Journal of Bioscience and BioengineeringVOL. 110 No. 1, 12–17, 2010

www.elsevier.com/locate/jbiosc

Enhancing the thermostability of α-glucosidase from Thermoanaerobactertengcongensis MB4 by single proline substitution

Cheng Zhou,1,2 Yanfen Xue,1 and Yanhe Ma1,⁎

⁎ CorrespondE-mail add

1389-1723/$doi:10.1016/

State Key Lab of Microbial Resources, Institute of Microbiology, Chinese Academy of Science, Beijing 100101, China1 and The Graduate School,Chinese Academy of Sciences, Beijing 100049, China2

Received 11 August 2009; accepted 3 December 2009Available online 3 January 2010

Thermostability can be increased by introducing prolines at suitable sites in target proteins. In this study, we comparedfive thermostable α-glucosidases and the moderate thermostable α-glucosidase (TtGluA) from Thermoanaerobactertengcongensis MB4. Based on the amino acid sequence alignment, four sites (Leu152, Asn208, Lys285, and Thr430) of TtGluAwere chosen for proline substitution to improve its thermostability. Thermostability of mutants L152P, K285P, and T430Pincreased evidently, but no thermostability improvement was observed for N208P. Compared to the wild-type enzyme, T5015 ofT430P had a rise of 2 °C without distinct loss of activity. However, T5015 values of L152P and K285P increased 2 °C and 10.5 °C,respectively, while retaining activity of only 26.6% and 24.9% of wild-type enzyme. The Km of L152P, K285P, T430P andwild-type enzyme was 1.61, 0.32, 1.64, and 1.08 mM, respectively. These indicate that the selected sites are not only importantfor the thermostability but also related to the substrate binding and catalytic activity of TtGluA. The CD spectra analysis of theimproved mutants and wild-type enzyme showed no distinct changes in their secondary structures. Combining analysis ofsecondary structure prediction and 3D structure modeling, the proline substitution at the three sites stabilized TtGluApossibly by reducing the flexibility of loop and random coil or (and) increasing the hydrophobic effect at these strategicregions with no evident structure change.

© 2009, The Society for Biotechnology, Japan. All rights reserved.

[Key words: α-Glucosidase; Thermoanaerobacter tengcongensis; Proline substitution; Thermostability improving; Structure modeling]

α-Glucosidases are exoenzymes hydrolyzing terminal glycosidicbonds and releasing α-glucose from the non-reducing end ofsaccharides chain (1). They are amylolytic enzymes involved in thelast step of starch degradation and are the second most importantenzymes during the early stages of raw starch hydrolysis (2). α-Glucosidases are typically used in food processing, fermentation andalcohol production in industry, where they are very important in theprocesses of hydrolyzing starch to produce fermentable sugars. Thethermostability of α-glucosidase is important because most of theindustrial processes, such as the conversion of starch to fermentablesugars during the industrial production of ethanol, typically take placeat the temperature of 65–73 °C (3). Thermolabile α-glucosidases notonly reduce the efficiency of starch breakdown at the high tempera-tures used for starch gelatinization. The requirement for cooling thestarch to a more favorable temperature for enzymatic hydrolysis alsowastes energy. Thus, improving the thermostability of thermolabile ormesophilic α-glucosidase is required for its industrial application.

There aremany protein engineeringmethodswhich can be used toincrease protein stability, although increasing the thermostability ofenzymes by rational design has no general principle yet. It wasobserved that the introduction of prolines at carefully selected sites intarget proteins was effective in molecular stabilization (4). At these

ing author. Tel.: +86 10 64807590; fax: +86 10 64807616.ress: mayanhe@im.ac.cn (Y. Ma).

- see front matter © 2009, The Society for Biotechnology, Japan. Alj.jbiosc.2009.12.002

sites, the proline residues restrict backbone bond rotation because oftheir pyrrolidine rings thereby decreasing the entropy during proteinunfolding by reducing the numbers of unfolded conformations thatcan be sampled by the protein (5). This proline theory was suggestedby Suzuki et al. (6, 7) based on the studies of thermostable oligo-1, 6-glycosidases from Bacillus. The thermostability of several glycosidaseshad been successfully improved according to this theory (3, 8–13),although there was no crystal structure available.

α-Glucosidase (TtGluA) from Thermoanaerobacter tengcongensisMB4 is a moderate thermostable enzyme with potential applicationfor maltooligosaccharides hydrolysis and isomaltooligosaccharidessynthesis in industry. However, it is not stable at the temperatureabove 70 °C where most industrial processes take place. In this study,we compared the deduced amino acid sequences of TtGluA and otherrelated thermostable α-glucosidases and investigated the effect ofsubstituting residues in the wild-type enzyme with proline on thethermostability of TtGluA.

MATERIALS AND METHODS

Bacterial strains, plasmids, restriction enzymes, and chemicals The obli-gately anaerobic and extremely thermophilic bacterium, T. tengcongensis MB4, wasfrom China General Microbiological Culture Collection Center (CGMCC). The expressionvector pET-28a was from Novagen (USA). Escherichia coli XL1-Blue and E. coli BL21(DE3) competent cells were from Stratagene (USA). All the restriction enzymes werefrom TaKaRa (Japan). The isopropyl-β-D-thiogalactopyranoside (IPTG), kanamycin andimidazole were from Merck (USA). All other chemicals were of reagent grade.

l rights reserved.

PROLINE SUBSTITUTION OF T. TENGCONGENSIS a-GLUCOSIDASE 13VOL. 110, 2010

Construction of wild-type expression plasmids The genomic DNA of T.tengcongensisMB4 was extracted by the bacteria genomic DNA extraction kit (Tiangen,China). The α-glucosidase gene TtgluA (GenBank No. AAM23323) was amplified fromthe genomic DNA by PCR with the primer pair of the forward (5′-GCTAGCTAG-CATGCTTCAAAGAAC-3′, the underline indicates the Nhe I site) and the reverse (5′-CCGGAATTCCTATTTCACTACAATC-3′, the underline indicates the EcoR I site) and pfuDNA polymerase. The PCR product was purified using the Gel Extraction Kit (OMEGABio-tek, USA) and then digested with Nhe I and EcoR I before ligating to pET28a vector.The resulted recombinant plasmid, pET28a-TtgluA, was transformed into E. coli XL1-Blue for cloning and verification through sequencing. Then the verified pET28a-TtgluAwas transformed into E. coli BL21 (DE3) for expression.

Alignments of α-glucosidase sequences from different thermophilic bac-teria Homologues of TtGluA were searched in the GenBank™ database using BLAST(Basic Local Alignment Search Tool). Sequences from thermophilic bacteria with highidentities to TtGluA were chosen for sequence alignment. Sequence alignments wereperformed by ClustalX and the results were formatted by Espript 2.2.

Proline substitution mutagenesis Proline substitution mutagenesis was doneusing the QuikChange® Site-Directed Mutagenesis Kit (Stratagene, USA). Therecombinant plasmid, pET28a-TtgluA, was used as the template. The primers formutagenesis are shown in Table 1. The mutagenesis PCR products were transformedinto E. coli XL1-Blue directly after Dpn I digestion. DNA sequencing was performed bySinoGenoMax Co., Ltd, China. The correct mutant plasmid was then transformed into E.coli BL21 (DE3) for expression.

Protein expression and purification E. coli BL21 (DE3) harboring the pET28a-TtgluA and mutant plasmids were cultured in 0.5 l of LB medium containing kanamycin(60 μg ml−1) until OD600 reached 0.6. Isopropyl-β-D-thiogalactopyranoside (IPTG) wasadded to a final concentration of 1mM, and the cells were cultivated further at 37 °C for5 hours.

Cells were harvested by centrifugation at 6000×g at 4 °C for 15 minutes, washedwith binding buffer (20 mM sodium phosphate buffer containing 500 mM NaCl and5 mM imidazole, pH 7.9) and then suspended in 50 ml of the same buffer. Thesuspended cells were disrupted by sonication and the supernatant was obtained bycentrifugation at 12,000×g for 20 minutes at 4 °C. The supernatant was incubated at65 °C for 30 minutes to denature the thermolabile E. coli proteins, and then centrifugedat 12,000×g for 30 minutes at 4 °C. The supernatant was loaded onto His·Bind column(5PKG) (Novogen, Germany) which was pre-equilibrated with the binding buffer. Thecolumn was washed with 10 ml of binding buffer and subsequently with 15 ml ofwashing buffer (20 mM sodium phosphate buffer containing 500 mM NaCl and 60 mMimidazole, pH 7.9). Finally, the protein was eluted with 3 ml of elution buffer (20 mMphosphate buffer containing 500 mM NaCl and 1 M imidazole, pH 7.9). The obtainedprotein solution was desalted by desalting column (GE Healthcare, England) with20 mM sodium phosphate buffer (pH 7.5) and then loaded onto Superdex 10/300 (GEHealthcare, England) equilibrated with 20 mM phosphate buffer containing 150 mMNaCl (pH 7.5). Elution was performed with the equilibration buffer at a flow rate of0.6 ml min−1 and the desired protein fraction were collected. The eluted protein wasdesalted again with 20 mM sodium phosphate buffer (pH 7.5). The proteinconcentration was defined with the Bio-Rad Protein Assay Reagent (Bio-Rad, USA)using bovine serum albumin as a standard. The purity of protein was examined by SDS–PAGE.

Standard enzyme assays The α-Glucosidase activity was determined bymeasuring the release of p-nitrophenol from the substrate p-nitrophenyl-α-D-glucopyranoside (pNPG, Sigma, USA). Assay mixtures (0.5 ml) containing 0.5 mMpNPG in 0.1 M citric acid–0.2 M sodium phosphate buffer (pH 5.5), were incubated at60 °C for 10 minutes with 0.02 mg ml−1 purified enzyme. The reaction was terminatedby addition of 0.5 ml of 1 M Na2CO3 solution. The absorbance of the liberated p-nitrophenol was measured at 410 nm. One unit of activity was defined as the amount ofenzyme liberating 1 μmol of p-nitrophenol in 1 min.

Thermostability testing of wild-type and mutant α-glucosidases T5015 andΔT5015 were determined according to the method of Eijsink et al. (14, 15) to evaluatethe thermostability of wild-type and mutated α-glucosidases. T5015 is the temperatureat which 50% of the initial activity remained after 15 minutes of incubation. ΔT5015

represents the difference in T5015 of the mutants from that of wild-type TtGluA. TheT5015 was detected by incubating 0.4 mg ml−1 of purified α-glucosidase for 15 minutesat different temperatures in 20 mM sodium phosphate buffer (pH 7.5). The residualactivity was assessed by standard assay method.

Circular dichroism spectroscopy Circular dichroism (CD) spectra of purerecombinant wild-type TtGluA and mutant enzymes were measured by the Jasco J-810spectropolarimeter (Jasco, Japan) over wavelength ranging from 190 to 260 nm underconstant nitrogen flush. The bandwidth was 1 nm. All spectra were recorded at roomtemperature, and three scans were averaged and blank-subtracted to give the spectra.

TABLE 1. Primers used for site-direc

Amino acid substitution Base change (codon) Forward prim

L152P CTT→CCT CAATCAAACTACCAAGCCN208P AAT→CCT ATTACTTCATTTACGGACCK285P AAA→CCA GGAACAAGGGTACTTTTCCT430P ACT→CCT GCCTTTTAAAGGAAAGACCT

Secondary structure prediction and 3D structure modeling Secondarystructure predictionwas done by the Protean program of DNASTAR. The tertiary structurewasmodeled by SWISS-MODEL serverwhich is a fully automatedmethod that generates athree-dimensional (3D) structural model by searching the homologous structures in theRCSB protein data bank (PDB) online. Figures were developed using Pymol 0.99. Thereliability of the modeled structure was checked by PROCHECK software online.

RESULTS AND DISCUSSION

Mutagenesis sites selection TtGluA is amoderate thermostableα-glucosidase belonging to family GH31 with a half-life of 50 min at65 °C. But it is not stable at temperature above 70 °C. To increase itsthermostability, proline substitution mutagenesis was considered.One strategy for identifying thermostabilising mutations is thecomparison of more stable proteins with less stable ones, with thegoal of identifying amino acid sequence patterns that correlate withthermostability (16). Another strategy is to compare the proteinsequences with high identities from the same genus bacteria and findthe conserved prolines possibly related to thermostability (11, 13). Inthis study, the selection of targets for mutagenesis was based on bothstrategies where comparisons of the amino acid sequences of twogroups of microbial thermostable α-glucosidases were made. Thesestrategies are especially useful since no three-dimensional structureinformation of TtGluA is available.

The first group consisted of two characterized thermostable α-glucosidases from Sulfolobus solfataricus P2 (MalA, AAK43151) andBacillus thermoamyloliquefaciens KP1071 (BthG II, BAA76396). Thehalf-life of MalA and BthG II is 35 hours at 85 °C (17) and 10 minutesat 81 °C (18) respectively. These two enzymes are thus morethermostable than TtGluA which has a half-life of less than 10minutes at 80 °C. Like TtGluA, these twoα-glucosidases also belong tothe GH31 family. And they both have the amino acid sequenceidentities of more than 35% with TtGluA. The portion of the sequencesalignment with TtGluA is shown in Fig. 1. Three sites (Leu152, Asn208,and Lys285) were selected from this alignment. In these three sites,the α-glucosidases of S. solfataricus P2 and B. thermoamyloliquefaciensKP1071 are prolines but TtGluA is not (Fig. 1A). This same rule ofmutagenesis targets selection was also used in the directed evolutionof several other glycosidases (3, 8).

The second group contained the α-glucosidases which show highidentities to TtGluA. Three α-glucosidases from the same genus ofThermoanaerobacter, Thermoanaerobacter sp. X514 (GenBank no.ABY91330), Thermoanaerobacter pseudethanolicus ATCC 33223 (Gen-Bank no. ABY93679), and Thermoanaerobacter ethanolicus (GenBankno. ABR26230) were selected. These three Thermoanaerobacter α-glucosidases have the highest sequence identities of about 78% withTtGluA. However, the functions of these three proteins are yet to beinvestigated. There are totally 26 prolines at the same sites of the fourenzymes except for the amino acid at the site 430 in which all threeThermoanaerobacter α-glucosidases are prolines but TtGluA isthreonine (Fig. 1B). We thus projected that proline substitution ofT430 may affect the thermostability of TtGluA. This proline substitu-tion sites selecting rule was also used in the directed evolution ofBacillus amylases and oligo-1, 6-glucosidases (13, 19).

Secondary structure prediction analysis Although there wasno three-dimensional structure information of TtGluA, the secondarystructure of TtGluA was predicted using the program Protean ofDNASTAR to evaluate the selected mutagenesis sites. Based on the

ted mutagenesis in this study.

er (5′→3′) Reverse primer (5′→3′)

TTTGTATGAGTCATATCC GGATATGACTCATACAAAGGCTTGGTAGTTTGATTGTGACATAAAAGAGGTAG CTACCTCTTTTATGTCAGGTCCGTAAATGAAGTAATAAATTATAAGGAAATGC GCATTTCCTTATAATTTGGAAAAGTACCCTTGTTCCAATGAAAGGCCCTTTATC GATAAAGGGCCTTTCATTAGGTCTTTCCTTTAAAAGGC

FIG. 1. Partial alignment of the amino acid sequences of α-glucosidases from different sources. (A) Alignment of characterized thermostable α-glucosidases and TtGluA. (B)Alignment of Thermoanaerobacter α-glucosidases and TtGluA. The filled triangles showed the selected sites for mutagenesis.

14 ZHOU ET AL. J. BIOSCI. BIOENG.,

Chou–Fasman model, the threonine at position 430 is at the secondsite of a β-turn containing four amino acids and L152 locates at a longrandom coil between two β-sheets. N208 is at a turn between a β-sheet andα-helix while K285 locates at a long loop including a β-turnof eight residues from 280 to 287. Proline residues are known torestrict backbone bond rotation because of their pyrrolidine. Suzuki(7) proposed a “proline rule” which states that many prolines at thesecond position of β-turn and at some appropriate sites of loops orrandom coils make proteins more thermostable. This rule had beenvalidated by many enzymes in past years (5, 10, 12, 20, 21).Therefore, it is probable that the addition of a proline at these selectedsites of T. tengcongensis α-glucosidase would make the enzyme morethermostable.

Expression and purification of TtGluA and mutant enzymesThe four residues (L152, N208, K285, and T430) of TtGluA wereindividually replaced by proline using site-directed mutagenesis.Unexpectedly, the expression of soluble protein of L152P and N208Pincreased, while in contrary that of K285P decreased markedly (Fig.2A). The expression of T430 showed no distinction with wild-typeenzyme. All the five enzymes were purified by a three-step processincluding heating treatment, His-Tag affinity column and gel-filtrationchromatography. As shown in Fig. 2B, the purified recombinant

FIG. 2. SDS–PAGE of recombinant wild-type and mutant enzymes expressed in E. coli BL21 (Dpurification. Lanes 1 and 6, TtGluA; lanes 2 and 7, L152P; lanes 3 and 8, N208P; lanes 4 and

mutant enzymes were homogeneity with molecular mass of approx-imately 89 kDa, similar to the wild-type enzyme.

Thermostability and hydrolytic activity To test the thermo-stability preliminarily, the purified enzymes were incubated attemperatures ranging from 70 °C to 78 °C for 15 minutes. The resultsshowed that the thermostability of L152P, K285P and T430Psignificantly increased in comparison with wild-type enzyme (Fig.3). The activity of wild-type enzyme decreased rapidly when theincubation temperature was more than 74 °C and almost no activityretained after 15 minutes of incubation at 77 °C. K285P, L152P, andT430P had about 67%, 25%, and 20% residual activity after 15 minutesof incubation at 77 °C respectively. N208P was more thermolabilethan wild-type enzyme with only 20% residual activity afterincubation at 74 °C compared to 32% residual activity for the wild-type enzyme. To further determine the T5015 values of the mutantsand wild-type enzyme, the incubation temperatures ranging from72 °C to 85 °C with an interval of 0.5 °C were used. The resultsshowed that the T5015 values of L152P, K285P, and T430P were 75 °C,83.5 °C, and 75 °C, with ΔT5015 of 2 °C, 10.5 °C, and 2 °C, respectively,while the T5015 of wild-type enzyme was 73 °C. T5015 value of N208Pwas 72.5 °C with a decrease of 0.5 °C in comparison to wild-typeenzyme.

E3). (A) Supernatant of crude enzyme extraction. (B) Purified enzyme after three steps9, K285P; lanes 5 and 10, T430P. M: molecular weight marker.

FIG. 3. The effect of the mutations on the thermal stability of the enzymatic activity ofTtGluA. The thermal stability of the enzymes was determined by monitoring theirresidual enzymatic activity after 15 minutes of incubation at varying temperatures.Enzymatic activity was assayed using pNPG as substrate. Values are expressed as theaverages of three experiments.

PROLINE SUBSTITUTION OF T. TENGCONGENSIS a-GLUCOSIDASE 15VOL. 110, 2010

Taken together, the results presented above indicated that thesubstitution of L152P, K285P, and T430P all enhanced the thermo-stability of TtGluA. These were in accordance with the predictionanalysis of secondary structure. However, the thermostability ofN208P decreased, indicating that this site was not appropriate forproline substitution to increase thermostability. On the other hand,the improved thermostability of T430P and analysis of the deducedamino acid sequences of Thermoanaerobacter sp. X514, T. pseudetha-nolicus, T. ethanolicus thermostable α-glucosidases and TtGluAindicated that the presence of the conserved proline at position 430in genus Thermoanaerobacter α-glycosidase perhaps represents anevolutionary trend towards a more thermostable enzyme.

The hydrolytic activity and kinetic parameters of the improvedmutants (L152P, K285P, and T430P) and wild-type enzyme weredetermined at 60 °C. The results are summarized in Table 2. The L152Pand K285Pmutants had less than 30% specific activity of the wild-typeenzyme, which indicated that the thermostability increasing of thesetwo mutants was accompanied by significant loss of activity. This isprobably due to the higher rigidity of the two sites where the prolinesubstitutions were located. This also indicated that these two sitesperhaps were close to the catalytic region. The T430Pmutant retainedabout 92% of wild-type activity. The Km of L152P and T430P are 1.5fold of wild-type enzyme while that of K285P is only one-third ofwild-type enzyme. This means that the mutations L152P and T430Pdecreased the substrate binding affinity while the mutation K285Pincreased the substrate binding affinity. The kcat of L152P and K285Pdecreased 36% and 74%, respectively, but that of T430P was about 1.1-fold of wild-type enzyme. Consequently, kcat/Km of L152P and K285Pdecreased to 43% and 87% of wild-type enzyme, respectively. Thus, notonly the specific activity but also the catalytic efficiency of these twomutants decreased. Although the specific activity and kcat of T430Pdid not change distinctly, the kcat/Km of T430P decreased to 72% ofwild-type enzyme because of the increased Km. These results showedthat themutations affected not only thermostability but also substratebinding and catalytic efficiency of TtGluA.

TABLE 2. Specific activity and kinetic parameters for wild-type and improved mutantsusing pNPG as substrate.

Enzyme Specific activity(m-unit mg−1)

Km

(mM)kcat(s−1)

kcat/Km

(s−1 mM−1)

Wild-type 362.2 1.08 6.64 6.15L152P 96.3 1.61 4.27 2.65K285P 90.2 0.32 1.72 5.37T430P 332.9 1.64 7.25 4.42

Modeled structure analysis of improved mutants The CDspectra were analyzed to determine whether the mutation affectedthe secondary structure. The results showed that there were nodistinct changes between wild-type enzyme and the three mutantswith increased thermostability (Fig. 4). This meant that thethermostability improvements were not caused by secondary struc-tural change. To further understand the possible mechanism ofimproved thermostability in these TtGluAmutations, the 3D structureof TtGluA was modeled by SWISS-MODEL software online, using thestructure of α-glucosidase (MalA) from S. solfataricus P2 as template.MalA belongs to family GH31 and shows the highest identity withTtGluA in the PDB database. As in MalA, the modeled structure ofTtGluA contained three domains: the catalytic domain, the N-terminaldomain, and the C-terminal domain (Fig. 5A). The overall quality ofthe model provided by the SWISS-MODEL server was assessedthrough evaluation of the Ramachandran plot. PROCHECK was usedto analyze the Ramachandran plot of the model by comparing it withthe Ramachandran plots of 118 representative structures solved athigh resolution of at least 2.0 Å and have an R-factor no greater than20%. In total, 84.2% residues were in the most favored regions and14.9% residues were in the additional allowed regions. Only fiveresidues were in generously allowed and disallowed regions of theplot. The model was thus evaluated to be of good quality based onthese results.

In this modeled structure of TtGluA, T430 is located at a loopbetween a β-sheet and anα-helix which are part of the (β/α)8-barrelthat forms the central catalytic region (Fig. 5B). This is similar to thepredicted result of secondary structure. K285 is located at a long loopincluding a β-turn of eight residues from 280 to 287. This loopconnects another β-sheet andα-helix of the central catalytic region inthe modeled 3D structure (Fig. 5C). The difference in comparison tothe loop harboring T430 is that there is a small helix between this β-sheet and α-helix. Proline in the polypeptide chain has lessconformational freedom compared to other amino acids, as thepyrrolidine ring of proline imposes rigid constraints on the N-Cαrotation (22, 23). So the T430P and K285P mutations probablydecrease the flexibility of loops carrying these mutations because ofconformation changes due to the pyrrolidine, which stabilized thestructure of TtGluA at high temperature. These prolines were alsofound in Bacillus thermoglucosidasius and Bacillus cereus oligo-1, 6-glucosidases (9, 10). Furthermore, for K285P, lysine is a hydrophilicresidue while proline is a hydrophobic amino acid. The hydrophobicside chain of proline can interact with the adjacent hydrophobiccavity, giving the turn a more fixed tertiary structure (6). Thisprobably was the second reason for the increased thermostability ofmutation K285P. In previous studies, the hydrophobic amino acids

FIG. 4. Far-UV CD spectra of TtGluA and improvedmutants in 20mM sodium phosphatebuffer (pH7.5). Each spectrum is the mean of three independent acquisitions.

FIG. 5. Modeled 3D structure of TtGluA. The structure wasmodeled by SWISS-MODEL online. The PDB file of 2g3nB ofMalAwas used as template. Figures were developed using Pymol0.99. (A) Overall 3D structure of TtGluA. (B) Location of Thr430. (C) Location of Lys285. (D) Location of Leu152.

16 ZHOU ET AL. J. BIOSCI. BIOENG.,

content was found to be marginally higher in thermophilies than thatin mesophilies as these residues can increase rigidity and hydropho-bicity of proteins (24). This indicated the role of hydrophobic effectthat destabilizes the unfolded forms and it will increase with thetemperature (25, 26). Thus, the increased thermostability of K285Pwas probably caused by the two combined action of flexibilityreduction and hydrophobic effect, giving K285P the highest T5015 valueamong the three improved mutants.

Based on the modeled structure, T430 is close to the bottom of theβ-barrel but a little far from the active sites and regions for substraterecognition and binding which are located at the top of the (β/α)8-barrel (Fig. 5A). Mutation T430P should thus not to perturb theseregions; however, this mutation increased the Km of TtGluA. This isprobably because the conformational change caused by mutationT430P induces conformational changes in other regions being close tothe substrate binding area, thus affecting kinetic properties of theprotein. Similar observation was also reported in Clostridiumbeijerinckii alcohol dehydrogenase (27). K285 is close to D267 andI268 which are conserved residues based on the comparison ofmodeled structure to family GH31 enzymes (Fig. 5C). In MalA, thesetwo corresponding residues are D212 and I213, which are thought tobe related to substrate binding and recognition (28). We postulatethat the mutation K285P possibly alter the interaction of K285 withD267 and I268 due to hydrophobic side chain of proline, whichaffected the substrate binding and recognition of TtGluA resulting inKm and specific activity changes.

In the modeled 3D structure, L152 is located at a long random coilbetween two β-sheets according to the secondary structure predic-tion (Fig. 5D). Because of less conformational freedom of proline, thepyrrolidine ring restricts the available conformational space of theresidue preceding proline (9). So the proline substitution of leucine at152 possibly enhanced the rigidity of the long random coil whichconsequently increased the thermostability of TtGluA. This phenom-enon was also observed in Streptomyces diastaticus xylose isomerase(21) and Bacillus oligo-1, 6-glucosidase (9-11). On the other hand,L152 is close to two other residues, D141 and Y154 in the modeled 3Dstructure (Fig. 5D), which are highly conserved in GH31 familyenzymes. In MalA, these two residues are D87 and Y99 which areconsidered to contribute to the environment around the active sitesand the N-terminal domain and might be implicated directly insubstrate binding and/or maintenance of the active site architecture(28). Leucine is more hydrophobic than proline, thus prolinesubstitution of leucine possibly reduced the hydrophobic interactionof residue at 152with D141 and Y154. This can affect the environmentaround the active sites and the substrate binding region, therebyresulting in the changes of the enzyme activity and kinetics.

In conclusion, three mutants of α-glucosidase (TtGluA) from T.tengcongensis MB4 with improved themostability were successfullyobtained by proline substitution mutagenesis in this work. Especially,mutation T430P increased thermostability without distinct loss ofspecific activity and catalytic efficiency. This shows the importance ofproline residues for the thermostability of TtGluA and makes TtGluA

PROLINE SUBSTITUTION OF T. TENGCONGENSIS a-GLUCOSIDASE 17VOL. 110, 2010

potentially more useful in industrial processes of starch hydrolysis. Allthe three mutations that increase thermostability are located at theflexible regions such as loops and random coils, which indicated that itis possible to achieve additional stabilization of TtGluA by introducingmutations at these flexible regions. The kinetic results also showedthat the selected mutagenesis sites not only are important forthermostability but also related to the substrate binding and catalyticactivity of TtGluA. This indicates that these sites are perhaps thecandidate mutation sites for improving activity of TtGluA.

ACKNOWLEDGMENTS

This work was financially supported by NSFC (30621005) andMinistry of Sciences and Technology of China (863 programs:2007AA021306 and 973 programs: 2007CB707801).

References

1. Krasikov, V. V., Karelov, D. V., and Firsov, L. M.: α-Glucosidase, Biochemistry(Moscow), 66, 267–281 (2001).

2. Emmanuel, L., Stefan, J., Bernard, H., and Abdel, B.: Thermophilic archaealamylolytic enzymes, Enzyme Microb. Tech., 26, 3–14 (2000).

3. Muslin, E. H., Clark, S. E., and Henson, C. A.: The effect of proline insertions on thethermostability of a barley α-glucosidase, Protein Eng., 15, 29–33 (2002).

4. Nakamura, S., Tanaka, T., Richey, Y. Y., and Nakai, S.: Improving thethermostability of Bacillus stearothermophilus neutral protease by introducingproline into the active site helix, Protein Engng., 10, 1263–1269 (1997).

5. Matthews, B. W., Nicholson, H., and Becktel, W. J.: Enhanced proteinthermostability from site-directed mutations that decrease the entropy ofunfolding, Proc. Natl. Acad. Sci. U. S. A., 84, 6663–6667 (1987).

6. Suzuki, Y., Oishi, K., Nakano, H., and Nagayama, T.: A strong correlation betweenthe increase in number of proline residues and the rise in thermostability of fiveBacillus oligo-1, 6-glucosdiases, Appl. Microbiol. Biotechnol., 26, 546–551 (1987).

7. Suzuki, Y.: A general principle of increasing protein thermostability, Proc. Jpn.Acad., Ser. B, 65, 146–148 (1989).

8. Suzuki, Y., Hatagaki, K., and Oda, H.: A hyperthermostable pullulanase producedby an extreme thermophile, Bacillus flavocaldarius KP1228, and evidence for theproline theory of increasing protein thermostability, Appl. Microbiol. Biotechnol.,34, 707–714 (1991).

9. Watanabe, K., Chishiro, K., Kitamura, K., and Suzuki, Y.: Proline residuesresponsible for thermostability occur with high frequency in the loop regions of anextremely thermostable oligo-1,6-glucosidase from Bacillus thermoglucosidasiusKP1006, J. Biol. Chem., 266, 24287–24294 (1991).

10. Watanabe, K, Masuda, T, Ohashi, H, Mihara, H, and Suzuki, Y.: Multiple prolinesubstitution cumulatively thermostabilize Bacillus cereus ATCC7064 oligo-1, 6-glucosidase, Eur. J. Biochem., 226, 277–283 (1994).

11. Kashiwabara, S. I., Matsuki, Y., Kishimoto, T., and Suzuki, Y.: Clustered prolineresidues around the active-site cleft in thermostable oligo-1, 6-glucosidase of Ba-cillus flavocaldarius KP1228, Biosci. Biotechnol. Biochem., 62, 1093–1102 (1998).

12. Martin, J. A., Pedro, M. C., and Clark, F. F.: Stabilization of Aspergillus awamoriglucoamylase by proline substitution and combining stabilizing mutations, ProteinEng., 11, 783–788 (1998).

13. Kazuaki, I., Tadahiro, O., Kaori, I. K., Yasuhiro, H., Hiroyuki, A., Keiji, E., Hiroshi,H., Katsuya, O., Shuji, K., and Susumu, I.: Thermostabilization by prolinesubstitution in an alkaline liquefying α-amylase from Bacillus sp. strain KSM-1378, Biosci. Biotechnol. Biochem., 63, 1535–1540 (1999).

14. Eijsink, V. G., Van den, B. B., Vriend, G., Berendsen, H. J., and Venema, G.:Thermostability of Bacillus subtilis neutral protease, Biochem. Int., 24, 517–525(1991).

15. Eijsink, V. G., Vriend, G., Van der, V. B., Hazes, B., Van den, B. B., and Venema, G.:Effects of changing the interaction between subdomains on the thermostability ofBacillus neutral proteases, Proteins, 14, 224–236 (1992).

16. Haney, P. J., Badger, J. H., Buldak, G. L., Reich, C. I., Woese, C. R., and Olsen, G. J.:Thermal adaptation analyzed by comparison of protein sequences frommesophilicand extremely thermophilic Methanococcus species, Proc. Natl. Acad. Sci. U. S. A.,96, 3578–3583 (1999).

17. Michael, R. and Paul, B.: Purification and characterization of a maltase from theextremely thermophilic Crenarchaeote Sulfolobus solfataricus, J. Bacteriol., 177,482–485 (1995).

18. Suzuki, Y., Nobiki, M., Matsuda, M., and Sawai, T.: Bacillus thermoamyloliquefa-ciens KP1071 α-glucosidase II is a thermostable Mr 540 000 homohexameric α-glucosidase with both exo-α-1, 4-glucosidase and oligo-1, 6-glucosidase activities,Eur. J. Biochem., 245, 129–136 (1997).

19. Kunihiko, W., Kazuhisa, K., and Yuzuru, S.: Analysis of the critical sites for proteinthermostabilization by proline substitution in oligo-1, 6-glucosidase from Bacilluscoagulans ATCC 7050 and the evolutionary consideration of proline residues, Appl.Environ. Microb., 62, 2066–2073 (1996).

20. Hardy, F., Vriend, G., Veltman, O. R., Vinne, B., Venema, G., and Eijsink, V. G. H.:Stabilization of Bacillus stearothermophilus neutral protease by introduction ofprolines, FEBS Lett., 317, 89–92 (1993).

21. Zhu, G. P., Xu, C., Teng, M. K., Tao, L. M., Zhu, X. Y., Wu, C. J., Hang, J., Niu, L. W.,and Wang, Y. Z.: Increasing the thermostability of D-xylose isomerase byintroduction of a proline into the turn of a random coil, Protein Eng., 12,635–638 (1999).

22. Schimmel, P. R. and Flory, P. J.: Conformational energies and configurationalstatistics of copolypetides containing L-proline, J. Mol. Biol., 34, 105–120 (1968).

23. MacArthur, M. W. and Thornton, J. M.: Influence of proline residues on proteinconformation, J. Mol. Biol., 218, 397–412 (1991).

24. Chakravarty, S. and Varadarajan, R.: Elucidation of determinants of proteinstability through genome sequence analysis, FEBS, 470, 65–69 (2000).

25. Ikai, A.: Thermostability and aliphatic index of globular proteins, J. Biochem(Tokyo), 88, 1895–1898 (1980).

26. Britton, K. L., Baker, P. J., Borges, K. M., Engel, P. C., Pasquo, A., Rice, D. W., Robb,F. T., Scandurra, R., Satillman, T. J., and Yip, K. S.: Insights into thermalstabilityfrom a comparison of the glutamate dehydrogenases from Pyrococcus furiosus andThermococcus litoralis, Eur. J. Biochem., 229, 688–695 (1995).

27. Oren, B., Moshe, P., Yael, H., Yakov, K., Felix, F. A., Joseph, K., and Yigal, B.:Enhanced thermal stability of Clostridium beijerinckii alcohol dehydrogenase afterstrategic substitution of amino acid residues with pralines from the homologousthermophilic Thermoanaerobacter, Protein Sci., 7, 1156–1163 (1998).

28. Heidi, A. E., Leila, L. L., Martin, W., Gordon, L., Paul, B., and Sine, L.: Structure ofthe Sulfolobus solfataricus α-glucosidase: implications for domain conservationand substrate recognition in GH31, J. Mol. Biol., 358, 1106–1124 (2006).