Characterization of a Novel β-l-Arabinofuranosidase in Bifidobacterium longum

Background: β-l-Arabinofuranosyl linkages are found in many plant biopolymers, but the degradation enzyme has never been found. Results: A novel β-l-arabinofuranosidase was found in Bifidobacterium longum. Conclusion: β-l-Arabinofuranosidase plays a key role in Bifidobacterium longum for β-l-arabinooligosaccharide usage. Significance: The members of the DUF1680 family might be used for the degradation of plant biopolymers. Pfam DUF1680 (PF07944) is an uncharacterized protein family conserved in many species of bacteria, actinomycetes, fungi, and plants. Previously, we cloned and characterized the hypBA2 gene as a β-l-arabinobiosidase in Bifidobacterium longum JCM 1217. In this study, we cloned a DUF1680 family member, the hypBA1 gene, which constitutes a gene cluster with hypBA2. HypBA1 is a novel β-l-arabinofuranosidase that liberates l-arabinose from the l-arabinofuranose (Araf)-β1,2-Araf disaccharide. HypBA1 also transglycosylates 1-alkanols with retention of the anomeric configuration. Mutagenesis and azide rescue experiments indicated that Glu-338 is a critical residue for catalytic activity. This study provides the first characterization of a DUF1680 family member, which defines a new family of glycoside hydrolases, the glycoside hydrolase family 127.

Recently, we cloned a hypBA2 gene that encodes a novel ␤-Larabinobiosidase from Bifidobacterium longum JCM 1217 on the basis of the sequence of BL0421 from B. longum NCC2705, which belongs to the glycoside hydrolase (GH) family 121 (10). The enzyme releases Araf-␤1,2-Araf disaccharide (␤-Ara 2 ) from Araf-␤1,2-Araf-␤1,2-Araf␤-Hyp (Ara 3 -Hyp). Because released ␤-Ara 2 should be hydrolyzed by its own enzyme for assimilation, we predicted that B. longum has a gene encoding ␤-L-arabinofuranosidase. BL0422 is part of a gene cluster with BL0421 and BL0420 and contains a domain of unknown function (DUF) 1680 family in the Pfam database (PF07944), which is a large family annotated as putative glycosyl hydrolases of unknown function.
In this study, we cloned the gene of a BL0422 ortholog from B. longum JCM 1217 and characterized the recombinant protein as a novel ␤-L-arabinofuranosidase. This is the first report of the characterization of a DUF1680 family member.
Expression and Purification of Recombinant HypBA1-The genomic DNA of B. longum JCM 1217 was extracted using a FastPure DNA kit (Takara) and then used for PCR amplification of the gene for the BL0422 ortholog, hypBA1. The forward (5Ј-AAGGAGATATACATATGAACGTTACAATCACTT-CCC-3Ј) and reverse (5Ј-TGCTCGAGTGCGGCCGCTCGA-CGCTGGAAGACA-3Ј) primers were designed from nucleotides 4 -22 and 1959 -1974, respectively, of BL0422 from B. longum NCC2705 to generate a C-terminal His 6 -tagged recombinant protein. The PCR amplification product of hypBA1 was cloned into the pET-23b vector (Novagen) with the In-Fusion Advantage PCR cloning kit (Clontech). The full-length hypBA1 gene was sequenced on an ABI 3100 DNA sequencer with a Big-Dye Terminator 3.1 cycle sequencing kit (Applied Biosystems). The resulting pET23b-hypBA1 plasmid was transformed into Escherichia coli BL21 (DE3) cells, which were then grown at 20°C by using the Overnight Express Autoinduction System (Novagen). Subsequently, the cell cultures were centrifuged, and the resultant pellet was resuspended in Bug-Buster protein extraction reagent (Novagen). The His-tagged proteins were purified on TALON metal affinity resin (Clontech), desalted by dialysis with a cellulose membrane (Wako), and concentrated using a 10-kDa ultrafiltration membrane (Millipore).
The pH dependence of enzyme activity was determined between pH 3.5 and 8.0 by using the following buffers: 50 mM sodium acetate (pH 3.5-6.0), 50 mM MES (pH 5.5-7.0), and 50 mM sodium phosphate (pH 6.5-8.0). The effect of temperature on enzyme activity was examined using 50 mM sodium acetate buffer (pH 4.5) at 15-50°C.
Kinetic Analysis-The kinetic parameters of HypBA1 were determined using 0 -750 M ␤-Ara 2 , cis-Ara 2 -Hyp-DNS, and cis-Ara-Hyp-DNS as the substrates. In the case of ␤-Ara 2 , the 40-l reaction mixture was incubated at 37°C for 10 min and then stopped by adding 10 l of 500 mM NaOH. The amount of liberated L-arabinose was quantified by HPAEC-PAD as described above, using an L-arabinose standard. In the case of cis-Ara-Hyp-DNS, liberated cis-Hyp-DNS were analyzed according to the same procedure used for cis-Ara 2 -Hyp-DNS.
Transglycosylation of Ara 2 -Hyp-The transglycosylation reactions were performed using Ara 2 -Hyp as a donor and 1-alkanols as acceptors. Thirty nanomoles of Ara 2 -Hyp was incubated at 37°C for 3 h with 340 milliunits⅐ml Ϫ1 of HypBA1 in 100 l of 50 mM sodium acetate buffer (pH 4.5) with 5 mM TCEP and 20% methanol, ethanol, or 1-propanol as an acceptor. Subsequently, the reaction products were analyzed by TLC with a 2:1:1 solvent mixture of ethyl acetate/acetic acid/water (v/v/v). The sugars were visualized by spraying an orcinol-sulfate reagent onto the plate (12). For structural analysis, the transglycosylation product from the reaction in 20% methanol was purified by HPLC on a Cosmosil Sugar-D (4.6 ϫ 250 mm, Nacalai) column at 30°C with a mobile phase of acetonitrile and water (75:25, v/v) and a constant flow rate (1.0 ml⅐min Ϫ1 ). The elution was monitored by a refractive index detector (RI-8022, TOSOH), and the fraction that contained the transglycosylation product was collected. 1 H and 13 C NMR spectra were measured with a JMM-ECA600KS spectrometer (JEOL).
Site-directed Mutagenesis and Chemical Rescue-The QuikChange site-directed mutagenesis kit (Stratagene) was used to introduce amino acid substitutions into HypBA1 by using the primers shown in Table 1. After confirmation of the desired mutations by DNA sequencing, these mutant enzymes were expressed and purified according to the same procedure used for the wild-type enzyme. The effect of external nucleophile of the E338A mutant was investigated by adding 0 -400 mM of sodium azide in 40 l of 50 mM sodium acetate buffer (pH 4.5), 7.5 g of E338A mutant, and 25 M cis-Ara 2 -Hyp-DNS as a substrate. After incubating at 37°C for 1 h, the reaction was stopped by adding 10 l of 5% trichloroacetic acid and then analyzed by HPLC as described above.
Bacterial  (13) containing 0.25% ␤-Ara 2 , glucose, or L-arabinose. The bacteria were cultured for 3 days at 37°C under anaerobic conditions. The bacterial growth was judged from the decreased pH of the culture solution (14).
Assays of Bacterial Enzyme Activities-The cell cultures were centrifuged at 17,000 ϫ g for 20 min, and the resultant pellets were washed with 50 mM Tris-HCl buffer (pH 6.8). Afterward, they were resuspended in 50 mM Tris-HCl buffer (pH 6.8) and sonicated with a Sonifier 250 (Branson). The cell lysates were incubated with 25 M cis-Ara 2 -Hyp-DNS for 16 h at 37°C and then analyzed by HPLC.

RESULTS
Expression and Purification of HypBA1-HypBA1 consisted of 658 amino acid residues exhibiting 98.9% identity with that of BL0422 and coincided with that of BLLJ_0211 from B. longum JCM 1217, for which the complete genome sequence is available (15). The recombinant HypBA1 protein was expressed at 20°C as a soluble protein. SDS-PAGE showed that the purified recombinant HypBA1 protein migrated as a single band with an apparent molecular mass of 74 kDa (Fig. 2), which was in agreement with its calculated molecular mass of 74,329 Da. The final yield of the purified enzyme was 140 mg/liter of culture.

TABLE 1 The primers used for site-directed mutagenesis
The positions of the mutated sequences are underlined.

Name
Sequence of oligonucleotide primers specific activity of the purified enzyme was 2.1 units⅐mg Ϫ1 protein. The kinetic parameters for ␤-Ara 2 , cis-Ara 2 -Hyp-DNS, and cis-Ara-Hyp-DNS are summarized in Table 2. The K m and k cat values for ␤-Ara 2 and cis-Ara 2 -Hyp-DNS were within the same range, but the k cat value for cis-Ara-Hyp-DNS was 480fold lower than that of cis-Ara 2 -Hyp-DNS. Consequently, the k cat /K m ratio of cis-Ara-Hyp-DNS was 670-fold lower than that of cis-Ara 2 -Hyp-DNS. HPAEC-PAD analysis showed that L-ar-abinose was released from Ara 2 -Hyp and then the liberated Ara-Hyp gradually hydrolyzed to L-arabinose and Hyp (Fig.  5A). Likewise, both cis-and trans-Ara 2 -Hyp-DNS also hydrolyzed to Ara-Hyp-DNS, which then hydrolyzed to Hyp-DNS (Fig. 7). Under the conditions in which Ara 3 -Hyp could be degraded by HypBA2 and HypBA1 (Fig. 8A, lane 4), the reactivities for the glycoproteins were tested. Liberated sugars were detected from carrot extensin and potato lectin by HypBA2 but not by HypBA1 (Fig. 8). Furthermore, HypBA1 did not act on pNP-galacto-, gluco-, and xylo-pyranosides. The substrate specificity is summarized in Table 3. These results suggested that HypBA1 reacts with the liberated ␤-L-arabinooligosaccharides. Consequently, we classified the enzyme as an exo-acting ␤-L-arabinofuranosidase. The cleavage sites for HypBA1 are shown in Fig. 1.
Transglycosylation Activity of HypBA1-When 1-alkanols were used as the acceptors, the transglycosylation products were detected on TLC (Fig. 9A). The purified transglycosylation product (methyl L-arabinofuranoside) was hydrolyzed to L-arabinose by the HypBA1 treatment (Fig. 9B), which indicates that the methanol was linked by the ␤-anomeric form. The structure of this product was determined by 1 H and 13 C NMR ( Fig. 10 and Table 4). The 1 H NMR spectrum showed the anomeric proton as a doublet at 4.74 ppm with coupling constant J 1,2 ϭ 4.8 Hz. Furthermore, the 13 C NMR spectra revealed that the transglycosylation product was found to be consistent with   a methyl ␤-L-arabinofuranoside (Ara-Me) (16). These data indicated that HypBA1 is a retaining enzyme. Sequence Analysis of HypBA1-HypBA1 consisted of 658 amino acids that included DUF1680 without other sequence motifs (Fig. 11). HypBA1 was 38 -98% identical to other DUF1680 members from bifidobacteria ( Fig. 11 and supplemental Fig. S1). Duplicated DUF1680 members were found in the sequences of almost all Bifidobacterium species. HypBA1 (BLLJ_0211) constitutes a gene cluster with HypBA2 (BLLJ_0212) and a GH43 family member (BLLJ_0213) (Fig. 11). The gene cluster was conserved in B. longum NCC2705, B. longum subsp. infantis 157F, and Bifidobacterium pseudocatenulatum DSM 20438. In addition, the gene cluster without the GH43 family member was conserved in Bifidobacterium catenulatum DSM 16992 and Bifidobacterium dentium ATCC 27678.
Critical Amino Acid Residues of HypBA1-The candidate acidic amino acid residues were selected for site-directed mutagenesis studies based on multiple alignments and the HMM logo of the DUF1680 family in the Pfam database (17). Alanine substitutions were introduced at the positions of Glu-322, Glu-338, and Glu-366, which are highly conserved among the HypBA1 homologs (indicated as asterisks in supplemental Fig. S1). The mutant enzymes were purified for the determination of specific activities. The E322A and E338A mutant enzymes were recovered in the soluble fractions with Bug-Buster. The E338A mutant enzyme exhibited a significant decrease in activity (0.0013%), and the E322A mutant showed 1.5% of the activity relative to the wild-type enzyme ( Table 5). The E366A mutant enzyme was insoluble, and only a small amount of protein was recovered. Nonetheless, it exhibited 16% relative activity compared with the wild-type enzyme. The effect of external nucleophile on the activity of the E338A mutant was further investigated by using different concentrations of sodium azide. The activity of the mutant was rescued by the addition of azide (Fig. 12). In the presence of 200 mM sodium azide, the enzymatic activity was 33-fold greater than in the absence of external nucleophile. We also confirmed azide rescue by ␤-Ara 2 as a substrate, whereas the glycosyl azide product was not observed on HPAEC-PAD and TLC (data not shown).
In Vitro Fermentability of ␤-Ara 2 by B. longum-First, lysates of bifidobacterial cells grown in GAM were used as the enzyme     (Fig. 13A). Moreover, enzymatic activity was not observed in the culture medium or in the bacterial cell suspensions for all Bifidobacterium strains described above (data not shown). The PYF medium containing 0.25% ␤-Ara 2 was utilized as a carbohydrate source by B. longum JCM Potato lectin was reacted without enzyme (2a), with HypBA2-C⌬486 (2b), or with HypBA1 (2c). Extensin was reacted without enzyme (3a), with HypBA2-C⌬486 (3b) or with HypBA1 (3c).

TABLE 3 Substrate specificity of HypBA1
Substrates Activity    Table 4.
1217 but not by B. adolescentis JCM 1275 (Table 6). In addition, ␤-Ara 2 in the PYF medium was utilized by the fermentation of B. longum JCM 1217 (Fig. 13B). Furthermore, ␤-L-arabinofuranosidase activity was found in the cell lysate of B. longum JCM 1217 grown on ␤-Ara 2 but not in the lysate of cultures grown in media containing glucose and L-arabinose (Fig. 13C). These data suggested that ␤-Ara 2 is metabolized by ␤-L-arabinofuranosidase in B. longum.

DISCUSSION
The DUF1680 family has 597 members distributed among 315 species of enteric bacteria (i.e. Bifidobacterium, Bacteroides, Salmonella, Clostridium, and Escherichia), plant-pathogenic Xanthomonas, actinomycetes, fungi, and plants, as shown in the Pfam database. The members of this family are hypothetical proteins of unknown function and have no sequence similarity with other glycoside hydrolase families. In this study, we cloned the gene encoding a member of the DUF1680 family and characterized its product as a novel ␤-L-arabinofuranosidase.
Therefore, we propose that the enzyme be assigned to a new family of glycoside hydrolases, the glycoside hydrolase family 127.   ␤-Ara 2 was a suitable substrate for HypBA1 as well as Ara 3 -Hyp and Ara 2 -Hyp, which contain the Araf-␤1,2-Araf structure at the nonreducing terminal. In extensins, ␤-L-arabinooligosaccharides are in close existence on repetitive Ser-Hyp 4 motifs and contribute to protease resistance. It is thought that Hyplinked ␤-L-arabinooligosaccharides do not occur naturally in the normal environment. Furthermore, HypBA1 did not directly release L-arabinose from extensin or potato lectin (Fig.  8). In addition, we showed that ␤-Ara 2 was used as a carbohydrate source for B. longum, with enzymatic activity detected in the cell lysate (Table 6 and Fig. 13). Interestingly, the enzymatic activity was not detected in cells grown in the presence of L-arabinose or glucose. The amino acid sequence of HypBA1 lacks both a secretory signal and a transmembrane domain. Collectively, these results indicate that HypBA1 is an intracellular enzyme that degrades HypBA2-released ␤-Ara 2 , as schematically summarized in Figs. 14 and 15.
Recently, Fukuda et al. (15) reported that B. longum has an advanced ability for fructose uptake and acetate production, with the released acetate improving the intestinal defense mediated by epithelial cells. In addition to fructose, L-arabinose is a naturally found common carbohydrate and is found as a component of biopolymers such as hemicellulose and pectin. B. longum JCM 1217 encodes a number of candidates for the ␣-L-arabinofuranosidase gene, 11 members of the GH43 gene family, and 4 members of the GH51 gene family. Several reports indicate that B. longum has the ability to grow on L-arabinose and ␣-L-arabinooligosaccharides (14,23,(25)(26)(27). We showed that B. longum also uses ␤-Ara 2 as a carbohydrate source (Table  6). Several ␣and ␤-L-arabinooligosaccharide degradation enzymes in B. longum might be involved in L-arabinose acquisition from plant polymers in the large intestine.
HypBA1 was identified as a retaining glycoside hydrolase, as described above. Hydrolysis by retaining glycoside hydrolases proceeds through a double-displacement mechanism with two catalytic residues. The catalytic residues typically utilized are either aspartate or glutamate residues. In the chemical rescue study, E338A mutant was rescued by the addition of azide, which suggests that Glu-338 is a catalytic residue for HypBA1. However, no glycosyl azide product was formed in the reaction mixture. A water molecule activated by azide ion might be reactivated E338A mutant without glycosyl azide production, as shown in GH43 ␤-xylosidase and GH14 ␤-amylase (28,29).