This article has been accepted for publication and undergone full peer review but has not been through the copyediting,

typesetting, pagination and proofreading process, which may lead to differences between this version and the Version of

Record. Please cite this article as doi: 10.1002/bab.1288.

This article is protected by copyright. All rights reserved. 1

Effect of acidic amino acids engineered into the active site cleft of

Thermopolyspora flexuosa GH11 xylanase

He Li and Ossi Turunen*

Department of Biotechnology and Chemical Technology, School of Chemical Technology,

Aalto University, P.O. Box 16100, 00076 Aalto, Finland

*Corresponding author: Tel.: +358 50 539 6499. Email address: Ossi.Turunen@aalto.fi.

Running title: pH engineering of GH11 xylanase

This article is protected by copyright. All rights reserved. 2

Synopsis

Thermopolyspora flexuosa GH11 xylanase (XYN11A) shows optimal activity at pH 6-7 and

75–80 oC. We studied how mutation to aspartic acid (N46D and V48D) in the vicinity of the

catalytic acid/base affects the pH activity of highly thermophilic GH11 xylanase. Both

mutations shifted the pH activity profile towards acidic pH. In general, the Km values were

lower at pH 4–5 than at pH 6, and in line with this, there was a slightly faster rate in hydrolysis

of xylotetraose at pH 4 than at pH 6. The N46D mutation and also lower pH in XYN11A

increased the hydrolysis of xylotriose. The Km value increased remarkably (from 2.5 to 11.6

mg/ml) due to V48D, which indicates the weakening of the binding affinity of the substrate to

the active site. Xylotetraose functioned well as substrate for other enzymes, but with lowered

reaction rate for V48D. Both N46D and V48D increased the enzyme inactivation by ionic

liquid [emim]OAc. In conclusion, the pH activity profile could be shifted to acidic pH due to

an effect from two different directions, but the tightly packed GH11 active site can cause steric

problems for the mutations.

This article is protected by copyright. All rights reserved. 3

1. Introduction

The thermostability of family GH11 xylanases has been studied widely by various methods

due to the industrial relevance of this enzyme group in feed application, bleaching of pulp in

paper industry, and processing of lignocellulose [1-7]. Another central feature for enzyme

engineering studies has been activity at low or high pH. In addition to the pH requirements of

biotechnological applications, an advantage of acidic pH is the preventing of microbial

growth [8, 9].

The structure of GH11 xylanases comprises a single α-helix and two strongly twisted β-

sheets forming a large cleft. Glycoside hydrolases, like xylanases, have two carboxylic acid

side chains (Glu or Asp) in their catalytic cleft, one functioning as a nucleophile and the other

as an acid/base catalyst. The pKa of nucleophile determines the acidic limb of the bell-like

pH-activity graph and pKa of acid/base determines the alkaline limb [10]. Therefore, the pKa

of acid/base has to be much higher than that of the nucleophile. The pH-activity range can be

shifted to acidic or alkaline pH by changing the pKa of the catalytic residues. Several studies

have explored the rules of how the pH-activity profile could be shifted in the desired

direction [11-13]. In particular, there are many reports on increasing activity and stability at

alkaline or acidic pH in GH11 xylanases [14-24]. Mutations located close to the catalytic

glutamic acid residues in xylanases modulate the pH dependence activity [25-28]. A mutation

of asparagine (having a hydrogen bond with acid/base) to aspartic acid shifted the pH activity

to acidic pH, and aspartic acid in this position is found in several acidic GH11 xylanases [8,

25, 26, 29].

Characterization of the recombinant enzyme and crystal structure of highly thermostable

Thermopolyspora flexuosa XYN11A have been reported in earlier studies [30-31]. XYN11A

is one of the most thermostable known GH11 xylanases with potential in applications

requiring high stability and activity at high temperatures. T. flexuosa XYN11A also has a

potential for the production of xylooligosaccharides due to its high activity and hydrolysis

pattern, producing xylobiose and xylotriose [32]. In this study, we investigated mutations of

Thermopolyspora flexuosa XYN11A introducing acidic residues into the catalytic cleft close

to the catalytic residues. In addition, we investigated how additional acidic amino acids in the

catalytic cleft affect the tolerance to ionic liquid, 1-ethyl-3-methylimidazole acetate

([emim]OAc) that dissolves cellulose.

This article is protected by copyright. All rights reserved. 4

2. Material and methods

2.1. Mutant xylanases

The E. coli expression vector pALK1405 (kindly provided by Marja Paloheimo, ROAL,

Rajamäki. Finland) contained the 24 kDa catalytic core of T. flexuosa xylanase XYN11A

[31]. The xylanase was expressed from pALK1405 under the control of Bacillus

amyloliquefaciens alpha-amylase promoter and the protein was secreted into the medium

through the B. amyloliquefaciens alpha-amylase signal sequence. Swiss-PdbViewer

(http://www.expasy.ch/spdbv/) [33] was used to examine the structure of XYN11A (1M4W)

for designing the mutations into the active site. T. flexuosa XYN11A wild type and the

mutants N46D, V48D, and L31E were created by PCR-based site-directed mutagenesis, as

described earlier [34]. Further, the structural modelling of the mutants was done using

SWISS-MODEL automatic server (http://swissmodel.expasy.org/).

2.2. Production of enzymes

The enzymes were expressed in E. coli XL1 Blue cells with cultivation in Luria-Bertani (LB)

broth containing 125mg/l ampicillin. The secreted enzyme was collected from the culture

broth and used without further purification in the stability and activity experiments. It turned

out that the E. coli cells contained a substantial amount of the enzyme intracellularly (or in

the periplasmic place) and this source was used for protein purification. To purify the

enzymes for kinetic experiments, the E. coli cells were suspended in 20 mM Tris-HCl (pH

7.5), sonicated, and the supernatant was used for protein purification in Äkta Purifier (GE

Healthcare), first by using hydrophobic interaction chromatography (HIC) with Hiload 16/10

Phenyl Sepharose High Performance column. The HIC binding buffer was 20 mM Tris-HCl

with 30% ammonium sulphate and 1 mM magnesium chloride, pH 7.5, and the elution buffer

was the same but without 30% ammonium sulphate. Anion exchange chromatography was

performed with Hiload 16/10 Q Sepharose High Performance column. The anion exchange

binding buffer was 20 mM Tris-HCl with 15 mM sodium chloride, pH 7.5, and the elution

buffer contained 1.0 M sodium chloride. Gel filtration chromatography was performed with

Hiload 16/60 Superdex 200 prep grade column. The gel filtration buffer was 20 mM sodium

acetate with 0.15 M sodium chloride and pH 7.5. The proteins were concentrated by Vivaspin

tubes (Sartorious Stedim Biotech). The protein concentration was determined by

This article is protected by copyright. All rights reserved. 5

densitometry of the SDS-PAGE gels by Image J software (National Institutes of Health)

using BSA as the standard.

2.3. Enzyme assays

The enzyme activity was determined by the 3, 5-dinitrosalisylic acid (DNS) method, in which

the amount of reducing sugars liberated from the hydrolyzed substrate (solubilized 1%

birchwood xylan) was measured at an absorbance of 540 nm [35]. The optimum temperature

was determined at pH 5 with 1% birchwood xylan as substrate in 50 mM citrate phosphate

buffer in the temperature range from 30°C to 100°C, with an incubation time of 30 min. The

optimum pH for enzyme activity was measured in the pH range from 3 to 7 in 50 mM citrate

phosphate buffer and pH 8 in 50 mM Tris-HCl buffer, with incubation for 30 min at 60oC.

For the residual activity assay, the enzyme samples were incubated without substrate at

different temperatures ranging from 50oC to 100

oC at pH 6 for 30 min; the remaining activity

was measured thereafter.

Purified enzymes were used for measuring the kinetic parameters with and without 5%

[emim]OAc at 60oC in 50 mM citrate-phosphate buffer (pH 6) with 0.1–1.5% birchwood

xylan. Km and Vmax were calculated by employing hyperbolic regression analysis (program

Hyper32). D-xylose was used as the standard, and bovine serum albumin (0.1 mg/ml BSA)

was used as a stabilizer in the reactions. The reaction time was 30 min and each experiment

was conducted thrice.

The cleavage products were analysed using HPLC (Waters Alliance 2695) with lead

column from Bio-Rad (Aminex HPX-87P, 300x78 mm column) and a detector (Waters 2414

refractive index detector). 300 µL of reaction mixture contained 0.5 units of enzyme mixed

with 1% birchwood xylan as substrate in 50 mM citrate-phosphate buffer (pH 4, 5, or 6) and

0.1 mg/mL BSA. The enzyme reaction was conducted at 50oC for 3, 6, or 24 hours. The used

enzyme amount (U/reaction) was based on a 30-min assay at 60oC. One unit is defined as 1

µmol reducing groups produced per min. The reaction was stopped by adding 300 µL of 0.25

M H2SO4, spinning for 30 min, removing the cell pellet, and then adding 300 µL of 0.5 M

NaOH to neutralize the supernatant, and finally filtering the solution. Xylose, xylobiose,

xylotriose, and xylotetraose were used as the standards. All experiments were repeated thrice.

This article is protected by copyright. All rights reserved. 6

2.4. Docking of [emim] cation to XYN11A

The binding of [emim] cation to XYN11A (PDB structure 1M4W) and its modelled mutants

N46D and V48D were analysed using SwissDock, as reported previously [36, 37]. [Emim]

cation was energy-minimised by MM2 in ChemBio3D Ultra 12.0 (CambridgeSoft) before

docking. Accurate mode was used in SwissDock.

3. Results

3.1 The mutants

The crystal structure of T. flexuosa XYN11A (1M4W) was used to design the mutations in

Swiss-Pdb Viewer [30]. Two catalytic residues Glu87 (catalytic nucleophile) and Glu176

(catalytic acid/base) located on both sides of the catalytic cleft participate in the double

displacement catalytic mechanism [4,30]. We used XYN11A to study how negatively

charged amino acid introduced into nearby positions affect catalytic residues. It is known

from other members of the GH11 family that when a residue corresponding to N46 in T.

flexuosa XYN11A is changed to aspartic acid, the pH activity profile of the enzyme is shifted

to acidic pH [25, 26, 38]. We tested how this mutation (N46D) and another similar nearby

mutation (V48D) affect the properties in a highly thermostable GH11 xylanase. The goal with

V48D mutation was to see how much the direction in which the negative charge is located

determines the effect on the pH activity profile. A third mutation that was at a distance (L31E)

close to the bottom of the active site cleft functioned as a control of a change to acidic amino

acid, but in longer distance from the catalytic amino acids.

3.2. Effect of mutations on thermal properties of XYN11A

The mutant enzymes were produced in E. coli and purified to 80–90% purity (not shown).

The masses of the XYN11A wild type and its mutants were approximately 23-25 kDa.

The effect of mutations on thermal properties of XYN11A was tested by incubating the

enzymes for 30 min at different temperatures without substrate and then measuring the

residual activity. This assay indicated that all mutations reduced the thermostability of

XYN11A, although the effect of N46D was small (Figure 1). The reason why N46D showed

higher stability than V48D and L31E could be that the latter two mutations are in an

embedded position. Then, charge of Asp at position 46 has more freedom, whereas the

charges at positions 31 and 48 may have destabilizing contacts to the nearby atoms. The

reason why N46D and L31E showed increase in residual activity at 90-100oC is not clear.

This article is protected by copyright. All rights reserved. 7

Xylanases have occasionally shown this kind of residual activity graphs. One possible reason

is the binding of denatured protein molecules to the remaining active enzymes, which could

subsequently protect the enzyme from further denaturation.

The apparent temperature optimum was determined by 30-min assay at different

temperatures (Figure 2). Although N46D was slightly less thermostable than the wild type, its

temperature-dependent activity profile was similar to the wild type with the apparent

temperature optimum at approximately 78oC. The mutation V48D decreased the apparent

temperature optimum approximately 10oC (Figure 2), whereas the decrease of thermostability

in residual activity assay was approximately 15oC (Figure 1). The mutation of Leu31 to Glu

in a buried position decreased both thermostability and apparent temperature optimum over

15oC. These examples indicated that mutations in the active site or close to it may easily

decrease thermostability, as observed earlier with xylose isomerase [39]. In a GH10 xylanase,

active site mutations also increased activity at high temperature [7].

3.3. Effect of mutations on catalytic activity and pH-activity profile

Purified enzymes were used to measure the specific activity (U/mg enzyme), which was

conducted at 60°C with 1% birchwood xylan as the substrate. At pH 6, the specific activities

were 2134±124 U/mg for XYN11A, 1666±92 U/mg for N46D, 133±3 U/mg for V48D, and

1273±51 U/mg for L31E. Therefore, N46D and L31E led to a rather moderate decrease in the

specific activity at pH 6, whereas the mutation V48D caused a substantial decrease in the

activity. Although V48D caused a drop in the activity, the reason probably is not the presence

of inactivated protein species in the protein sample. Last purification step was size exclusion

chromatography and there was seen only one major protein peak of xylanase, indicating that

there were not present significant amounts of inactivated protein aggregates.

The wild-type enzyme displayed over 50% specific activity in pH 4.5–9.0, with the

optimum pH at approximately 6.5 (Figure 3). The substitution of aspartic acid to the position

46 or 48 in the active site decreased the pH optimum by approximately 1.2 and 1.8 pH units,

respectively. The pH profile of L31E experienced only a small shift to acidic direction in the

alkaline limb of the graph. When compared to the wild type only N46D caused an increase in

the activity at an acidic pH (3–4).

This article is protected by copyright. All rights reserved. 8

3.4. Effect of mutations, pH and [emim]OAc on kinetic parameters

Kinetic parameters were determined at pH 6 by using purified enzymes with and without

[emim]OAc (Table 1). Without [emim]OAc, Km and Vmax were 2.5 mg/ml and 3110 U/mg

for XYN11A, 1.9 mg/ml and 1582 U/mg for N46D, 11.6 mg/ml and 315 U/mg for V48D,

and 2.1 mg/ml and 2300 U/mg for L31E, respectively. It was observed that aspartic acid at

position 48 (V48D) had a substantial effect on catalytic efficiency; Km increased by 4.6-fold

and Vmax decreased by 9.9-fold. Aspartic acid at position 46 (N46D) and glutamic acid at

position 31 (L31E) impaired the catalytic efficiency only slightly due to a decrease in Vmax.

We also conducted kinetic experiments at pH 5 and 4. We observed a trend that Km values

were lower at pH 5 and 4 than at pH 6. While at pH 5 and 6, the wild-type enzyme showed

higher catalytic efficiency (Vmax/Km; Table I) than N46D, and at pH 4, N46D showed a

higher catalytic efficiency than XYN11A in line with the pH-activity profile (Figure 3).

The effect of ionic liquid on kinetic parameters was tested with 5% [emim]OAc solution at

pH 6. The results indicated that Km became significantly higher in all enzymes (2.3–3.3-fold),

whereas Vmax remained close to the same level than without ionic liquid, except in the

mutants V48D and L31E, which experienced 42% and 19% decrease in the Vmax value,

respectively. We observed earlier that typically [emim]OAc increases Km, whereas Vmax

remains at the same level. While both aspartic acids (N46D, V48D) increased the

vulnerability of the enzyme to the detrimental effect of the ionic liquid, the effect of V48D

was stronger.

We used molecular docking with SwissDock to identify possible reasons for the differing

effect of [emim]OAc on the N46D and V48D mutants (Figure. S1). [Emim] cation was

energy-minimized by MM2 force field in ChemBio3D Ultra 12.0 and docked to the 1M4W

structure and the modelled N46D and V48D structures. While there are many potential

binding sites in the active site cleft for the [emim] cation, the identified poses with highest

binding energy (G) were located close to the catalytic residues (Figure S1). Their exact

positions depended on the nearby aspartic acids (N46D and V48D). The additional aspartic

acids slightly increased the level of highest binding energies that were -7.5 kcal/mol for the

wild type (1M4W), -8.0 kcal/mol for N46D, and -8.2 kcal/mol for V48D (both in downward

and upward position of Asp48 side chain; see Figures 4 and S1). On the other hand, these

results may not fully show that the slightly higher binding energy of V48D when compared to

N46D is associated with a stronger sensitivity to [emim]OAc, but it is not eliminated. It was

observed that in the modelled position when the side chain of Asp48 was located towards the

This article is protected by copyright. All rights reserved. 9

bottom of the active site cleft, the [emim] cation formed a hydrophobic stacking interaction

with Trp20 (Figure S1), probably increasing the binding affinity. Trp20 is a substrate binding

side chain that appears to easily form stacking interactions with the [emim] cation [37].

Docking results indicated that the additional acidic side chains may change the positions of

the [emim] cations that are trapped with a highest binding energy to the active site (Figure

S1). Acetate ion was also docked to the wild type structure, and only a couple of binding

places were detected at the edge of the active site. Therefore, acetate is not likely to hinder

the substrate binding.

3.5. Effect of mutations on hydrolysis pattern

The analysis of xylan hydrolysis products (xylose, xylobiose, xylotriose, xylotetraose) by

HPLC showed that xylobiose and xylotriose were the main products in the used conditions

(Table 2). XYN11A appeared not to hydrolyze xylotriose at pH 6, but did so slowly at pH 4.

N46D hydrolysed xylotriose slowly at pH 6. It was also observed that there was a slightly

faster rate in hydrolysis of xylotetraose at pH 4 than at pH 6 (Table 2). Xylobiose was not

hydrolysed under any conditions, and as a function of time its amount was increasing. The

mutation V48D significantly affected the hydrolysis pattern; xylotriose was not hydrolysed at

all and xylotetraose was hydrolysed slower than with XYN11A and N46D. In conclusion, the

analysis of the hydrolysis products from birchwood xylan indicated that acidic amino acid

(instead of asparagine) at position 46 in less than 4Å distance from the substrate increases the

affinity of substrate to the active site. A similar effect was caused by lower pH.

4. Discussion

The mutations in the active site canyon or close to it are likely to considerably affect the

activity profile of enzymes, although there is also a risk that many of these mutations are

harmful [38-40]. Mutations also reveal detailed roles of active site residues [25, 26, 41, 42].

In this study, we explored how mutations close to the catalytic residues affect the properties

of a highly thermostable GH11 xylanase with optimum activity at 75–80oC. The active site of

GH11 xylanases is known rather well. The catalytic nucleophile with low pKa (4.6 in Glu78

of Bacillus circulans GH11 xylanase, BCX) controls the acidic limb of bell-like pH profile

and acid/base with high pKa (6.7 in Glu172 of BCX) controls the alkaline limb [25]. The

change of asparagine to aspartic acid at the position corresponding to N46 in T. flexuosa

This article is protected by copyright. All rights reserved. 10

XYN11A shifted the pH optimum from 5.7 to 4.6 in BCX (N35D) [25]. Inversely, in the

acidic Aspergillus kawachii GH11 xylanase, the change of native Asp37 to Asn increased the

optimal activity from pH 2-3 to pH 5 [8]. In T. flexuosa XYN11A, the shift of pH optimum

was similar to that in BCX—1.2 pH units to acidic pH from pH 6.5 to pH 5.3. However, the

pH optima of XYN11A and the N46D-mutant are 0.7-0.8 pH units higher than those of the

corresponding BCX enzymes. The nearest side chains surrounding the catalytic residues

inside the active site cleft (below 8Å distance) are exactly the same in these two proteins.

Therefore, the difference in the pH optimum is caused by differences in the long-distance

effects and/or minor structural differences in the nearby areas. In contrast to BCX, the

absolute activity level was lowered in XYN11A by 40%, whereas the same mutation caused

20% increase in activity in BCX [25].

In XYN11A, Asp46 is in close distance (3.3 Å) from the acid/base (Glu176) and it has been

concluded from the BCX studies that this aspartic acid (Asp35 with pKa 3.7 in BCX) has to

be protonated for the enzyme to be active, which happens at low pH. Otherwise it prevents

the functioning of the acid/base. Therefore, the pH activity profile becomes more acidic.

Alignment of xylotetraose to the active site of XYN11A (Figure 4) indicated that Asp46 is

also in a rather appropriate position to protonate the glycosidic oxygen; in addition, the

substrate apparently approaches Asp46 before Glu176. However, the distance between Asp46

and Glu87 (nucleophile) is 7 Å, whereas the distance between Glu87 and Glu176 is 5.6 Å.

Therefore, it is evident that the close positioning between C1 of the substrate xylose (Figure 4)

and Glu87 necessitates that Glu176 remains the acid/base catalyst, in line with the findings in

BCX [25].

We tested how another mutation to aspartic acid, V48D—near Glu176, but in a lower

position in the catalytic cleft (Figure 4)—affects catalytic properties. While the side chain of

Asp46 has substantial free space around it, Asp48 is packed against neighbouring side chains

on the bottom of the catalytic cleft below the -1 and -2 binding sites of the substrate. The

modelling indicated that Asp48 is at a distance of approximately 4 Å from Glu176 (Figure 4),

although torsion of the side chain can bring the terminal oxygen to a closer position (~3 Å),

but apparently without hydrogen bonding to Glu176. The mutation dramatically decreases the

activity. Since Tyr78 is located in front of Asp48, it is possible that Asp48 slightly shifts the

position of Tyr78 and then Glu87 that has a hydrogen bond to Tyr78; thus, possibly

This article is protected by copyright. All rights reserved. 11

disturbing the enzyme activity (Figure 4). Moreover, the mutation V48D removes a

stabilizing hydrophobic interaction between Val48 and Tyr78. Another possibility is that the

additional acidic side chain in this position changes the electrostatic balance in a disturbing

manner. Despite of the large drop in the activity level, the pH activity profile remained wide

and the pH optimum shifted approximately 1.8 pH units to acidic pH (optimum at pH 4.7).

Therefore, the effect of V48D on pH activity profile was slightly stronger than the effect of

N46D. The shift to acidic pH is so large that it is likely that the enzyme is active only when

Asp48 is protonated.

When the hydrolysis pattern was analyzed, it was observed that unlike the wild type, the

mutant V48D does not hydrolyze xylotriose. The hydrolysis of xylotetraose was not

prevented. This finding, like the elevated Km, indicated that V48D impairs the binding of

substrate to the active site. Consequently, xylotriose cannot bind at all in a catalytically

favourable way, whereas xylotetraose can still bind but with lower strength. Molecular

modelling indicates that V48D may not interact directly with the substrate, because SWISS-

MODEL placed the side chain downwards (Figure 4). Then, its effect is likely to be caused

by the reorganization of the tightly packed nearby side chain matrix, which then changes the

positions of the substrate binding side chains. However, if Asp48 is located upwards, then it

could prevent the descending of the substrate to binding position in this site. In both

alternatives the binding at -1 and -2 binding sites is likely to be disturbed. Despite the strong

negative effect on the activity level, the mutation affected the pH profile in a similar manner

to N46D.

Since the negatively charged amino acids on the protein surface have been proposed to

protect from the harmful effect of biomass-dissolving ionic liquids [43], we tested how the

engineered acidic amino acids in the active site affect the tolerance to [emim]OAc. Both the

mutation N46D and V48D increased the inactivation of enzyme activity by [emim]OAc.

However, the effect was stronger in V48D and was seen in impaired Vmax, whereas the effect

was smaller in N46D and it was seen in impaired Km. This difference indicated that the

mechanism of the effect was different in the two mutations. Interaction of [emim]OAc or the

[emim] cation with Asp46 located in an upper position probably increased the competitive

inhibition of substrate binding by [emim]OAc. On the contrary, the interaction of ionic liquid

with Asp48 located deeper in the active site cleft probably caused disturbing steric or

This article is protected by copyright. All rights reserved. 12

electrostatic effects influencing the catalytic reaction, but did not increase the competitive

inhibition. Changes in the charged amino acids close to the active site can have an effect on

the catalytic rate [39]. There are also other possibilities than a change to aspartic acid near to

the catalytic residues in modulating the pH activity [44]. Since both mutations are close to the

catalytic amino acids, we used molecular docking with SwissDock to study further the

possible reason for the effect of mutations on the sensitivity to [emim]OAc. SwissDock,

which counts charges in calculating interactions, indicated that the introduced acidic side

chains attract the positively charged [emim] cation, probably being one reason for the effect

on activity.

In conclusion, the pH-modulating close-distance effect can come from different directions,

but the side chain requires sufficient free space around it in order not to disturb the catalytic

activity. In these respects, the wider active site cleft in GH10 xylanases allows the presence

of a higher number of bulky charged side chains in the active site cleft, thereby profoundly

affecting the pH activity profile [40]. Nevertheless, the narrow GH11 active can also harbour

some mutations that modify catalytic properties.

5. Acknowledgements

We thank the Graduate School of Chemical Engineering for the financial support and

Johanna Aura for technical assistance.

This article is protected by copyright. All rights reserved. 13

6. References

[1] Jänis, J., Turunen, O., Leisola, M., Derrick, P.J., Rouvinen, J., Vainiotalo, P. (2004)

Biochemistry 43, 9556-9566.

[2] Xiong, H., von Weymarn, N., Leisola, M., Turunen, O. (2004) Process Biochem. 39, 731-36.

[3] Collins, T., Gerday, C., Feller, G. (2005) FEMS Microbiol. Rev. 29, 3-23.

[4] Paës, G., Berrin, J.G., Beaugrand, J. (2012) Biotechnol. Adv. 30, 564-592.

[5] Wang, Y., Fu, Z., Huang, H., Zhang, H., Yao, B., Xiong, H., Turunen, O. (2012) Bioresour.

Technol. 112, 275-279.

[6] Li, H., Kankaanpää, A., Xiong, H., Hummel, M., Sixta, H., Ojamo, H., Turunen, O. (2013)

Enzyme Microb. Technol. 53, 414-419.

[7] Wang, K., Luo, H., Tian, J., Turunen, O., Huang, H., Shi, P., Hua, H., Wang, C., Wang, S., Yao,

B. (2014) Appl. Environ. Microbiol. 80, 2158-2165.

[8] Fushinobu, S., Ito, K., Konno, M., Wakagi, T., Matsuzawa, H. (1998) Protein Eng. 11, 1121-

1128.

[9] Xiong, H., Fenel, F., Leisola, M., Turunen, O. (2004) Extremophiles 8, 393-400.

[10] Mclntosh, L.P., Hand, G., Johnson, P.E., Joshi, M.D., Körner, M., Plesniak, L.A., Ziser, L.,

Wakarchuk, W.W., Withers, S.G. (1996) Biochemistry 35, 9958-9966.

[11] Shirai, T., Suzuki, A., Yamane, T., Ashida, T., Kobayashi, T., Hitomi, J., Ito, S. (1997) Protein

Eng. 10, 627-634.

[12] Nielsen, J.E., Borchert, T.V., Vriend, G. (2001) Protein Eng. 14, 505-512.

[13] Dubnovitsky, A.P., Kapetaniou, E.G., Papageorgiou, A.C. (2005) Protein Sci. 14, 97-110.

[14] Chen, Y.L., Tang, T.Y., Cheng, K.J. (2001) Can. J. Microbiol. 47, 1088-1094.

[15] Inami, M., Morokuma, C., Sugio, A., Tamanoi, H., Yatsunami, R., Nakamura, S. (2003)

Nucleic Acids Res. Suppl. 3, 315-316.

[16] Turunen, O., Jänis, J., Fenel, F., Leisola, M. (2004) Methods Enzymol. 388, 156-167.

[17] Shibuya, H., Kaneko, S., Hayashi, K. A. (2005) Biosci. Biotechnol. Biochem. 69, 1492-1497.

[18] Umemoto, H., Ihsanawati, Inami, M., Yatsunami, R., Fukui, T., Kumasaka, T., Tanaka, N.,

Nakamura, S. (2007) Nucleic Acids Symp. Ser. (Oxf). 51, 461-462.

[19] Umemoto, H., Ihsanawati, Inami, M., Yatsunami, R., Fukui, T., Kumasaka, T., Tanaka, N.,

Nakamura, S. (2009) Biosci. Biotechnol. Biochem. 73, 965-967.

[20] Wang, Q., Xia, T. (2008) Biotechnol. Lett. 30, 937-944.

This article is protected by copyright. All rights reserved. 14

[21] Yang, J.H., Park, J.Y., Kim, S.H., Yoo, Y.J. (2008) J. Biotechnol. 133, 294-300.

[22] Stephens, D.E., Singh, S., Permaul, K. (2009) Error-prone PCR of a fungal xylanase for

improvement of its alkaline and thermal stability. FEMS Microbiol. Lett. 293, 42–47.

[23] Al Balaa, B., Brijs, K., Gebruers, K., Vandenhaute, J., Wouters, J., Housen, I. (2009) Bioresour.

Technol. 100, 6465-6471.

[24] Beliën, T., Joye, I.J., Delcour, J.A., Courtin, C.M. (2009) Protein Eng. Des. Sel. 22, 587-596.

[25] Joshi, M.D., Sidhu, G., Pot, I., Brayer, G.D., Withers, S.G., Mclntosh, L.P. (2000) J. Mol. Biol.

299, 255-279.

[26] Joshi, M.D., Sidhu, G., Nielsen, J.E., Brayer, G.D., Withers, S.G., Mclntosh, L.P. (2001)

Biochemistry 40, 10115-10139.

[27] De Lemos Esteves, F., Gouders, T., Lamotte-Brasseur, J., Rigali, S., Frère, J.M. (2005) Protein

Sci. 14, 292-302.

[28] Tanaka, H., Okuno, T., Moriyama, S., Muguruma, M., Ohta, K. (2004) J. Biosci. Bioeng. 98,

338-343.

[29] Krengel, U., Dijkstra, B.W. (1996) J. Mol. Biol. 263, 70-78.

[30] Hakulinen, N., Turunen, O., Jänis, J., Leisola, M., Rouvinen, J. (2003). Eur. J. Biochem. 270,

1399-1412.

[31] Leskinen, S., Mäntylä, A., Fagerström, R., Vehmaanperä, J., Lantto, R., Paloheimo, M.,

Suominen, P. (2005) Appl. Microbiol. Biotechnol. 67, 495-505.

[32] Zhang, J., Siika-Aho, M., Puranen, T., Tang, M., Tenkanen, M., Viikari, L. (2011) Biotechnol.

Biofuels 4, 12.

[33] Guex, N., Peitsch, M.C. (1997) Electrophoresis 18, 2714–23.

[34] Turunen, O., Etuaho, K., Fenel, F., Vehmaanperä, J., Wu, X., Rouvinen, J., Leisola, M.A.

(2001) J. Biotechnol. 88, 37-46.

[35] Bailey, M.J., Biely, P., Poutanen, K. (1992) J. Biotechnol. 23, 257–270.

[36] Grosdidier, A., Zoete, V., Michielin, O. (2011) Nucleic Acids Res. 39, W270-277.

This article is protected by copyright. All rights reserved. 15

[37] Chawachart, N., Anbarasan,

S., Turunen,

S., Li,

H., Khanongnuch,

C., Hummel,

M., Sixta,

H.,

Granström, T., Lumyong, S. and Turunen, O. (2014) Extremophiles DOI 10.1007/s00792-

014-0679-0.

[38] Kongsted, J., Ryde, U., Wydra, J., Jensen, J.H. (2007) Biochemistry 46, 13581-13592.

[39] Karimäki, J., Parkkinen, T., Santa, H., Pastinen, O., Leisola, M., Rouvinen, J., Turunen, O.

(2004) Protein Eng. Des. Sel. 17, 861–869.

[40] Morley, K.L., Kazlauskas, R.J. (2005) Improving enzyme properties: when are closer mutations

better? Trends Biotechnol. 23, 231-237.

[41] Mardo, K., Visnapuu, T., Vija, H., Elmi, T., Alamäe, T. (2014) Biotechnol. Appl. Biochem. 61,

11-22.

[42] Roberge, M., Shareck, F., Morosoli, R., Kluepfel, D., Dupont, C. (1998) Protein Eng. 11, 399-

404.

[43] Zhang, T., Datta, S., Eichler, J., Ivanova, N., Axen, S. D., Kerfeld, C. A., Chen, F., Kyrpides,

N., Hugenholtz, P., Cheng, J. F., Sale, K. L., Simmons, B., Rubin, E. (2011) Green Chemistry

13, 2083-2090.

[44] Ludwiczek, M.L., D'Angelo, I., Yalloway, G.N., Brockerman, J.A., Okon, M., Nielsen, J.E.,

Strynadka, N.C., Withers, S.G., McIntosh, L.P. (2013) Biochemistry 52, 3138-3156.

This article is protected by copyright. All rights reserved. 16

Fig. 1 Residual activity of T. flexuosa XYN11A and its mutants as a function of temperature.

The enzyme samples were incubated without substrate at different temperatures for 30 min

and after that the remaining activity was measured. Symbols: XYN11A, circle; N46D,

diamond; V48D, triangle; L31E, square.

This article is protected by copyright. All rights reserved. 17

Fig. 2 Temperature dependent activity of T. flexuosa XYN11A wild type and its mutants at

pH 5. Symbols: XYN11A, circle; N46D, diamond; V48D, triangle; L31E, square.

This article is protected by copyright. All rights reserved. 18

Fig. 3 pH dependent activity of T. flexuosa XYN11A wild type and its mutants at 60 oC. The

activity is shown as U/mg enzyme. Symbols: XYN11A, circle; N46D, diamond; V48D,

triangle; L31E, square.

This article is protected by copyright. All rights reserved. 19

Fig. 4 Structure of the catalytic site of T. flexuosa XYN11A with mutations N46D and V48D

and aligned xylotetraose. The catalytic residues (nucleophile Glu87 and acid/base Glu176)

and the nearby two tyrosines are shown. Space-filling of Asp48, Tyr78, Glu87, Tyr89, and

Glu176 is shown. The glycosidic oxygen approaching the acid/base is indicated by an asterisk.

Glycerol (GOL) from the crystal structure 1M4W is also shown. -1 and -2 binding sites of the

active site are indicated. The figure was created with PyMol (http://www.pymol.org/).

This article is protected by copyright. All rights reserved. 20

Table 1 Kinetic parameters of T. flexuosa XYN11A and its mutants.

Km(mg/ml) Vmax (U/mg) Vmax/Km

pH6

XYN11A 2.5 ±0.1 3110 ±31 1244

N46D 1.9 ±0.1 1582 ±58 833

V48D 11.6 ±0.9 315 ±32 27

L31E 2.1 ±0.2 2300 ±72 1095

pH5

XYN11A 1.5 ±0.1 1799 ±75 1199

N46D 1.4 ±0.1 1500 ±21 1071

V48D 9.0 ±0.6 373 ±28 41

pH4

XYN11A 1.6 ±0.1 772 ±52 483

N46D 1.7 ±0.2 1047 ±13 616

With 5% [emim]OAc

pH6

XYN11A 6.1 ±0.1 2945 ±28 483

N46D 6.3 ±1.3 1564 ±120 248

V48D 28.1 ±2.8 184 ±6 7

L31E 4.8 ±0.5 1855 ±143 386

This article is protected by copyright. All rights reserved. 21

Table 2 Pattern of xylan hydrolysis products by XYN11A, N46D and V48D. The values for

xylose (X), xylobiose (X2), xylotriose (X3), and xylotetraose (X4) are expressed as a

percentage of the total of the four carbohydrates. Hydrolysis of xylan (10 mg/ml) was done at

50oC with 0,1 mg/ml BSA. The used amount of enzyme was 0.5 U/reaction in each pH. The

total is the sum for the amounts of X, X2, X3, and X4.

Enzyme

type

pH Time

(h)

X

(%)

X2

(%)

X3

(%)

X4

(%)

Total

(g/L)

XYN11A 6 3 0.6 ±0.8 26.6 ±3.0 57.3 ±0.7 15.5 ±4.2 2.7 ±0.1

6 6 2.4 ±0.9 30.1 ±1.5 56.4 ±0.3 11.2 ±1.8 2.9 ±0.1

6 24 4.6 ±2.2 39.2 ±1.4 56.3 ±0.9 0 3.1 ±0.2

N46D 6 3 1.4 ±1.0 29.7 ±3.8 55.6 ±1.6 13.3 ±5.3 2.8 ±0.1

6 6 3.0 ±1.5 33.7 ±2.5 54.7 ±0.5 8.7 ±3.1 3.0 ±0.1

6 24 7.0 ±3.9 41.8 ±2.8 51.2 ±3.3 0 3.3 ±0.1

XYN11A 4 3 2.3 ±0.6 37.9 ±3.2 56.1 ±2.7 3.8 ±4.7 2.8 ±0.2

4 6 2.4 ±1.0 42.5 ±3.8 55.0 ±3.8 0 3.2 ±0.1

4 24 4.9 ±0.9 49.4 ±3.2 45.8 ±3.0 0 3.8 ±0.3

N46D 4 3 3.1 ±0.6 38.6 ±3.4 54.7 ±3.0 1.3 ±2.1 3.0 ±0.2

4 6 4.1 ±0.9 44.3 ±3.6 51.6 ±3.6 0 3.4 ±0.1

4 24 8.4 ±1.0 51.8 ±3.4 39.8 ±3.1 0 4.1 ±0.3

V48D 5 3 0 12.9 ±1.6 56.3 ±1.5 30.8 ±3.0 2.4 ±0.1

5 6 0 15.2 ±2.0 56.2 ±2.4 28.5 ±4.3 2.7 ±0.2

5 24 0 24.2 ±2.3 57.2 ±1.0 18.6 ±2.6 3.1 ±0.2