Characterization of two thermophilic cellulases from Talaromyces leycettanus JCM12802 and their synergistic action on cellulose hydrolysis

Talaromyces leycettanus JCM12802 is a great producer of thermophilic glycoside hydrolases (GHs). In this study, two cellulases (TlCel5A and TlCel6A) belonging to GH5 and GH6 respectively were expressed in Pichia pastoris and functionally characterized. The enzymes had acidic and thermophilic properties, showing optimal activities at pH 3.5–4.5 and 75–80°C, and retained stable at temperatures up to 60°C and over a broad pH range of 2.0−8.0. TlCel5A and TlCel6A acted against several cellulose substrates with varied activities (3,101.1 vs. 92.9 U/mg to barley β-glucan, 3,905.6 U/mg vs. 109.0 U/mg to lichenan, and 840.3 and 0.09 U/mg to CMC-Na). When using Avicel, phosphoric acid swollen cellulose (PASC) or steam-exploded corn straw (SECS) as the substrate, combination of TlCel5A and TlCel6A showed significant synergistic action, releasing more reduced sugars (1.08–2.87 mM) than the individual enzymes. These two cellulases may represent potential enzyme additives for the efficient biomass conversion and bioethanol production.


Cloning of the cellulase-encoding genes
The full length cellulase-encoding genes, Tlcel5A and Tlcel6A, were identified in the genome sequence of T. leycettanus JCM12802, and obtained by PCR amplification with specific primers GH5F/GH5R and GH6F/GH6R (Table 1). After three days' growth in complex medium, the total RNA was extracted from the mycelia of T. leycettanus JCM12802 using the Qiagen RNeasy plant mini kit. cDNAs were then synthesized in vitro using the ReverTra Ace-a-™ kit. The cDNAs coding for the mature TlCel5A and TlCel6A were amplified with primers GH5F1/ GH5R1 and GH6F1/GH6R1 harboring restriction sites, respectively. The PCR products with appropriate sizes were inserted into plasmid pGEM-T Easy and then transformed into E. coli Trans1-T1. Positive clones were confirmed by PCR and DNA sequencing.

Sequence and structure analysis
The nucleotide and amino acid sequences were used for BLAST analysis at NCBI (http:// www.ncbi.nlm.nih.gov/BLAST/). Simplified sequence alignment and prediction of the isoelectric point and molecular weight were conducted by using the Vector NTI Suite 10.0 software. Signal peptide prediction was carried out with SignalP 4.1 Server (http://www.cbs.dtu.dk/ services/SignalP/). Multiple alignment of protein sequences was performed with the ClustalX 2.1 (http://www.clustal.org, [31], and the results were demonstrated by ESPript 3.0 (http:// espript.ibcp.fr/ESPript/cgi-bin/ESPript.cgi [32]. The NetNglyc server was used to predict the putative N-glycosylation sites (http://www.cbs.dtu.dk/services/NetNGlyc/). And the modeled structures were predicted and visualized by using the SWISS-MODEL [33] and Pymol [34].

Heterogeneous expression in P. pastoris
Heterologous expression of TlCel5A and TlCel6A was performed in P. pastoris according to the manual of the Pichia Expression kit (Invitrogen) with some modifications. The cDNA fragments coding for the mature proteins of TlCel5A and TlCel6A were digested with restriction enzymes SnabI or EcoRI and NotI and subcloned into plasmid pPIC9 to yield expression vectors pPIC9-Tlcel5A and pPIC9-Tlcel6A, respectively. The recombinant plasmids were then linearized with BglII and individually transformed into P. pastoris GS115 competent cells by electroporation. Transformants were selected on MD plates, and positive transformants were grown in 10 mL shake tubes containing 3 mL BMGY for 2 days and 1 mL BMMY for another 2 days (220 rpm and 30˚C). The culture supernatants were collected by centrifugation at 5,000 rpm for 10 min, followed by cellulase activity assay as described below. The transformants with highest cellulase activities were selected for flask-level fermentation.
The most active transformants of TlCel5A and TlCel6A were grown in 1 L shake-flasks containing 400 mL BMGY at 220 rpm and 30˚C for 2 days, respectively. The cultures were pelleted by centrifugation and resuspended in 200 mL of BMMY for 3-day-growth at 30˚C and 200 rpm. Methanol was then added at the final concentration of 0.5% every 24 h for continuous induction of 3 days. The culture supernatants were collected by centrifugation at 4500× g for further analysis.

Purification of recombinant TlCel5A and TlCel6A
The viva flow 200 ultrafiltration membrane system (Sartorius, Germany) with 5 kDa cut-off was used for the concentration and buffer exchange (to 0.1 M Tris-HCl, pH 8.0) of the crude enzymes. The recombinant proteins were desalted by HiTrapTM Desalting column and purified using the HiTrap Q Sepharose XL FPLC column (GE Healthcare) pre-equilibrated with 0.1 M Tris-HCl (pH 8.0). The gradient NaCl of 0-0.7 M at the flow rate of 3 mL/min was used to elute the proteins. Fractions showing cellulase activities were pooled and further desalted with a 5 kDa molecular cut-off concentration tube (Millipore) using 0.1 M McIlvaine buffer (pH 3.5 or 4.5). The purified proteins were separated on 12% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) for purity and molecular mass analysis. Protein concentration was determined by using the Bradford method.

Deglycosylation of the purified recombinant TlCel5A and TlCel6A
To remove N-glycosylation, the purified recombinant TlCel5A and TlCel6A were treated with Endo H for 2 h at 37˚C following the manufacturer's instructions (New England Biolabs), and checked by SDS-PAGE.

Enzyme activity assays
The cellulase activities were determined using 1% (w/v) CMC-Na, barley β-glucan or lichenan as the substrate. The reaction mixtures containing 100 μL of properly diluted enzyme solution and 900 μL substrate solution in 0.1 M McIlvaine buffer (pH 3.5 and 4.5) were incubated at 75˚C or 80˚C (optimum temperature) for 10 min. When using 5 mg/mL Avicel as the substrate, the reaction period was lengthened to 60 min. The amounts of reducing sugars were determined via the 3,5-dinitrosalicylic acid (DNS) method [35]. One unit (U) of enzyme activity was defined as the amount of enzyme producing 1 μmol of reducing sugar per minute under the assay conditions.

Biochemical characterization of TlCel5A and TlCel6A
CMC-Na, barley β-glucan and PASC were used as the substrates for enzyme characterization. The optimal pH was determined at 75˚C for TlCel5A and 80˚C for TlCel6A, respectively, in 0.1 M McIlvaine buffer with pH ranging from 2.0 to 8.0. To test the pH stability, the purified enzymes were pre-incubated at 37˚C for 60 min in buffers with pH ranging from 1.0 to 12.0 (0.1 M KCl-HCl for pH 1.0-2.0, 0.1 M McIlvaine buffer for 2.0 to 8.0, and 0.1 M glycine-NaOH for pH 9.0-12.0). The residual enzyme activities were determined at 75˚C and pH 3.5 for TlCel5A and 80˚C (barley β-glucan) or 70˚C (PASC) and pH 4.5 for TlCel6A, respectively. To determine the optimal temperature, the cellulase activities were determined at temperatures ranging from 30˚C to 90˚C and pH 3.5 for TlCel5A or pH 4.5 for TlCel6A for 10 min, respectively. For thermal stability assays, TlCel5A and TlCel6A were incubated at temperatures of 70˚C, 75˚C or 80˚C and pH 3.5 or 4.5 for different periods of time. For half-life determination, 0.05 mg of TlCel5A or 0.2 mg of TlCel6A was incubated at 60˚C, 65˚C or 70˚C for 0.5-24 h. Residual activities were determined as described above.
The kinetic parameters (K m , V max and k cat /K m ) of TlCel5A were determined by incubating the enzyme with 0.25 to 10 mg/mL CMC-Na at pH 3.5 and 75˚C for 5 min. For TlCel6A, the kinetic parameters were determined at pH 4.5 and 80˚C for 5 min with 0.25-10.0 mg/mL barley β-glucan as the substrate. The enzyme activities were determined by using the DNS method. The kinetic constants were calculated using the Lineweaver-Burk plots by GraphPad Prism 6.0 (http://www.graphpad.com/scientific-software/prism/).

Analysis of the cellooligosaccharides hydrolysis products
The hydrolysis products of cellooligosaccharides by TlCel5A and TlCel6A were detected by high-performance anion exchange chromatography (HPAEC, model 2500, Dionex, Sunnyvale, CA). Purified TlCel5A (0.05 U) or TlCel6A (0.02 U) was added into 1 mL of cellooligosaccharide solution containing 200 μg of cellotetraose, cellopentaose, or cellohexaose and incubated at 60˚C for 0 min, 1 min, 5 min, 30 min, 1 h, 4 h or 5 h. After enzyme inactivation by boiling water bath, the hydrolysates were diluted 100 times with ddH 2 O, and 100 μL of each sample was injected into the column of HPAEC. The oligosaccharides were eluted by 100 mM NaOH. The standards consisted of glucose, cellobiose, cellotriose, cellotetraose, cellopentaose, and cellohexaose. The amount of each hydrolysate was calculated based on the peak area. One enzyme unit (U) was defined as the amount of enzyme to hydrolyze 1 μmol of substrate per minute under the assay conditions. The catalytic efficiency values (k cat /K m ) of TlCel6A against cellooligosaccharides were calculated according to Xia et al. (2016).

Synergistic action of TlCel5A and TlCel6A on cellulose hydrolysis
To determine the hydrolytic capabilities of TlCel5A and TlCel6A to degrade cellulose substrates, the enzymes were combined at different ratios, and the commercial β-glucosidase from Aspergillus fumigatu (Sigma-Aldrich) was added at the concentration of 10% (mol β-glucosidase/mol total enzyme) to convert cellobiose to glucose. When using Avicel as the substrate, TlCel5A and TlCel6A were combined at the ratios of 0.4:3.6, 0.8:3.2, 1. Based on the results above, the best ratio of TlCel5A and TlCel6A was used for further enzyme combinations. When using SECS or Avicel as the substrate, 8 μM of TlCel5A or TlCel6A alone, or 4 μM TlCel5A and 4 μM TlCel6A was incubated with 5 mg/mL substrate at pH 4.0 and 60˚C for 12 or 18 h. With 2.5 mg/mL PASC as the substrate, 2 μM of TlCel5A or TlCel6A alone or 1 μM TlCel5A and 1 μM TlCel6A was added. The amounts of reducing sugars released were determined by using the DNS method. The amounts of glucose released were analyzed by a one-factor ANOVA of SPSS19.0 to assess the synergistic effects of TlCel5A and TlCel6A. Statistical differences were considered to be significant at P < 0.05.

Nucleotide sequence accession numbers
The nucleotide sequences of TlCel5A and TlCel6A from T. leycettanus JCM12802 have been deposited in the GenBank database under the accession numbers MG993208 and MG993209, respectively.

Gene cloning and sequence analysis
The full-length Tlcel5A and Tlcel6A contained 1464 bp and 1898 bp, respectively, which were interrupted by four (67, 56, 57, and 54 bp) and seven (69, 75, 78, 69, 71, 70, and 71 bp) introns. The cDNAs were 1230 bp and 1398 bp in lengths, and encoded two polypeptides of 409 and 465 amino acids, respectively. A putative signal peptide of 18 amino acids was predicted at the N-termini, and two (Asn36 and Asn240 in TlCel5A) and three (Asn293, Asn307 and Asn423 in TlCel6A) potential N-glycosylation sites were predicted in deduced TlCel5A and TlCel6A.
BLAST analysis indicated that TlCel5A is an endoglucanase of GH5 while TlCel6A is a cellobiohydrolase of GH6, respectively. Both enzymes consist of a cellulose-binding module (CBM), a linker region and a catalytic domain. The CBM of TlCel5A is at the C-terminus while that of TlCel6A is at the N-terminus. The deduced TlCel5A showed the highest protein identity of 76% to an EG from Aspergillus udagawae (GAO86105.1) and 77% with the catalytic domain of structure-resolved EG (1GZJ) from Thermoascus aurantiacus. Deduced TlCel6A shared the highest identity of 99% with a GH6 CBH from T. leycettanus (CDF76448.1).

Expression and purification of the recombinant TlCel5A and TlCel6A
Recombinant TlCel5A and TlCel6A were successfully expressed in P. pastoris GS115. With barley β-glucan as the substrate, the highest activities of 2.02 U/mL and 1.86 U/mL, were detected in the positive transformants of TlCel5A and TlCel6A, respectively. After large-scale fermentation for 144 h, the culture supernatants were collected, concentrated and purified to electrophoretic homogeneity by anion exchange chromatography. SDS-PAGE revealed that the purified recombinant TlCel5A and TlCel6A had apparent molecular masses of approximately 45 kDa and 65 kDa (Fig 1), which were higher than their theoretical values (42.5 kDa and 47.0 kDa). After Endo H treatment, the molecular mass of TlCel5A remained the same, while the TlCel6A migrated as a single band of about 60 kDa. Other post-translational modification like O-glycosylation might account for the extra molecular masses.

Enzymatic properties of purified recombinant TlCel5A and TlCel6A
The enzymatic properties of TlCel5A were determined with CMC-Na, and the TlCel6A activities were determined by using barley β-glucan and PASC as the substrates. TlCel5A had the optimal pH of 3.5 (Fig 2A) and retained more than 90% activity at pH 2.0 to 10.0 (Fig 2B). The pH optima of TlCel6A towards barley β-glucan and PASC were both 4.5 (Fig 2A), but retained stable over a narrower pH range of 2.0 to 8.0 (Fig 2B). Both enzymes were thermophilic: TlCel5A was optimally active at 75˚C, and remained more than 90% activity at 80˚C, while TlCel6A showed the maximum activity at 80˚C to barley β-glucan and 70˚C to PASC (Fig 2A). The enzymes were highly stable at 60˚C (Fig 2D). When increased the temperature to 70˚C, TlCel5A retained most of the activity after 1-h incubation, while TlCel6A lost approximately 50% and 40% activities towards barley β-glucan and PASC, respectively, after 30 min (Fig 2E   Fig 1. SDS-PAGE (12% polyacrylamide gel)
The specific activities of TlCel6A on lichenan, barley β-glucan and PASC were 109.0 ± 4.5 U/mg, 92.9 ± 1.4 U/mg and 4.0 ± 0.1 U/mg. When using Avicel and CMC-Na as the substrate, TlCel6A showed much lower activities (0.12 U/mg and 0.09 U/mg). No activity was detected on pNP-substrates. The K m , V max , and k cat /K m values of TlCel6A using barley β-glucan as the substrate were determined to be 2.1 mg/mL, 196.1 μmol/min/mg, and 73.2 mL/s/mg, respectively. For cellooligosaccharide hydrolysis, TlCel5A had low specific activity on cellotetraose, cellopentaose (15.2 U/mg) and cellohexaose (16.1 U/mg). And TlCel6A had high activities on cellotetraose (186.7 U/mg), cellopentaose (99.6 U/mg), and cellohexaose (138.9 U/mg). The relative low activity on cellopentaose might be ascribed to the preference of TlCel6A for cellooligosaccharides containing even numbers of glucose units.

Hydrolysis products of cellooligosaccharides
The hydrolysis products of TlCel5A and TlCel6A on cellooligosaccharides (cellotetraose, cellopentaose, or cellohexaose) were analyzed by using the method of HPLC (Table 2)

Synergistic action of TlCel5A and TlCel6A on cellulose hydrolysis
The hydrolytic capabilities of TlCel5A, TlCel6A and their combinations at different concentrations (0.4-4.0 μM) towards Avicel were determined after 24-h incubation at 60˚C. As shown in Fig 3A, the enzyme combinations released more glucose than the individual enzyme or the sum of the individuals. The highest amount of glucose (4.00 mM) was released by the enzyme combination of 2 μM TlCel5A and 2 μM TlCel6A. Thus the enzyme ratio of 1:1 was used for further studies.
The combination of TlCel5A and TlCel6A released 3.72 mM, 4.33 mM, and 7.87 mM of glucose from Avicel at 18 h, SECS at 12 h, and PASC at 60 min, respectively, which were significantly higher (P < 0.05) than that released by TlCel5A or TlCel6A individually from Avicel (1.08 and 2.64 mM; Fig 3B), SECS (1.46 and 1.96 mM; Fig 3C), and PASC (5.33 and 5.51 mM; Fig 3D). The results indicated that TlCel5A and TlCel6A have synergistic action on cellulose degradation.

Discussion
The endoglucanases of GH5 and cellobiohydrolases of GH6 are crucial components of the cellulase complex for biomass conversion. These cellulases generally have a temperature optimum of 40-70˚C and retain stable at 50-60˚C [9,36,37]. To improve the hydrolytic efficiency and reduce the cost, thermostable cellulases with high activities are much favorable. In the present study, two cellulases (TlCel5A and TlCel6A) from T. leycettanus JCM12802 were heterologously produced in P. pastoris and functionally characterized. The recombinant enzymes were found to be thermophilic and highly active, and showed great efficiency in the bioconversion of various cellulose substrates.
In comparison to their close homologues, TlCel5A and TlCel6A showed adaptability to higher temperatures. For example, the optimal temperature of TlCel5A is 75˚C, which is at least 5˚C higher than those characterized EGs (with 60-77% sequence identities) of Penicillium brasilianum (Cel5C; [38], Aspergillus niger (AnCel5A; [39], and Thermoascus aurantiacus (EGI; [17]. One exception is the TeEgl5A from T. emersonii CBS394.64, which exhibits optimal activity at 90˚C [12]. However, the TlCel5A activity (840 U/mg) on CMC-Na was much higher than that of TeEgl5A (624 U/mg), EGI (336 U/mg) and Cel5C (51 U/mg). The temperature optima of TlCel6A on barley β-glucan and PASC are 80˚C and 70˚C, much higher than or similar to its homologues 40-70˚C [26,40,41]. These excellent properties make TlCel5A and TlCel6A much potential for the application in the biofuel industry.
The hydrolytic capabilities of TlCel5A and TlCel6A in combination with a commercial glucosidase were then assessed. These enzymes employ endo-and exo-mode of actions [42][43][44], and acted synergistically on the hydrolysis of different cellulose substrates. When added the TlCel5A and TlCel6A at the ratio of 1:1, it produced the highest amount of glucose (up to 4 Characterization of two thermophilic cellulases and their synergistic action on cellulose hydrolysis mM). There are significant differences in the amounts of reducing sugars released by TlCel5A, TlCel6A and their combination. The synergy effect of combined TlCel5A and TlCel6A with Avicel as the substrate is greater than that of Ctendo45 and CtCel6 [19], which might be ascribed to the enzymatic properties of different enzyme components. Although TlCel5A and TlCel6A individually showed higher activities on PASC than on Avicel, the synergy effect of TlCel5A and TlCel6A was lower on PASC. Similar results have been reported by Zhang and Lynd [45]. For natural substrate like SECS, TlCel5A and TlCel6A showed greater synergy than the individual enzymes. Of the two enzymes, TlCel6A had higher hydrolysis activity than TlCel5A towards crystalline Avicel, but had no difference in hydrolysis of PASC and steamexploded corn straw. The reason might be that TlCel5A contributes to the increased hydrolytic efficiency of TlCel6A by producing new chain ends and smoothing the physical obstacles on the surface of cellulose, while TlCel6A depolymerizes some tight cellulose structure to promote the TlCel5A activity [46]. Therefore, the enzyme combination of TlCel5A and TlCel6A is more favorable to degrade natural biomass with higher hydrolysis efficiency. Characterization of two thermophilic cellulases and their synergistic action on cellulose hydrolysis It has been reported that N-linked glycan is one important component of the processive machinery of cellobiohydrolases and plays a role in the enzyme activity [47]. Although some studies indicated that N-glycosylation has no significant effect on enzyme activity, and it does influence the protein stability [48]. According to the theory of [47], the position of N-glycosylation on the protein surface plays the key role. There are three potential N-glycosylation sites (Asn293, Asn307 and Asn423) in TlCel6A. Since the glycan linked to Asn293 is located in close proximity to the entrance of the active site tunnel, this long glycan chain might block the cellulose molecules from the tunnel and finally decrease the activity. On the other hand, Asn423 is located on the active loop, and the glycans attached to this residue might form interactions with the cellulose molecules and boost the enzyme activity. Further site-directed mutagenesis at these potential N-glycosylation sites will be conducted to reveal the underlying mechanisms.

Conclusions
Two acidic cellulases TlCel5A and TlCel6A were identified in T. leycettanus JCM12802 and heterologously produced in P. pastoris. In comparison to close homologues, these cellulases are superior in higher activities and greater thermostability. Their combination at the ratio of 1:1 showed distinguished synergy on Avicel, PASC and steam-exploded corn straw. Thus TlCel5A and TlCel6A may represent great candidates for the industrial conversion of lignocellulose.