Yeast GH30 Xylanase from Sugiyamaella lignohabitans Is a Glucuronoxylanase with Auxiliary Xylobiohydrolase Activity

Xylanases are the enzymes that catalyze the breakdown of the main hemicellulose present in plant cell walls. They have attracted attention due to their biotechnological potential for the preparation of industrially interesting products from lignocellulose. While many xylanases have been characterized from bacteria and filamentous fungi, information on yeast xylanases is scarce and no yeast xylanase belonging to glycoside hydrolase (GH) family 30 has been described so far. Here, we cloned, expressed and characterized GH30 xylanase SlXyn30A from the yeast Sugiyamaella lignohabitans. The enzyme is active on glucuronoxylan (8.4 U/mg) and rhodymenan (linear β-1,4-1,3-xylan) (3.1 U/mg) while its activity on arabinoxylan is very low (0.03 U/mg). From glucuronoxylan SlXyn30A releases a series of acidic xylooligosaccharides of general formula MeGlcA2Xyln. These products, which are typical for GH30-specific glucuronoxylanases, are subsequently shortened at the non-reducing end, from which xylobiose moieties are liberated. Xylobiohydrolase activity was also observed during the hydrolysis of various xylooligosaccharides. SlXyn30A thus expands the group of glucuronoxylanases/xylobiohydrolases which has been hitherto represented only by several fungal GH30-7 members.

Based on amino acid sequence alignment ( Figure S1, Supplementary Materials), two catalytic residues were identified: Glu199 as an acid/base and Glu292 as a nucleophile. The alignment also revealed that SlXyn30A displays structural features of the GH30-7 enzymes including a longer β2-α2 loop, a lack of α6-helix, and a presence of β-strands β8A and β8B [12]. Moreover, SlXyn30A contains several structural elements typical for GH30-7 glucuronoxylanases/xylobiohydrolases ( Figure 1). The first is a presence of Arg46 which was suggested to be responsible for a MeGlcA recognition and glucuronoxylanase activity of TcXyn30B [4]. Although the 3D structure of another GH30-7 glucuronoxylanase, TtXyn30A, did not confirm such a role of the arginine, mutational studies of TtXyn30A indicated its importance [9]. The second structural feature considered to be responsible for xylobiohydrolase activity is the length and amino acid sequence of the β2-α2 loop. This loop is of the same length in SlXyn30A, TcXyn30B and TtXyn30A, and is longer than in other GH30-7 enzymes (e.g., TrXynVI; Figure 1). Asn93 in TcXyn3B and Asp78 in TtXyn30A in this loop were shown to interact with Xylp residue accommodated in the −2a subsite [4,9]. The corresponding residue in SlXyn30A is Asn90 which may play a similar role. other GH30-7 enzymes (e.g., TrXynVI; Figure 1). Asn93 in TcXyn3B and Asp78 in TtXyn30A in this loop were shown to interact with Xylp residue accommodated in the −2a subsite [4,9]. The corresponding residue in SlXyn30A is Asn90 which may play a similar role.  [7], and Talaromyces leycettanus TlXyn30A [10]. The secondary structure elements and numbering of TcXyn30B are shown on top, numbering of SlXyn30A is at the bottom. Arginine suggested to be responsible for the recognition of MeGlcA substitution is shown in blue and is marked by a blue up triangle. Longer β2-α2 loop in xylobiohydrolases is highlighted in yellow and the residue interacting with Xylp moiety occupying the −2a subsite is brown and marked by a brown down triangle.

Recombinant Strain Selection
Four clones with integrated pPICZαA vector carrying the Slxyn30A gene were selected from the ZeocinTM (250 mg/L) plate. The clones were individually cultivated in shake flasks, and after 120 h of induction, supernatant was screened by SDS-PAGE electrophoresis for the presence of 50 kDa protein which should correspond to a mature enzyme. The expected protein was confirmed in three transformants ( Figure S2). The molecular mass of the enzyme was, however, a little bit higher (58 kDa), presumably due to glycosylation. According to NetNGlyc server, the amino acid sequence of SlXyn30A contains 4 potential N-glycosylation sites (N86, N114, N252, N306). The first one is conserved in TcXyn30B where is actually glycosylated [2]. Three transformants were then tested for xylanase activity and the transformant with the highest activity (transformant 4) was used for the determination of the catalytic properties of SlXyn30A.

Thermal and pH Optima and Stability
SlXyn30A showed a temperature optimum of 50 °C, showing 71% and 24% of maximal activity at 40 °C and 60 °C, respectively (Figure 2a). pH optimum was around 3.5 and the enzyme was active in acidic range of pH, keeping only 6.6% of the maximal activity at pH 6 ( Figure 2b). The enzyme was stable at temperatures up to 50 °C, while at 60 °C it completely lost its activity within 30 min. T. reesei TrXynIV (AAP64786.1), Acremonium alcalophilum AaXyn30A [7], and Talaromyces leycettanus TlXyn30A [10]. The secondary structure elements and numbering of TcXyn30B are shown on top, numbering of SlXyn30A is at the bottom. Arginine suggested to be responsible for the recognition of MeGlcA substitution is shown in blue and is marked by a blue up triangle. Longer β2-α2 loop in xylobiohydrolases is highlighted in yellow and the residue interacting with Xylp moiety occupying the −2a subsite is brown and marked by a brown down triangle.

Recombinant Strain Selection
Four clones with integrated pPICZαA vector carrying the Slxyn30A gene were selected from the ZeocinTM (250 mg/L) plate. The clones were individually cultivated in shake flasks, and after 120 h of induction, supernatant was screened by SDS-PAGE electrophoresis for the presence of 50 kDa protein which should correspond to a mature enzyme. The expected protein was confirmed in three transformants ( Figure S2). The molecular mass of the enzyme was, however, a little bit higher (58 kDa), presumably due to glycosylation. According to NetNGlyc server, the amino acid sequence of SlXyn30A contains 4 potential N-glycosylation sites (N86, N114, N252, N306). The first one is conserved in TcXyn30B where is actually glycosylated [2]. Three transformants were then tested for xylanase activity and the transformant with the highest activity (transformant 4) was used for the determination of the catalytic properties of SlXyn30A.

Thermal and pH Optima and Stability
SlXyn30A showed a temperature optimum of 50 • C, showing 71% and 24% of maximal activity at 40 • C and 60 • C, respectively ( Figure 2a). pH optimum was around 3.5 and the enzyme was active in acidic range of pH, keeping only 6.6% of the maximal activity at pH 6 ( Figure 2b). The enzyme was stable at temperatures up to 50 • C, while at 60 • C it completely lost its activity within 30 min.

Hydrolysis of Polysaccharides
SlXyn30A showed the highest specific activity on glucuronoxylan (GX) (8.4 U/mg), while activity on linear β-1,3-1,4-xylan (rhodymenan, Rho) was about 2.5 times lower (3.1 U/mg). Specific activity on arabinoxylan was extremely low (0.03 U/mg). TLC analysis of hydrolysis products showed that GX was initially cleaved to a series of acidic xylooligosaccharides (XOs) (Figure 3) which were partly shortened after a prolonged incubation. The acidic XOs shortening was accompanied by xylobiose (Xyl2) formation, which was detectable after 1 h, indicating xylobiohydrolase activity of the enzyme. Later, after 5 h, xylotetraose (Xyl4) also appeared as a result of transglycosylation reaction. After an application of β-xylosidase to 24 h hydrolysate, the acidic XOs were hydrolyzed to MeGlcA 2 Xyl2 which indicates that all acidic XOs had MeGlcA substitution on the second xylopyranosyl (Xylp) residue from the reducing end. This mode of GX hydrolysis is typical for the GH30 glucuronoxylanases [13][14][15]. Glucuronoxylanase activity accompanied by xylobiohydrolase activity has been already described for other GH30-7 enzymes The acidic XOs shortening was accompanied by xylobiose (Xyl 2 ) formation, which was detectable after 1 h, indicating xylobiohydrolase activity of the enzyme. Later, after 5 h, xylotetraose (Xyl 4 ) also appeared as a result of transglycosylation reaction. After an application of β-xylosidase to 24 h hydrolysate, the acidic XOs were hydrolyzed to MeGlcA 2 Xyl 2 which indicates that all acidic XOs had MeGlcA substitution on the second xylopyranosyl (Xylp) residue from the reducing end. This mode of GX hydrolysis is typical for the GH30 glucuronoxylanases [13][14][15]. Glucuronoxylanase activity accompanied by xylobiohydrolase activity has been already described for other GH30-7 enzymes TcXyn30B and TtXyn30A [2,3]. Our results indicate that SlXyn30A might also be a glucuronoxylanase/xylobiohydrolase. However, the acidic XOs in the 24 h hydrolysate were not shortened exclusively to MeGlcA 2 Xyl 2 and MeGlcA 2 Xyl 3 , as was observed in the case of TtXyn30A [3], but longer acidic products remained in the hydrolysate even after 5-day incubation or after an addition of fresh enzyme. In this regard, SlXyn30A resembles much more TcXyn30B than TtXyn30A. MALDI-ToF analysis of the 5-day hydrolysate confirmed the presence of a broad spectrum of acidic XOs from MeGlcAXyl 2 to MeGlcAXyl 12 (MeGlcAXyl 2 -MeGlcAXyl 5 prevailing) ( Figure 4). Xyl 2 was predominant neutral XO but traces of longer XOs up to DP10 were also observed. Kinetic parameters determined for GX were K m 16.8 mg/mL, k cat 20.2 s −1 and k cat /K m 1.2 mL/mg·s. Rho was hydrolyzed to a mixture of β-1,4-linked XOs and β-1,3-1,4-XOs, among which the most predominant were Xyl 2 , Xyl 4 , Xylβ1-3Xylβ1-4Xyl and Xylβ1-4Xylβ1-3Xylβ1-4Xyl ( Figure 3). β-Xylosidase added to the 24 h-hydrolysate of Rho cleaved all β-1,4-linked XOs to Xyl and all the β-1,3-1,4-linked XOs to Xylβ1-3Xylβ1-4Xyl. The final extent of hydrolysis seems to be higher for Rho than GX. Hydrolysis of AraX was very weak ( Figure 3). In this case, Xyl 2 was a predominant product accompanied by a few Ara-substituted XOs, most probably having the Ara substitution on the non-reducing end (they were not attacked by β-xylosidase).

Hydrolysis of Oligosaccharides
Xylobiohydrolase activity of SlXyn30A was also observed on various XOs. The main hydrolysis product released from Xyl4 was Xyl2, but Xyl6 was also formed, presumably through transglycosylation reaction (Figure 5a). Xyl5 was mainly cleaved to Xyl2 and Xyl3, and Xyl6 to Xyl2 and Xyl4. In both cases, the transglycosylation products with DP higher by two than the substrate were observed during the early stages of reaction ( Figure 5b). Xyl3 was the worst substrate, being cleaved to Xyl2 and Xyl only after a prolonged incubation. Transglycosylation products were also formed during the action of SlXyn30A on 4-nitrophenyl glycosides of β-1,4-xylobiose (Xyl2-NP) and β-1,4-xylotriose (Xyl3-NP). Xyl2 release was accompanied by a liberation of 4-nitrophenol from Xyl2-NP and 4-nitrophenyl xyloside from Xyl3-NP. This mode of action has unambiguously confirmed that xylobiose moiety is released from the non-reducing end of the substrates. Compared to linear β-1,4-

Hydrolysis of Oligosaccharides
Xylobiohydrolase activity of SlXyn30A was also observed on various XOs. The main hydrolysis product released from Xyl 4 was Xyl 2, but Xyl 6 was also formed, presumably through transglycosylation reaction (Figure 5a). Xyl 5 was mainly cleaved to Xyl 2 and Xyl 3 , and Xyl 6 to Xyl 2 and Xyl 4 . In both cases, the transglycosylation products with DP higher by two than the substrate were observed during the early stages of reaction (Figure 5b). Xyl 3 was the worst substrate, being cleaved to Xyl 2 and Xyl only after a prolonged incubation. Transglycosylation products were also formed during the action of SlXyn30A on 4-nitrophenyl glycosides of β-1,4-xylobiose (Xyl 2 -NP) and β-1,4-xylotriose (Xyl 3 -NP). Xyl 2 release was accompanied by a liberation of 4-nitrophenol from Xyl 2 -NP and 4-nitrophenyl xyloside from Xyl 3 -NP. This mode of action has unambiguously confirmed that xylobiose moiety is released from the non-reducing end of the substrates. Compared to linear β-1,4-XOs and the corresponding NP-glycosides, MeGlcA-substituted XOs of the same DP were cleaved faster ( Figure 6). About 60% of MeGlcA 3 Xyl 4 was hydrolyzed after 5 min of the reaction, when only about 10% of Xyl 4 or Xyl 3 -NP were converted. Specific activities were 0.037 U/mg for Xyl 3 , 0.222 U/mg for Xyl 4 , 1.9 U/mg for MeGlcA 3 Xyl 4 and 56.7 U/mg for MeGlcA 3 Xyl 3 showing the preference of the enzyme for MeGlcA-substituted substrates. It should be noted that the transglycosylation reaction was not observed during the processing of the acidic XOs MeGlcA 3 Xyl 4 and MeGlcA 3 Xyl 3 .
XOs and the corresponding NP-glycosides, MeGlcA-substituted XOs of the same DP were cleaved faster ( Figure 6). About 60% of MeGlcA 3 Xyl4 was hydrolyzed after 5 min of the reaction, when only about 10% of Xyl4 or Xyl3-NP were converted. Specific activities were 0.037 U/mg for Xyl3, 0.222 U/mg for Xyl4, 1.9 U/mg for MeGlcA 3 Xyl4 and 56.7 U/mg for MeGlcA 3 Xyl3 showing the preference of the enzyme for MeGlcA-substituted substrates. It should be noted that the transglycosylation reaction was not observed during the processing of the acidic XOs MeGlcA 3 Xyl4 and MeGlcA 3 Xyl3.  The activity of SlXyn30A was further tested on oligosaccharides containing β-1,3-or β-1,2-linkages or arabinosyl substitution (Figure 7, compounds 1-23). The enzyme was able to slowly release methanol from methyl β-1,4-xylobioside (2) indicating that Xylp unit accommodated in the +1 subsite is not indispensable for enzyme activity. Methyl β-1,4- XOs and the corresponding NP-glycosides, MeGlcA-substituted XOs of the same DP were cleaved faster ( Figure 6). About 60% of MeGlcA 3 Xyl4 was hydrolyzed after 5 min of the reaction, when only about 10% of Xyl4 or Xyl3-NP were converted. Specific activities were 0.037 U/mg for Xyl3, 0.222 U/mg for Xyl4, 1.9 U/mg for MeGlcA 3 Xyl4 and 56.7 U/mg for MeGlcA 3 Xyl3 showing the preference of the enzyme for MeGlcA-substituted substrates. It should be noted that the transglycosylation reaction was not observed during the processing of the acidic XOs MeGlcA 3 Xyl4 and MeGlcA 3 Xyl3.  The activity of SlXyn30A was further tested on oligosaccharides containing β-1,3-or β-1,2-linkages or arabinosyl substitution (Figure 7, compounds 1-23). The enzyme was able to slowly release methanol from methyl β-1,4-xylobioside (2) indicating that Xylp unit accommodated in the +1 subsite is not indispensable for enzyme activity. Methyl β-1,4- The activity of SlXyn30A was further tested on oligosaccharides containing β-1,3or β-1,2-linkages or arabinosyl substitution (Figure 7, compounds 1-23). The enzyme was able to slowly release methanol from methyl β-1,4-xylobioside (2) indicating that Xylp unit accommodated in the +1 subsite is not indispensable for enzyme activity. Methyl β-1,4xylotrioside (4) was hydrolyzed much faster than (2), and exclusively to Xyl 2 and Xyl-Me in accordance with xylobiohydrolase activity of the enzyme. SlXyn30A was able to cleave very slowly also β-1,2and β-1,3-linkages in X4X2XMe (5) and X4X3XMe (7), the former being cleaved faster but significantly slower than (4). However, if Xylp residues at the non-reducing end are connected by α-1,4-linkage, the substances (6,8) are not hydrolyzed. The compounds, in which the substitution would occur on Xylp unit accommodated in the −1 subsite, were not hydrolyzed (9)(10)(11)(12). On the other hand, some substitutions of Xylp in the −2a subsite were tolerated. The tolerance was influenced by three factors: (1) position of  (3) whether the compound was further elongated at the non-reducing end, i.e., it carried additional Xylp residue that was accommodated in the subsite −3. If the Xylp unit accommodated in the −2a subsite is decorated at position 2 by the MeGlcA, the substrate (MeGlcA 3 Xyl 3 , 16) is readily hydrolyzed. When the MeGlcA substitution is replaced by α-L-arabinofuranose, the substrate (A2X4X4X, 17) is slowly attacked, but a change for β-D-xylopyranose (X2X4XMe, 13) caused a resistance to the enzyme attack. The tolerance at position 3 is greater since A3X4XMe (18) and X3X4XMe (14) as well as Xα3X4XMe (15) were slowly processed. SlXyn30A was also able to cleave the substrate (19) with doubly 2,3-O-arabinosylated Xylp residue accommodated in the −2a subsite. However, if the compounds with the substitution on Xylp residue in the −2a subsite are by one Xylp longer at the non-reducing end (i.e., they occupy also the −3 subsite), the hydrolysis is slowed down (20)(21)(22) or even abolished (23).
in accordance with xylobiohydrolase activity of the enzyme. SlXyn30A was able to cleave very slowly also β-1,2-and β-1,3-linkages in X4X2XMe (5) and X4X3XMe (7), the former being cleaved faster but significantly slower than (4). However, if Xylp residues at the nonreducing end are connected by α-1,4-linkage, the substances (6,8) are not hydrolyzed. The compounds, in which the substitution would occur on Xylp unit accommodated in the −1 subsite, were not hydrolyzed (9)(10)(11)(12). On the other hand, some substitutions of Xylp in the −2a subsite were tolerated. The tolerance was influenced by three factors: (1) position of the decoration (2 and/or 3); (2) the nature of the substituent; and (3) whether the compound was further elongated at the non-reducing end, i.e., it carried additional Xylp residue that was accommodated in the subsite −3. If the Xylp unit accommodated in the −2a subsite is decorated at position 2 by the MeGlcA, the substrate (MeGlcA 3 Xyl3, 16) is readily hydrolyzed. When the MeGlcA substitution is replaced by α-L-arabinofuranose, the substrate (A2X4X4X, 17) is slowly attacked, but a change for β-D-xylopyranose (X2X4XMe, 13) caused a resistance to the enzyme attack. The tolerance at position 3 is greater since A3X4XMe (18) and X3X4XMe (14) as well as Xα3X4XMe (15) were slowly processed. SlXyn30A was also able to cleave the substrate (19) with doubly 2,3-O-arabinosylated Xylp residue accommodated in the −2a subsite. However, if the compounds with the substitution on Xylp residue in the −2a subsite are by one Xylp longer at the non-reducing end (i.e., they occupy also the −3 subsite), the hydrolysis is slowed down (20)(21)(22) or even abolished (23).

Discussion
Many xylanases have been isolated and characterized from bacteria and filamentous fungi while a number of characterized yeast xylanases is limited and no yeast GH30 xylanase has been described so far. Catalytic properties of eukaryotic GH30 xylanases belonging to GH30-7 subfamily, where SlXyn30A is also grouped, are diverse. It was, Figure 7. Various XOs tested as the substrates for SlXyn30A. The site of cleavage is denoted by an arrow, X marks compounds which were not attacked.

Discussion
Many xylanases have been isolated and characterized from bacteria and filamentous fungi while a number of characterized yeast xylanases is limited and no yeast GH30 xylanase has been described so far. Catalytic properties of eukaryotic GH30 xylanases belonging to GH30-7 subfamily, where SlXyn30A is also grouped, are diverse. It was, therefore, interesting to determine the specificity of the yeast xylanase, which may be related to a biotechnological potential of the yeast due to its reported ability to convert xylose to xylitol, ethanol or organic acids [26,27]. The GH30 xylanase SlXyn30A was cloned, expressed, and characterized. SlXyn30A showed the highest amino acid similarity to glucuronoxylanases/xylobiohydrolases TtXyn30A and TcXyn30B. SlXyn30A, similarly to these fungal enzymes, contains Arg46 which was shown to play a role in MeGlcA recognition [4,9]. Another aspect of similarity between the three enzymes is the exact length of β2-α2 loop, which may affect the occupation of the −3 subsite. Moreover, Asn90 in this loop corresponds to Asn93 in TcXyn3B and Asp78 in TtXyn30A which are supposed to play a role in xylobiohydrolase activity of the GH30-7 glucuronoxylanases [4,9]. All these structural features of SlXyn30A are in accordance with its biochemical properties. Glucuronoxylan was the best substrate among the heteroxylans studied and during its hydrolysis, the MeGlcA residue was accommodated in the −2b subsite of the enzyme, yielding acidic XOs of general formula MeGlcA 2 Xyl n . Such a hydrolysis of GX is typical for GH30 glucuronoxylanases [13][14][15]. Later, the acidic XOs were partially shortened by SlXyn30A through a liberation of Xyl 2 from the non-reducing end, similarly to GH30 xylobiohydrolases [7,8] and glucuronoxylanses/xylobiohydrolases [2,3]. The shortening of the acidic XOs was not complete and the aldouronic acids of a medium size persisted in the SlXyn30A hydrolysate. From this point of view, SlXyn30A most resembles TcXyn30B, which shows an essentially identical end-stage hydrolysis profile [2], and slightly differs from TtXyn30A, which liberated the acidic XOs and Xyl 2 simultaneously, and Xyl 2 , MeGlcA 2 Xyl 2 and MeGlcA 2 Xyl 3 were the only final products of GX hydrolysis [3]. In addition, kinetic parameters of SlXyn30A on GX are also similar to those of TcXyn30B [2].
The xylobiohydrolase activity of SlXyn30A was even more pronounced during the hydrolysis of rhodymenan. It was reflected in an accumulation of not only Xyl 2 , but also isomeric xylotriose Xylβ1-3Xylβ1-4Xyl and isomeric xylotetraose Xylβ1-4Xylβ1-3Xylβ1-4Xyl. The release of Xylβ1-3Xylβ1-4Xyl is in agreement with the ability of SlXyn30A to cleave X3X4XMe (Figure 7, 14) to X3X4X and methanol. Hydrolysis of X4X3XMe is very slow (not finished after 145 h of hydrolysis) indicating a very limited ability of the enzyme to hydrolyze β-1,3-xylosidic linkage. This is in accordance with the presence of Xylβ1-4Xylβ1-3Xylβ1-4Xyl in 24 h hydrolysate of Rho. In contrast, xylobiohydrolases AaXyn30A and HcXyn30A are able to cleave β-1,3-linkages much more efficiently and therefore Xylβ1-4Xylβ1-3Xylβ1-4Xyl was not accumulated but cleaved to two molecules of Xyl 2 in their hydrolysates [7,8].
Linear β-1,4-XOs were processed by SlXyn30A in the same way as was described for TcXyn30B and TtXyn30A [2,3]. Xyl 2 was the main product and XOs longer by two xylose units were formed via transglycosylation. For SlXyn30A, Xyl 3 was a much worse substrate than Xyl 4 , suggesting that an occupation of the +2 subsite has a significant positive effect on the enzyme activity.
The ability of SlXyn30A to recognize MeGlcA substitution was confirmed by a comparison of specific activities on linear XOs and the corresponding XOs decorated by MeGlcA. Specific activity on MeGlcA 3 Xyl 3 was about 1500 times higher than on Xyl 3 . On the other hand, MeGlcA 3 Xyl 4 was only about 8 times better substrate than Xyl 4 . MeGlcA 3 Xyl 3 was hydrolyzed about 30 times faster than MeGlcA 3 Xyl 4 . These data clearly indicate the preference of the enzyme for the MeGlcA-decorated substrates and for the substrates not occupying the −3 subsite of the enzyme. The latter preference was confirmed also during the hydrolysis of various methyl glycosides and arabinoxylooligosaccharides when elongation of the XO chain to the −3 subsite of the enzyme (Figure 7, 16 vs. 20, 17 vs. 21, 18 vs.  22, 19 vs. 23) caused a slowdown or an abolishment of the reaction. The evaluation of these compounds allowed us to draw the following conclusions on the requirement of SlXyn30A on the structure of the substrates. First, SlXyn30A does not tolerate any substitution on Xylp residue accommodated in the −1 subsite. Xylose residue accommodated in the −2a subsite may be substituted at position 2 by MeGlcA, which improves the activity, and by α-L-arabinofuranose but not by β-D-xylopyranose. Position 3 may be both arabinosylated and xylosylated, but the activity on such substrates is lower compared to unsubstituted XOs. The elongation of the substrate main chain that results in an occupation of the −3 subsite may dramatically decrease the hydrolysis rate.
Four transformants from YPD plates with 250 mg/L of ZeocinTM were selected and screened for a recombinant SlXyn30A production. Additionally, 500 mL shake-flasks with 100 mL of BMGY medium were inoculated with a single P. pastoris colony and cultivated at 30 • C and 200 rpm for 22 h. The induction of enzyme was carried out as reported previously [41]. The cells were harvested by centrifugation (7197× g, 10 • C, 5 min), resuspended in 6 mL of sterile distilled water and transferred to 100 mL of BMMH medium with 0.5% (v/v) methanol. The cells were then cultivated at 30 • C and 200 rpm for 120 h, and methanol (100 µL) was added 2 times per day. After termination of cultivation, biomass was centrifuged (7197× g, 10 • C, 5 min) and the supernatants were concentrated and desalted on Microcon centrifugal filter devices (10 kDa cut-off, Millipore) and used for enzyme characterization. Protein concentration was determined by Bradford method using BSA as a standard [42].

Determination of pH and Temperature Optimum and Temperature Stability
pH optimum was determined at 50 • C using 10 mg·mL −1 solution of GX in 40 mM Britton-Robinson buffer (pH 2.0-8.0), 15 min incubation, and 64.4 nM SlXyn30A. Temperature optimum was determined in the same manner in 50 mM sodium acetate buffer, pH 3.5, and temperatures ranging from 23 • C to 60 • C. Temperature stability was tested in 50 mM sodium acetate buffer, pH 3.5, at 40-60 • C for up to 5 h. During the incubation, aliquots were taken at different time points and the residual activity was immediately determined as described above (10 mg·mL −1 GX, pH 3.5, 50 • C, 15 min).
100% acetonitrile. The elution was isocratic at a flow rate of 0.5 mL·min −1 with a mixture of mobile phases A and B in a ratio of 30:70. Specific activities were determined on 1 mM Xyl 3 , Xyl 4 , MeGlcA 3 Xyl 4 and MeGlcA 3 Xyl 3 in 50 mM sodium acetate buffer, pH 3.5, at 50 • C using 1.3 µM SlXyn30A, and calculated on the basis of the amount of liberated Xyl 2 (linear XOs, in the case of Xyl 4 divided by two) or Xyl (MeGlcA 3 Xyl 4 and MeGlcA 3 Xyl 3 ).

Conclusions
The first GH30 xylanase originating from a yeast has been cloned, expressed and characterized. The enzyme SlXyn30A from S. lignohabitans is a glucuronoxylanase with auxiliary xylobiohydrolase activity. In addition to hardwood glucuronoxylan, it efficiently depolymerizes linear β-1,3-β-1,4-xylan but not cereal arabinoxylan. Its amino acid sequence has the highest similarity to the fungal bifunctional GH30-7 enzymes TcXyn30B and TtXyn30A which also display glucuronoxylanase and xylobiohydrolase activities. Catalytic properties of SlXyn30A also resemble those of TcXyn30B and TtXyn30A including the recognition of MeGlcA side chain in the −2b subsite, no substitution of xylose occupying the subsite −1, and certain flexibility of decoration of xylopyranosyl unit bound in the −2a subsite. Further characterization of new xylanases from different yeast species will help us to reveal how the yeasts cope with xylan degradation in nature and to better evaluate their biotechnological potential. The crystal structure of SlXyn30A with appropriate ligands would improve our knowledge of how GH30-7 glucuronoxylanases/xylobiohydrolases switch between endo-and exo-activities which is not yet fully understood.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author.