Characterization of a Thermoacidophilic l-Arabinose Isomerase from Alicyclobacillus acidocaldarius: Role of Lys-269 in pH Optimum

Abstract

The araA gene encoding l-arabinose isomerase (AI) from the thermoacidophilic bacterium Alicyclobacillus acidocaldarius was cloned, sequenced, and expressed in Escherichia coli. Analysis of the sequence revealed that the open reading frame of the araA gene consists of 1,491 bp that encodes a protein of 497 amino acid residues with a calculated molecular mass of 56,043 Da. Comparison of the deduced amino acid sequence of A. acidocaldarius AI (AAAI) with other AIs demonstrated that AAAI has 97% and 66% identities (99% and 83% similarities) to Geobacillus stearothermophilus AI (GSAI) and Bacillus halodurans AI (BHAI), respectively. The recombinant AAAI was purified to homogeneity by heat treatment, ion-exchange chromatography, and gel filtration. The purified enzyme showed maximal activity at pH 6.0 to 6.5 and 65°C under the assay conditions used, and it required divalent cations such as Mn²⁺, Co²⁺, and Mg²⁺ for its activity. The isoelectric point (pI) of the enzyme was about 5.0 (calculated pI of 5.5). The apparent K_m values of the recombinant AAAI for l-arabinose and d-galactose were 48.0 mM (V_max, 35.5 U/mg) and 129 mM (V_max, 7.5 U/mg), respectively, at pH 6 and 65°C. Interestingly, although the biochemical properties of AAAI are quite similar to those of GSAI and BHAI, the three AIs from A. acidocaldarius (pH 6), G. stearothermophilus (pH 7), and B. halodurans (pH 8) exhibited different pH activity profiles. Based on alignment of the amino acid sequences of these homologous AIs, we propose that the Lys-269 residue of AAAI may be responsible for the ability of the enzyme to act at low pH. To verify the role of Lys-269, we prepared the mutants AAAI-K269E and BHAI-E268K by site-directed mutagenesis and compared their kinetic parameters with those of wild-type AIs at various pHs. The pH optima of both AAAI-K269E and BHAI-E268K were rendered by 1.0 units (pH 6 to 7 and 8 to 7, respectively) compared to the wild-type enzymes. In addition, the catalytic efficiency (k_cat/K_m) of each mutant at different pHs was significantly affected by an increase or decrease in V_max. From these results, we propose that the position corresponding to the Lys-269 residue of AAAI could play an important role in the determination of the pH optima of homologous AIs.

l-Arabinose isomerase (AI) (EC 5.3.1.4) is an intracellular enzyme that catalyzes the reversible isomerization of l-arabinose to l-ribulose involved in either the pentose phosphate or the phosphoketolase pathway (12, 30). Thermophilic AIs have been reported to possess a catalytic activity for the conversion of d-galactose to d-tagatose (16, 23, 25). This bifunctional activity could be exploited industrially for the production of d-tagatose, a novel and natural sweetener (4, 18). Indeed, in recent years, a biological process using thermophilic AIs for the production of d-tagatose from d-galactose resulting from the degradation of lactose-rich whey by β-galactosidase has been developed (14). Generally, isomerization performed at higher temperatures (>70°C) offers several advantages, such as higher conversion yield, faster reaction rate, and decreased viscosity of the substrate in the product stream. However, higher-temperature processes introduce undesired effects like browning and unwanted by-product formation (29). In order to overcome these problems, a thermostable AI with an acidic pH optimum (pH_opt) would be desirable and crucial for industrial applications.

A number of directed evolution studies of enzymes applied to industrial processes have been performed (1, 6). Such attempts have aimed to improve physical characteristics, such as pH optimum, temperature optimum, and thermostability, as well as catalytic activity towards the target substrate. However, since random directed evolution methods require high-throughput screening systems, limitations to improving enzyme properties for many industries still remain (20). As an alternative, three-dimensional (3-D) structure-based rational engineering has been also extensively performed (15, 19, 34, 38). Nonetheless, although recent progress has been made using both methods in engineering enzymes that can operate at a lower pH, attempts to rationally engineer enzymes have been limited to those enzymes whose structures are known (35).

In recent years, protein engineering via semirational design has been challenged by amino acid sequence comparisons (27, 39). The latter approach is based on the hypothesis that at a given position in an amino acid sequence alignment of homologous proteins, the nonconsensus amino acids rather than the consensus amino acids are responsible for the distinct characteristics of each protein. By applying the consensus concept, several thermolabile proteins have been successfully engineered to attain thermostability (26).

To dates, several acidic proteins from Alicyclobacillus acidocaldarius, a thermoacidophilic bacterium that grows optimally at 50 to 60°C and at pH 3 to 4 (7, 42), has been investigated for pH-related properties (9, 10, 33, 34, 41). For example, although AIs from Bacillus species such as B. subtilis and B. halodurans and from Geobacillus stearothermophilus are highly conserved in terms of amino acid sequences (>60%), they showed different pH optima (23). Thus, AI from A. acidocaldarius could serve as a good model to investigate amino acid residues essential for pH optimum in order to engineer AIs compatible with acidic conditions. Toward this end, we have cloned, sequenced, and expressed the araA gene from the thermoacidophilic bacterium A. acidocaldarius. Together with previously well-characterized AIs from the alkalophilic B. halodurans and the neutrophilic G. stearothermophilus, comparative biochemical studies were performed. Based on sequence alignment, site-directed mutagenesis was carried out to generate mutants possessing a modified pH optimum. From this simple rational approach based on the consensus concept, AI with an improved acidic pH performance more favorable for the production of d-tagatose could be developed. Moreover, our findings provide insight into the molecular evolution of AIs for environmental adaptation.

MATERIALS AND METHODS

Materials.

Ex-Taq DNA polymerase, deoxynucleoside triphosphate (dNTP), and chemicals for PCR were obtained from Takara Biomedicals, and the pGEM-T Easy vector and T4 DNA ligase were obtained from Promega. All columns for protein purification came from Pharmacia, while Genomictip and a plasmid miniprep kit were from QIAGEN. The pET-22b(+) expression vector was purchased from Novagen, electrophoresis reagents were purchased from Bio-Rad, and all chemicals used for enzyme assays and characterization were obtained from Sigma. Oligonucleotides were synthesized by Cosmo.

Bacterial strains and culture conditions.

A. acidocaldarius (ATCC 43030) was obtained from the American Type Culture Collection (ATCC), Manassas, Va. A. acidocaldarius was grown in an ATCC medium (573 Bacillus medium, pH 4.0) containing 1 g of l-arabinose, 1 g of yeast extract, 1.3 g of (NH₄)₂SO₄, 0.37 g of KH₂PO₄, 0.25 g of MgSO₄ · 7H₂O, 0.07 g of CaCl₂ · 2H₂O, and 0.02 g of FeCl₃ per liter at 55°C. Escherichia coli DH5α was used as a host for the construction of expression vectors, and E. coli BL21(DE3) was used as the host for expression. Both strains were grown in Luria-Bertani (LB) medium with ampicillin (100 μg/ml) in a rotary shaker at 37°C. The plasmid pGEM-T Easy vector was used as a cloning and sequencing vector, and pET-22b(+) was used for expression.

Cloning and sequencing of the araA gene from A. acidocaldarius.

In order to obtain an internal sequence of the AI-encoding araA gene from A. acidocaldarius, several putative araA genes obtained from the microbial genome sequence in GenBank were aligned. Two conserved sequences were identified and used to design the degenerate primers DaraAF (5′-TGGMTGCAYACNTTYTCDCCDKCNAARATGTGGAT-3′) and DaraAR (5′-TANGTGTARTCYTCCATRAANGANGT-3′) (Fig. 1). Genomic DNA was isolated from A. acidocaldarius using a genomic DNA extraction kit (QIAGEN) according to the manufacturer's instructions. The internal gene (770 bp) encoding AI from A. acidocaldarius was amplified by PCR using genomic DNA as a template. PCR mixtures (50 μl) contained 20 ng of genomic DNA, 0.2 μM each primer, 0.2 mM dNTP mix, and 2.5 U of Ex-Taq DNA polymerase in 1× PCR buffer (2 mM MgCl₂). After an initial denaturation step for 4 min at 94°C, the DNA was amplified by 30 cycles, each consisting of 30 s of denaturation at 94°C, 30 s of annealing at 60°C, and 1 min of extension at 72°C, followed by a final 5-min extension step at 72°C. The 770-bp PCR product was cloned into the pGEM-T Easy vector and transformed into E. coli DH5α competent cells. Transformants containing the pGEM-T Easy vector harboring the partial gene encoding A. acidocaldarius AI (AAAI) were selected on LB medium-ampicillin plates containing 0.01% 5-bromo-4-chloro-3-indolyl-β-d-galactopyranoside (X-Gal). Plasmid DNA was isolated from the transformants and sequenced.

FIG. 1.

Alignment of AAAI, GSAI, and BHAI amino acid sequences. The alignment was generated using CLUSTAL X (37). The residues of AIs involved in α-helices or β-strands are indicated by H and E, respectively. Secondary structure prediction was performed by using the Jpred (5) and the PSIPRED (13) programs. The amino acid residues for mutagenesis are indicated by a box, and the amino acid sequences used to prepare degenerate primers for cloning are shown in boldface type. AAAI, A. acidocaldarius AI (GenBank accession number DQ013256); GSAI, G. stearothermophilus AI (accession number AAD45718); BHAI, B. halodurans AI (accession number BAB05592).

A pair of primers for inverse PCR (40), IaraAF (5′-CGGTTGATGAAGGTCATGGCGGATGGC-3′) and IaraAR (5′-CAATAACGGTTTTCGCAGTTCCAAAAG-3′), was designed from the sequence of the 770-bp PCR product. Genomic DNA (100 ng) was partially digested with Sau3AI for 2 min at 37°C, and the restriction endonuclease was inactivated by heating to 80°C for 20 min. The 2- to 4-kb DNA fragments were purified by using a QIAEX II gel extraction kit (QIAGEN). The digested and purified DNA solutions were diluted to 3 ng of DNA per μl (50 μl) and circularized with 2 U of T4 DNA ligase per μl in ligase buffer containing 1 mM ATP for 16 h at 16°C. Inverse PCR mixtures (100 μl) contained 30 ng of DNA, 0.5 μM primers, and 0.5 mM dNTP mix in 1× PCR buffer and 5 U of Ex-Taq DNA polymerase. The reaction mixtures were transferred directly to a thermal cycler preheated to 94°C. Thermal cycler parameters were as follows: an initial denaturation step of 2 min at 94°C; 26 cycles of 30 s at 94°C, 30 s at 60°C, and 3 min at 72°C; and a final extension step of 10 min at 72°C. Inverse PCR products were purified using the QIAEX II gel extraction kit (QIAGEN) and cloned into the pGEM-T Easy vector and sequenced using the same procedure as described above. A final PCR amplicon containing the entire araA coding sequence was prepared using a forward/reverse primer pair, 5′-GGATCCGCTGTCATTACGTCCTTATGAA-3′, with an introduced BamHI site (underlined), and 5′-CCAAGCTTTCACCGCCCCCGCCAAAA-3′, with an introduced HindIII site (underlined). PCR was performed in 1× PCR buffer containing 2 mM MgCl₂, 20 ng of DNA, 10 pmol of primers, 200 μM dNTP mix, and 2.5 U of Ex-Taq DNA polymerase in a total volume of 50 μl. After an initial denaturation step for 4 min at 94°C, each DNA was amplified for 30 cycles of 30 s of denaturation at 94°C, 30 s of annealing at 60°C, and 1 min of extension at 72°C each, followed by a final extension step of 5 min at 72°C. The PCR product was cloned into the pGEM-T easy vector and transformed into E. coli DH5α. Transformants were selected on LB-ampicillin plates containing 0.01% X-Gal. Plasmid DNA was isolated from clones with inserts and digested with BamHI and HindIII. The araA gene of A. acidocaldarius was ligated into the pET-22b(+) vector to yield pET-AAAI, and the products were transformed into E. coli BL21. The pET-BHAI vector harboring the araA gene of B. halodurans was prepared as described previously (23).

Expression and purification of AIs.

For expression of the recombinant wild-type and mutant enzymes, the transformed E. coli BL21 cells were grown in 1 liter of LB medium at 37°C containing 100 μg of ampicillin per ml, induced in mid-exponential phase (A₆₀₀ = 0.5 ∼ 0.6) with 1 mM of isopropyl-β-d-thiogalactopyranoside (IPTG), grown for an additional 4 h, and harvested by centrifugation (10,000 × g, 20 min, 4°C).

To purify recombinant wild-type AAAI and mutant AAAI-K269E, the centrifuged cell pellets were resuspended in 20 mM Tris-HCl buffer (pH 8.0) and disrupted by sonication. The lysates were spun (12,000 × g, 20 min) to remove cell debris, and the supernatants were heated at 70°C for 15 min. After centrifugation (12,000 × g, 20 min), the supernatants of AAAI and AAAI-K269E were filtered through a 0.2-μm filter and loaded onto a Q-Sepharose column (High-prep 16/10 Q-XL, 20 ml; Pharmacia) equilibrated with 20 mM Tris-HCl buffer (pH 8.0). Elution was performed with a linear gradient of NaCl (0 to 500 mM) at a flow rate of 3 ml/min. Protein fractions containing AI activity were pooled and passed through a Superdex 200 column (HiLoad 16/60 prep grade; Pharmacia) preequilibrated with 20 mM Tris-HCl buffer (pH 7.5) and 150 mM NaCl. The purified enzymes were dialyzed against 10 mM potassium phosphate (AAAI, pH 6.5; AAAI-K269E, pH 7.5) and stored at 4°C. Wild-type B. halodurans AI (BHAI) and mutant BHAI-E268K were purified as described previously (23). The purified BHAI-E268K was dialyzed against 10 mM potassium phosphate (pH 7.5). All purification steps were carried out using an Δkta purifier system (Pharmacia) at 4°C. Protein concentrations were determined by the bicinchoninic acid method with bovine serum albumin as a standard (36).

Purity and size of proteins were analyzed by 12% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (PAGE) and visualized with Coomassie blue (22), and native PAGE and isoelectric focusing were performed as described previously (23, 25). N-terminal sequencing of the purified protein was performed at the Analytical Core Facility, Tufts University (Boston, Mass.).

Enzyme activity assay.

An AI assay was performed as described previously (8). Unless otherwise noted, standard reaction mixtures contained 50 mM phosphate buffer (pH 6.0 for AAAI and pH 7.0 for AAAI-K269E) or 50 mM HEPES buffer (pH 8.0 for BHAI and pH 7.0 for BHAI-E268K at room temperature), 1 mM MnCl₂, 1 mM CoCl₂, 20 μl of enzyme preparation at a suitable dilution, 0.1 M l-arabinose (d-galactose), and distilled water to a final volume of 125 μl. Assay mixtures for AAAI/AAAI-K269E and BHAI/BHAI-E268K were incubated for 20 min at 65 and 50°C, respectively, and stopped by cooling on ice. The ribulose (tagatose) formed was quantified by the cysteine-sulfuric acid-carbazole method, and the A₅₆₀ was measured. One unit of isomerase activity is defined as the amount of enzyme that produces 1 μmol of product per min under these conditions.

Site-directed mutagenesis.

AAAI and BHAI mutants were created by site-directed mutagenesis using pET-AAAI and pET-BHAI as the DNA template. Site-directed mutants were generated by PCR with the QuickChange kit (Stratagene, La Jolla, Calif.) and primers specific for each mutation. The forward/reverse primer pair of AAAI-K269E (5′-AAAGCCTTCCTGGAGGACGGGAATTTT-3′ and 5′-AAAATTCCCGTCCTCCAGGAAGGCTTT-3′) and another forward/reverse primer pair of BHAI-E268K (5′-AAGGAGTTCTTGAAAGAAGGGGGCTAC-3′ and 5′-GTAGCCCCCTTCTTTCAAGAACTCCTT-3′) were designed to introduce the K269E and E268K mutations into the AI genes of A. acidocaldarius and B. halodurans, respectively (underlined nucleotides represent those nucleotides changed by point mutations). The mutagenesis was carried out using 10 ng template DNA and a PCR program of 5 min at 95°C followed by 16 cycles of 30 s at 95°C, 1 min at 55°C, and 8 min at 68°C. The amplified plasmids were digested with DpnI and transformed into E. coli DH5α cells. Clones were sequenced to verify the presence of the mutation. Upon identification of a single coding region mutation, the mutated plasmids were reintroduced into E. coli BL21(DE3) cells and expressed as described above.

pH and temperature studies.

To determine the pH optimum, AI activity was measured with l-arabinose as a substrate as described above, except that the reaction buffer was replaced by 50 mM sodium acetate buffer (pH 3.5 to 6.0), 50 mM potassium phosphate buffer (pH 6.0 to 7.5), and 50 mM Tris-HCl buffer (pH 7.0 to 9.0). For pH stability studies, the enzyme was incubated at various pHs for different periods of time. All pHs were adjusted at room temperature, and the ΔpK_a/ΔT_s (ΔT_S being the change in temperature) for each buffer was taken into account in analyzing the results. The residual AI activity was measured under the above-described assay conditions at 65°C.

Kinetics.

The kinetic parameters of wild-type and mutant AIs were determined in the same reaction mixtures as described above for pH study, except that AI was assayed with a 3-min incubation to obtain the initial reaction rate. The concentration of l-arabinose or d-galactose ranged from 5 to 300 mM. Kinetic data were analyzed using Lineweaver-Burk plots to determine V_max and K_m.

Circular dichroism.

Circular dichroism (CD) measurements were performed with a Jasco J-810 spectropolarimeter with a Peltier temperature-controlled cuvette holder. Wild-type and mutant enzymes (0.5 mg/ml) were preincubated at 25°C in 20 mM sodium acetate buffer (pH 4 to 6), 20 mM potassium phosphate buffer (pH 6 and 7), and 20 mM Tris-HCl buffer (pH 8 and 9). The CD spectra of enzyme samples in a cuvette with a 0.1-cm path length were recorded in the far-UV region (190 to 250 nm). Scans were collected at 0.1-nm intervals with a 1-nm bandwidth five times. Each spectrum was corrected by subtracting the spectrum of the solution containing the used buffer.

High-performance ionic chromatography.

High-performance ionic chromatography (HPIC; Dionex) was performed using a Carbopac PA1 column (0.4 by 25 cm; Dionex) equipped with an electrochemical detector (ED40; Dionex) as described previously (23). d-Galactose and d-tagatose (Sigma) were separated by isocratic elution in 20 mM NaOH solution at a flow rate of 0.3 ml/min.

Nucleotide sequence accession number.

The GenBank/EMBL accession number for the A. acidocaldarius araA gene is DQ013256.

RESULTS

Isolation and cloning of the araA gene from A. acidocaldarius.

In order to amplify the corresponding araA gene from A. acidocaldarius ATCC 43030, the degenerate oligonucleotides DaraAF and DaraAR were designed on the basis of the internal conserved amino acid sequences of the araA genes from available microbial genome sequences in the GenBank databases (Fig. 1). With this primer pair, a 771-bp PCR product was obtained by PCR from the A. acidocaldarius chromosomal DNA, and its nucleotide sequence was determined by DNA sequencing (data not shown). The internal DNA sequence was used in inverse PCRs designed to obtain the full length of the araA gene from A. acidocaldarius. Primers IaraAF and IaraAR amplified approximately 2.5- to 4-kb DNA fragments, which were cloned into the pGEM-T Easy vector, sequenced, and shown to contain the full-length araA gene. The intact araA gene encoding AI from A. acidocaldarius was obtained by PCR with primers araAF and araAR, and the resulting DNA was cloned into the pET 22b(+) vector to generate pET-AAAI. Analysis of the sequence demonstrated that the 1,494-bp araA gene of A. acidocaldarius encoded a 497-amino-acid polypeptide with a calculated molecular mass of 56,043 Da. Comparison of the presumed AAAI amino acid sequence with those of other AIs showed that the enzyme has the highest identity (96 to 97%) to AIs from closely related thermophilic bacteria such as G. stearothermophilus and Geobacillus kaustophilus. AAAI also showed high similarity (>80%) and identity (>61%) to other AIs from Bacillus species such as B. halodurans, B. subtilis, and B. licheniformis.

Overexpression, purification, and characterization of AAAI.

The recombinant AAAI was purified to homogeneity by heat treatment followed by ion-exchange chromatography and gel filtration as described in Materials and Methods. The 14 N-terminal amino acid residues of the purified enzyme were determined to be MDIGINSDPLSLRP, which was identical to the deduced amino acid sequence of the araA gene from A.acidocaldarius. The molecular mass of the purified AAAI wasestimated to be 57 kDa by sodium dodecyl sulfate-PAGE, consistent with the molecular mass (56,043 Da) calculated from the presumptive amino acid sequence. However, native PAGE and gel filtration chromatography on Superdex 200 suggested that the native form of AAAI is a homotetramer with a molecular mass of about 224 kDa. Its isoelectric point (pI) was determined to be 5.0 (calculated pI of 5.51) on the isoelectric focusing gel (Table 1).

TABLE 1.

Comparison of biochemical properties of AAAI and other AIs from the genus of Bacillus

Property	AAAI	GSAI	BHAI
TempReact (°C)b	65	70	50
pHopt	6	7-7.5	7.5-8
Apparent Mr (103)
Oligomeric structure	α4	α4	α4
Subunit	57 (56.04)a	57 (56.09)a	57 (56.32)a
pI	5.0	5.6	5.8
Metal ion requirement for AI activity	Mn2+, Co2+, Mg2+	Mn2+, Co2+, Mg2+	Mn2+
Reference or source	This study	23	23

^aParentheses indicate calculated molecular weights of the deduced amino acid sequences.

^bTemp_React, reaction temperature.

The pH dependence of the recombinant enzyme was determined in the standard reaction mixtures after 20 min of incubation at various pHs (Fig. 2A). The pH optimum of AAAI was 6.0 to 6.5 at 65°C, and AAAI exhibited more than 50% of the maximal activity in the pH range of 5.2 to 8.0. As shown in Fig. 2B, the pH stability study demonstrated that AAAI retained its original activity even after 48 h of incubation at pH 6.0 and that its half-life at pH 5.0 was about 10 h. Thus, the purified AAAI is the most acidophilic AI of AIs reported so far.

FIG. 2.

pH optimum (A) and stability (B) of recombinant AAAI. (A) The enzyme was incubated under the standard assay conditions except for the reaction buffer: ▪, 0.1 M sodium acetate (pH 3.5 to 6.0); ○, 0.1 M potassium phosphate (ph 6.0 to 7.5); ▴, 0.1 M Tris-HCl (ph 7.0 to 9.0). (B) After various periods of incubation of the enzyme in different pH buffers (□, pH 4.0; •, pH 5.0; ▵, pH 6.0; ○, pH 7.0; ▪, pH 8.0) at room temperature, the residual activity of each aliquot was determined as described in Materials and Methods.

The isolated recombinant AAAI activity was activated by the addition of 1 mM Mn²⁺ (2.3-fold), 1 mM Co²⁺ (1.9-fold), and 1 mM Mg²⁺ (1.6-fold), whereas, as expected, the addition of 1 mM EDTA completely inhibited the AI activity. This is consistent with the observation that thermophilic AIs have an absolute metal requirement for activity (23, 24).

Kinetic studies of wild-type and mutant AIs.

Based on the primary sequence alignment of the AIs from three aerobic bacteria, A. acidocaldarius (acidophile), G. stearothermophilus (neutrophile), and B. halodurans (alkalophile), it was found that AAAI only differs from GSAI by 11 amino acid residues. Since, at pH 6.0, the AAAI has a pH optimum more acidic than those of GSAI (pH 7.0 to 7.5) and BHAI (pH 7.5 to 8.0), we proposed that the Lys-269 residue of AAAI could be responsible for the acidic pH optimum. Thus, we obtained AAAI-K269E and BHAI-E268K mutants using site-directed mutagenesis to investigate the role of Lys residue, where Glu-268 of BHAI is the corresponding residue to Lys-269 of AAAI, in the pH optimum of AI. In addition, the kinetic parameters and pH profiles of the mutants at different pHs were compared with those of the wild-type enzymes.

The K_m values of AAAI for l-arabinose and d-galactose as substrates at 65°C and at pH 6 were 48 mM (V_max, 35.5 U/mg) and 129.0 mM (V_max, 7.5 U/mg), respectively. These values are almost the same as those of GSAI at pH 7.0, indicating that AAAI has a quite similar substrate affinity to GSAI except for the pH optimum (23). Interestingly, the apparent K_m values of AAAI and AAAI-K269E for l-arabinose as a substrate were not significantly affected by the reaction pHs (Table 2). In addition, the K_m values of AAAI-K269E and BHAI-E268K were not largely different from those of their respective wild-type enzymes, whereas the V_max values changed with pH as shown in Fig. 3. In the case of the wild-type AIs, AAAI and BHAI exhibited V_max values at pH 6 and >8, respectively. However, the optimal pHs for their respective mutants, AAAI-K269E and BHAI-E268K, were shifted to neutral pHs around pH 7.0. In addition, the V_max value of AAAI-K269E was lower than that of AAAI by 1.7-fold at pH 6.0, whereas it was elevated 1.3-fold at pH 7.0. Although the V_max value of BHAI-E268K was little different than that of BHAI (1.03-fold at pH 8.0), the mutant had a 1.4-fold higher V_max value than the wild-type at pH 7.0. From these results, the difference in the k_cat/K_m value at each pH seemed to be dependent on the V_max value. Moreover, it appears that the Lys-269 residue of AAAI, corresponding to Glu-268 of BHAI, might play a key role in the determination of pH optimum for each AI.

TABLE 2.

pH dependence of the kinetic parameters of wild-type AIs and their mutants

Enzyme	Km (mM)			Vmax (U/mg)			kcatb (min−1)			kcat/Km (mM−1 · min−1)
	pH 6	pH 7	pH 8	pH 6	pH 7	pH 8	pH 6	pH 7	pH 8	pH 6	pH 7	pH 8
AAAI	48.0 ± 1.0a	46.2 ± 0.7	47.5 ± 1.3	35.5 ± 2.0	30.6 ± 0.8	26.8 ± 1.7	1,989	1,713	1,498	41.5	37.1	31.6
AAAI-K269E	57.9 ± 1.2	53.8 ± 3.3	52.5 ± 2.3	21.0 ± 1.2	39.9 ± 2.8	27.6 ± 0.8	1,177	2,232	1,548	20.3	41.5	29.5
BHAI	38.2 ± 8.9	39.7 ± 2.0	43.8 ± 0.2	17.0 ± 1.4	25.3 ± 1.7	33.2 ± 4.8	953	1,419	1,862	25.0	35.7	42.5
BHAI-E268K	56.1 ± 7.0	43.7 ± 0.6	46.0 ± 3.6	25.8 ± 1.9	35.3 ± 0.6	31.7 ± 2.6	1,447	1,976	1,778	25.8	42.8	37.5

^aData are means ± standard deviations.

^bk_cat is the number of substrate molecules reacted per active site per min.

FIG. 3.

Effect of pH on the AI activities of the wild-type (wild) (closed symbols) and mutant (open symbols) enzymes. The V_max values were obtained from the kinetic studies (see Materials and Methods).

d-Tagatose production at different pHs.

To investigate the effect of pH on the activity and conversion yield of each AI, AAAI and AAAI-K269E were assessed for their abilities to convert d-galactose to d-tagatose at pH 6.0 and 7.0 (Fig. 4). As shown in Fig. 4, only d-galactose and d-tagatose peaks were obtained at pH 6.0. However, undesirable small peaks were noted and a slight browning color was observed in the product solution at pH 7.0, indicating that the reaction mixture incubated at pH 7 contained by-products, while few were detected at pH 6. After 6 h of incubation at 60°C, the isomerization yields of d-galactose to d-tagatose by wild-type AAAI at pH 6.0 and 7.0 were 44.0 and 34.2%, respectively. However, AAAI-K269E, having a V_max that was two times higher at pH 7 than at pH 6, showed little difference, with conversion yields of 36.8% (pH 6.0) and 36.2% (pH 7.0), probably due to the formation of by-products and subsequently reduced conversion yield.

FIG. 4.

HPIC (Dionex) analysis of reaction mixtures after incubation of AAAI (A and B) and its mutant, AAAI-K269E (C and D), with 10mM d-galactose as a substrate. The reaction mixtures (1 ml) contained 40 mM potassium phosphate buffer (pH 6 and 7), 10 mM d-galactose, 1mM Mn²⁺, 1 mM Co²⁺, and 0.5 mg of AI. After 12 h of incubation at 60°C, the reaction mixtures were filtered and analyzed.

DISCUSSION

In food industries, considerable effort is being directed at replacing high-calorie sugar with a variety of natural and artificial sweeteners. Because of various health benefits, d-tagatose, a natural ketohexose, has been one of the most attractive candidates for a sucrose substitute (18, 28). In light of this, production of d-tagatose is achieved not only by a chemical process but also by a biological procedure using microbial AIs (4, 18). Indeed, the latter approach could be more favorable from an environmental viewpoint, given the series of complicated acid-base-treated processes resulting in a large amount of by-products associated with the former approach. Accordingly, several thermostable AIs have been isolated and characterized from various extremophiles (14, 16, 17, 23, 25). However, such enzymes are still required to perform catalytic activity under conditions for which they have not evolved. For instance, catalysis at such high temperatures and neutral pHs causes browning and other unwanted side reactions (18, 29). Ideally, industry requires AIs that are able to perform a highly efficient enzymatic process with high yield at moderately high temperatures and slightly low pHs.

To assess the effect of pH on conversion yields and formation of by-products, we performed isomerization of d-galactose to d-tagatose using wild-type and mutant AIs at 60°C and at different pHs (pH 6 and 7) and compared the conversion yields obtained by HPIC analysis (Fig. 4). From these results, it is evident that acidic conditions would be more favorable and advantageous for the production of d-tagatose. Therefore, since AAAI catalyzes the conversion of d-galactose to d-tagatose with an acidic pH optimum of 6.0 at 60°C, it could be a potentially attractive source for industrial application.

Compared with other thermophilic AIs, AAAI showed very similar biochemical properties except for pH optimum. Interestingly, the amino acid sequences of AI homologues including AAAI are highly conserved over the entire length of the protein, as shown in Fig. 1. Moreover, it is known that the 3-D structures of very distant members of a protein family (i.e., with very different sequences) may be practically superimposable, demonstrating that they can adopt essentially the same 3-D structure and perform the same biological function (2, 3). The identities (similarities) of the amino acid sequences of many variants can therefore be exploited to identify those regions that are absolutely critical for the structure and function of the given group of orthologous proteins (27). As such, it can be inferred that differences between the amino acid sequences of members of a given family could account for the unique properties of family members. Therefore, if the specific region responsible for a desired property was identified, the enzyme could be modified to render it suitable for industrial purposes. In early studies with fungal phytases, this sequence alignment-based approach, known as a simple semirational approach, was used to improve enzyme thermostability (26).

The different pH optima (AAAI, pH 6.0 to 6.5; GSAI, pH 7.0 to 7.5; BHAI, pH 7.5 to 8.0) of two thermophilic AIs and one alkalophilic AI could thus be explained by differences in primary amino acid sequences. As shown in Fig. 3, Lys-269 of AAAI, corresponding to Gln-269 of GSAI and Glu-268 of BHAI, seems to be responsible for the acidic optimum pH ofAAAI. To verify the hypothesis, the K269E mutant from AAAI (pH_opt of 6.0) and the E268K mutant from BHAI (pH_opt of 8.0) were prepared by site-directed mutagenesis. Comparative analysis of the pH activity profiles of the wild-type and mutant forms revealed the importance of Lys-269 in the pH tolerance of AAAI (Fig. 3). The replacement of Lys-269 with a Glu residue in AAAI to mimic Glu-268 of BHAI resulted in a significant decrease in catalytic activity at a low pH, with the AAAI-K269E mutant showing a neutral pH optimum, reflecting an approximately 1 unit change from pH 6.0 to 7.0. In contrast, the BHAI-E268K mutant showed a neutral pH optimum (pH 7.0) relative to the original pH optimum of the wild-type BHAI (pH 8.0). To exclude the possibility that conformational changes between the mutant and the wild-type proteins could affect the pH profiles, CD spectrum analysis in the far-UV region around 200 to 250 nm was performed for both versions of the enzymes (Fig. 5). Although the CD spectra in the far-UV region of 200 to 230 nm vanished at less than pH 5.0, those of AAAI and the K269E mutant did not undergo any significant change from pH 6.0 to 8.0, indicating that no significant conformational change exists between them in the pH range showing optimal catalytic activity (Fig. 5A and B). Therefore, we conclude that the decreased activity of the K269E mutant at a low pH is not due to significant structural changes in the protein.

FIG. 5.

CD spectra of the wild-type and mutant enzymes of AAAI and BHAI at different pHs. (A) AAAI; (B) AAAI-K269E; (C) BHAI; (D) BHAI-E268K. See Materials and Methods for conditions of data collection.

To investigate the role of the Lys-269 in AAAI in detail, we carried out kinetic analysis with the wild-type and mutant forms of AAAI and BHAI. Although the K_m values of AAAI-K269E were slightly higher than those of AAAI, there is little significant difference between K_m values at pH 6 to 8, respectively. However, the V_max value of the AAAI-K269E mutant at pH 6 was decreased by 1.95-fold relative to that of the K269E mutant at pH 7, resulting in a decrease in the k_cat/K_m value. In addition, the BHAI-E268K mutant also showed a decrease in V_max between pH 7 and 8, indicating that the mutation of Lys-269 of AAAI (corresponding to Glu-268 of BHAI) affected the k_cat/K_m value for each AI due to the difference in the V_max values between them. From the above-described results, we suggest that positively and/or negatively charged amino acid residues corresponding to Lys-269 of AAAI and Glu-268 of BHAI play a key role in the pH optimum of AI. Furthermore, we propose that the character of the residue at this position is highly correlated to the protonation of the active site of AI, affecting the level of enzyme-substrate complex.

Together with a notable role of Lys-269 of AAAI in the alteration of the pH activity profile, we have investigated the relationships between physical properties and amino acid sequences of AIs by alignment of their amino acid sequences. As shown in Fig. 6, a phylogenetic tree was constructed using all 17 available AI-homologous sequences by the neighbor-joining and p-distance methods. Similar topologies resulted from different tree-building methods (neighbor-joining versus minimum evolution) and distance estimations (p distance versus Poisson correction), showing the tree's robustness. The tree contains two major branches, one including AIs from thermophilic/mesophilic strains such as G. kaustophilus, G. stearothermophilus, A. acidocaldarius, B. subtilis, and B. licheniformis and the other including AIs from other sources, including hyperthermophiles. Bootstrap values indicate that the sequences in the thermophilic/mesophilic Bacillus-containing branch are well conserved (99%). Nevertheless, despite the low degree of sequence variability between these AI homologues, they show distinct physiological properties such as pH and temperature. Thus, small changes in amino acid sequences could change the characteristic of an enzyme in terms of its environmental adaptation. In particular, it appears that the character of the side chain found at the equivalent position of AAAI Lys-269 in the various Bacillus proteins dictates the pH optimum of the enzyme. Therefore, mutation at this position could be a simple strategy for changing enzymatic properties.

FIG. 6.

Phylogenetic tree of AI homologues. Bootstrap values are indicated at the branch points. The bar indicates a branch length equivalent of 0.1 changes per amino acid. All sequences were obtained from the TIGR website except for the G. stearothermophilus and T. neapolitana sequences, which were obtained from GenBank. Abbreviations: AAAI, Alicyclobacillus acidocaldarius; BH1873, Bacillus halodurans; BL0272, Bifidobacterium longum; BL00352, Bacillus licheniformis; Bsu2876, Bacillus subtilis; CAC1346, Clostridium acetobutylicum; b0062, Escherichia coli; GK1904, Geobacillus kaustophilus; AAD45718, Geobacillus stearothermophilus; lp3554, Lactobacillus plantarum; OB2797, Oceanobacillus iheyensis; MSMEG, Mycobacterium smegmatis; SMb21420, Sinorhizobium meliloti; STM0102, Salmonella enterica serovar Typhimurium; TM0276, Thermotoga maritima; AAK18729, Thermotoga neapolitana; YPO2253, Yersinia pestis. Amino acid sequences were aligned with Vector NTI AlignX software (Suite 9.0.0; Invitrogen, Carlsbad, Calif.). Phylogenetic trees were constructed with the neighbor-joining method (32) and the minimum evolution method (31), both from MEGA software version 3.0 (21). The p-distance and Poisson correction substitution models were used in both tree-building methods. Bootstrap values were calculated based on 1,000 replicates of the data (11).

In summary, we have isolated, overexpressed, and characterized an acidothermophilic AI from the acidophilic bacterium A. acidocaldarius. Using site-directed mutagenesis techniques, comparative studies with the wild-type and mutant AIs were also performed to identify the role of Lys-269 of AAAI in determining pH optimum. From this study, we suggest that the amino acid residue at positions corresponding to Lys-269 of AAAI may play a significant role in the pH optimum of homologous AIs, possibly reflecting molecular evolution for environmental adaptation.