ␤-1,4-glycosidic bonds between xylose residues in xylan (12); family 11 xylanases display a jelly roll fold including a single ␣-helix and two twisted ␤-sheets (13,14). In common with all GH11 xylanases, the structure of B. subtilis xylanase A (XynA) resembles a right hand, where the ␤-strands form the finger and palm domains that constitute the active site cleft, access to which is regulated by movements of a flexible ␤-turn structure denominated as the "thumb" domain (15,16).
Various applications of laccases, xylanases, and laccase/xylanase combinations have been reported for improved processing of lignocellulosic feed stocks. These include the reduction in the levels of chlorine-based oxidants in the paper and pulp bleaching process with the consequent reduction in pollutants (17,18) and the use of these enzymes to improve the biochemical and rheological properties of wheat flour, gluten, rye, oats, and other cereals for the baking industry (19,20). Synergism between laccase and xylanase may also benefit the animal feed sector by increasing the digestibility of enzyme-treated lignocellulosic material in the animal diet (21). Furthermore, these enzymes can be used to generate biofuels from biomass during the delignification process that is necessary to produce second generation ethanol from lignocellulose feed stocks (22,23).
During evolution, genes that encode proteins with related functions have frequently undergone fusion events, generating multifunctional enzymes (24,25). Biochemists and bioengineers share a common goal of understanding protein evolution, and protein engineering methods have been used to mimic processes observed in nature (26). Engineering multidomains enzymes that are capable of catalyzing two or more reactions is a potential strategy to reduce enzyme costs in industrial processes, as including multiple catalytic properties in a single polypeptide simplifies production and purification operations for the recombinant product. Moreover, when enzymes are fused to form chimera, the physical proximity between the resulting catalytic domains may increase the rate of the reactions, and chimeric enzymes frequently have an improved performance as compared with the activities of the separate parental enzymes (27). Hybrid enzymes are created by recruiting functions of existing enzymes and fusing them by structurebased design (28); therefore, these enzymes (or their fragments) can potentially serve as building blocks for the creation of polypeptides with combinations of catalytic activities not found in nature (29).
Chimeric proteins have been constructed for various purposes, including improved stability (30,31), the creation of new substrate specificities (32,33), and increasing substrate affinity for enhanced biosensor sensitivity (34,35). The creation of bifunctional enzymes acting on lignocellulosic materials has been shown to be a practical option for the improvement of enzymes involved in various processes, including biomass degradation (36). A key aspect in the construction of enzyme chimeras is the preservation or improvement of the catalytic characteristics of the parental proteins. A widely used approach involves the "end-to-end" fusion between the N and C termini of the parental enzymes; however, this technique may result in nonfunctional chimeras either due to misfolding or to restrictions resulting from steric hindrance between the domains (37). In some cases, these restrictions may be overcome by the insertion of a loop at the fusion site to increase inter-domain flexibility and maintain functionality of the parental enzymes (38,39).
A key factor for success in the creation of enzyme chimeras is the compatibility between the two catalytic activities with respect to their physicochemical requirements such as pH and temperature optima. Here we present the design, construction, and characterization of two novel multifunctional enzymes that combine laccase and xylanase activities. With the aim of improving the catalytic overlap of the two activities in the chimeras, swapping of the xylanase domain was used to introduce a thermophilic variant, which improved the compatibility between the catalytic functions.

EXPERIMENTAL PROCEDURES
Rational Design of the Bifunctional Enzyme-We reasoned that the surface loops in the B. subtilis laccase (CotA) were the most favored sites for insertion of the endo-␤-1,4-xylanase (XynA) to obtain a chimera with minimum disruption of the native conformation of the parental enzymes. A survey of the three-dimensional structure of the crystal structure of the CotA (PDB code 1GSK) restricted the insertion sites on the basis of possible steric hindrance between the two protein molecules or unfavorable orientation of the active sites after fusion of the two polypeptides. In addition, the location of the insertion was further restricted by considering loops with elevated crystallographic B-factors with the aim of maintaining flexibility in the inter-domain linking regions. On the basis of this analysis, the XynA was inserted between residues Ser-216 and Pro-217 in the laccase using the high resolution structures of xylanase (PDB code 1XXN (15)) and laccase from B. subtilis (PDB code 1GSK (40)) as templates for building a structural model of the CotA-XynA chimera by comparative modeling techniques with the program Modeler (24). The structural model was validated utilizing the program Procheck (41), which showed that 91% of residues were in highly favored regions of the Ramachandran graph.
Molecular Dynamics Simulations-Molecular dynamics simulations (MDS) were performed using the GROMACS simulation package version 4.5.3 (42) with the GROMOS96 (43A2) force field (43). The crystal structure of the B. subtilis CotA laccase (PDB entry 1GSK) and the CotA-XylA chimera (constructed as described above) were used as the starting structures. The simulation boxes for the laccase (x ϭ 10.0 nm; y ϭ 8.6 nm; z ϭ 9.2 nm) and the chimera (x ϭ 13.80 nm; y ϭ 10.0 nm, z ϭ 9.8 nm) included five and four Na ϩ ions, respectively, to maintain electrical neutrality. The "leapfrog" algorithm (44) was used to integrate the equations of motion with a time step of 2.0 fs over a total time of 40.0 ns. In all simulations, periodic boundary conditions and a minimum image convention were applied, and the atom neighbor list (45) within 1.4 nm was updated every 10 steps. The Particle-Mesh Ewald technique (46) was used for the treatment of electrostatic interactions, with a Particle-Mesh Ewald order of 4 and a Fourier spacing of 0.12. The system was maintained at a constant temperature of 300 K using the V-rescale thermostat with a coupling time (42) of 0.2 ps. Both simulations were performed with the numbervolume-temperature (NVT) ensemble (47). Bond lengths involving hydrogen atoms were constrained by means of the LINCS algorithm (48), and the SETTLE algorithm (49) was used to constrain water molecules. The imidazole groups of histidines (and their equivalents in the chimera) form contacts with the four Cu 2ϩ ions and were protonated (using either HISA or HISB in the GROMOS force field). To maintain the nitrogen to Cu 2ϩ contacts and histidine positions as observed in the laccase crystal structure, the positions of the four Cu 2ϩ ions together with the nitrogen, sulfur, and carbon atoms in side chains in immediate contact (His-106, -152, -154, -604, -607, -609, -676, and -682; Cys-677 and Ile-679) were maintained in their original positions during the simulation. The root mean square deviation (r.m.s.d.) 2 was calculated using the g_rms program within the GROMACS 4.5.3 package. The interaction potential was obtained from the simulation trajectory using the program mdrun (with the "rerun" option) considering only short range potentials at a distance of Ͻ1.0 nm.
Construction of the Chimera-Genomic DNA of the PAP1158 strain of Bacillus subtilis (BGSC 1A1 code 168) was used as the template for PCR amplification of the xynA gene encoding endo-␤-1,4-xylanase as described previously (50). The laccase gene, cotA (GenBank TM accession number 936023 (51)), was amplified from genomic DNA by PCR using primers P1 (5Ј-aaggaaagcttgctagcACACTTGAAAAATTTGTGGAT-GCTCT-3Ј) that included the 5Ј-end of the cotA (uppercase) and restriction sites (underlined) for HindIII (boldface) and NheI (italic), and P6 (5Ј-GTAATACGATAAATAGGATC-CAAAGGTTTCATAAAG-3Ј), encoding the 3Ј-end of the cotA and including a BamHI site (underlined). The 1624-bp PCR product was digested with HindIII and BamHI, cloned into plasmid pT7T3-18U (Amersham Biosciences) to yield pTcotA, and after full nucleotide sequencing was subsequently subcloned into pET28a(ϩ) (Novagen) using the NheI and BamHI restriction sites to generate pETCotA. Cloning of the endo-1,4-␤-xylanase from B. subtilis (xynA) and creation of a thermophilic variant of the same enzyme (XynAG3) by directed evolution have been previously described (50,52). Both the native XynA and XynAG3 coding sequences were subcloned into the NheI and BamHI sites of pET28a(ϩ) to create pETXynA and pETXynAG3, respectively.
The scheme for the construction of the chimeric enzymes is presented in Fig. 1. The cotA gene was divided into two fragments as follows: a 661-nucleotide fragment encoding the N-terminal 216 amino acids up to Ser-216 (fragment A), and the rest of gene encoding Pro-217 to the C-terminal residue (fragment B). Fragment A was amplified by PCR using the pTcotA as a template with primers P1 and P2 (5Ј-ggtcactgCA-GAAGGGTTTTCCGGTGCGCT-3Ј, PstI site underlined), which introduced a PstI restriction site at the 3Ј extremity of the amplified fragment. Fragment B was amplified with P5 (5Ј-agtggggCCCTCACTGCCTAATCCTTCAAT-3Ј) and P6 primers, where P5 introduced an ApaI site (underlined). Fragment A was digested with HindIII and PstI and fragment B with ApaI and BamHI. The parental XynA and XynAG3 xylanase variants were amplified by PCR with P3 (5Ј-agctcctgcaGC-CAGTACAGACTACTGGCAAA-3Ј) and P4 (5Ј-cccacgggC-CCCCACACTGTTACGTTAGAA-3Ј) primers, using the cloned genes in pT7T3-18U as a template, which introduced restriction sites (underlined) for PstI and ApaI, respectively. The amplified xylanase fragment was digested with PstI and ApaI and mixed with digested fragments A and B, and the mixture ligated into HindIII-and BamHI-digested pT7T3-18U creating the full-length pTcot-xyn. The final construct was sequenced, and the HindIII-BamHI fragment was subcloned into pET28a(ϩ).
Expression and Purification of Recombinant Enzymes-Both the XynA and XynAG3 were expressed in E. coli BL21 (DE3) pLysS transformed with pETXynA or pETXynAG3 grown in HDM medium containing (per liter) 25 g of yeast extract, 15 g of tryptone, 1.2 g of MgSO 4 , supplemented with 40 g⅐ml Ϫ1 kanamycin and 34 g⅐ml Ϫ1 chloramphenicol. The cells were grown at 37°C to A 600 of 0.6, and expression was induced with 1 mM isopropyl ␤-D-thiogalactopyranoside for 5 h. Cells were harvested by centrifugation (8000 ϫ g, 4°C, 5 min), and the cell pellets were retained for further processing. The parental CotA and CotA-XynA and CotA-XynAG3 chimeras were expressed as described previously (53), using E. coli Rosetta (DE3) grown in HDM with kanamycin (40 g⅐ml Ϫ1 ) and chloramphenicol (34 g⅐ml Ϫ1 ). After 24 h of induction, the cells were harvested by centrifugation (8000 ϫ g, 4°C, 5 min). For all the recombinant enzymes, whole-cell extracts were prepared from cell pellets by ultrasonication in 4% of the original culture volume of lysis buffer (100 mM HEPES (pH 7.5), 50 mM NaCl, 0.5 mM phenylmethylsulfonyl fluoride, 1% Triton X-100, and 10 mM imidazole). The cell extracts were cooled on ice and cleared of cell debris by centrifugation (10,000 ϫ g, 4°C, 30 min). The supernatants were loaded on an immobilized metal affinity column (GE Healthcare) pre-equilibrated with a buffer containing 100 mM HEPES, 50 mM NaCl, and 10 mM imidazole (pH 7.5). The column was washed with buffer containing 100 mM HEPES (pH 7.5), 300 mM NaCl, and 40 mM imidazole for xylanases, 50 mM for laccase, and 60 mM for the chimeras, until no further reduction in the A 280 was observed. Protein was eluted with 300 mM imidazole, and protein samples were dialyzed against 20 mM Tris-HCl (pH 8.0) and 200 mM NaCl and stored at 4°C for future use. Laccase and chimeras were concentrated with a 50-kDa cutoff membrane (Centricon, Millipore, São Paulo, Brazil), and protein concentrations were determined by measuring the A 280 .
The effect of pH on xylan hydrolysis by the purified XynA and XynAG3 was measured using the same buffers as for the laccase assays, with 1% (w/v) oat spelt xylan substrate (Sigma) as described previously (50). The effect of temperature on xylanase activity was conducted at temperatures between 20 and 85°C in 100 mM MES-HCl, pH 6.5, for the native xylanase and the CotA-XynA fusion and in 100 mM MOPS-NaOH, pH 7.0 for the XynAG3 xylanase and CotA-XynAG3 fusion. Thermostability was determined by incubation of ϳ60 g of the purified enzymes at 70°C, and residual activity was measured in aliquots collected at different times.
Kinetic parameters for the purified laccase and laccase/xylanase chimeras were determined at 37°C in 100 mM sodium acetate buffer (pH 4.5) using increasing concentrations of ABTS (5-200 M) in 1-cm cuvettes. Reactions were initiated by addition of 20 nM of the purified proteins, and the rates of oxidation obtained from the linear portion of the reaction curve by the slopes (⌬A/⌬t) were measured from 96 to 150 s. The xylanase kinetic parameters were determined using the oat spelt xylan substrate in concentrations ranging from 1 to 12 mg/ml. Reactions were performed at the temperature and pH optima and initiated by addition of 30 nM of the purified enzymes. All enzymatic activities were determined in triplicate, and the maximum velocity (V max ), apparent dissociation constant (K m ), and catalytic constant (k cat ) were calculated by nonlinear regression fitting of the data to the semi-logarithmic form of the Hill equation using the software SigrafW (55). Data are reported as the mean of duplicate determinations, which differed by less than 5%, as verified by the Tukey test.
Biobleaching of Cellulose Pulp-Dried cellulose pulp (20 g) obtained from Eucalyptus grandis was prepared at 10% consistency in 100 mM NaCl and 50 mM MES (pH 6). A total of 5 nmol of enzyme (or enzyme mixture)/g of dried pulp and 1 mM ABTS were added, and the samples were incubated in sealed polyethylene bags at 60°C for 24 h. Subsequently, the treated pulp was filtered through a Büchner funnel, and the absorbance of the filtrate was measured at 237 nm to determine phenolic chromophore release (56). The pulp was washed with water and submitted to alkaline extraction using 1.5% NaOH at 60°C for 1 h, followed by washing with 1 liter of distilled water. The number of the washed pulp was determined, which was defined as the volume of a 0.1 N KMnO 4 solution consumed by 1.0 g of dried pulp under standard conditions. Control experiments were performed in which pulps were treated under the same conditions but in the absence of enzyme. All the procedures and calculations were carried out according to TAPPI (Atlanta, GA) methodology (57).
UV-visible Spectroscopy-UV-visible absorption spectra of laccase and laccase/xylanase fusions were measured at 25°C between 300 and 800 nm in 1.0-cm quartz cuvettes using a Cary 50 UV-visible spectrophotometer. Measurements were performed using ϳ2 mg/ml of protein in 10 mM HEPES buffer (pH 7.5) and 100 mM NaCl.
Electron Paramagnetic Resonance-Low temperature (10 K) EPR experiments were carried out in a Bruker ELEXSYS E580 spectrometer operated at X-band (9.5 GHz). The temperature was controlled using an Oxford ITC 503 cryostat. Aliquots of 50 g in 50 l of each sample were drawn into a quartz EPR tube and quickly frozen in liquid nitrogen prior to the experiment. Experimental conditions were set to optimize the signal-tonoise ratio without saturating and/or distorting the EPR spectra. Experimental parameters were as follows: modulation frequency, 100 kHz; modulation amplitude, 0.4 millitesla; microwave power, 8 milliwatts. Theoretical spectra were obtained by simulating the EPR spectra using a spin Hamiltonian taking into account the electron Zeeman (g-value) and hyperfine (A-value) interactions as implemented in the Easy-Spin software (58). Additional broadening mechanisms are available in the program and were used to achieve the best agreement between experimental and calculated spectra.
Small Angle X-ray Scattering-Small angle x-ray scattering (SAXS) data were collected using a 165-mm MarCCD detector on the D02A/SAXS2 beamline at the Brazilian Synchrotron Light Laboratory. The radiation wavelength was 1.48 Å, and the sample-to-detector distance was 1415.9 mm to give a scattering vector-range from 0.10 to 2.3 nm Ϫ1 (q ϭ 4sin/, where 2 is the scattering angle). Samples of the CotA-XynA chimera at 2.0 mg/ml in 20 mM Tris-HCl (pH 7.0) and 50 mM NaCl were centrifuged for 10 min at 20,000 ϫ g at 4°C and filtered to remove any aggregates. Two sequential 600-s exposures were compared with monitor radiation damage. Background scattering was subtracted from the protein scattering pattern, which was then normalized and corrected. Buffer base lines were collected under the same conditions immediately after the protein sample to guarantee an accurate solvent correction. Data analysis was performed using GNOM (59). A series of models of the low resolution envelope of the chimera were determined using ab initio modeling as implemented in the program DAMMIN (60), and a final averaged model was generated from these envelopes using the DAMAVER (61) suite of programs. The SAXS envelope and the chimera structure from the molecular dynamics simulations were superimposed using the program SUPCOMB (62).

Design, Construction, and Expression of Chimeric Genes-
The fusion of the 1642-bp laccase (CotA) with either the 555-bp xylanase (XynA) or the thermostable variant (XynAG3) were performed by insertion of the xylanase into a surface loop of the laccase. The resulting chimeric constructs of 2197 bp contain a central region composed of the XynA or XynAG3 sequence flanked by the regions of the CotA encoding the N-terminal residues 1-216 (forming the 5Ј region of the chimera) and the Bifunctional Laccase-Xylanase Enzymes DECEMBER 16, 2011 • VOLUME 286 • NUMBER 50
Catalytic Activities of the Parental and Chimeric Enzymes-The chimeric enzymes displayed both xylanase and laccase catalytic activities, and the effect of temperature and pH of the parental and bifunctional enzymes are presented in Fig. 3. Although the maximum laccase activity was observed at pH 4.5 for the parental laccase and for both the CotA-XynA and CotA-XynAG3 enzymes (Fig. 3a), the chimeric enzymes showed a higher activity than the parental enzyme at pH 3.0 and 5.0. A slight decrease from 80 to 75°C in the optimum temperature of the laccase activity was observed in the chimeric enzymes (Fig.   3b). In the thermotolerance assays of laccase activity (Fig. 3c), the parental enzyme and both chimeras were thermoactivated, and during the first 10 min of incubation at 80°C the catalytic activity increased 2.5-and 6-fold for the parental and chimeric enzymes, respectively. After activation, the half-lives of the laccase activity in the parental CotA, the CotA-XynA, and the CotA-XynAG3 were 154, 133, and 138 min, respectively, when incubated at 80°C.
As shown in Fig. 3d, the maximum xylanase activity for parental native xylanase (XynA) and the CotA-XynA chimera was pH 6.5, as compared with the thermostable xylanase (XynAG3) and the CotA-XynAG3 fusion, which presented a maximum activity at pH 7.5. The effect of temperature on the xylanase activity was measured at the pH at which maximum activity was observed for each enzyme, and Fig. 3e shows that the temperature optima for the XynA and CotA-XynA were 55 and 60°C, respectively, whereas the optimum xylanase activity of the XynAG3 and CotA-XynAG3 was 70°C. The thermostability of the XynAG3 and CotA-XynAG3 enzymes was measured at 70°C (Fig. 3f), and under these conditions the parental XynAG3 lost 90% of its activity after 10 min, and after 30 min no activity was observed. In contrast, the chimeric CotA-XynAG3 retained 60% activity after 10 min and ϳ20% activity after 120 min. XynAG3 and CotA-XynAG3 half-lives (t1 ⁄ 2 ) at 70°C were 2.2 and 21 min, respectively, showing that the xylanase activity of chimera was more thermotolerant than the parental enzyme.
Kinetic Parameters of the Parental and Chimeric Enzymes-The kinetic properties of the purified parental and bifunctional enzymes were compared, and Table 1 presents the values for the apparent dissociation constant (K m ), turnover numbers (k cat ), and catalytic efficiency (k cat /K m ). The measured xylanase activity of the CotA-XynA and CotA-XynAG3 is essentially the same as that of the parental enzymes XynA and XynAG3, respectively. In contrast, the maximum of the laccase activity (V max ) of the CotA-XynA and CotA-XynAG3 chimeras was enhanced by ϳ2-fold as compared with the parental CotA enzyme, and consequently the k cat and the catalytic efficiency (k cat /K m ) increased by the same factor. The K m values for the laccase activity of all enzymes were similar.
Effect of the Enzymes on Cellulose Pulp Bleaching-The effect of the chimeric enzyme on the bleaching of cellulose pulp was evaluated by measurement of the absorbance at 237 nm and by reduction in the number. The absorbance of the pulp filtrate at 237 nm is a measure of the release of phenolic compounds liberated from the pulp (56), and Fig. 4a presents the A 237 values of the filtrate after alkaline treatment of pulp after treatment either with each of the parental enzymes (XynAG3 and CotA), with an equimolar mixture of the two enzymes (CotA ϩ XynAG3), or with the chimera (CotA-XynAG3). The effect of XynAG3 alone was ϳ10-fold lower than the CotA alone, and the effect of a mixture of the two parental enzymes was essentially additive demonstrating that the CotA activity played the major role in mobilizing aromatic chromophores. The release of phenolic compounds from pulp by the CotA-XynAG3 chimera was ϳ2.3-fold higher than the parental enzyme mixture, showing that the effects of the activities of the chimeric enzyme were synergistic. The 50% increase in the k cat value of the laccase activity in the chimera as compared with the parental CotA Unique restriction sites for PstI and ApaI were introduced at the junctions between the regions to facilitate domain exchange. The creation of the PstI site (introduced by the P2 and P3 primers) and ApaI site (P4 and P5 primers) resulted in the insertion of an alanine (A) and a glycine (G), respectively. For the construction of CotA-XynAG3 enzyme, the native xylanase (XynA) was replaced by thermophilic xylanase (XynAG3), using the PstI and ApaI sites. Sites for HindIII (H) and NheI (N) were included in primer P1, and for BamHI (B) in primer P6 were used to subclone the entire construct between plasmids. partially explains the increased effect, and additional factors such as the physical proximity of the two catalytic domains must also contribute to the full synergistic effect of the chimera. Fig. 4b presents the numbers of the treated pulps after alkaline extraction and shows a 46.6% reduction using the CotA-XynAG3 chimera, as compared with a 42.3% reduction with an equimolar mixture of the parental enzymes (CotA ϩ XynAG3). The effect of the parental enzyme mixture was similar to that of the xylanase (XynAG3) alone, which may be due to ability of the enzyme to modify the fiber and to expose the less accessible lignin fraction that is insoluble at the pH used for the enzyme assay, but it is extracted at high pH in the alkaline extraction step (63).  Table S1. The formation of the inter-domain interface and the contribution of individual regions to the stability of the final conformation was accompanied by monitoring the calculated potentials of these interactions (Fig. 5). The decrease in the potential energy of the interaction between the laccase and xylanase domains of the chimera occurs in a stepwise manner and is essentially complete by 13 ns, reaching a final value of Ϫ165 Ϯ 10 kcal⅐mol Ϫ1 in the final 5 ns of the simulation. The interface is formed after 13 ns and is presented in Fig. 6 (left-hand box), and it reveals that two loop regions in    (40)), indicating that this loop is highly flexible in the parental CotA enzyme. During the approximation of the two domains in the MDS, this loop participates in the formation of the inter-domain interface contributing Ϫ88 Ϯ 7 kcal⅐mol Ϫ1 to the stability of the final structure. After 13 ns, the potential is unchanged, indicating the conformation of the loop has reached a stable conformation. The r.m.s.d. profile of the loop 1 residues are highly variable during the initial 12 ns of the simulation and thereafter are significantly reduced, which is consistent with the immobilization of loop 1 on formation of the extensive inter-domain contacts. The concomitant decrease in potential that accompanies interface formation and the large energy difference between the parental and chimeric enzymes together with the low final r.m.s.d. value all suggest a major role for loop 1 in the formation and stabilization of the inter-domain interface.
The interaction potential profile of loop 2 residues with the xylanase domain shows a gradual decrease from an initial value of 0 to Ϫ45 Ϯ 5 kcal⅐mol Ϫ1 in the final 5 ns of the simulation, indicating that during the approximation between the two domains the conformation of the loop is gradually altered. A potential calculation of the interaction of the loop 2 residues with all other residues in both domains of the chimera shows that on reaching the final position the total energy stabilization is approximately Ϫ60 kcal⅐mol Ϫ1 .
The movement of loop 2 in the chimera is highlighted in Fig.  6 (right-hand box), which compares the positions in the parental CotA (shown in yellow) and in the chimeric enzyme (shown in green). In the chimera, the altered position of the loop 2 results in a widening of the surface cleft that is adjacent to the external T1 Cu 2ϩ , resulting in improved access of the substrate that acts as the electron donor to this copper ion during the catalytic cycle.
Small Angle X-ray Scattering-The overall structural features predicted from the MDS were validated using SAXS. The x-ray scattering curve and the distance distribution function p(r) of the CotA-XynA chimera are shown in Fig. 7, A and B, and are consistent with a maximum dimension of 107 Å and a gyration radius of 32.10 Ϯ 0.16 Å. The SAXS-derived molecular shape of the CotA-XynA indicates that the chimeric protein is a monomer in solution and presents an elongated shape. The dimensions estimated from the SAXS data corroborate with those calculated from the average of 26 structural conformations sampled at 1-ns intervals over the 15-40-ns range of the MDS simulations. Superposition of the averaged MDS structure with the SAXS envelope (Fig. 7C) showed a good shape complementarity as assessed by the normalized spatial discrepancy (or parameter d) value of 1.37.
UV-visible Spectroscopy of Laccases-The MDS predicts an alteration in the position of the residues in loop 1, and this event is therefore likely to be associated with changes in the environment of the T1 copper. The T1 and T3 coppers of the parental CotA and chimeric enzymes were observed by UV-visible absorbance spectroscopy. The type 1 (T1) copper site is distinguished by an intense S 3 Cu(d x 2 Ϫy 2) charge transfer absorption band at around 600 nm, which confers the intense blue color to the parental CotA (7) and the chimeric enzymes. The T3, or coupled binuclear copper site, consists of a pair of tightly coupled copper ions with maximum absorbance at 330 nm (⑀ ϳ4000 M Ϫ1 cm Ϫ1 ) originating from a bridging hydroxide (53). The purified parental CotA, CotA-XynA, and CotA-XynAG3 chimeras all exhibited the blue color characteristic of the laccases, and the absorption spectra of the purified enzymes showed a band at 600 nm and a shoulder at ϳ330 nm corresponding to the T1 and T3 copper centers, respectively (Fig. 8).
In these experiments, the CotA-XynA and CotA-XynAG3 chimeras showed an increased absorption at 600 nm relative to the parental CotA. Because the A 600 may depend on the copper content of the enzymes, atomic absorption spectroscopy measurements were performed on all recombinant enzymes resulting in a ratio of 3.7 and 3.8 mol of copper/mol of protein for the parental CotA and chimeric enzymes, respectively. These results indicate that the copper sites in all the enzymes are nearly fully occupied and therefore that the increased A 600 in the chimeric enzymes is due to an alteration in the environment of the T1 copper center.
Electron Paramagnetic Resonance-EPR spectra of the parental CotA, CotA-XynA, and CotA-XynAG3 together with the best theoretical simulation obtained in each case are shown in Fig. 9. All spectra show two distinguishable components, which can be assigned to the T1 and T2 copper centers present in the protein structures and observed in previous studies (53,64,65). The parameters characterizing the electron Zeeman (g min , g med , and g max ) and hyperfine (A max values) are shown in Table 2. From these EPR spectra and simulations, it can be seen that the component attributed to T1 centers (dotted arrow in Fig. 9) could be fitted with the same magnetic parameters ( Table 2) for all three proteins. Conversely, the g-values of the T2 centers showed significant changes when comparing the parental CotA enzyme with the CotA-XynA and CotA-XynAG3 chimeras, where the major alterations are higher g min and lower g max values in the EPR spectra of both the chimeric enzymes. Overall. the changes in g-values tend to bring their values closer than observed for the parental CotA suggesting a less distorted symmetry around the copper ion in the T2 centers of CotA-XynA and CotA-XynAG3. The magnitudes of the hyperfine coupling between the copper electron and nuclear spins are not disturbed in any of the proteins.

DISCUSSION
Rational design methods were successfully applied to create two bifunctional enzymes with xylanase-laccase activity that demonstrated catalytic properties similar to the parental enzymes. The recombinant parental (CotA, XynA, and  XynAG3) and chimeric enzymes (CotA-XynA and CotA-XynAG3) were produced in E. coli at reasonable yields (more than 10 mg/liter of cell culture) and purified to homogeneity by nickel-affinity chromatography (Fig. 2). As observed previously (50,65), the recombinant parental CotA and XynA enzymes presented laccase or xylanase activities, and the CotA-XynA and CotA-XynAG3 chimera showed both laccase and xylanase activities, which demonstrated that the proteins retained catalytic function despite the insertion of an autonomous domain level folding units. These results are in accord with previous studies showing that the sequential order of secondary structure elements is not crucial for in vivo protein folding (28,66).
Results from the biochemical assays confirm that the parental enzymes and the chimeras had maximum catalytic activities at similar pH values (Fig. 3). However, both the chimeras showed an increase of laccase activity of 40% at pH 3.0 and 5.0 in relation to the parental enzyme (Fig. 3a), and the parental xylanases were more active at slightly higher pH values (50% activity at pH 9.0) (Fig. 3d). These differences can be attributed to changes in the microenvironments of the active sites that may lead to changes in the pK values of catalytic residues or by conformational changes resulting from the fusion of the enzymes (67). The differences between the effect of pH and temperature on laccase and xylanase activities of the parental enzymes posed an initial challenge to the fusion of these enzymes. At the pH optimum of one parental enzyme, less than 20% activity is observed in the other (Fig. 3, a and d), leading to potential compatibility problems for the combined activities in the chimeras. Nevertheless, after the fusion of the CotA with native XynA, the overlap in pH dependence for the two activities improves considerably, and at pH 5.0 the CotA-XynA chimera presents 96% of the maximum laccase activity and ϳ50% of the maximum xylanase activity.
Possible protein engineering strategies include in vitro evolution strategies to improve one or both of the individual enzymes prior to their fusion in the chimera. We evaluated the viability of this by constructing a laccase-xylanase chimera using a thermostable variant of the XynA (denominated XynAG3) that was previously created by directed evolution and that contains four mutations located on the surface of the molecule (52). Comparison of the effect of temperature on the xylanase activity of the CotA-XynA and CotA-XynAG3 showed consistent results. The temperature optimum is between 55 and 60°C for the parental XynA, which is maintained in the CotA-XynA chimera, and both the parental XynAG3 and CotA-XynAG3 showed an optimum temperature of 70°C (Fig.  3e). When the temperature compatibility of the laccase and xylanase activities is compared, the CotA-XynA chimera at 60°C presents 100% xylanase activity and 50% of laccase activity as compared with the CotA-XynAG3 chimera, which at 75°C presents 90% maximum xylanase activity and 100% laccase activity. These results demonstrate the viability of improving the compatibility between the temperature optima of the enzymes by protein engineering strategies.
The thermotolerance experiments for the laccase activity showed that both chimeric enzymes were similar to the parental CotA and that in all proteins the laccase activity was heatactivated, resulting in an increase in the laccase activity by ϳ2-fold (in the case of the chimeras) or 3-fold (for the parental CotA). The thermal activation phenomenon was observed in protein that had been previously stored at 4°C and subsequently heated to 70°C during the experiment. Protein that had been heated to 70°C for 10 min and subsequently cooled did not present the thermal activation effect (data not shown). This demonstrates that the conformational changes in the protein that results in the thermal activation are not reversible, and we suggest that the effect is due to a slow annealing of the recombinant protein in solution to adopt the final native structure that shows maximal catalytic activity. This suggestion is consistent with previous explanations of the thermal activation  effect, which correlated the effect to conformational changes that modulated the properties of the catalytic residues (or cofactors) such as pK a , exposure to solvent, or altered molecular interactions between the enzyme and substrate (68). The chimeric enzymes significantly increased the thermotolerance of the xylanase activity at higher temperatures and moreover proved to be resistant to proteolytic attack in vivo and, in addition, improved the solubility of the recombinant proteins. This stability may be partly due to the contacts established between the two domains and the junction of two connecting points in the rational design. The approximation between the two catalytic domains of the CotA-XynA chimera is a striking feature observed in the molecular dynamics simulations, and the close contact was confirmed experimentally in the SAXS experiments. The approximation between the two domains results in extensive inter-domain contacts that are mediated primarily through loops one and two in the laccase domain. The significantly reduced potentials calculated in the molecular dynamics simulations suggest that these inter-domain contacts make a significant contribution to the stabilization of the chimeric proteins. In end-to-end fusions, this stabilization usually is absent. In addition, the N-and C-terminal regions are known to influence the thermostability of the CotA laccase (69,70), and in the rational design strategy used in this study, both these regions were unaltered in the chimeric enzymes.
The kinetic characterization showed that the kinetic parameters of the xylanase activity showed no changes between the parental xylanases and chimeras, demonstrating that the stabilization of the xylanase domain in the fusion proteins did not alter the catalytic properties of the parental XynA and XynAG3. In contrast, although no significant changes in K m values of the laccase activity were observed, the chimeric enzymes showed a laccase catalytic efficiency that was ϳ2-fold greater as compared with the parental CotA, which was due to an increase in V max . The observed K m and k cat values for the laccase activity in this study were lower than previously reported values (53,65), possibly due to differences in the mathematical models adopted for determination of the catalytic parameters, differences in the methods of protein concentration and determination between the different studies, differences in the methodology of protein expression, and differences in the O 2 concentration in the catalytic assays or storage of the enzyme.
The bifunctional enzyme had a pronounced effect on cellulose kraft pulp, both increasing the mobilization of phenolic chromophores in a synergistic manner and reducing the number. These effects indicate a significant reduction in the lignin content of the enzyme-treated pulp, which demonstrates the potential for biobleaching applications of cellulose kraft pulp. The release of phenolic compounds detects the soluble lignin, which showed a direct relationship with the laccase activity of the enzymes. Xylan is integrated into the polymer matrix of the pulp along with lignin and cellulose, and xylan degradation by xylanase results in greater exposure of lignin, facilitating the attack of laccase or the ABTS radical to release phenolic compounds from the pulp fibers. Some features of the CotA-XynAG3 chimera, which are likely to contribute to the improved effect against cellulose pulp, include the thermosta-bility of the xylanase domain, the higher catalytic efficiency of laccase domain, and the emergence of synergic behavior when compared with the separate parental enzymes. Furthermore, the heterologous production of the stable bifunctional CotA-XynAG3 containing laccase and xylanase activities in a single polypeptide offers a potential benefit in cost reduction as compared with the production of two separate enzymes.
The molecular and dynamics simulations showed that the catalytic domains maintained the individual structural characteristics and flexibility of the parental enzymes as seen by the calculated r.m.s.d. values (supplemental Fig. S1). The simulations also revealed that as a consequence of the approximation between the two catalytic domains, the loop regions in the laccase domain that contribute to the inter-domain interface undergo significant conformational changes in relation to the parental CotA. Of particular interest is the change in position of loop 2 (residues Leu-558 to Lys-573), which contributes to the formation of a substrate binding cleft in the laccase molecule. In the chimeric enzymes, the interactions of residues Glu-565, Tyr-566, and Arg-577 in the laccase domain with residues in the xylanase domain result in a dislocation of the entire loop region and a broadening of the surface cleft. We suggest that this structural change may result in an increased substrate access and therefore be responsible for the increased V max observed experimentally.
The change in the conformation of loop 2 observed in the molecular dynamics simulations results in an increased solvent access to the copper atom in the T1 site, and it is consistent with the changes in the UV-visible spectra observed in the chimeric protein. The EPR signal of the copper atom in the T1 sites of the CotA and the two chimeras is unaltered, indicating that despite the increased exposure of the T1 site to solvent the configuration of the atoms in the coordination sphere of the Cu 2ϩ remained unchanged. This implies that the increased substrate access due to the movement of loop 2 does not perturb the residues involved in the Cu 2ϩ ion binding, and superposition of the parental CotA with the CotA-XynA chimera indeed show that the conformation of the loop regions surrounding the T1 site are conserved. Although concomitant alterations in the results from UV-visible and EPR spectroscopy of the CotA are frequently observed (53), it is noteworthy that these two techniques measure discrete properties of the environment surrounding the T1 copper. Indeed, changes in the UV-visible spectrum in the absence of alterations in the EPR signal from the T1 copper have been observed in previous studies of the CotA laccase (64).
The catalytic cycle of the CotA involves the sequential transfer of a total of four electrons derived from the oxidation of four substrate molecules to the dioxygen electron acceptor producing two water molecules and involves a defined series of redox reactions in the copper cluster (71). Although no evidence was found for a change of the copper in the T3 site, the EPR results unambiguously identified alterations in the configuration of the copper in the T2 site in the chimeric enzymes. The T2 copper appears to be involved in binding the hydroxyl ion intermediates during the reduction of dioxygen (71), and therefore alterations in the environment around the T2 copper are likely to influence the catalytic cycle and may contribute to the increase in V max observed in the chimeric enzymes. However, because of the unsuitability of the GROMOS force field with respect to calculations involving the copper atoms, no attempt was made to correlate the molecular dynamics simulations in the region of the copper cluster with the alterations in the EPR spectra.
The rational design approach presented in this work proved to be an efficient tool for the creation of laccase-xylanase chimera, and the modular nature of the bifunctional enzymes was demonstrated by substitution of the native xylanase domain by a thermostable mutant protein that retained its catalytic properties after incorporation into the chimera. This demonstrates that enzyme fusion methods may be combined with in vitro evolution techniques to create enzymes with compatible catalytic properties for subsequent fusion to create multifunctional enzymes. A potential difficulty facing further progress was revealed by this study and lies in the observation that the catalytic domains of the fusion enzymes present altered catalytic properties when compared with the parental enzymes, and it highlights the need for more effective rational design strategies for the fusion enzymes. Despite these uncertainties, this study provides a glimpse of the rich opportunities for the application of protein engineering to improve the compatibility of the catalytic properties of enzymes for multifunctional fusions that can be exploited for various biotechnological applications.