Val216 decides the substrate specificity of a-glucosidase in Saccharomyces cerevisiae Keizo Yamamoto 1 , Akifumi Nakayama 2 , Yuka Yamamoto 1, * and Shiro Tabata 1 1 Department of Chemistry, Nara Medical University, Japan; 2 Nara Prefectural Institute for Hygiene and Environment, Japan Differences in the s ubstrate s pecificity o f a-glucosidases should be due to the differences in the su bstrate binding and the catalytic domains of the enzymes. To elucidate such differences of enzymes hydrolyzing a-1,4- and a-1,6-glu- cosidic linkages, two a-glucosidases, maltase and isomaltase, from Saccharomyces c erevisiae were cloned a nd analyzed. The cloned yeast isomaltase and maltase consisted of 589 and 584 amino a cid residues, respectively. There w as 72.1% sequence identity with 165 amino acid a lterations between the two a- glucosida ses. These two a-glucosidase genes were subcloned into the pKP1500 expression vector and expressed in Escherichia coli. The purified a-glucosidases showed the same substrate specificities as those of their parent native glucosidases. Chimeric enzymes constructed from isomaltase by exchanging with maltase fragments were characterized by their substrate specificities. When the con- sensus region I I, which is one of the f our regions conserved in family 13 (a-amylase family), is replaced with the maltase type, the chimeric enzymes a lter to hydrolyze maltose. T hree amino a cid r esidues i n consensus region I I w ere d ifferent i n the t wo a-glucosidases. Thus, we modified Val216, G ly217, and Ser218 of isom altase to the m altase-type amino acids by site-directed mutagenesis. The Val216 mutant was altered to hydrolyze both maltose and isomaltose but neither the Gly217 nor the Ser218 mutant changed their substrate specificity, indicating that Val216 is an important residue discriminating the a-1,4- and 1,6-glucosidic linkages of substrates. Keywords: family 13; a-glucosidase; Saccharomyces cere- visiae; s ite-directed mutagenesis; substrate s pecificity. Glucosyl hydrolases (EC 3.2.1 ) are key e nzymes of carbohydrate metabolism that were found in the three major kingdoms, and are categorized into 57 structural families [1,2]. Family 13 (a-amylase family) includes enzymes such as a-amylase, a-glucosidase, pullulanase, cyclodextrin glucanotransferase, and 1,4-a- D -glucan branching enzyme, specifically acting on a-1,4- and a-1 ,6-O-glucosidic linkages [ 1]. Many primary structures of the members of family 13 from various origins are now available, and have been compared to each other. The existence of four highly conserved regions (regions I–IV) and three acidic residues located in the conserved regions as catalytic residues has been r eported [3–7]. Further more, computing s econdary structure analysis indicated that specific structural features of the catalytic (b/a) 8 -barrel domain exist in these enzymes [8–10]. The relationship of sequence and structure to substrate specificity i n family 13 enzymes, particularly a-amylase, cyclomaltodextrinase, and neo pullulanase, has been well studied [11–13]. Despite the fact that many a-glucosidases with diverse substrate specificities h ave been purified and cloned from mammals, plants, and m icroorganisms, i t i s s till not clear which amino acid residues of a-glucosidase recognize t he difference between a-1,4- a nd a-1,6-glucosidic bonds contained in saccharides. Yeast contains two a-glucosidases, a-1,4-glucosidase (E.C. 3.2.1.20, maltase) and oligo-1,6-glucosidase (E.C. 3.2.1.10, isomaltase), which act preferentially on maltose or isomaltose and methyl a- D -glucopyranoside (a-mg), respectively. The expression of these e nzymes is controlled by different polymeric genes, MAL or MGL, separately [14–16]. Maltase (the MAL6 product of Saccharomyces carlsbergensis) preferentially hydrolyzed maltose but neither isomaltose nor a-mg, whereas isomaltase hydrolyzes isomaltose and a-mg but not maltose [17,18]. Thus, we focused on the structure–function relationship of the two a-glucosidases f rom Saccharomyces as a model in respect of the difference in t heir substrate s pecificities. The yeast genome directory which was constructed by Goffeau et al. revealed t he existence of many homologo us open reading frames of a-glucosidase [19]. The comp lete nucleotide sequence of the MAL gene of Sacchar omyces has been determined [20], w hereas it is not clear w hich open reading frame corresponds to the MGL gene. In this study, we cloned the genes encoding isomaltase and m altase by means of a RT-PCR method and expressed them in Escherichia coli. Subsequently, f rom a comparison of the p rimary structures of the two a-glucosidases, chimeric enzymes were constructed by exchange parts of maltase and isomaltase genes including any one of the four conserved Correspondence to K. Yamamoto, Department of Chemistry, Nara Medical University, Shijo, Kashihara, Nara 634–8521, Japan. Fax/Tel.: +81 744 29 8810, E-mail: kama@naramed-u.ac.jp Abbreviations: a-mg, methyl a- D -glucopyranoside; a-pNPG, p-nitro- phenyl a- D -glucopyranoside. Enzymes: a-1,4-glucosidase (mal tase) (E.C. 3.2.1.20); oligo- 1,6-glucosidase ( isomaltase) (E.C. 3.2.1.10). *Present address: Department o f General Medicine, Nara Medical University, J apan. (Received 2 April 2004, r evised 17 J une 2004, accepted 5 J uly 2004) Eur. J. Biochem. 271, 3414–3420 (2004) Ó FEBS 2004 doi:10.1111/j.1432-1033.2004.04276.x regions f or family 13 members. In addition, we constructed mutants of isomaltase by replacing three amino acid residues after Asp215 in consensus region II with residues of the maltase type using site-directed mutagenesis. The substrate specificities o f all mutant enzymes were examined. We found that one amino acid residue in consensus region II decided the substrate specificity of isomaltase. Materials and methods Materials The yeast strains used w ere Saccharomyces cerevisiae D-346 (ATCC 56960) and 727–14C (ATCC 56959). The bacterial strains and plasmids used were Escherichia coli JM109, KP3998 [21], pUC18 and pKP1500 [21]. Hydroxyapatite (Gigapite) was purchased from Seika- gaku Kogyo and hydroxyapatite (Micro-Prep Ceramic Hydroxyapatite, type I) was from B io-Rad. Maltose, isomaltose, a-mg, and p-nitrophenyl a- D -glucopyranoside were from Nakalai T esque, Japan. A site-directed muta- genesis kit (Quickchange TM ) was obtained from S tratagene and the bicinchoninic acid protein assay reagent was from Pierce Chemicals. La-Taq polymerase was purchased from Takara Syuzo and restriction endonucleases and T4 DNA ligase were from Takara S yuzo, or New En gland Biolabs. Reverse transcriptase was used from the Expand TM Reverse Transcriptase kit from Boehringer Mannheim. The Marathon kit was purchased from Clontech Labor- atory. Oligonucleotides were synthesized by Takara Syuzo Custom Service. Assay method for enzyme activity a-glucosidase activity was determined by measuring the release of p-nitrophenol from p-nitrophenyl a- D -glucopyranoside ( a-pNPG) accord- ing t o the method described previously [22]. When maltose, isomaltose, and methyl a- D -glucopyranoside (a-mg) were used as substrates, the enzyme activity was determined as the rate of hydrolysis of the substrate by measuring the release of glucose according to the enzymatic method of NADP + reduction using hexokinase and glucose-6-phos- phate dehydrogenase [23]. Cloning of the isomaltase gene from S. cerevisiae The production of maltase and isomaltase of S. cerevisiae was induced by adding maltose and a-mg to the culture medium, respectively [ 22]. In t he case of the cloning of the isomaltase gene, total RNA w as prepared from S . cerevisiae D-346 grown o n a medium including 3% (w/v) a-mg by the method of Chomczynski & S acchi [24]. The mRNA was purified from the to tal RNA using Oligotex-dT 30 (Super) (Takara Syuzo) according to the manufacturer’s instruc- tions. Double-stranded cDNA was constructed from poly(A) RNA with the oligo-dT primer u sing the Expand TM Reverse Transcriptase kit. The N-terminal amino acid sequence, TISSAHPETEPK, which was determined from purified yeast isomaltase, matched the ORF YGR287c on chromosome VII of S. cerev isiae [19]. Therefore, the iso- maltase gene was amplified from the c DNA library by PCR using t he N-terminal sequence of ORF YGR287c and oligo- dT as primers . The 1 .8 kb RT-PCR product was ligated to plasmid pUC 18 after digestion with SmaI and introduced into E. coli JM109. The insert was sequenced by the dideoxy method [25] using t he Dye T erminator C ycle Sequencing F S Ready Reaction kit (Applied Biosystems). To verify the 5 ¢ end sequence, 5¢ RACE was performed using the Marathon kit with the AP1 primer and a gene-specific pr imer (5¢-AGATTGCCTTTCTACAGTCTTCATTC-3¢) accord- ing to the manufacturer’s p rotocol. The 5¢-RACE product was sequenced by a direct sequencing method. Subcloning into the pKP1500 expression vector Forward and reverse primers were designed based on the 5¢-and3¢-terminal nucleotide sequences of the isomaltase gene (MGL) for cloning into plasmid pKP1500. The forward p rimer 5¢-ATGACTATTTCTTCTGCACAT CCAGAGACAGAAC-3¢ con tains the initiation codon, while the reverse primer 5¢-CTTTCTGCAGACTCA TTCGCTGATATATATTC-3¢ linked a PstI restriction site to the termination codon. PCR was carried out on the isomaltase gene cloned above. PCR products were digested with PstI. Simultaneously, pKP1500 was digested with EcoRI and PstI, and then the EcoRI site was blunted by the use of a Blunting kit (Takara Syuzo). The vector and the insert MGL gene were ligated with T4 ligase followed by transformation into E. coli JM109. The cells were plated on Luria–Bertani agar supplemented with 40 lgÆmL )1 5-bromo-4-chloro-3-indolyl-a- D -glucopyranoside (Boehrin- ger Mannheim), 50 lgÆmL )1 ampicillin, and 1 m M isopro- pyl t hio-b- D -galactoside and then one da y later several b lue colonies appeared. One of the clones expressing isomaltase was selected. The plasmid containing the isomaltose gene was designated pYIM. Cloning of the maltase gene from S. cerevisiae For cloning of the maltase gene, S. cerevisiae 727–14C was grown in medium containing 3% (w/v) maltose. The mRNA and cDNA were pr epared using the same procedure as described above. The gene-specific primers were synthesized based on the information of Hong & Marmur [20]. The reverse primer was modified by introducing a HindIII site seven bases down- stream from the s top codon. PCR with La-Taq polymerase was carried out on the cDNA prepared from S. cerevisiae 727–14C. The 1 .8 kb PCR fragment w as digest ed wit h HindIII. Plasmid pKP1500 was digested with EcoRI and HindIII, and then the Eco RI site was blunted by the use of a Blunting kit (Takara Syuzo). The fragment was inserted into the pKP1500 vector and the resulting plasmid was intro- duced into E. coli KP3998. Several transformants containing the 1.8 kb insert were selected and sequenced. The plasmid carrying the maltase gene was designated p YMA. Expression of recombinant enzymes in E. coli and purification of the enzymes The E. coli transformant carrying pYIM (or pYMA) was inoculated into PYG medium [21] supplemented with 50 lgÆmL )1 of ampicillin and i ncubated at 37 °C. Isopropyl thio-b- D -galactoside (final 1 m M ) was added when cell density at A 660 reached 0.5 and the culture was further incubated for 12 h. Ó FEBS 2004 Substrate specificity of a-glucosidase (Eur. J. Biochem. 271) 3415 Cells were resuspended in 50 m M Tris/HCl buffer (pH 7 .5) a nd sonicated. The cell-free extract was applied to a Q AE-Toyopearl column equilibrated with 50 m M Tris/ HCl (pH 7.5) and the column was washed with the same buffer containing 20 m M NaCl. The enzyme was eluted with a linear gradient o f NaCl (20–150 m M ) in the same buffer. Active fractions were pooled and applied to a column of Gigapite equilibrated with 20 m M sodium phosphate buffer (pH 7 .0). The enzyme was eluted with a linear gradient of sodium phosphate buffer up t o 1 50 m M . T he act ive fractions were collected and dialyzed against 40 m M sodium phosphate buffer ( pH 6.8), t hen subjected to c hromato- graphy on hydroxyapatite (Micro-Prep Ceramic Hydroxy- apatite type I). The purified enzyme was eluted at 250 m M phosphate buffer (pH 6.8) by a linear gradient of 40–320 m M phosphate. Construction of chimeric enzymes from recombinant maltase and isomaltase Chimeric enzymes were constructed by exchanging nucleo- tide fra gm ents between t he maltase and isomaltase genes at a single restriction site on the plasmid or by inserting a fragment which was introduced at a unique restriction site by PCR. Chimeric enzymes MAa/IMb and IMa/MAb were con- structed by exchanging two Mun I/BglII fragments of pYIM and pYMA which were cleaved at single restriction sites with both of these restriction enzymes. The chimeric enzyme, Mun/Bpu was constructed by inserting a fragment, which was amplified by PCR with the forward p ri- mer 5¢-AGAAGCCATT GCTGAGCAATTTTTGTTC-3¢ (underlining i ndicates t he Bpu1102I restriction site) and t he reverse primer 5¢-AAA AAGCTTGCACTAATTTTATTT GAC-3¢ (underliningindicates the HindIII restriction site and stop codon, respectively) and pYMA as a template, into IMa/ MAb at Bpu1102I/HindIII. Other c himeric enzymes, M un/ Bst, Mun/Pst, and Pst/Bst were constructed by the same method described for the Mun/Bpu c himera. The chimeric enzymes are shown in a sche matic diagram in Fig. 2. Site-directed mutagenesis Site-directed mutagenesis (Asp215 fi Ala,Val216 fi Th r, Gly217 fi Ala, and Ser218 fi Gly of isomaltase) was carried out by the use of the Quick Change TM Site-Directed Mutagenesis k it and DNA from pYIM as a template and two additional mutagenic oligonucleotide primers for each amino acid substitution according to t he instruction man- ual. The sites to which the mutation was introduced were sequenced to confirm that only the expected mutation had occurred. Results and Discussion Cloning of yeast a-glucosidases Two a-glucosidase genes, encoding isomaltase and maltase, were isolated from an S. cerevisiae cDNA library using the PCR technique. In the case of isomaltase, the N-terminal amino a cid sequence, TISSAHPETEPK, matched O RF YGR287c on chromosome VII of S. cerevisiae [19]. More- over, s ix peptides, i ncluding the N -terminal amino acid sequence (TISSAHPETEPK, GSAWTFDEK, NGPRI HEFHQEM, LYTSASR, FRYNLVP, and TLKW PWEGR) obtained from t he native isomaltase were in accord with their nucleotide sequences of the ORF. B ased on this in formation, a 1.8 kb fragment was a mplified from the cDNA library by PCR using 5¢-sequence o f the ORF and oligo d T as p rimers, and was inserted into pUC18. Sequen- cing of the 1 .8 kb insert revealed an open r eading frame o f 1770 bp including a stop codon, TGA. The 5 89 amino acid protein deduced from the ORF was c omp letely identical to the amino acid sequence deduced from ORF Y GR287c. The sequence data for isomaltase is availab le from the DNA Data Bank of Japan with accession number AB109221. The entire coding region of the insert was amplified by PCR and subcloned into the pKP1500 expression vector, and the resulting plasmid was introduced into E. coli JM109. Expression of the gene was screened by a plate assay using 5-bromo-4-chloro-3-indolyl-a- D -glucopyrano- side. Several blue colonies were found to hydrolyze a-pNPG. The expression of isomaltase in these clones w as confirmed by their ability to hydrolyze isomaltose and a-mg but not maltose. The plasmid containing the isomaltase gene was designated pYIM. The maltase gene was also isolated f rom the DNA library of S. cerevisiae by PCR using gene specific primers. The amplified 1.8 kb fragment was inserted into plasmid pKP1500 and the resulting plasmid was transformed into E. coli. KP3998. DNA sequence analysis of the fragment gave 100% identity to the MAL6 gene [15]. The plasmid containing the maltase gene was designated pYMA. Figure 1 s hows a comparison of amino a cid s equences between maltase and isomaltase. T here is 72.1% of sequence identity with 165 amino acid alterations. Assessment of recombinant enzymes in comparison with native a-glucosidases We assessed the tw o recomb inant a-glucosidases in terms of substrate specificity and immunological identity and com- pared them to their native enzymes. The two recombinant a-glucosidases showed the same substrate specificities as those of their parent glucosidases, namely, maltase hydro- lyzed maltose but not isomaltose and a-mg, whereas isomaltase hydrolyzed isomaltose and a-mg but not malt- ose. Upon double i mmunodiffusion, rabbit antiserum against native isomaltase produced a single precipitation line without spurs with recombinant isomaltase (data not shown). When the two recombinant enzymes reacted with antisera against n ative maltase and isomaltase, the recom- binant enzymes showed the s ame dos e–response a s t he native enzymes by antiserum neutralization (data not shown). These results indicate that the two recombinant a-glucosidases a re identical to their parent enzymes. Substrate specificities of chimeric enzymes The c omparison of the primary structures of the members of family 13 from various origins has revealed the e xistence of four highly conserved regions I, II, III, and IV [3–7]. Thus, for the design o f chimeric a-glucosidases, the a-glucosidase gen es i n the two plasmids, pYMA and pYIM, 3416 K. Yamamoto et al.(Eur. J. Biochem. 271) Ó FEBS 2004 were divided into fi ve portions taking into account the four consensus regions. F igure 2 is a schematic representation of a number o f t he chimeric enzymes. Chimeric enzymes were characterized based on substrate s pecificities for maltase, isomaltase, a-mg, sucrose, and a-pNPG, and t he K m for a-pNPG. MAa/IMb a nd IMa/MAb w ere c onstructed by a recombination o f the N-terminal fragment containing consensus region I of isomaltase and maltase, respectively. The recombination had no effect on either the substrate specificities o r the K m for a-pNPG (Table 1). In the M un/ Bam chimera, the amino acids from 488 to the C-terminus of IMa/MAb were substituted b y the corresponding amino acids of m altase ( residues 4 85–584). The subst itution of the C-terminal fragment of IMa/MAb also had no effect on the substrate specificities. We further dissected the C-terminal region of IMa/MAb by preparing chimeras with switch- over points at r esidues 332 an d 231 (Mun/Bpu and Mun/ Bst, respectively). The specific activity for isomaltose of Mun/Bpu and Mun/Bst were about 10 and 80 times lower than that of isomaltase, respectively. The K m for a-pNPG of Mun/Bpu was the same a s that of isomaltase, whereas the K m for a-pNPG of M un/Bst was about 50 times l ower than that of isomaltase. T hus, fragments including consensus regions III and IV may affect the substrate affinity of the a-glucosidases. To investigate the role of the fragment containing consensus r egion II, two c himeras, Mun/Pst a nd Pst/Bst, were constructed. In the Mun/Pst chimera, a 27 amino acid fragment of Mun/Bst including consensus Fig. 1. Comparison of amino a cid sequences between maltase and is oma ltase. Identical and similar amino acid resi dues are designated by *andÆ, respectively. Four highly co nserved regions of family 13 are underlined. Fig. 2. Schematic diagram of the chimeric en zymes. Isomaltase se- quenceisrepresentedasanopenbarandmaltasesequenceisrepre- sented as a s haded bar. MunI, Pst I, BstBI, and Bpu1102I are r estriction sites used for the construction o f chimeric enzymes. I, II, III, and IV indicate the l ocation of f o ur highly c onserved regions of f amily 13. Ó FEBS 2004 Substrate specificity of a-glucosidase (Eur. J. Biochem. 271) 3417 region II was replaced by the corresponding fragment of pYMA. The substrate s pecificities of M un/Pst changed completely to those of maltase type. However, the charac- teristics of Pst/Bst which contained only the 27 amino acid fragment of pYIM in pYMA were t he same as those of Mun/Bst. Therefore, these results indicate that the fragment including consensus region II c ontributes to the determin- ation of the substrate specificity of a-glucosidase. Site-directed mutagenesis There were s ix amino acid d ifferences between the two a-glucosidases in the fragment including consensus region II. Three out of the six alterations were similar, thus, we targeted the other three amino acid residues in consensus region II for s ite-directed mutagenesis. The V al216, Gly217, andSer218inconsensusregionIIofisomaltasewere substituted to the corresponding amino acid residues of maltase, Thr, Ala, and Gly, respectively. The mutant enzymes G217A and S218G did not exhibit different substrate specificity to that of isomaltase but their K m for a-pNPG tended toward m altase (Table 2). Mutant V216T could hydrolyze the a-1,4-glucosidic linkage retaining the isomaltase type s ubstrate s pecificity a nd its hydrolyzing ratio of maltose/isomaltose was 1 : 1. As shown in T able 2, doubly and triply mutated enzymes including V216T (V216T/G217A, V216T/S218G, and V216T/G217A/ Table 1. Substrate specificities of the chimeric e nzymes. The enzyme was incubated with 0.5 M substrate in 100 lLof0.1 M sodium phosphate buffer, pH 7.0 at 30 °C for 5 min. The rea ction was s topped by addition of 100 lLof0.5 M Tris/HCl buffe r, pH 7.5, then r eleased glucose w as assayed. For a-pNPG, an increase of absorbanc e at 41 0 n m was me asured in 5 m M a-pNPG in 0 .1 M sodium ph osphate b uffer, pH 7.0 at 30 °C. Enzyme Specific activity (lmolÆmin )1 Æmg )1 enzyme) K m for a-pNPG (m M ) Maltose Isomaltose a-mg a-pNPG Maltase 70.0 0.00 0.00 132 0.31 Isomaltase 0.00 46.0 48.0 92.0 2.13 MAa/IMb 36.6 0.00 0.00 126 0.30 IMa/MAb 0.00 30.0 21.0 57.0 1.26 Mun/Bpu 0.00 4.40 2.30 34.0 3.32 Mun/Bst 0.00 0.69 0.23 5.30 0.045 Mun/Pst 34.0 0.00 0.00 98.0 0.15 Pst/Bst 0.00 0.46 0.23 5.70 0.043 Table 2. Kinetic parameters of wild-type isomaltase and site-directed mutants. TheconsensusregionIIofisomaltasewasmutatedtothe maltase type by site-directed mutagenesis. For example, V216T was made by exch angin g Val216 of isomaltase with Thr of maltase. Enzyme Specific activity (lmolÆmin )1 Æmg )1 enzyme) K m for a-pNPG (m M ) Maltose Isomaltose Isomaltase 0.00 45.8 2.13 Maltase 68.7 0.00 0.31 D215A 0.00 0.00 ND V216T 16.5 16.5 0.59 G217A 0.00 16.0 0.53 S218G 0.00 27.5 0.84 V216T/G217A 36.6 6.41 0.61 V216T/S218G 21.1 6.87 0.40 G217A/S218G 0.00 22.9 0.66 V216T/G217A/S218G 6.18 0.57 0.49 Fig. 3. Sequence alignment of a-glucosidases of known substrate spe- cificityintheconsensusregionII.Asp residue of the c atalytic nucleo- phile is labeled with an arrow and the next residue is highlighted in bold. Shown are: Sce D-346, S. cerevisiae isomaltase (this study); Bt h, B. thermoglucosidasius oligo-1,6-glucosidase [27]; Bce, B. ce reus suc- rase-isomaltase [28]; Bco, B. coagulans sucrase-isomaltase [29]; Bsp1, Bacillus sp. D G0303 a-glucosidase [30] , Bsp2, Basillus sp . F5 sucrase - isomaltase [31]; Bfl, B. flav ocaldarius oligo-1,6-glucosidase [32]; Spn, Streptococcus pneumoniae a-1,6-glucosidase [33]; Bsu, B. subtilis suc- rase-isomaltase-maltase [34]; B sp3, Bacillus sp. a-glucosidase [35], T cu, Thermomonospora curvata a-glucosidase [36] ; Sce727–14C, S. cerevis- iae maltase (this study); Sca, S. c arlsbergensis maltase [20]; C al, C. albicans maltase [37]; H po, Hansenula polymorpha maltase [38]. 3418 K. Yamamoto et al.(Eur. J. Biochem. 271) Ó FEBS 2004 S218G) exhibited a change in the h ydrolyzing ratio o f maltose/isomaltoseto5:1,3:1,and10:1,respectively. These facts indicate that the three residues in consensus region II, particularly Val, plays an importan t role in distinguishing between the a-glucosidic linkages of a-1,4 and a-1,6. McCarter and Withers [26] indicated that Asp214 on the consensus region II of maltase is the catalytic nucleophile. Because the Asp214 of maltase is equivalent to the Asp215 of isomaltase, a mutant with the residue altered to Ala was tested for its activity on a-pNPG. None of the mutants including D215A had activity on a-pNPG and a-mg although the proteins were detected with antiserum against isomaltase by immunoblotting (data not shown). T hus, the Asp215 of isomaltase is one of three active acidic residues which are completely conserved in a-glucosidase group. Amino acid sequence alignment Figure 3 shows the amino acid sequence alignment of the consensusregionIIofa-glucosidases o f known substrate specificity. In the case of a-glucosidases hydrolyzing the a-1,6-glucosidic linkage, the amino acid residue following the catalytic nucleophile is Val. On the other hand, the corresponding residue of a-glucosidases which acting on the a-1,4-glucosidic linkage but does not a-1,6-linkage is Thr. X-ray crystallographic analysis of B. cereus oligo- 1,6-glucosidase revealed t hat Val200 following t he cata- lytic nucleophile Asp199 locates on the long loop region followed by Nb4, and the side chain of Val200 faces toward the inside of the c atalytic cleft [39]. Figure 4 shows the h ypothetical structure of the active site of S. cerevisiae isomaltase in complex with isomaltose or maltose using the crystal structure of B. cereus oligo-1,6-glucosidase [39] as the starting model. In the case of wild-type isomaltase, isomaltose fit to the active site, whereas maltose cannot bind to the active site because the side chain of Val216 interfere with binding of a 4-linked glucose. The CG1 of Val216 is too close to the O3¢ of maltose. On the other hand, both isomaltose a nd maltose can bind to th e V216T mutant because the steric hindrance between OG1 of Thr216 and O3¢ of maltose is canceled by the rotation of the s ide c hain of Thr216. The results indicate that the amino acid residue just after the catalytic nucleophile in consensus region II must b e involved in the recognition of a-glucosidic linkages. In conclusion, this work was successful in identifying the region and residue important in the determination of the substrate specificity of a-glucosidases. The identification of V216T and doubly and triply mutated enzymes altered in substrate specificity w ill serve as a bas is for p rogress toward further understanding the structure-function relationship of family 13 a-glucosidases. References 1. Henrissat, B. (1991) A c lassifocation of glycosyl hydrolases based on ami n o acid s equence similarities. Biochem. J. 28 0 , 309–316. 2. Henrisaat, B. (1996) Updating the sequence-based c lassification of glucosyl hydrolases. Bioc hem. J. 316, 695–696. 3. Matsuura, Y., Kusunoki. M., Harada, W. & Kakudo, M. (19 84) Structure and possible catalytic residues of Taka-amylase A. J. Bi ochem. 95, 697 –702. 4. Nakajima, R., Imanaka, T. & Aiba, S. (1986) Comparison of amino acid s equenc es of eleven d ifferent a-amylases. Appl. Microbiol. Bio technol. 23, 355–360. 5. Svensson, B. (1988) Regional distant sequence homology between amylases, a-glucosidases, and transglucanosylases. FEBS Lett. 230, 7 2–76. 6. Machius, M., Wiegand, G. & H uber, R . (1995) Crystal s tructure of calcium-depleted Bacillus licheniformis a-amylase at 2.2 A ˚ Resolution. J. Mol. Biol. 246 , 545–559. 7. Brayer, G .D., Luo, Y. & Withers, S.G. (1995) The structure of human pancreatic a-amylase at 1.8 A ˚ resolution an d c omparisons with related en zymes. Protein Sci. 4, 1730 –1742. 8. Jespersen,H.M.,MacGregor,E.A.,Sierks,M.R.&Svensson,B. (1991) Comparison of the domain-level o rganization of starch hydrolases and related e nzymes. Biochem. J. 28 0, 51–55. 9. Jespersen, H.M., MacGregor, E.A., Henrissat, B., Sierkes, M.R. & Svensson, B. (1993) Starch and glycogen-debranching and Fig. 4. Hypothetical model structure o f the ac tive site of isomaltase in complex with i somaltose or m altose. (A) W ild type. (B) V216T m utant. The models were co nstructed by the use o f the program HOMOLOGY from the Insight II (Accelrys Inc., San D iego,CA,USA).Thecrystalstructureof B. cereus oligo-1,6-glucosidase was used as the starting model. Isomaltose and maltose a re colored dark gray and pale gray, respectively. F igures were produced with MOLSCRIPT [40] and RENDER from the Raster3D package [ 41]. Ó FEBS 2004 Substrate specificity of a-glucosidase (Eur. J. Biochem. 271) 3419 branching enz ymes: P redic tion o f s t ructural f eatures of the cata- lytic (b/a) 8 -barrel domain a nd evoluti onary rela tionship to other amylolytic e nzymes. J. Prote in Chem. 12 , 791–805. 10. Svensson, B. (1994) Regional sequence alignment for extentions from b-strands 4 a nd 7 of catalytic (b/a) 8 -barreldomaininglyc- osyl hydrolase s family 13. Plant Mol . Biol. 25, 141–157. 11. Takata, H., Kuriki, T ., Okuda, S., Takesada, Y., I izuka, M., Mimamiura, N. & Imanaka, T. (1992) Action of neopullulanase catalyzes both hydrolysis and transglycosilation at a-(1-4)- and a-(1-6)-glucosidic linkages. J. Biol. Chem. 267, 18447–18452. 12. Ibuka, A., T onozuka, T., Matsuzawa, H. & Sakai, H. (1998) Conversion of neopullulanase-a-amylase from Thermoactinomy- ces vulgaris R-74 into an amylopullulanase-type enzyme. J. Bio- chem. 123, 275–282. 13. MacGregor, E.A., Janecek, S. & Svensson, B. (2001) Relationship of sequence and structure to specificity in the a-amylase family of enzymes. Biochim. Biop hys. Acta 1546, 1–20. 14. Carlson, M. (1987) Regulation of sugar utilization in Saccharo- myces species. J. Bacteriol. 169, 4873–4877. 15.Vanoni,M.,Sollitti,P.,Goldenthal,M.&Marmur,J.(1989) Structure and regulation of the multigene family controlling maltose fermentation in budding yeast. Prog. Nucleic Acid Res. Mol. Biol. 37, 281–322. 16. Johnson, M. & Carlson, M. (1992) Regulation of carbon and phosphate utilization. In The Molecular and C elullar Biology of Yeast Saccharomyces: Gene Expression (Johns, E .W ., Prngle, J.R. & Broach, J., eds), pp. 193–281. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. 17. Khan, N.A. & Eaton, N.R. (1 967) Purification and characteriza- tion of maltase and a-methyl glucosidase from yeast. Biochim. Biophys. A cta 146, 173–180. 18. Needleman, R.B., F ede roff, H.J., Eccleshall, T.R., Buchferer, B. & Marmur, J . (1978) Purific ation and c haracterization of a-glucosi- dase from Sacchromyces carlsbergensis. Biochemistry 37, 4567– 4661. 19.Goffeau,A.,Aert,M.L.,Agostini-Carbone,A.,Ahmed,M., Aigle, L., Alberghina, K., Albermann, M., Albers, M., Aldea, D., Alexandraki, G., et al. (1997) Th e yeast g eno me directory . Nature 387 (Suppl ), 1–105. 20. Hong, H.S. & Marmur, J. (1986) Primary s equence o f t he m altase gene of th e MAL6 l o cus of Saccharomyces carlsbergensis. Gene 41, 75–84. 21. Miki, T., Yasukochi, T., Nagatani, H., Furuno, M., Orita, T., Yamada, H., Imoto, T. & Horiuchi, T. (1987) Construction of a plasmid vector for the regulatable high level expression of eukaryotic ge nes in Escherichia c oli: an application to over- production of chicken lysozyme. Protein Eng. 1, 327 –332. 22. Tabata, S., Ide, T., Umemura, Y. & Trii, K. ( 1984) Purification and c h aracterization of a-glucosidasesproducedbySaccharomy- ces in response to three d istinct maltose genes. Biochim. Biophys. Acta 797, 231–238. 23.Kunst,A.,Draeger,B.&Ziegenhorn,J.(1984)Methods of Enzymatic Analysis, Vol. VI, 3rd edn. (Bergmeyer, H.U., ed), pp. 163–172. Ve rlag Chemie, Weinheim. 24. Chomczynski, P. & Sacchi, N. (1977) Single-step method of RNA isolation by acid guanidinium thiocyanate–phenol–ch loroform extraction. Anal. B iochem. 162, 256– 159. 25. Sanger, F., Nicolen, S. & C oulson, A.R. (1977) DN A sequencing with chain-t erminating inhibitors. Pro c. Natl A cad. Sci . USA 74, 5463–5467. 26. McCarter, J .D. & Withers, S.G. (1996) Unequivocal identification of Asp-214 as the catalytic nucleophile of Sac charomyces cerevisiae a-glucosidase using 5-fluoro glyco syl fl uorides. J. Bi ol. Chem. 271, 6889–6894. 27. Watanabe,K.,Chiishiro,K.,Kitamura,K.&Suzuki,Y.(1991) Proline residues responsible for t hermostability o ccur with high frequency in the loop regions of an extremely thermostable oligo- 1,6-glucosidase from Bacillus thermoglucosidasius KP1006. J. Bio l. Chem. 266, 24287–24294. 28. Suzuki, Y., Aoki, R. & Hayashi, H. (1982) Assignment of a p-nitrophenyl-a- D -glucopyranoside-hydro lyzing a-glucosidase of Bacillus cereus ATCC 7064 to an exo-oligo-1,6-glucosidase. Biochim. Bi ophys. Acta 704, 476–483. 29. Suzuki, Y . & Tomura, Y. (1986) Purification and characterization of Bacillus c oagulans oligo-1,6-glucosidase. Eur. J. Bioc hem. 15 8, 77–83. 30. Lee, Y E. (2000) Cloning and characterization of a-glucosidase gene from thermophilic Bacillus sp. DG0303. J. Microbiol. Bio- technol. 10, 244–250. 31. Yamamoto, M. & H orikoshi, K. (1990) Nucleotide sequence of alkalophilic Bacillus olig o-1,6-glucosidase geneandtheproperties ofthegeneproductinEscherichia c oli HB101. Denpun Kag aku 37, 137–144. 32. Kashiwabara, S., M atsuki, Y ., Kishimoto, T. & Suzuki, Y. ( 1998) Clustered proline r esid ues a roun d t he active-site cleft in therm o- stable oligo-1,6 glucosidase of Bacillus flavocaldarius KP1228. Biosci. B iotechnol. B iochem. 62 , 1093–1102. 33. Coffey, T.J., Enright, M .C., Daniels, M., M orona, J.K., Morona, R., H ryniewicz, W., Paton, J.C. & Spratt, B.G . (1998) Recombinational exchanges at the capsular polysaccharide bio- synthetic locus lead to frequent serotype changes among natural isolates of Stre ptococcus pne umoniae. Mol. Microbiol. 27 , 73–83. 34. Scho ¨ nert, S., Buder, T. & D ahl, M.K. (1998) Identification and enzymatic characterization of the maltose-inducible a-glucosidase MalL (sucras e-isomaltase-maltase) of Bacillus subtilis. J. Bacteriol. 180, 2 574–2578. 35. Nakao, M., N akayama, T., Kakudo, A., Inohara, M., Harada, M., O mura, F. & Shibano, Y. (1994) Struc ture and expression of a gene coding for t hermostable a-glucosidase with a broad substrate specificity from Bacillus sp. SAM1606. Eur. J. Biochem. 220, 293– 300. 36. Janda, L., Pavelka, P., Tichy, P., Spizek, J. & P etricek, M. (1997) Production and properties of a-glucosidase from the t hermotol- erant bacterium Thermomonospora curvata. J. Appl. Micobiol. 83, 470–476. 37. Geber, A., Williamson, P.R., R ex, J.H., Sweeney, E.C. & B ennett, J.E. (1992) Cloning and characterization of a Candida albicans maltase gene involve d in s ucrose utilization. J. Ba cteriol. 174, 6992–6996. 38. Liiv, L., Pa ¨ rn, P. & Alama ¨ e, T. (2001) Cloning of maltase gene from a m ethylotrophic yeast, Hansenula polymorpha. Gene 265, 77–85. 39. Watanabe, K., Hata, Y., Kizaki, H., Katsub e, Y. & Suzuki, Y. (1997) The r efined crystal structure of Bacillus cereus oligo- 1,6-glucosidase at 2.0 A ˚ resolution: structural characterization of proline-substitution sites for prote in t h ermostab ilizatio n. J. Mo l. Biol. 269 , 142–153. 40. Kuraulis, P.J. (1991) MOLSCRIPT : a pro gram to produce both detailed and s chemat ic plots o f p roteins. J. Appl . Crystallog. 23, 946–950. 41. Merritt, E.A. & Murphy, M .E.P. (1994) RASTER 3 D , a program for photorealistic molecular graphics, Version 2.0. Acta Crystallog. Sect. D 50, 869–873. 3420 K. Yamamoto et al.(Eur. J. Biochem. 271) Ó FEBS 2004 . because the side chain of Val216 interfere with binding of a 4-linked glucose. The CG1 of Val216 is too close to the O3¢ of maltose. On the other hand, both. shows the amino acid sequence alignment of the consensusregionIIofa-glucosidases o f known substrate specificity. In the case of a-glucosidases hydrolyzing the a-1,6-glucosidic