A novel Swollenin from Talaromyces leycettanus JCM12802 enhances cellulase hydrolysis towards various substrates

Background: Swollenins exist within some fungal species such as Trichoderma reesei and Aspergillus niger, and are candidate accessory enzymes for the biodegradation of cellulosic substrates. Swollenins theorized role in biological processes is as promoters in the rearrangement of non-covalent interaction networks formed by cell wall polysaccharides. Such network rearrangements though to be responsible for expanding accessibility for hydrolytic enzyme degradation processes. We hereby report in detail the characterization of a recombinant swollenin and its disruptive activity on cellulosic substrates and its synergistic effects while interacting with cellulases. Results: We have expressed a novel swollenin gene Tlswo consists of an open reading frame encoding 503 amino acids to be identi�ed from Talaromyces leycettanus JCM12802 and successfully expressed it in both Trichoderma reesei and Pichia pastoris. Similar to other fungal swollenins, TlSWO is composed of a N-terminal family 1 carbohydrate binding module (CBM1) followed by a Ser/Thr rich linker connected to an expansin-like domain, followed by a family of 45 endoglucanase-like domains and a group-2 grass pollen allergen domain. From our experimental results, TlSWO demonstrated disruptive activity on Avicel and displayed a synergistic effect with cellobiohydrolases, enhancing its hydrolytic performance up to 132%. We also explored the activity of TlSWO on several model substrates as well as pretreated biomass. Our results indicate that TlSWO is capable of releasing reducing sugars from lichenan, barley β-glucan, carboxymethyl cellulose sodium (CMC-Na) and laminarin. The speci�c activity of TlSWO towards the substates mentioned above is 9.0 ± 0.100 U/mg, 8.9 ± 0.100U/mg, 2.3 ± 0.002 U/mg and 0.79 ± 0.002 U/mg, respectively. Moreover, further experimentation con�rms the fungal swollenin TlSWO exhibits maximum activity at pH 4.0 and 50 ℃ . Conclusion: This study reports on a novel swollenin


Background
Lignocellulosic biomass has been considered as an alternative to fossil fuels, however challenges still remain for biomass feedstocks such as this to be properly processed and converted into fuels or other chemical production purposes [1].A successful pretreatment process is measured two ways.First, by the highly e cient recovery of carbohydrates from within the raw materials.And second, by minimizing the formation of toxic and inhibiting compounds to prevent unwanted health risks and environmental hazards [2].The enzymatic deconstruction of biomass that follows pretreatment can be enhanced by using a number of non-glycoside hydrolase accessory enzymes.These enzymes include but not limited to expansins, loosenins, ceratoplatanin proteins, and certain other types of carbohydrate binding modules [3][4][5][6].Among these proteins, expansins have been widely found in plant cell walls and studies have shown that they are responsible for loosening the cellular wall [7].
Studies done by Mcqueen-Mason have proposed that expansins have the ability to disrupt hydrogen bonding between plant cell wall polysaccharides without hydrolyzing them rst [8].In 2002, the rst expansin-like protein SWOI from fungi was discovered in Trichoderma reesei [9] and because swollenins from fungi in many studies are shown to have sequence similarity to expansins, they are now often referred to as expansin-like proteins.This study also concluded that the SWOI protein was able to disrupt cotton bers and lter paper structures on a microscopic level without detectable reducing sugars [9].
Over the past decade, scientists have identi ed more than 10 types of swollenins [10][11][12][13].And recently, a new database, Expansin engineering Database (ExED, https://exed.biocatnet.de), has been released to the public [14].In general, expansins have a length of no more than 250 amino acids which constitute a twodomain structure.The primary domain resembles the glycoside hydrolase family 45 (GH45) and such homology preserves certain sequence features of the GH45 catalytic site [15].The second domain has a characteristic at aromatic-rich surface and is homologous to group-2 grass pollen allergens.In some studies, this domain was proposed to function as a CBM [16,17].The two domains within the expansin are interconnected by a short linker, and both domains are required for plant cell-wall loosening activity.
However, there are functional differences from structural discrepancies between swollenins and expansins.For example, swollenins have an additional CBM domain, allowing a display of homology to fungal cellulases in the N-terminal [18].In other cellulases, CBMs direct the enzymes to bind on cellulosic surfaces and thus enhance lignocellulose degradation [19][20][21].Also, O-glycosylation of linkers may also have some effect on the binding of the enzyme to surfaces [22].
Previous studies have also described the synergy of swollenin with glycoside hydrolases in releasing soluble sugars from substrates.As a prominent example, Zhou et al. were able to successfully express swollenin SWO2 in Aspergillus niger, showing that the simultaneous incubation of SWO2 with cellulases resulted in a signi cant increase in synergistic activity in cellulose hydrolysis.This synergistic activity was improved even further when the cellulose was pretreated with swollenin rst [12].Other investigations focused on the synergy of swollenin with xylanase by Santos et al. found that a swollenin ThSWO from T. harzianum created a rough and amorphous surface on Avicel and also displayed a high synergistic effect with a commercial xylanase from T. viride, enhancing its hydrolytic performance up to 147% [11].Furthermore, Anthony et al. indicated in their study that a chimeric enzyme with swollenin from T. reesei fused with feruloyl esterase A from A. niger was also found to produce a signi cant increase in ferulic acid released from lignocellulose samples [23].Other studies demonstrated that swollenins also have the capability to release reducing sugars from cellulosic materials.For example, swollenin SWO2 from T. pseudokoningii and AfSwo1 from Aspergillus fumigatus were both found to exhibit very low level of endoglucanase activity [12,24].Recently, Andberg et al. demonstrated swollenin SWOI from T. reesei expressed activity on substrates containing β-1,4 glucosidic bonds, hypothesized a unique mode of mechanistic action with similarities in both endoglucanases and cellobiohydrolases [25].
In our previous work, we identi ed a thermophilic T. leycettanus strain JCM12802 which is an excellent CAZyme source [26][27][28][29].In this study, we present a novel swollenin protein TlSWO that was cloned from T. leycettanus JCM12802, and successfully expressed in T. reesei and Pichia pastoris.We have examined TlSWO activity performance on various substrates, demonstrating TlSWO's capability in releasing reducing sugar from barley β-glucan, lichenan, laminarin, and carboxymethyl cellulose sodium (CMC-Na).In addition, we show that TlSWO was able reduce the particle size and the surface structure of Avicel while signi cantly increase synergistic activity by up to 72.2% on pre-treated corn stover (PCS).

Identi cation and characterization of Tswo gene
The ORF of putative Tswo consists of six exons, encoding a protein (TlSWO) of 503 amino acids and a signal peptide at the cleavage site between amino acids 20 and 21.The cloning sequence of putative TlSWO was submitted to NCBI GenBank as MT180127.The predicted product shows the highest sequence similarity to the amino acid sequence of the known swollenin which are from Aspergillus fumigatus or Penicillium oxalicum.Further analysis by PROSITE demonstrated that TlSWO consists of three domains, fungal-type carbohydrate-binding module family 1 (CBM1) (amino acids 23-59), family 45 endoglucanase-like domain of expansin (Expansin_EG45) (amino acids 206-388) and a cellulose-bindinglike domain of expansin (Expansin_CBD) (amino acids 400-492), which are typical of the swollenins from fungi (Fig. 1).In the CBM1 of TlSWO, six cysteines were highly conserved.Disul de bond prediction by DiANNA showed that there are three disul de bonds in the CBM1 of TlSWO (Cys4-Cys21, Cys11-Cys28, Cys22-Cys28).CBM1 and Expansin_EG45 are connected by a Serine -Threonine rich linker domain.Although the function of linker has been studied in other cellulases, it is not clear if the linker plays any similar role in swollenins.Sequence alignment of swollenins showed TlSWO maintained a conserved HMD (histidine, methionine, aspartic acid) catalytic motif of the GH45 cellulase (HFD, histidine, phenylalanine, aspartic acid), which is part of the active site (Figure 1).In the GH45 HFD, the aspartic acid is the proton donor during the catalytic process.However, the other residue of the catalytic active site, aspartic acid, is absent from both swollenin and part of GH45 cellulases.In the C-terminal region of TlSWO, the expansin_CBD is homologous to pollen allergen.There are also several conserved aromatic amino acids in the sequence, i.e., Y400, Y401, F402, W429, Y447, W450, Y496 and F503, which may play a key role in binding substrates (Figure 1).Expression of TlSWO in P. pastoris GS115 and T. reesei AST1116 Expression of TlSWO was preformed using the aox1 promoter and eno promoter in P. pastoris and T. reesei respectively.Following high-throughput fungal expression and screening, a large set of colonies resulting in SDS bands after transformation were identi ed.The results revealed that recombinant TlSWO (483 amino acid residues, 51.1 kDa and a theoretical isoelectric point of 4.45) was successfully expressed using P. pastoris and T. reesei (Figure S1).The puri ed swollenin protein migrated as an about ∼80 kDa protein in SDS-PAGE.Due to the difference between the SDS-band and the theoretical molecular weight, a single band was analyzed by MALDI-TOF MS and the trypsin-digested peptide sequences were matched to the deduced amino acid sequences of TlSWO (Figure S2).Sequence prediction suggested that there are ve N-glycan sites on TlSWO (Asn35, Asn154, Asn249, Asn366 and Asn436).After Endo H digestion, TlSWO's molecular weight decreased to ∼72 kDa which is still higher than theoretical MW (Figure S1A).We speculate that the rest of the molecular weight increase was caused by heavy O-glycan glycosylation in the linker region which is rich in serines and threonines and reported for other proteins expressed in these hosts.

Activity of TlSWO on different substrates
TlSWO's cellulolytic activity was measured with lichenan, barley β-glucan, CMC-Na, laminarin, avicel, glucomannan, the xylanase activity was measured with birchwood xylan, and the mannase activity was measured with locust bean gum.All reactions were carried out overnight.Our results show that TlSWO only has signi cant activity on lichenan, barley β-glucan, glucomannan, and CMC-Na, and a very low activity on laminarin.TlSWO showed the highest activity on lichenan (9.0 ± 0.100 U/mg) and barley βglucan (8.9 ± 0.100 U/mg), followed by CMC-Na (2.3 ± 0.002 U/mg).In contrast, only slight activity was observed against laminarin (0.79 ± 0.002 U/mg) (Figure 2).These ndings suggest that TlSWO mainly acts on cellulose rich substrates and shows a preference to the substrates that contain 1,4 linkages.

Mode of action of TlSWO
The TlSWO's mode of action was assessed using the aforementioned four substrates (Figure 3).As a result, CMC-Na was hydrolyzed into cellobiose and small amount of cellotriose (Figure 3).Analysis of the hydrolysis products of lichenan, barley β-glucan showed that TlSWO preferentially hydrolyzed these two substrates into products with different degrees of polymerization, including cellobiose and cellopentose, followed by cellohexose and cellotetrose (Figure 3).We detected no sugar release after incubating TlSWO with laminarin (Figure 3).These results suggested that TlSWO may have a function of endo-cellulase.

Effect of temperature and pH on TlSWO
The effect of pH and temperature on TlSWO activity were further investigated with lichenan as a substrate.We determined the optimum pH is 4.0 while the enzyme displayed activity across a broad pH range of 2.0-12.0(Figure 4A).As for pH stability, TlSWO retained more than 80% of its activity within the pH range 2.0-9.0 after incubation at 37 ℃ for 1 h.In constrast, TlSWO lost 30% of its activity after incubation at pH 10.0-12.0 after 1 h (Figure 4B).In addition, TlSWO reached optimal activity at 50 ℃ and retained more than 90% of its activity within the temperature range of 40-60 ℃.However, after heating to 70 ℃ or above, its activity falls off very rapidly (Figure 4C).Also, as shown in gure 4D , TlSWO maintained stable activity between 37 ℃ and 50 ℃ after 1 hour of incubation, however, its activity quickly decreased to 40% after a 10 minute incubation as temperature was raised to 70 ℃.

Disruptive action of TlSWO on Avicel
The disruptive effect of TlSWO on Avicel was evaluated using light microscopy (LM) and scanning electron microscopy (SEM).LM analysis showed that after incubation with varied amounts of TlSWO for 24 h, the Avicel's physical structure tend to be signi cantly different from the untreated Avicel (Figure 5).As we increased the amount of TlSWO, Avicel was disrupted into smaller particles.The sample that was pretreated with 300 mg of TlSWO for 12 h was taken out for further analysis using SEM.As we can see, TlSWO created a rough and amorphic surface on Avicel compared with the unpretreated sample (Figure S3).

Synergism between TlSWO and cellulases
To test the capacity of TlSWO to enhance biomass hydrolysis by enzymatic cocktail, we hydrolyzed pretreated biomass using cellulases alone rst, followed by treatment using both cellulases and TlSWO.Biomass degradation experiments were performed using β-glucosidase (EC 3.2.1.21),cellobiohydrolase (EC 3.2.1.91)and endoglucanase (EC 3.2.1.4) in the presence of TlSWO.The reactions with BSA and without TlSWO were used as controls.The total protein used was 13 mg protein/g of glucan in all reactions.Since endoglucanases randomly cleave the internal β-1,4-glycosidic bonds creating new reducing ends, this allows cellobiohydrolases to continuously act on the chain termini to release cellobiose and β-glucosidase that hydrolyzes cellobiose into glucose [30], therefore only the production of glucose was compared in different reactions.
Our results showed that TlSWO exhibited signi cant synergetic effects on cellobiohydrolase Cel7A when using PCS as the substrate.Initially, we measured the reaction's glucose quantity containing only 13 (mg protein/g of glucan) of TlSWO.No glucose was detected in the reaction, suggesting that PCS could not be hydrolyzed to monomers by TlSWO alone (Figure 6).When the reaction contained Cel7A and βglucosidase individually PCS conversion achieved an increase from 8.9% to 16.4% within the 24 h to 120 h period.When Cel7A and β-glucosidase were combined with 2 (mg protein/g of glucan) TlSWO the increase changed from 11.2% at 24 h to 26.4% at 120 h.Even though TlSWO action alone did not lead to any detectable levels of released glucose, the enzyme addition to Cel7A and β-glucosidase led to a signi cantly enhanced hydrolytic activity of the cocktail on PCS.
PASC and CNC exist in amorphous and crystalline forms, respectively, which may affect their binding with TlSWO.Therefore, the action of TlSWO for the two substrates were further examined.Similar to the results of PCS, TlSWO could not release any sugars from PASC and CNC when used individually (Figure 7).When utilizing Cel7A and β-glucosidase the conversion results showed that PASC conversion rate reached from 9.9% at 24 h, to 40.2% at 120 h (Figure 7A).However, when Cel7A and β-glucosidase were supplemented by TlSWO the conversion rates increased to from 33.0% at 24 h to 72.2% at 12 h.Looking at the 120-hour data, the conversion rate of PASC increased by approximately 32% compared to Cel7A and β-glucosidase individually, suggesting that TlSWO exhibited signi cant synergetic effects with Cel7A.In addition, the synergetic effects of TlSWO and endoglucanases were also explored during this research.When using endoglucanase and β-glucosidase individually the conversion rate increase was from 51.4% at 24 h to 85.6% at 120 h.When used in combination with TlSWO, the conversion rate reached into from 58.6% at 24 h to 85.7% at 120 h.Although these conversion rates are slightly higher than that of using endoglucanase and β-glucosidase individually, the conversion rate at 120 h was 86.4% is within the margin of error when compared to 85.6%.Therefore, we conclude that the addition of TlSWO did not provide a signi cant increase in enzymatic hydrolysis of PASC when used in combination with endoglucanase.
The cellulose conversion rate of CNC was overall signi cantly lower than that of PASC.Our results showed that on CNC, the Cel7A and β-glucosidase achieved a conversion rate from 28.2% at 24 h to 59.3% at 120 h (Figure 7B), which was increased further when TlSWO was added.The glucose yields obtained by the cocktail enzyme systems containing TlSWO and Cel7A and β-glucosidase were 31.9% at 24 h to 68.9% at 120 h , higher than all other cases when compared with the control group.Therefore, we conclude that TlSWO showed a synergistic effect with processive cellobiohydrolases.However, when used in combination with endoglucanases, TlSWO did not present any signi cant synergistic effect.After 120 h of enzymatic hydrolysis, glucose yield released from CNC was 16.4% with combined use of TlSWO, endoglucanase and β-glucosidase, which corresponds well with the hydrolysis rate of 14.9% using endoglucanase and β-glucosidase individually.Comparing with our PASC results, we hypothesize TlSWO will also act on amorphous cellulose more e ciently than their respective crystalline forms.

Discussion
Several previous studies initially showed the disruptive capabilities of SWOI on plant cell wall structures without leaving any traceable amount of reducing sugars.However, subsequent research con rmed that SWOI does exhibits some hydrolytic activity on cellulosic substrates with features of both endoglucanases and cellobiohydrolases [9,25].Indeed, similarly to SWOI, the other two swollenins, AfSWO1 from A. fumigatus, SWO2 from T. pseudokoningii have also showed hydrolytic activity on various substrates, thereby suggesting that these proteins interact with cellulose or hemicellulose [12,24].
In this study, TlSWO from T. leycettanus JCM12802 was found to have similar functionalities as other fungal swollenins while sharing a relatively high sequence identity with SWOI, AfSWO1 and SWO2 (64.5%, 73.7% and 63.2%, respectively).Although TlSWO has the highest activity on lichenan and barley β-glucan substrates, both contain β-1,4 and β-1,3 linkages, its activity becomes very minimal on laminarin, a substrate that only contains β-1,3 linkages.Indicating activity on β-1,4 linkages is the primary mechanism of action.According to previously reports, expansins have higher similarity with GH45 subfamily C enzymes than with other members of the GH45 family [31,32].Studies indicating expansins with higher similarity with GH45 subfamily C enzymes than others have a motif termed HFD as part of its active site, which is also present in TlSWO (HMD).The aspartic acid in this motif plays the role of proton donor in GH45.Nevertheless, other key residues that are critical for catalytic activities are absent in both expansins and swollenins.This evidence suggests expansins and swollenins may have adopted a inverted mechanism during the catalytic process.In 2015, Nakamura et al. proposed that PcCel45A which belongs to the GH45 subfamily C, uses an imidic acid form of asparagine residue as a general base in the "Newton's cradle" proton relay catalytic mechanism [33].This proposal sheds light on some mechanistic implications of expansin's catalytic process.Evolutionary process' horizontal gene transfer (HGT) drives differences between the fungal swollenins and the plant/bacterial expansins in sequence [34].These differecnes could make PcCel45A's catalytic mechanism theory unfeasible when applied to fungal swollenins.
Fungal swollenin sizes are roughly twice as large as plant and bacterial swollenins due to their D1 and D2 domains containing extra sequence insertions such as the additional N-terminal CBM with linkers.CBMs can increase the concentration of their parent enzyme substrate surface, leading to more rapid degradation of polysaccharide [35,36].Primary amino acid sequence analysis using BLAST indicates that TlSWO contains an N-terminal CBM region (amino acid residues 21-59) that shows the highest similarity towards the fungal GH6 family 1 CBMs.Six typical conserved cysteines which could form three pairs of disul de bonds.Also three conserved aromatic residues which are typical in GH6 and GH7 cellulases CBMs are presented in the TlSWO's CBM (Trp28, Tyr54, and Tyr55).These sequences' key roles are highlighted in their importance in cellulases stability and activity during the reaction.Moreover, it has been demonstrated that the length of linker is crucial for the activity of cellulases [37].The linker region of TlSWO is over 140 amino acids, longer than that of most of reported swollenins and fungal cellulases [38].Although CBM and linker have been well studied in cellulases, little is currently known about the actual role of these two regions in swollenins.
We have shown that when treated with TlSWO, avicel's smooth surface structure transitions into a rough texture.This is consistent with previous studies suggesting a cell wall modi cation activity usually expressed by swollenin interference [39,40].Previous studies have also shown that swollenins can synergize with other enzymes such as cellulases and xylanases [10,11,24].In this study we explored TlSWO's ability to boost cellulosic substrates hydrolysis by different enzymes.When the cellulose was incubated with TlSWO and cellobiohydrolases, a greater increase in glucose yields was observed.However, we observed no signi cant synergistic effect between TlSWO and endoglucanases.These results differ from a previous report, in which swollenin exhibited strong synergistic interaction with endoglucanases [3].We therefore propose that TlSWO has better synergistic activity with cellobiohydrolases.As the classical, C 1 -C x model hypothesized the process of cellulose degradation (C 1 : non-hydrolytic component, C x : endo-or exo-acting cellulases).One proposal using this model hypothesizes C 1 can disrupt cellulose by displacing hydrogen bonds in the micro bril, leading to a more available structure for C x [4,41].Eibinger et al. and Kang et al. speculated swollenins, much like endoglucanases can play the possible role of C 1 enzymes due to their disruptive activity in the enzymatic sacchari cation of lignocellulosic substrates, although the action mode of various C 1 enzymes needs to be further explored [10,18,42].Our results of PASC and CNC treated with or without TlSWO were used to compare the effects of swollenin on different crystallinity materials.We show that when PASC was incubated with TlSWO and cellobiohydrolases, the total glucose concentration was increased by 32%, compared to incubating with only cellobiohydrolases.However, when using CNC as the substrate, there is no signi cant difference between the groups with or without TlSWO, suggesting swollenin has some additional ability to bind and disrupt the amorphous cellulose than the crystalline cellulose.This may be due to its endoglucanase like activity as demonstrated on the model substrates we previously used, or possibly because it is able to bind amorphous substrates more easily than crystalline substrates due to its CBM.

Conclusions
In the present study, we report TlSWO, from T. leycettanus JCM12802.TlSWO as an acidic and mesophilic swollenin, showing activity towards lichenan, barley β-glucan, carboxymethyl cellulose sodium and laminarin.A greater increase in glucose yield was observed when clean cellulose substrates were incubated with TlSWO and cellobiohydrolases.Moreover, TlSWO exhibited synergetic effects on cellobiohydrolase when using PCS and PASC as substrate.However, there were no observed signi cant synergistic effect between TlSWO and endoglucanases, suggesting that TlSWO has better coordination with cellobiohydrolases.
Comparing to chemical pretreatment, biological pretreatment utilizing enzyme has massive research potential given the advantages such as low energy consumption, environmentally friendliness, and lower production cost.Different lignocellulosic biomass needs different types of pretreatments because the structural features of cellulose plays an important role in enzyme hydrolysis rate that affects outcomes such as polymerization degree, cellulose crystallization arrangement, surface area accessibility, particles size, and the existence of hemicelluloses and lignin [2,43].We show that TlSWO directly caused the alteration of the cellulose structure, consequentially increasing its hydrolysis rate.By optimized preconditioning and molecular design, TlSWO could be a promising additive for improving lignocellulosic biomass generation performance observing hydrolysis e ciency.

Methods
Strains and Plasmids T. leycettanus JCM12802 as the donor strain was purchased from Japan Collection of Microorganisms RIKEN BioResource Center (Tsukuba, Japan).Escherichia coli Trans I-T1 (TransGen, Beijing, China) was used for routine gene cloning.T. reesei AST1116 and P. pastoris GS115 (Invitrogen, Carlsbad, CA, USA) were both used as the host for gene expression.The plasmids of pPIC9 (Invitrogen) and pTrEno were used for driving Tlswo gene expression in P. pastoris and in T. reesei.The plasmid construction of pTrEno was according to the description of Linger et al. [44].

Gene cloning and recombinant protein expression
T. leycettanus JCM12802 was cultured at 42 ℃ for 3-5 days in medium containing 5.0 g/L (NH 4 ) 2 SO 4 , 1.0 g/L KH 2 PO 4 , 0.5 g/L MgSO 4 •7H 2 O, 0.2 g/L CaCl 2 , 10.0 mg/L FeSO 4 •7H 2 O, 30.0 g/L wheat bran, 30.0 g/L soybean meal, and 30.0 g/L corncob.Genomic DNA of T. leycettanus JCM12802 was extracted with the DNA isolation kit (Tiangen) following the manufacturer instructions and was used as a template for PCR ampli cation.Total RNA isolation and rst strand cDNA synthesis were carried out as previously described [50].PCR ampli cation was carried out using Fastpfu DNA polymerase and which purchased from TransGen.The fragment of Tlswo corresponding to the 21-503 amino acid sequence was ampli ed with primers P1 (5'-GCTCGTGCTCAGAGCAGCTGTGCAGG-3') and P2 (5'-CTAGATTACCTAGGTAAACTGCACC-3').The ampli ed PCR product was cloned into pPIC9 and pTrEno vector using Gibson Assembly (New England Biolabs, Ipswich, MA, USA) according to the manufacturer's protocol for cloning and expression.Escherichia coli Trans I-T1 (TransGen, Beijing, China) which used for routine gene cloning was grown at 37 °C overnight in Luria-Bertani (LB) medium supplemented with 50 mg/mL of ampicillin (Sigma, St. Louis, MO).When using P. pastoris GS115 as the expression host, the recombinant plasmids of pPIC9-Tlswo was linearized with BglII (New England Biolabs, UK) and transformed into the expression host via electroporation.The positive transformants were screened on minimal dextrose medium (MD) at 30 ℃ for 3 or 4 days until the single colonies appeared and placed into shaking tubes for enzyme production according to the protocol described in the Pichia Expression Kit (Invitrogen).The large-scale fermentation was performed as the previously described [30].The recombinant P. pastoris GS115 containing pPIC9-Tlswo was grown at 30 ℃ in 400 mL BMGY medium in a 1 L shaking ask for 48 hours.Cells were collected and re-suspended again with 200 mL buffered methanol-complex medium (BMMY) medium with 0.5% (v/v) methanol and cultured at 30 °C for 72 h with shaking (200 rpm).Methanol was added into the medium every 24 hours.When using T. reesei AST1116 as the expression host, the recombinant plasmids of pTrEno-Tlswo were linearized with SbfI (New England Biolabs, UK) and then transformed into T.reesei AST1116 via electroporation.Potato dextrose (PD) plates were used for spore production and PDHX plates (PD plates with hygromycin and TritonX-100 at a nal concentration of 100 μg/mL and 0.1% respectively) were used for the screening of potential T. reesei which allowed to grow for 2 to 3 days at 30 °C.Growth medium for TlSWO expression was Mandels and Andreotti medium with 5% glucose (MAG).Subsequently complete medium lactose (CML) was used for the overexpression of the transformants.The medium protocols of MAG and CML were performed following published protocol [44].For large-scale fermentation, the positive transformants spore stocks were streaked on potato dextrose agar plates and allowed to grow 2 to 3 days.Then the spore was extracted from the plate and deposited into 1.0 L of MAG medium in a 2.8 L shake ask.The culture was grown at 28 °C with at 225 RPM for 24 h, after which the entire 1.0 L was transferred to 7.0 L of the same medium in a bioreactor.
Then the culture broth were extracted for further analysis by SDS-PAGE and activity assay.Culture broths were clari ed via centrifugation and transferred to microcentrifuge tubes.Broths were diluted 3:1 in 4×LDS sample buffer (Life Technologies Corp., Carlsbad, CA, USA) with 50 μL/mL β-mercaptoethanol as a reducing agent.Samples were incubated at 95 °C for 5 min prior to loading onto NuPAGE SDS gels with MOPS buffer, electrophoresed at 200 V constant for approximately 40 min.

Protein puri cation
Fermentation broths were harvested and sequentially vacuum-ltered.This ltered broth was then concentrated by tangential ultra ltration with a 10 kDa MWCO.The broths were roughly concentrated to 100 mL.The nal concentrated volume was exchanged with at least 2.0 L of 20 mM Bis-Tris pH 6.5 to remove residual peptides and other low molecular weight debris.The following puri cation steps were then performed according to previous publications [44].The crude enzyme was puri ed through hydrophobic interaction chromatography (HIC) using a 26/10 Phenyl Sepharose Fast Flow column (GE Healthcare, Chicago, USA).Then the protein went through anion exchange chromatography using a 10/100 anion exchange column packed with Source 15Q (GE Healthcare, Chicago, USA) followed by HIC using a Source15 iso 10/100 column (GE Healthcare, Chicago, USA) and SEC using a 26/60 Superdex 75 column (GE Healthcare, Chicago, USA) and 20 mM acetate buffer pH 5.0, 100 mM NaCl as mobile phase.SDS-PAGE were performed to assess purity of the protein TlSWO.Electrophoretically separated on a 12% SDS-PAGE and visualized by Coomasie Blue staining.Protein concentration was measured with a NanoDrop 2000 Spectrophotometer (Thermo Fisher Scienti c Inc., Rockford, USA) and the Bradford protein assay kit (Bio-Rad).

Effect of pH and temperature on TlSWO activity
Further, the effects of pH and temperature on the activities of TlSWO were measured and compared.To determine the optimum pH of TlSWO, the activities were assayed with 1% lichenan (w/v) in buffer of different pH, 100 mM glycine-HCl (pH 1.0-3.0),McIlvaine buffer (pH 3.0-8.0)and glycine-NaOH (pH 9.0-12.0).For pH stability, TlSWO was preincubated at 37 °C for 1 h in buffers of different pH (1.0-12.0)and subjected to the residual activity assay.For the optimum temperature of TlSWO was determined at optimal pH over the temperature range from 30 °C to 80 °C.The thermostability assay of TlSWO (100 μg/mL) was carried out by preincubating at 37 °C, 50 °C, 60 °C or 70 °C for 0-60 min, and aliquots of 100 μL were withdrawn at different time points for residual activity assay.

Light and scanning electron microscopic analyses
Avicel PH-101was used as a solid cellulosic substrate.Ten milligrams of Avicel were incubated with different amount of puri ed TlSWO in 100 mM citric acid-Na 2 HPO 4 buffer (pH 4.0).The experiment was carried out on a rotary shaker at 40 °C for different time intervals.Control experiments without TlSWO were also performed under the same conditions.The physical structure of Avicel bers was initially observed using light microscopy (Olympus TH4-200, Japan).Subsequently, photomicrographs of the samples were captured using a scanning electron microscope (Hitachi SU8010, Tokyo, Japan) at a voltage of 15 kV.

Synergism between TlSWO and cellulases
The substrate used in this work were NREL dilute acid pre-treated corn stover (PCS) P120927, cellulose nanocrystals (CNCs), phosphoric acid swollen cellulose (PASC) and each substrate equivalent to 8.5 mg of glucan.For CNC preparations from Avicel, about 2 g of Avicel were added to pre-heated HCl at 80 °C.
Then run the acid hydrolysis for 4 hours, stirring every 15 minutes with a glass or Te on rod followed by centrifugation several times, 1600×g for 10 minutes.Decant supernatant and resuspend pellet in DI water until the pH reached 5.0.The CNCs were then suspended in the supernatant after centrifugation.Collect the translucent supernatant and resuspend pellet in DI water and shake vigorously to break up clumps until the supernatant is no longer translucent.The reaction mixtures were carried out in triplicate vials at 40 °C and each substrate was suspended in 20 mM sodium acetate buffer, pH 5.0.The enzyme cocktail comprised endoglucanase I from Trichoderma longibrachiatum (Megazyme Co., Bray, Ireland), cellobiohydrolases Cel7A from Penicillium funiculosum and β-glucosidase from Aspergillus niger (Megazyme Co., Bray, Ireland) at a concentration (mg protein/g of glucan) of 2, 13 and 1, respectively.The reaction was followed for 120 h, with sampling every 24 h. 100 μL samples containing both solids and liquid are removed from the mixtures and diluted for sugar analysis by high performance liquid chromatography (HPLC) with a BioRad HPX-87H column.Control experiments using BSA were also performed under the same conditions as mentioned above.

Additional Files
The substrate speci city of the TlSWO.Hydrolysis reactions on 1% barley β-glucan, 1% CMC-Na, 1% laminarin and 0.5% lichenan, were carried out overnight at pH 4.0, 50 °C.The linkage type of each substrate was labeled.

Supplementary Files
This is a list of supplementary les associated with this preprint.Click to download.

Fig. S1
Fig. S1 Production and puri cation of recombinant TlSWO in Pichia pastoris and Trichoderma reesei.A: SDS-PAGE analysis of the culture supernatant of TlSWO produced by Pichia pastoris.Lane 1: the crude enzyme of TlSWO; lane 2, the puri ed TlSWO; lane 3, the deglycosylated TlSWO; B: SDS-PAGE analysis of the culture supernatant of TlSWO produced by Trichoderma reesei.Lane 1: the crude enzyme of positive control contained TrCel7; lane 2, the crude enzyme of TlSWO; lane 3, the puri ed TlSWO.

Figures
Figures

Figure 1 Sequence
Figure 1

Figure 1 Sequence
Figure 1

3
Hydrolysis capacity of TlSWO to degrade CMC-Na, lichenan, barley β-glucan and laminarin.All the reactions were carried out overnight at pH 4.0, 37 °C using the enzyme in a nal concentration of 100 μg/mL.

Figure 5 Light
Figure 5

Figure 5 Light
Figure 5

Figure 7 Conversion
Figure 7 FigureS1.pdfFigureS1.pdfFigureS2.pdfFigureS2.pdfFigureS3.pdfFigureS3.pdf the National Renewable Energy Laboratory for the U.S. Department of Energy (DOE) under Contract No. DE-AC36-08GO28308.Funding provided by U.S. Department of Energy O ce of Energy E ciency and Renewable Energy Bioenergy Technologies O ce.The views expressed in the article do not necessarily represent the views of the DOE or the U.S. Government.We thank the U.S. Department of Energy Bioenergy Technologies O ce for funding under Contract DE-AC36-08GO28308 with the National Renewable Energy Laboratory. 47. Petersen TN, Brunak S, Heijne GV, Nielsen HH.SignalP 4.0: discriminating signal peptides from transmembrane regions.Nat Methods.2011;8(10):785-6.4 .Hulo N, Bairoch A, Bulliard V, Cerutti L, De Castro E, Langendijk-Genevaux PS, et al.The PROSITE database.Nucleic Acids Res.2006;34(Database issue):D227-30.49.Ferrè F, Clote P. DiANNA: a web server for disul de connectivity prediction.Nucleic Acids Res.