- Email: [email protected]
Novel bifidobacterial glycosidases acting on sugar chains of mucin glycoproteins
JOURNAL OF BIOSCIENCE AND BIOENGINEERING Vol. 99, No. 5, 457–465. 2005 DOI: 10.1263/jbb.99.457
© 2005, The Society for Biotechnology, Japan
Novel Bifidobacterial Glycosidases Acting on Sugar Chains of Mucin Glycoproteins TAKANE KATAYAMA,1 KIYOTAKA FUJITA,1 AND KENJI YAMAMOTO1* Graduate School of Biostudies, Kyoto University, Oiwake-cho, Kitashirakawa, Sakyo-ku, Kyoto 606-8502, Japan1 Received 22 February 2005/Accepted 9 March 2005
Bifidobacterium bifidum was found to produce a specific 1,2-a-L-fucosidase. Its gene (afcA) has been cloned and the DNA sequence was determined. The AfcA protein consisting of 1959 amino acid residues with a predicted molecular mass of 205 kDa can be divided into three domains; the N-terminal function-unknown domain (576 aa), the catalytic domain (898 aa), and the C-terminal bacterial Ig-like domain (485 aa). The recombinant catalytic domain specifically hydrolyzed the terminal a-(1®2)-fucosidic linkages of various oligosaccharides and sugar chains of glycoproteins. The primary structure of the catalytic domain exhibited no similarity to those of any glycoside hydrolases but showed similarity to those of several hypothetical proteins in a database, which resulted in establishment of a novel glycoside hydrolase family (GH family 95). Several bifidobacteria were found to produce a specific endo-a-N-acetylgalactosaminidase, which is the endoglycosidase liberating the O-glycosidically linked galactosyl b1®3 N-acetylgalactosamine disaccharide from mucin glycoprotein. The molecular cloning of endo-a-N-acetylgalactosaminidase was carried out on Bifidobacterium longum based on the information in the database. The gene was found to comprise 1966 amino acid residues with a predicted molecular mass of 210 kDa. The recombinant protein released galactosyl b1®3 N-acetylgalactosamine disaccharide from natural glycoproteins. This enzyme of B. longum is believed to be involved in the catabolism of oligosaccharide of intestinal mucin glycoproteins. Both 1,2-a-L-fucosidase and endo-a-N-acetylgalactosaminidase are novel and specific enzymes acting on oligosaccharides that exist mainly in mucin glycoproteins. Thus, it is reasonable to conclude that bifidobacteria produce these enzymes to preferentially utilize the oligosaccharides present in the intestinal ecosystem. [Key words: bifidobacteria, Bifidobacterium bifidum, Bifidobacterium longum, 1,2-a-L-fucosidase, endo-a-N-acetylgalactosaminidase, mucin, O-linked oligosaccharide]
survive in the lower intestinal tract, bifidobacteria produce various kinds of exo- and endoglycosidases in surfacebound and/or extracellular forms, by which they can utilize diverse carbohydrates (6, 7). Recent genome sequence analysis of Bifidobacterium longum NCC2705 revealed that more than 8.5% of the total predicted proteins were involved in the degradation of oligo- and polysaccharides, perhaps reflecting the superior ability of this organism to adapt to its environment (8). The epithelial cells of the human intestine express and/or secrete mucin glycoproteins (9), which are assumed to play an important role in protecting enterocytes from chemical and physical damage as well as in preventing invasion by pathogens. Intestinal mucin glycoproteins contain a lot of O-linked oligosaccharides, on the non-reducing ends of which a-linked L-fucosyl residues are frequently found (10, 11). Such a-L-fucosyl residues are also present in glycolipids on the cell surface, submaxillary mucin, blood group substances, and oligosaccharides in human milk (11, 12).
The human gastrointestinal tract is inhabited by a vast and diverse community of microbes, and a well-balanced microflora is thought to be important for normal digestion and maintenance of the intestinal ecosystem (1). In the complex intestinal microflora, bifidobacteria, which are grampositive obligate anaerobes, are considered to be key commensals that promote a healthy intestinal tract because of their many beneficial effects on the host, such as regulation of the state of the intestine, reduction of harmful bacteria and toxic compounds, immuno modulation, and anticarcinogenic activity (2–5). Therefore, bifidobacteria have attracted a great deal of attention. Bifidobacteria naturally colonize the lower intestinal tract, an environment poor in mono- and disaccharides since such sugars are preferentially consumed by the host and microbes present in the upper intestinal tract. Therefore, in order to * Corresponding author. e-mail: [email protected] phone: +81-(0)75-753-6277 fax: +81-(0)75-753-6275 457
458
KATAYAMA ET AL.
Taking into consideration that several fecal mucin-degrading bacteria such as bifidobacteria, clostridia, and bacteroides are known to produce a-L-fucosidase (13–15), it was envisaged that a-L-fucosidase might be responsible for the selective colonization of the intestine by these bacteria. Meanwhile, intestinal mucin glycoproteins possess an abundant amount of O-glycosidic oligosaccharides linked glycosidically to the hydroxyl groups of serine or threonine residues of the protein (16). These oligosaccharides are also found sparsely in soluble glycoproteins. They are characterized by a core structure consisting of a galactosyl b1®3 N-acetylgalactosamine a1-serine/threonine residue, and are termed mucin type oligosaccharides. Endo-a-N-acetylgalactosaminidase (endo-a-GalNAc-ase; EC 3.2.1.97) is an endoglycosidase that catalyzes the hydrolysis of the O-glycosidic a-linkage between galactosyl b1®3 N-acetylgalactosamine (Galb1-3GalNAc) and a serine or threonine residue in mucin glycoproteins of various animal sources (17). Recently, we isolated and identified some bifidobacteriaspecific enzymes that are involved in the degradation of sugar chains of intestinal mucin. We also revealed that these enzymes belong to novel glycoside hydrolase families. In this review, we describe the molecular cloning and characterization of 1,2-a-L-fucosidase (EC 3.2.1.63) and endoa-GalNAc-ase from bifidobacteria, an intestinal colonizer strain, in order to obtain a better understanding of the catabolism of sugars in bifidobacteria and to examine the roles of sugar chain-degrading enzymes in intestinal bacteria. These unique enzymes are not normally found in microorganisms. I. MOLECULAR CLONING OF 1,2-a-L-FUCOSIDASE FROM BIFIDOBACTERIUM BIFIDUM JCM1254 To date, a-L-fucosidases that liberate terminal a-linked from the oligosaccharides of various glycoconjugates including mucin glycoprotein have been purified from several prokaryotic and eukaryotic sources (14, 15, 18–23), and the enzymes have been divided into two groups (19); one capable of hydrolyzing various types of fucosidic linkages as well as synthetic substrates, and another that is only active on the a-(1®2)-linkage, although a few reports have described enzymes that cannot be classified into either group (20, 21). In contrast to the dozens of studies on the purification of such enzymes, there is a paucity of reports regarding the cloning of a-L-fucosidase genes, and they are all derived from eukaryotic cells such as human and rat livers (24, 25), Canis familiaris (26), and Dictyostelium discoideum (27), which constitute glycoside hydrolase family 29 (GH family 29). To the best of our knowledge, there have been no reports regarding the isolation of a bacterial a-L-fucosidase gene, although the sequence of the 1,3-/4a-L-fucosidase gene from Streptomyces sp. has been deposited in GenBank (accession no. U39394). We have found that a few bifidobacteria strains produced 1,2-a-L-fucosidase (EC 3.2.1.63) in cell surface-bound and/or extracellular forms, using 2¢-fucosyllactose as a substrate. Among these bacteria, we chose B. bifidum JCM1254 and attempted to clone the 1,2-a-L-fucosidase gene (designated as afcA). The expression cloning was done using Escherichia coli L-fucose
J. BIOSCI. BIOENG.,
DH5a, a non-fucosidase-producing bacterium. A genomic library of B. bifidum JCM1254 constructed in E. coli DH5a was screened for the ability to hydrolyze the a-(1®2)-linkage of 2¢-fucosyllactose. One recombinant, designated SA3, was selected. Sequence analysis of plasmid pSA3 revealed that the cloned gene contained two large truncated open reading frames (designated as ORF1 and ORF2) (Fig. 1). The sense strands of ORF1 and ORF2 overlapped to a large extent in reverse, which is not surprising since such cases are sometimes found in the recently determined genome sequence of B. longum NCC2705 (8). In order to determine which ORF actually encodes 1,2-a-L-fucosidase, each ORF (the MfeI-AflII fragment) was placed under the control of the lac promoter and its expression was induced by the addition of IPTG. While no increase in 1,2-a-L-fucosidase activity was observed when ORF2 was induced, the activity was significantly elevated when ORF1 was induced, indicating that ORF1 encodes 1,2-a-L-fucosidase. The codon usage of the afcA gene was quite similar to that of other genes of B. bifidum in the database (Codon Usage Database in Kazusa DNA Research Institute), and Southern hybridization analysis with the 2.3-kb KpnI fragment as a specific probe revealed that the afcA gene exists as asingle-copy on the genome of B. bifidum JCM1254. Analysis of the primary structure by using the SignalP (28) and PSORT (29) programs revealed the presence of a signal peptide and a membrane anchor at the N-terminus and C-terminus, respectively (Fig. 1). A possible ribosome-binding site was located 6-bp from a probable initiation codon and a promoter-like sequence was also found in the upstream region. The AfcA protein consists of 1959 amino acid residues with a calculated molecular mass of 205 kDa (GenBank accession no. AY303700). II. DOMAIN STRUCTURE OF 1,2-a-L-FUCOSIDASE FROM B. BIFIDUM JCM1254 Since the primary structure of AfcA protein did not exhibit any similarity to those of known glycosidase families, we attempted to localize a catalytic domain that is essential for the hydrolysis of 2¢-fucosyllactose. The N-terminal and C-terminal deletion mutants were constructed and expressed, and then their activity was assessed. Consequently, the 576 N-terminal amino acid residues (911–2638 bases) and 485 C-terminal amino acid residues (5333–6790 bases) were found to be removable without loss of fucosidase activity. These results revealed that the region consisting of the 577th–1474th amino acid residues constitutes the catalytic domain (Fig. 1). The region consisting of the 1475th–1728th amino acid residues contained four repetitive sequences with immunoglobulin-like folds, the so-called bacterial Iglike domain B (Pfam 02368). Although the function of this domain is not clear, it is highly likely that this domain, 254 amino acids in length, at least acts to display the fucosidase domain of AfcA so that it protrudes from the cell surface, thereby enabling B. bifidum JCM1254 cells to gain access to and degrade the fucosyl residues present on the glycoconjugates of enterocytes. Neither sequence similarity to other ORFs nor a functional motif was found in the sequence of
VOL. 99, 2005
BIFIDOBACTERIAL GLYCOSIDASES ACTING ON SUGAR CHAINS
459
FIG. 1. Gene organization of the afcA locus of B. bifidum JCM1254 and schematic representation of the domain structure of AfcA. The top thin line is a restriction map, indicating the recognition sites of AflII (A), BamHI (B), EcoRI (E), KpnI (K), MfeI (Mf), and MluI (Ml). The thick lines indicate the lengths of the inserts of the plasmids, and the dashed line indicates the specific probe (1808–2888 bp) used for cloning of the upstream region of ORF1 (1–2888 bp, pSA19). The deduced amino acid sequence of ORF3 showed high similarity to the consensus sequence of Pfam 03577, thus it probably encodes a dipeptidase. The amino acid numbering of the schematic representation of AfcA starts at the probable initiation codon, and the domain responsible for 1,2-a-L-fucosidase activity (577–1474 aa) is depicted as a gray box. The region with Ig-like folds (1475–1728 aa) is shown by a stippled box. The black bars at the N-terminal and C-terminal ends indicate a signal peptide and a membrane anchor, respectively.
the N-terminal domain (1st–576th amino acid residues). III. CHARACTERIZATION OF FUCOSIDASE DOMAIN We attempted to overexpress and purify the AfcA protein, however, an intact form of the protein was not obtained, most likely due to the extremely long length of the protein. Therefore, the catalytic domain (Fuc domain) was expressed and purified as a C-terminal hexahistidine-tagged protein. The ability of the Fuc domain to hydrolyze the a-(1®2)-L-fucosidic linkage of 2¢-fucosyllactose was confirmed by ESI-MS and NMR analyses. The substrate specificity was examined using naturally occurring substrates. As shown in Table 1, of the milk oligosaccharides examined, the most readily hydrolyzed substrates were 2¢-fucosyllactose and lacto-N-fucopentaose I, both of which contain L-fucose bound to galactose through an a-(1®2)-linkage at the non-reducing termini. The enzyme exhibited very limited activity for the a-(1®3)-linked L-fucosyl residues of 3-fucosyllactose and lacto-N-fucopentaose V, and apparently had no action at all on the a-(1®4)-linkage of lacto-N-fucopentaose II. The enzyme could not hydrolyze 3-fucosylgalactose (Fuca1®3Gal) or 4-fucosyl-N-acetylglucosamine (Fuca1®4GlcNAc). Blood group H (II) active substance (Fuca1®2Galb1®4GlcNAc) was found to be a good substrate, however, the enzyme had very limited action on blood groups A and B active substances, which have a-(1®3)GalNAc and a-(1®3)-Gal residues, respectively, in addition to a-(1®2)-Fuc residues at the non-reducing termini. The a-(1®6)-linkage of 6-fucosyl-N,N¢-diacetylchitobiose was not cleaved by this enzyme. In addition to fucosecontaining oligosaccharides, glycoproteins with terminal
a-(1®2)-linked L-fucosyl residues, e.g., porcine gastric mucin (H), were also found to be substrates of the Fuc domain. On the other hand, the Fuc domain did not liberate fucose from any of the three artificial substrates examined (pNP-a-L-fucoside, pNP-b-L-fucoside, and 4-methylumbelliferyl-a-L-fucoside). The catalytic domain showed the maximum activity at pH 5, suggesting the involvement of acidic residue(s) in the catalysis, and was stable below 35°C for 30 min and at pH 6.5–7.5 for 12 h. The Km and Vmax values for 2¢-fucosyllactose were determined to be 0.53 mM and 1.6 mmol/min/mg, respectively. IV. STRUCTURAL SIMILARITY OF FUCOSIDASE DOMAIN TO HYPOTHETICAL PROTEINS IN A DATABASE The primary structure of the Fuc domain was subjected to a BLAST search, but no significant similarity was found with proteins with defined functions (30). Scrutiny of the AfcA sequence in reference to conserved motifs of family GH29 did not lead to the finding of any trails of the family. In a BLAST search, high scores were obtained with several hypothetical proteins, the amino acid sequences of which are compared and the highly homologous regions are shown in Fig. 2. Of particular interest is the finding that the Fuc domain is homologous to CPE1875 in the genome sequence of Clostridium perfringens strain 13 (31) since it has been shown that C. perfringens produces 1,2-a-L-fucosidase with a high molecular weight (ca 200 kDa) in the culture medium (14). The deduced amino acid sequence of CPE1875 has a typical signal sequence at the N-terminus (27), and its molecular weight is estimated to be 165 kDa. Thus, CPE1875 might encode 1,2-a-L-fucosidase. Other homologous pro-
460
J. BIOSCI. BIOENG.,
KATAYAMA ET AL.
TABLE 1.
Substrate specificity of the Fuc domain of AfcA from B. bifidum JCM1254
a The abbreviations used are: Gal, D-galactose; Glc, D-glucose; Fuc, L-fucose; GlcNAc, N-acetyl-D-glucosamine; and GalNAc, N-acetyl-Dgalactosamine. b The reaction mixture consisting of 100 mM sodium phosphate (pH 6.5), 2 mM substrate, and 20 mU of the purified Fuc domain, in a total volume of 100 ml, was incubated for 1 min–12 h at 30°C. The value obtained with 2'-fucosyllactose was taken as 100%. ND, Not detectable.
teins were found from the genome sequences of Streptococcus pneumoniae R6 (32), Bacteroides thetaiotaomicron VPI-5482 (33), Microbulbifer degradans 2-40, Bacillus halodurans C-125 (34), Xanthomonas campestris pv. campestris ATCC33913 (35), and Arabidopsis thaliana. Although it remains to be elucidated whether these homologues indeed show 1,2-a-L-fucosidase activity, the finding of a new conservative sequence indicates the existence of a novel glycoside hydrolase family (GH family 95, personal communication with Dr. B. Henrissat, Universites de Marseille, France). Although B. bifidum cannot ferment L-fucose, the presence of a-L-fucosidase does benefit the cells because, while B. bifidum cells are capable of degrading and fermenting 2¢-fucosyllactose, none of the bifidobacterial strains that do not secrete fucosidase could degrade 2¢-fucosyllactose without the aid of exogenously added fucosidase (purified Fuc
domain). These results suggest that, in preference to nonfucosidase-producing bacteria, B. bifidum are able to degrade several types of substrates present in the intestinal epithelium and mucosa, and imply the biological importance of a secretory fucosidase from gut microbes for the intestinal ecosystem. V. ISOLATION AND IDENTIFICATION OF ENDO-a-N-ACETYLGALACTOSAMINIDASE FROM B. LONGUM JCM1217 Endo-a-N-acetylgalactosaminidase (endo-a-GalNAc-ase; EC 3.2.1.97) catalyzes the hydrolysis of the O-glycosidic a-linkage between galactosyl b1-3 N-acetylgalactosamine (Galb1-3GalNAc) and a serine or threonine residue in mucins and mucin-type glycoproteins from various animal
VOL. 99, 2005
BIFIDOBACTERIAL GLYCOSIDASES ACTING ON SUGAR CHAINS
461
FIG. 2. Comparison of the amino acid sequences of the Fuc domain of AfcA with homologous proteins in the database. The primary structure of the Fuc domain was aligned with those of homologues using the Clustal W program, and highly homologous regions are shown. White letters on a black background indicate identical residues in more than four sequences, and conservative amino acid changes are shown in gray. B. bif, AfcA of B. bifidum JCM1254; S. pne, Streptococcus pneumoniae R6 (GenBank accession no. NP_359091); C. per, Clostridium perfringens 13 (GenBank accession no. NP_562791); M. deg, Microbulbifer degradans 2–40 (GenBank accession no. ZP_00064866); A. tha, Arabidopsis thaliana (GenBank accession no. AY125494); B. hal, Bacillus halodurans B-125 (GenBank accession no. NP_241708); X. axo, Xanthomonas axonopodis pv. campestris str. ATCC 33913 (GenBank accession no. NP_637123); and B. the, Bacteroides thetaiotaomicron VPI-5482 (GenBank accession no. NP_809923).
sources. The enzyme was first found from the culture fluid of C. perfringens type 33-4A (36), and then purified from the culture fluid of Diplococcus pneumoniae (now assigned as S. pneumoniae) type 1 (37, 38). We also purified the enzyme from the culture fluids of Alcaligenes sp. F-1906 (39) and Bacillus sp. A-198 (40) which were isolated from soil and identified. The enzyme has a strict substrate specificity, acting only on the a-linked disaccharide Galb13GalNAc. A similar enzyme was also detected in the culture fluid of Streptomyces sp. OH-11242 which could release longer oligosaccharides than disaccharide from porcine mucin (41). As these endo-a-GalNAc-ases liberate O-linked oligosaccharides from cell surface glycoproteins without causing damage, they are useful for the investigation of the structure and function of O-linked oligosaccharides on the cell surface (17). Recently, we found endo-a-GalNAc-ase activity in the culture fluids of several bifidobacteria. Using Galb13GalNAca1-p-nitrophenyl as a substrate, we determined the endo-a-GalNAc-ase activity in the culture fluids of various strains of bifidobacteria. As shown in Fig. 3, activity was found in the culture fluids of Bifidobacterium breve JCM1192, B. bifidum JCM1254, JCM7004, and ATCC29521, and B. longum JCM1217 and JCM7054. These results sug-
FIG. 3. TLC analysis of the reaction mixtures incubated with pNP -substrate and various Bifidobacteria cells. The cell pellets were incubated with 0.5 mM Galb1,3GalNAca1pNP in 10 mM acetate buffer (pH 5.0) for 2 h at 37°C. Lane 1, Galb1,3GalNAca1pNP; lane 2, Gal; lane 3, GalNAc; lane 4, Galb1,3GalNAc; lane 5, B. breve JCM1192; lane 6, B. infantis JCM1222; lane 7, B. bifidum JCM1254; lane 8, B. bifidum JCM7004; lane 9, B. bifidum ATCC29521; lane 10, B. pseudolongum JCM1205; lane 11, B. longum JCM1217; lane 12, B. longum JCM7054.
gest that the enzyme might be involved in the catabolism of oligosaccharides in bifidobacteria. We next attempted to elucidate the function of this enzyme in the bacteria, and
462
J. BIOSCI. BIOENG.,
KATAYAMA ET AL.
carried out the molecular cloning, expression, and characterization of bifidobacterial endo-a-GalNAc-ase. The above results suggested that bifidobacteria produced and secreted an endo-a-GalNAc-ase into the cell surface or extracellularly, and that they may have a signal peptide and cell-surface associated domains. There are several predicted endoglycosidases such as endo-b-xylanase, endo-b-N-acetylglucosaminidase, endo-a-arabinosidase, and arabinogalactan endo-b-galactosidases, which may be useful for the intake of extracellular carbohydrate polymers, based on information obtained from the genome sequencing database of B. longum NCC2705 (8). Based on the information concerning uncharacterized hypothetical proteins from the genome database, we searched for a hypothetical protein containing a secretion signal and transmembrane domain using BLAST and Pfam HMM searches, and assumed one potential sequence “AAN24297” (locus name, BL0464) to be endoa-GalNAc-ase. Then, using the sequence of BL0464 “narrowly conserved hypothetical protein” from B. longum NCC2705 (8), oligonucleotide primers were designed and a 6.0-kb PCR product was amplified from genomic DNA of B. longum JCM1217, which was confirmed to produce endoa-GalNAc-ase. The upstream and downstream regions of the gene were obtained using colony hybridization and PCR, respectively. The composite DNA encoded an open reading frame of 5901 bp corresponding to a protein of 1966 amino acids. The predicted peptide sequence of endo-a-GalNAcase showed 99.1% identity with BL0464. A region at the Nand C-termini of the enzyme contained a predicted signal peptide (amino acids 1–29) and a transmembrane domain (amino acids 1936–1963), respectively (Fig. 4). The coding sequence of endo-a-GalNAc-ase had no homologous domains of known GH families. Hypothetical proteins with homology to the endo-a-GalNAc-ase of B. longum JCM1217 were observed in the genome sequences of C. perfringens 13, S. pneumoniae R6, Streptomyces coelicolor A3, and Enterococcus faecalis V583 (Fig. 4). Some of these bacteria have been shown to produce endo-a-GalNAc-ase in their culture fluids (36, 37). A conserved region, which was found in the region of amino acids 590–1381 of the enzyme from B. longum JCM1217, is 32% to 54% identical to the corresponding hypothetical proteins. We expressed a fulllength of His-tagged enzyme in E. coli BL21(DE3) using the expression plasmid pET23d, and the soluble cell lysate exhibited endo-a-GalNAc-ase activity. The His-tagged endoa-GalNAc-ase was purified and the obtained enzyme preparation migrated as a single band corresponding to a protein of 210 kDa on SDS–PAGE. VI. VARIOUS PROPERTIES OF RECOMBINANT B. LONGUM JCM1217 ENDO-a-NACETYLGALACTOSAMINIDASE The optimal pH of the recombinant enzyme was around 5.0 at 37°C in 10 mM acetate buffer and the enzyme showed 68% and 73% of the maximum activity at pH 4.0 and 6.0, respectively. The enzyme was stable at temperatures up to 37°C and retained 73% activity at 45°C after incubating for 30 min in 10 mM acetate buffer (pH 5.0). The optimal temperature for the enzyme reaction was 60°C. The kinetic
parameters of the enzyme were determined at 37°C and pH 5.0, and the Km value and kcat value for Galb1®3GalNAca1pNP were 0.165 mM and 37.0 s–1, respectively. We examined the substrate specificity of the recombinant enzyme using synthetic pNP substrates and natural glycoproteins. TLC analysis of the reaction mixtures indicated that Galb13GalNAc disaccharide was liberated from asialofetuin and pNP substrate containing the core 1 structure (Galb1® 3GalNAca1). However, the hydrolysis products could not be detected in the presence of sialofetuin or any pNP substrate other than Galb1®3GalNAca1-pNP. Derensy-Dron et al. characterized b-1,3-galactosyl-Nacetylhexosamine phosphorylase (lacto-N-biose phosphorylase; EC 2.4.1.211) which has the ability to produce a-galactopyranose 1-phosphate (Gal-1-P) and N-acetylglucosamine in cell-free extracts of B. bifidum DSM20082 (42). This enzyme could also phosphorolyze Galb1®3GalNAc which is produced by endo-a-GalNAc-ase to Gal-1-P and N-acetylgalactosamine. Kitaoka et al. found that this enzyme was produced by B. longum JCM1217 as an intracellular enzyme (43). This result suggests the catabolism of the sugar chains of mucin glycoprotein by B. longum JCM1217 using endo-a-GalNAc-ase and lacto-N-biose phosphorylase. We also confirmed endo-a-GalNAc-ase activity in the cells of three B. bifidum strains. As mucin may be major oligosaccharide supplier in the human colon, these data suggest that several bifidobacteria possess a degradation system for mucin oligosaccharides. VII. REMARKS Since bifidobacteria are key commensals that promote a healthy intestinal tract, obtaining a better understanding of their specific metabolism is quite important for their effective and preferential colonization of the intestine. We have focused on the sugar catabolism of bifidobacteria, and have revealed, for the first time, the gene organizations and primary structures of 1,2-a-L-fucosidase and endo-a-N-acetylgalactosaminidase, both of which can act on and degrade naturally-occurring substrates including high-molecular-weight glycoproteins, indicating the physiological significance of these enzymes. In our laboratory, we have already cloned the gene encoding b-D-glucosidase, which has b-D-fucosidase activity, from B. breve clb (44). By using this recombinant enzyme, we synthesized a disaccharide that seemed to be specifically assimilated by bifidobacteria. The mechanism of catabolism of sugar in bifidobacteria seems to be very complicated. The accumulation of data on the genes of various glycosidases is necessary in order to elucidate the specific characteristics of bifidobacteria with respect to the assimilation of sugars. From this point of view, since 1,2-a-L-fucosidase and endo-a-N-acetylgalactosaminidase are unique enzymes that are not usually found in ordinary microorganisms, their utilization for specifically promoting the growth of bifidobacteria is being anticipated. Our results open the way to elucidating the roles of these specific enzymes in the intestinal ecosystem, and moreover, provide a basis upon which we hope new bifidus factors (prebiotics) will be developed in the near future.
VOL. 99, 2005
BIFIDOBACTERIAL GLYCOSIDASES ACTING ON SUGAR CHAINS
463
FIG. 4. Multiple alignment of the amino acid sequences of B. longum JCM1217 endo-a-GalNAc-ase with the homologous proteins. (A) Schematic diagram of endo-a-GalNAc-ase from B. longum JCM1217. FIVAR domain is presumably involved in binding of the protein to cell wall. (B) The gray boxes indicate comparatively conserved regions. (C) Multiple alignment of the amino acid sequences (shown in panel B as gray boxes). Identical residues and conserved substitutions are highlighted in black and dark gray, respectively. Asterisks indicate conserved acidic amino acids. The organisms and accession nos. (in brackets) are as follows: SC, S. coelicolor A3 (CAA20079); EF, E. faecalis V583 (AAO81568); BL1, B. longum JCM1217 (AY836679); BL2, B. longum NCC2705 (AAN24297); SP, S. pneumoniae R6 (AAK99132); and CP, C. perfringens 13 (BAB80399).
464
KATAYAMA ET AL.
ACKNOWLEDGMENTS This work was partly supported by a Grant-in-Aid for Scientific Research from Core Research for Evolutional Science and Technology (CREST) from the Japan Science and Technology Agency (JST), by a Grant-in-Aid for Scientific Research (B) from the Ministry of Education, Culture, Sports, Science and Technology, Japan, and by a Grant-in-Aid for Food Science Research from the Food Science Institute Foundation.
REFERENCES 1. Simon, G. L. and Gorbach, S. L.: The human intestinal microflora. Dig. Dis. Sci., 31, 147S–162S (1986). 2. Hooper, L. V. and Gordon, J. I.: Commensal host-bacterial relationships in the gut. Science, 292, 1115–1118 (2001). 3. Gibson, G. R. and Wang, X.: Regulatory effects of bifidobacteria on the growth of other colonic bacteria. J. Appl. Bacteriol., 77, 412–420 (1994). 4. Ouwehand, A., Isolauri, E., and Salminen, S.: The role of the intestinal microflora for the development of the immune system in early childhood. Eur. J. Nutr., 41, 32–37 (2002). 5. Femia, A. P., Luceri, C., Dolara, P., Giannini, A., Biggeri, A., Salvadori, M., Clune, Y., Collins, K. J., Paglierani, M., and Caderni, G.: Antitumorigenic activity of the prebiotic inulin enriched with oligofructose in combination with the probiotics Lactobacillus rhamnosus and Bifidobacterium lactis on azoxymethane-induced colon carcinogenesis in rats. Carcinogenesis, 23, 1953–1960 (2002). 6. Hoskins, L. C., Agustines, M., McKee, W. B., Boulding, E. T., Kriaris, M., and Niedermeyer, G.: Mucin degradation in human colon ecosystems. Isolation and properties of fecal strains that degrade ABH blood group antigens and oligosaccharides from mucin glycoproteins. J. Clin. Invest., 75, 944–953 (1985). 7. Larson, G., Falk, P., and Hoskins, L. C.: Degradation of human intestinal glycosphingolipids by extracellular glycosidases from mucin-degrading bacteria of the human fecal flora. J. Biol. Chem., 263, 10790–10798 (1988). 8. Schell, M. A., Karmirantzou, M., Snel, B. D., Vilanova, B., Berger, B., Pessi, G., Zwahlen, M. C., Desiere, F., Bork, P., Delley, M., Pridmore, R. D., and Arigoni, F.: The genome sequence of Bifidobacterium longum reflects its adaptation to the human gastrointestinal tract. Proc. Natl. Acad. Sci. USA, 99, 14422–14427 (2002). 9. Gum, J. R., Byrd, J. C., Hicks, J. W., Toribara, N. W., Lamport, D. T., and Kim, Y. S.: Molecular cloning of human intestinal mucin cDNAs. Sequence analysis and evidence for genetic polymorphism. J. Biol. Chem., 264, 6480– 6487 (1989). 10. Podolsky, D. K.: Oligosaccharide structures of human colonic mucin. J. Biol. Chem., 260, 8262–8271 (1985). 11. Slomiany, A., Zdebska, E., and Slomiany, B. L.: Structures of the neutral oligosaccharides isolated from A-active human gastric mucin. J. Biol. Chem., 259, 14743–14749 (1984). 12. Song, J.-F., Weng, M.-Q., Wu, S.-M., and Xia, Q.-C.: Analysis of neutral saccharides in human milk derivatized with 2-aminoacridone by capillary electrophoresis with laser-induced fluorescence detection. Anal. Biochem., 304, 126–129 (2002). 13. Salyers, A. A., West, S. E., Vercellotti, J. R., and Wilkins, T. D.: Fermentation of mucins and plant polysaccharides by anaerobic bacteria from the human colon. Appl. Environ. Microbiol., 34, 529–533 (1977). 14. Aminoff, D. and Furukawa, K.: Enzymes that destroy blood group specificity. I. Purification and properties of a-L-fucosidase from Clostridium perfringens. J. Biol. Chem., 245, 1659– 1669 (1970).
J. BIOSCI. BIOENG.,
15. Berg, J. O., Lindqvist, L., and Nord, C. E.: Purification of glycoside hydrolases from Bacteroides fragilis. Appl. Environ. Microbiol., 40, 40–47 (1980). 16. Marth, J. D.: O-Glycans, p. 101–113. In Varki, A., Cummings, R., Esko, J., Freeze, H., Hart, G., and Marth, J. (ed.), Essentials of glycobiology. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, New York (1999). 17. Kobata, A.: Use of endo- and exoglycosidases for structural studies of glycoconjugates. Anal. Biochem., 100, 1–14 (1979). 18. Yamamoto, K., Li, S.-C., and Li, Y.-T.: Microbial glycosidases, p. 497–511. In Ernst, B., Hart, G., and Sinay, P. (ed.), Carbohydrates in chemistry and biology. Wiley-VCH, New York (2000). 19. Bahl, O. P.: Glycosidases of Aspergillus niger. II. Purification and general properties of 1,2-a-L-fucosidase. J. Biol. Chem., 245, 299–304 (1970). 20. Tsuji, Y., Yamamoto, K., Tochikura, T., Seno, T., Ohkubo, Y., and Yamaguchi, H.: Purification and characterization of a-L-fucosidase from Bacillus circulans grown on porcine gastric mucin. J. Biochem., 107, 324–330 (1990). 21. Eneyskaya, E. V., Kulminskaya, A. A., Kalkkinen, N., Nifantiev, N. E., Arbatskii, N. P., Saenko, A. I., Chepurnaya, O. V., Arutyunyan, A. V., Shabalin, K. A., and Neustroev, K. N.: An a-L-fucosidase from Thermus sp. with unusually broad specificity. Glycoconj. J., 18, 827–834 (2001). 22. Ogata-Arakawa, M., Muramatsu, T., and Kobata, A.: a-LFucosidase from almond emulsin: characterization of the two enzymes with different specificities. Arch. Biochem. Biophys., 181, 353–358 (1977). 23. Khunsook, S., Alhadeff, J. A., and Bean, B. S.: Purification and characterization of human seminal plasma a-L-fucosidase. Mol. Hum. Reprod., 8, 221–227 (2002). 24. Fukushima, H., de Wet, J. R., and O’Brien, J. S.: Molecular cloning of a cDNA for human a-L-fucosidase. Proc. Natl. Acad. Sci. USA, 82, 1262–1265 (1985). 25. Fisher, K. J. and Aronson, N. N.: Isolation and sequence analysis of a cDNA encoding rat liver a-L-fucosidase. Biochem. J., 264, 695–701 (1989). 26. Occhiodoro, T. and Anson, D. S.: Isolation of the canine a-L-fucosidase cDNA and definition of the fucosidosis mutation in English Springer spaniels. Genome, 7, 271–274 (1996). 27. Muller-Taubenberger, A., Westphal, M., Noegel, A., and Gerisch, G.: A developmentally regulated gene product from Dictyostelium discoideum shows high homology to human a-L-fucosidase. FEBS Lett., 246, 185–192 (1989). 28. Nielsen, H., Engelbrecht, J., Brunak, S., and von Heijime, G.: Identification of prokaryotic and eukaryotic signal peptides and prediction of their cleavage sites. Protein Eng., 10, 1–6 (1997). 29. Nakai, K. and Horton, P.: PSORT: a program for detecting the sorting signals of proteins and predicting their subcellular localization. Trends Biochem. Sci., 24, 34–35 (1999). 30. Henrissat, B.: A classification of glycosyl hydrolases based on amino acid sequence similarities. Biochem. J., 280, 309– 316 (1991). 31. Shimizu, T., Ohtani, K., Hirakawa, H., Ohshima, K., Yamashita, A., Shiba, T., Ogasawara, N., Hattori, M., Kuhara, S., and Hayashi, H.: Complete genome sequence of Clostridium perfringens, an anaerobic flesh-eater. Proc. Natl. Acad. Sci. USA, 99, 996–1001 (2002). 32. Hoskins, J. A., Alborn, W., Arnold, J., Blaszczak, L., Burgett, S., DeHoff, B. S., Estrem, S., Fritz, L., Fu, D.-J., Fuller, W., and other 32 authors: Genome of the bacterium Streptococcus pneumoniae strain R6. J. Bacteriol., 183, 5709– 5717 (2001). 33. Xu, J., Bjursell, M. K., Himrod, J., Deng, S., Carmichael, L. K., Chiang, H. C., Hooper, L. V., and Gordon, J. I.: A genomic view of the human-Bacteroides thetaiotaomicron symbiosis. Science, 299, 2074–2076 (2003).
VOL. 99, 2005
BIFIDOBACTERIAL GLYCOSIDASES ACTING ON SUGAR CHAINS
34. Takami, H., Nakasone, K., Takaki, Y., Maeno, G., Sasaki, Y., Masui, N., Fuji, F., Hirama, C., Nakamura, Y., Ogasawara, N., Kuhara, S., and Horikoshi, K.: Complete genome sequence of the alkaliphilic bacterium Bacillus halodurans and genomic sequence comparison with Bacillus subtilis. Nucleic Acids Res., 28, 4317–4331 (2000). 35. da Silva, A. C. R., Ferro, J. A., Reinach, F. C., Farah, C. S., Furlan, L. R., Quaggio, R. B., Monteiro-Vitorello, C. B., Van Sluys, M. A., Almeida, N. F., Alves, L. M. C., and other 55 authors: Comparison of the genomes of two Xanthomonas pathogens with differing host specificities. Nature, 417, 459–463 (2002). 36. Huang, C. C. and Aminoff, D.: Enzymes that destroy blood group specificity. J. Biol. Chem., 247, 6737–6742 (1972). 37. Endo, Y. and Kobata, A.: Partial purification and characterization of an endo-a-N-acetylgalactosaminidase from the culture medium of Diplococcus pneumoniae. J. Biochem., 80, 1–8 (1976). 38. Bhavanandan, V. P., Umemoto, J., and Davidson, E. A.: Characterization of an endo-a-N-acetylgalactosaminidase from Diplococcus pneumoniae. Biochem. Biophys. Res. Commun., 70, 738–745 (1976). 39. Fan, J. Q., Kadowaki, S., Yamamoto, K., Kumagai, H., and Tochikura, T.: Purification and characterization of endoa-N-acetylgalactosaminidase from Alcaligenes sp. Agric. Biol. Chem., 52, 1715–1723 (1988).
465
40. Ashida, H., Yamamoto, K., Murata, T., Usui, T., and Kumagai, H.: Characterization of endo-a-N-acetylgalactosaminidase from Bacillus sp. and syntheses of neo-oligosaccharides using its transglycosylation activity. Arch. Biochem. Biophys., 373, 394–400 (2000). 41. Iwase, H., Ishii, I., Ishihara, K., Tanaka, Y., Omura, S., and Hotta, K.: Release of oligosaccharides possessing reduced-end N-acetylgalactosamine from mucus glycoprotein in Streptomyces sp. OH-11242 culture medium through action of endo-type glycosidase. Biochem. Biophys. Res. Commun., 151, 422–428 (1988). 42. Derensy-Dron, D., Krzewinski, F., Brassart, C., and Bouquelet, S.: b-1,3-Galactosyl-N-acetylhexosamine phosphorylase from Bifidobacterium bifidum DSM20082: characterization, partial purification and relation to mucin degradation. Biotechnol. Appl. Biochem., 29, 3–10 (1999). 43. Kitaoka, M., Tian, J., and Nishimoto, M.: A novel putative galactose operon involving lacto-N-biose phosphorylase found in Bifidobacterium longum. Appl. Environ. Microbiol. (2005). (in press) 44. Nunoura, N., Ohdan, K., Tanaka, K., Tamaki, H., Yano, T., Inui, M., Yukawa, H., Yamamoto, K., and Kumagai, H.: Cloning and nucleotide sequence of the b-D-glucosidase gene from Bifidobacterium breve clb, and expression of b-Dglucosidase activity in Escherichia coli. Biosci. Biotech. Biochem., 60, 2011–2018 (1996).
© 2005, The Society for Biotechnology, Japan
Novel Bifidobacterial Glycosidases Acting on Sugar Chains of Mucin Glycoproteins TAKANE KATAYAMA,1 KIYOTAKA FUJITA,1 AND KENJI YAMAMOTO1* Graduate School of Biostudies, Kyoto University, Oiwake-cho, Kitashirakawa, Sakyo-ku, Kyoto 606-8502, Japan1 Received 22 February 2005/Accepted 9 March 2005
Bifidobacterium bifidum was found to produce a specific 1,2-a-L-fucosidase. Its gene (afcA) has been cloned and the DNA sequence was determined. The AfcA protein consisting of 1959 amino acid residues with a predicted molecular mass of 205 kDa can be divided into three domains; the N-terminal function-unknown domain (576 aa), the catalytic domain (898 aa), and the C-terminal bacterial Ig-like domain (485 aa). The recombinant catalytic domain specifically hydrolyzed the terminal a-(1®2)-fucosidic linkages of various oligosaccharides and sugar chains of glycoproteins. The primary structure of the catalytic domain exhibited no similarity to those of any glycoside hydrolases but showed similarity to those of several hypothetical proteins in a database, which resulted in establishment of a novel glycoside hydrolase family (GH family 95). Several bifidobacteria were found to produce a specific endo-a-N-acetylgalactosaminidase, which is the endoglycosidase liberating the O-glycosidically linked galactosyl b1®3 N-acetylgalactosamine disaccharide from mucin glycoprotein. The molecular cloning of endo-a-N-acetylgalactosaminidase was carried out on Bifidobacterium longum based on the information in the database. The gene was found to comprise 1966 amino acid residues with a predicted molecular mass of 210 kDa. The recombinant protein released galactosyl b1®3 N-acetylgalactosamine disaccharide from natural glycoproteins. This enzyme of B. longum is believed to be involved in the catabolism of oligosaccharide of intestinal mucin glycoproteins. Both 1,2-a-L-fucosidase and endo-a-N-acetylgalactosaminidase are novel and specific enzymes acting on oligosaccharides that exist mainly in mucin glycoproteins. Thus, it is reasonable to conclude that bifidobacteria produce these enzymes to preferentially utilize the oligosaccharides present in the intestinal ecosystem. [Key words: bifidobacteria, Bifidobacterium bifidum, Bifidobacterium longum, 1,2-a-L-fucosidase, endo-a-N-acetylgalactosaminidase, mucin, O-linked oligosaccharide]
survive in the lower intestinal tract, bifidobacteria produce various kinds of exo- and endoglycosidases in surfacebound and/or extracellular forms, by which they can utilize diverse carbohydrates (6, 7). Recent genome sequence analysis of Bifidobacterium longum NCC2705 revealed that more than 8.5% of the total predicted proteins were involved in the degradation of oligo- and polysaccharides, perhaps reflecting the superior ability of this organism to adapt to its environment (8). The epithelial cells of the human intestine express and/or secrete mucin glycoproteins (9), which are assumed to play an important role in protecting enterocytes from chemical and physical damage as well as in preventing invasion by pathogens. Intestinal mucin glycoproteins contain a lot of O-linked oligosaccharides, on the non-reducing ends of which a-linked L-fucosyl residues are frequently found (10, 11). Such a-L-fucosyl residues are also present in glycolipids on the cell surface, submaxillary mucin, blood group substances, and oligosaccharides in human milk (11, 12).
The human gastrointestinal tract is inhabited by a vast and diverse community of microbes, and a well-balanced microflora is thought to be important for normal digestion and maintenance of the intestinal ecosystem (1). In the complex intestinal microflora, bifidobacteria, which are grampositive obligate anaerobes, are considered to be key commensals that promote a healthy intestinal tract because of their many beneficial effects on the host, such as regulation of the state of the intestine, reduction of harmful bacteria and toxic compounds, immuno modulation, and anticarcinogenic activity (2–5). Therefore, bifidobacteria have attracted a great deal of attention. Bifidobacteria naturally colonize the lower intestinal tract, an environment poor in mono- and disaccharides since such sugars are preferentially consumed by the host and microbes present in the upper intestinal tract. Therefore, in order to * Corresponding author. e-mail: [email protected] phone: +81-(0)75-753-6277 fax: +81-(0)75-753-6275 457
458
KATAYAMA ET AL.
Taking into consideration that several fecal mucin-degrading bacteria such as bifidobacteria, clostridia, and bacteroides are known to produce a-L-fucosidase (13–15), it was envisaged that a-L-fucosidase might be responsible for the selective colonization of the intestine by these bacteria. Meanwhile, intestinal mucin glycoproteins possess an abundant amount of O-glycosidic oligosaccharides linked glycosidically to the hydroxyl groups of serine or threonine residues of the protein (16). These oligosaccharides are also found sparsely in soluble glycoproteins. They are characterized by a core structure consisting of a galactosyl b1®3 N-acetylgalactosamine a1-serine/threonine residue, and are termed mucin type oligosaccharides. Endo-a-N-acetylgalactosaminidase (endo-a-GalNAc-ase; EC 3.2.1.97) is an endoglycosidase that catalyzes the hydrolysis of the O-glycosidic a-linkage between galactosyl b1®3 N-acetylgalactosamine (Galb1-3GalNAc) and a serine or threonine residue in mucin glycoproteins of various animal sources (17). Recently, we isolated and identified some bifidobacteriaspecific enzymes that are involved in the degradation of sugar chains of intestinal mucin. We also revealed that these enzymes belong to novel glycoside hydrolase families. In this review, we describe the molecular cloning and characterization of 1,2-a-L-fucosidase (EC 3.2.1.63) and endoa-GalNAc-ase from bifidobacteria, an intestinal colonizer strain, in order to obtain a better understanding of the catabolism of sugars in bifidobacteria and to examine the roles of sugar chain-degrading enzymes in intestinal bacteria. These unique enzymes are not normally found in microorganisms. I. MOLECULAR CLONING OF 1,2-a-L-FUCOSIDASE FROM BIFIDOBACTERIUM BIFIDUM JCM1254 To date, a-L-fucosidases that liberate terminal a-linked from the oligosaccharides of various glycoconjugates including mucin glycoprotein have been purified from several prokaryotic and eukaryotic sources (14, 15, 18–23), and the enzymes have been divided into two groups (19); one capable of hydrolyzing various types of fucosidic linkages as well as synthetic substrates, and another that is only active on the a-(1®2)-linkage, although a few reports have described enzymes that cannot be classified into either group (20, 21). In contrast to the dozens of studies on the purification of such enzymes, there is a paucity of reports regarding the cloning of a-L-fucosidase genes, and they are all derived from eukaryotic cells such as human and rat livers (24, 25), Canis familiaris (26), and Dictyostelium discoideum (27), which constitute glycoside hydrolase family 29 (GH family 29). To the best of our knowledge, there have been no reports regarding the isolation of a bacterial a-L-fucosidase gene, although the sequence of the 1,3-/4a-L-fucosidase gene from Streptomyces sp. has been deposited in GenBank (accession no. U39394). We have found that a few bifidobacteria strains produced 1,2-a-L-fucosidase (EC 3.2.1.63) in cell surface-bound and/or extracellular forms, using 2¢-fucosyllactose as a substrate. Among these bacteria, we chose B. bifidum JCM1254 and attempted to clone the 1,2-a-L-fucosidase gene (designated as afcA). The expression cloning was done using Escherichia coli L-fucose
J. BIOSCI. BIOENG.,
DH5a, a non-fucosidase-producing bacterium. A genomic library of B. bifidum JCM1254 constructed in E. coli DH5a was screened for the ability to hydrolyze the a-(1®2)-linkage of 2¢-fucosyllactose. One recombinant, designated SA3, was selected. Sequence analysis of plasmid pSA3 revealed that the cloned gene contained two large truncated open reading frames (designated as ORF1 and ORF2) (Fig. 1). The sense strands of ORF1 and ORF2 overlapped to a large extent in reverse, which is not surprising since such cases are sometimes found in the recently determined genome sequence of B. longum NCC2705 (8). In order to determine which ORF actually encodes 1,2-a-L-fucosidase, each ORF (the MfeI-AflII fragment) was placed under the control of the lac promoter and its expression was induced by the addition of IPTG. While no increase in 1,2-a-L-fucosidase activity was observed when ORF2 was induced, the activity was significantly elevated when ORF1 was induced, indicating that ORF1 encodes 1,2-a-L-fucosidase. The codon usage of the afcA gene was quite similar to that of other genes of B. bifidum in the database (Codon Usage Database in Kazusa DNA Research Institute), and Southern hybridization analysis with the 2.3-kb KpnI fragment as a specific probe revealed that the afcA gene exists as asingle-copy on the genome of B. bifidum JCM1254. Analysis of the primary structure by using the SignalP (28) and PSORT (29) programs revealed the presence of a signal peptide and a membrane anchor at the N-terminus and C-terminus, respectively (Fig. 1). A possible ribosome-binding site was located 6-bp from a probable initiation codon and a promoter-like sequence was also found in the upstream region. The AfcA protein consists of 1959 amino acid residues with a calculated molecular mass of 205 kDa (GenBank accession no. AY303700). II. DOMAIN STRUCTURE OF 1,2-a-L-FUCOSIDASE FROM B. BIFIDUM JCM1254 Since the primary structure of AfcA protein did not exhibit any similarity to those of known glycosidase families, we attempted to localize a catalytic domain that is essential for the hydrolysis of 2¢-fucosyllactose. The N-terminal and C-terminal deletion mutants were constructed and expressed, and then their activity was assessed. Consequently, the 576 N-terminal amino acid residues (911–2638 bases) and 485 C-terminal amino acid residues (5333–6790 bases) were found to be removable without loss of fucosidase activity. These results revealed that the region consisting of the 577th–1474th amino acid residues constitutes the catalytic domain (Fig. 1). The region consisting of the 1475th–1728th amino acid residues contained four repetitive sequences with immunoglobulin-like folds, the so-called bacterial Iglike domain B (Pfam 02368). Although the function of this domain is not clear, it is highly likely that this domain, 254 amino acids in length, at least acts to display the fucosidase domain of AfcA so that it protrudes from the cell surface, thereby enabling B. bifidum JCM1254 cells to gain access to and degrade the fucosyl residues present on the glycoconjugates of enterocytes. Neither sequence similarity to other ORFs nor a functional motif was found in the sequence of
VOL. 99, 2005
BIFIDOBACTERIAL GLYCOSIDASES ACTING ON SUGAR CHAINS
459
FIG. 1. Gene organization of the afcA locus of B. bifidum JCM1254 and schematic representation of the domain structure of AfcA. The top thin line is a restriction map, indicating the recognition sites of AflII (A), BamHI (B), EcoRI (E), KpnI (K), MfeI (Mf), and MluI (Ml). The thick lines indicate the lengths of the inserts of the plasmids, and the dashed line indicates the specific probe (1808–2888 bp) used for cloning of the upstream region of ORF1 (1–2888 bp, pSA19). The deduced amino acid sequence of ORF3 showed high similarity to the consensus sequence of Pfam 03577, thus it probably encodes a dipeptidase. The amino acid numbering of the schematic representation of AfcA starts at the probable initiation codon, and the domain responsible for 1,2-a-L-fucosidase activity (577–1474 aa) is depicted as a gray box. The region with Ig-like folds (1475–1728 aa) is shown by a stippled box. The black bars at the N-terminal and C-terminal ends indicate a signal peptide and a membrane anchor, respectively.
the N-terminal domain (1st–576th amino acid residues). III. CHARACTERIZATION OF FUCOSIDASE DOMAIN We attempted to overexpress and purify the AfcA protein, however, an intact form of the protein was not obtained, most likely due to the extremely long length of the protein. Therefore, the catalytic domain (Fuc domain) was expressed and purified as a C-terminal hexahistidine-tagged protein. The ability of the Fuc domain to hydrolyze the a-(1®2)-L-fucosidic linkage of 2¢-fucosyllactose was confirmed by ESI-MS and NMR analyses. The substrate specificity was examined using naturally occurring substrates. As shown in Table 1, of the milk oligosaccharides examined, the most readily hydrolyzed substrates were 2¢-fucosyllactose and lacto-N-fucopentaose I, both of which contain L-fucose bound to galactose through an a-(1®2)-linkage at the non-reducing termini. The enzyme exhibited very limited activity for the a-(1®3)-linked L-fucosyl residues of 3-fucosyllactose and lacto-N-fucopentaose V, and apparently had no action at all on the a-(1®4)-linkage of lacto-N-fucopentaose II. The enzyme could not hydrolyze 3-fucosylgalactose (Fuca1®3Gal) or 4-fucosyl-N-acetylglucosamine (Fuca1®4GlcNAc). Blood group H (II) active substance (Fuca1®2Galb1®4GlcNAc) was found to be a good substrate, however, the enzyme had very limited action on blood groups A and B active substances, which have a-(1®3)GalNAc and a-(1®3)-Gal residues, respectively, in addition to a-(1®2)-Fuc residues at the non-reducing termini. The a-(1®6)-linkage of 6-fucosyl-N,N¢-diacetylchitobiose was not cleaved by this enzyme. In addition to fucosecontaining oligosaccharides, glycoproteins with terminal
a-(1®2)-linked L-fucosyl residues, e.g., porcine gastric mucin (H), were also found to be substrates of the Fuc domain. On the other hand, the Fuc domain did not liberate fucose from any of the three artificial substrates examined (pNP-a-L-fucoside, pNP-b-L-fucoside, and 4-methylumbelliferyl-a-L-fucoside). The catalytic domain showed the maximum activity at pH 5, suggesting the involvement of acidic residue(s) in the catalysis, and was stable below 35°C for 30 min and at pH 6.5–7.5 for 12 h. The Km and Vmax values for 2¢-fucosyllactose were determined to be 0.53 mM and 1.6 mmol/min/mg, respectively. IV. STRUCTURAL SIMILARITY OF FUCOSIDASE DOMAIN TO HYPOTHETICAL PROTEINS IN A DATABASE The primary structure of the Fuc domain was subjected to a BLAST search, but no significant similarity was found with proteins with defined functions (30). Scrutiny of the AfcA sequence in reference to conserved motifs of family GH29 did not lead to the finding of any trails of the family. In a BLAST search, high scores were obtained with several hypothetical proteins, the amino acid sequences of which are compared and the highly homologous regions are shown in Fig. 2. Of particular interest is the finding that the Fuc domain is homologous to CPE1875 in the genome sequence of Clostridium perfringens strain 13 (31) since it has been shown that C. perfringens produces 1,2-a-L-fucosidase with a high molecular weight (ca 200 kDa) in the culture medium (14). The deduced amino acid sequence of CPE1875 has a typical signal sequence at the N-terminus (27), and its molecular weight is estimated to be 165 kDa. Thus, CPE1875 might encode 1,2-a-L-fucosidase. Other homologous pro-
460
J. BIOSCI. BIOENG.,
KATAYAMA ET AL.
TABLE 1.
Substrate specificity of the Fuc domain of AfcA from B. bifidum JCM1254
a The abbreviations used are: Gal, D-galactose; Glc, D-glucose; Fuc, L-fucose; GlcNAc, N-acetyl-D-glucosamine; and GalNAc, N-acetyl-Dgalactosamine. b The reaction mixture consisting of 100 mM sodium phosphate (pH 6.5), 2 mM substrate, and 20 mU of the purified Fuc domain, in a total volume of 100 ml, was incubated for 1 min–12 h at 30°C. The value obtained with 2'-fucosyllactose was taken as 100%. ND, Not detectable.
teins were found from the genome sequences of Streptococcus pneumoniae R6 (32), Bacteroides thetaiotaomicron VPI-5482 (33), Microbulbifer degradans 2-40, Bacillus halodurans C-125 (34), Xanthomonas campestris pv. campestris ATCC33913 (35), and Arabidopsis thaliana. Although it remains to be elucidated whether these homologues indeed show 1,2-a-L-fucosidase activity, the finding of a new conservative sequence indicates the existence of a novel glycoside hydrolase family (GH family 95, personal communication with Dr. B. Henrissat, Universites de Marseille, France). Although B. bifidum cannot ferment L-fucose, the presence of a-L-fucosidase does benefit the cells because, while B. bifidum cells are capable of degrading and fermenting 2¢-fucosyllactose, none of the bifidobacterial strains that do not secrete fucosidase could degrade 2¢-fucosyllactose without the aid of exogenously added fucosidase (purified Fuc
domain). These results suggest that, in preference to nonfucosidase-producing bacteria, B. bifidum are able to degrade several types of substrates present in the intestinal epithelium and mucosa, and imply the biological importance of a secretory fucosidase from gut microbes for the intestinal ecosystem. V. ISOLATION AND IDENTIFICATION OF ENDO-a-N-ACETYLGALACTOSAMINIDASE FROM B. LONGUM JCM1217 Endo-a-N-acetylgalactosaminidase (endo-a-GalNAc-ase; EC 3.2.1.97) catalyzes the hydrolysis of the O-glycosidic a-linkage between galactosyl b1-3 N-acetylgalactosamine (Galb1-3GalNAc) and a serine or threonine residue in mucins and mucin-type glycoproteins from various animal
VOL. 99, 2005
BIFIDOBACTERIAL GLYCOSIDASES ACTING ON SUGAR CHAINS
461
FIG. 2. Comparison of the amino acid sequences of the Fuc domain of AfcA with homologous proteins in the database. The primary structure of the Fuc domain was aligned with those of homologues using the Clustal W program, and highly homologous regions are shown. White letters on a black background indicate identical residues in more than four sequences, and conservative amino acid changes are shown in gray. B. bif, AfcA of B. bifidum JCM1254; S. pne, Streptococcus pneumoniae R6 (GenBank accession no. NP_359091); C. per, Clostridium perfringens 13 (GenBank accession no. NP_562791); M. deg, Microbulbifer degradans 2–40 (GenBank accession no. ZP_00064866); A. tha, Arabidopsis thaliana (GenBank accession no. AY125494); B. hal, Bacillus halodurans B-125 (GenBank accession no. NP_241708); X. axo, Xanthomonas axonopodis pv. campestris str. ATCC 33913 (GenBank accession no. NP_637123); and B. the, Bacteroides thetaiotaomicron VPI-5482 (GenBank accession no. NP_809923).
sources. The enzyme was first found from the culture fluid of C. perfringens type 33-4A (36), and then purified from the culture fluid of Diplococcus pneumoniae (now assigned as S. pneumoniae) type 1 (37, 38). We also purified the enzyme from the culture fluids of Alcaligenes sp. F-1906 (39) and Bacillus sp. A-198 (40) which were isolated from soil and identified. The enzyme has a strict substrate specificity, acting only on the a-linked disaccharide Galb13GalNAc. A similar enzyme was also detected in the culture fluid of Streptomyces sp. OH-11242 which could release longer oligosaccharides than disaccharide from porcine mucin (41). As these endo-a-GalNAc-ases liberate O-linked oligosaccharides from cell surface glycoproteins without causing damage, they are useful for the investigation of the structure and function of O-linked oligosaccharides on the cell surface (17). Recently, we found endo-a-GalNAc-ase activity in the culture fluids of several bifidobacteria. Using Galb13GalNAca1-p-nitrophenyl as a substrate, we determined the endo-a-GalNAc-ase activity in the culture fluids of various strains of bifidobacteria. As shown in Fig. 3, activity was found in the culture fluids of Bifidobacterium breve JCM1192, B. bifidum JCM1254, JCM7004, and ATCC29521, and B. longum JCM1217 and JCM7054. These results sug-
FIG. 3. TLC analysis of the reaction mixtures incubated with pNP -substrate and various Bifidobacteria cells. The cell pellets were incubated with 0.5 mM Galb1,3GalNAca1pNP in 10 mM acetate buffer (pH 5.0) for 2 h at 37°C. Lane 1, Galb1,3GalNAca1pNP; lane 2, Gal; lane 3, GalNAc; lane 4, Galb1,3GalNAc; lane 5, B. breve JCM1192; lane 6, B. infantis JCM1222; lane 7, B. bifidum JCM1254; lane 8, B. bifidum JCM7004; lane 9, B. bifidum ATCC29521; lane 10, B. pseudolongum JCM1205; lane 11, B. longum JCM1217; lane 12, B. longum JCM7054.
gest that the enzyme might be involved in the catabolism of oligosaccharides in bifidobacteria. We next attempted to elucidate the function of this enzyme in the bacteria, and
462
J. BIOSCI. BIOENG.,
KATAYAMA ET AL.
carried out the molecular cloning, expression, and characterization of bifidobacterial endo-a-GalNAc-ase. The above results suggested that bifidobacteria produced and secreted an endo-a-GalNAc-ase into the cell surface or extracellularly, and that they may have a signal peptide and cell-surface associated domains. There are several predicted endoglycosidases such as endo-b-xylanase, endo-b-N-acetylglucosaminidase, endo-a-arabinosidase, and arabinogalactan endo-b-galactosidases, which may be useful for the intake of extracellular carbohydrate polymers, based on information obtained from the genome sequencing database of B. longum NCC2705 (8). Based on the information concerning uncharacterized hypothetical proteins from the genome database, we searched for a hypothetical protein containing a secretion signal and transmembrane domain using BLAST and Pfam HMM searches, and assumed one potential sequence “AAN24297” (locus name, BL0464) to be endoa-GalNAc-ase. Then, using the sequence of BL0464 “narrowly conserved hypothetical protein” from B. longum NCC2705 (8), oligonucleotide primers were designed and a 6.0-kb PCR product was amplified from genomic DNA of B. longum JCM1217, which was confirmed to produce endoa-GalNAc-ase. The upstream and downstream regions of the gene were obtained using colony hybridization and PCR, respectively. The composite DNA encoded an open reading frame of 5901 bp corresponding to a protein of 1966 amino acids. The predicted peptide sequence of endo-a-GalNAcase showed 99.1% identity with BL0464. A region at the Nand C-termini of the enzyme contained a predicted signal peptide (amino acids 1–29) and a transmembrane domain (amino acids 1936–1963), respectively (Fig. 4). The coding sequence of endo-a-GalNAc-ase had no homologous domains of known GH families. Hypothetical proteins with homology to the endo-a-GalNAc-ase of B. longum JCM1217 were observed in the genome sequences of C. perfringens 13, S. pneumoniae R6, Streptomyces coelicolor A3, and Enterococcus faecalis V583 (Fig. 4). Some of these bacteria have been shown to produce endo-a-GalNAc-ase in their culture fluids (36, 37). A conserved region, which was found in the region of amino acids 590–1381 of the enzyme from B. longum JCM1217, is 32% to 54% identical to the corresponding hypothetical proteins. We expressed a fulllength of His-tagged enzyme in E. coli BL21(DE3) using the expression plasmid pET23d, and the soluble cell lysate exhibited endo-a-GalNAc-ase activity. The His-tagged endoa-GalNAc-ase was purified and the obtained enzyme preparation migrated as a single band corresponding to a protein of 210 kDa on SDS–PAGE. VI. VARIOUS PROPERTIES OF RECOMBINANT B. LONGUM JCM1217 ENDO-a-NACETYLGALACTOSAMINIDASE The optimal pH of the recombinant enzyme was around 5.0 at 37°C in 10 mM acetate buffer and the enzyme showed 68% and 73% of the maximum activity at pH 4.0 and 6.0, respectively. The enzyme was stable at temperatures up to 37°C and retained 73% activity at 45°C after incubating for 30 min in 10 mM acetate buffer (pH 5.0). The optimal temperature for the enzyme reaction was 60°C. The kinetic
parameters of the enzyme were determined at 37°C and pH 5.0, and the Km value and kcat value for Galb1®3GalNAca1pNP were 0.165 mM and 37.0 s–1, respectively. We examined the substrate specificity of the recombinant enzyme using synthetic pNP substrates and natural glycoproteins. TLC analysis of the reaction mixtures indicated that Galb13GalNAc disaccharide was liberated from asialofetuin and pNP substrate containing the core 1 structure (Galb1® 3GalNAca1). However, the hydrolysis products could not be detected in the presence of sialofetuin or any pNP substrate other than Galb1®3GalNAca1-pNP. Derensy-Dron et al. characterized b-1,3-galactosyl-Nacetylhexosamine phosphorylase (lacto-N-biose phosphorylase; EC 2.4.1.211) which has the ability to produce a-galactopyranose 1-phosphate (Gal-1-P) and N-acetylglucosamine in cell-free extracts of B. bifidum DSM20082 (42). This enzyme could also phosphorolyze Galb1®3GalNAc which is produced by endo-a-GalNAc-ase to Gal-1-P and N-acetylgalactosamine. Kitaoka et al. found that this enzyme was produced by B. longum JCM1217 as an intracellular enzyme (43). This result suggests the catabolism of the sugar chains of mucin glycoprotein by B. longum JCM1217 using endo-a-GalNAc-ase and lacto-N-biose phosphorylase. We also confirmed endo-a-GalNAc-ase activity in the cells of three B. bifidum strains. As mucin may be major oligosaccharide supplier in the human colon, these data suggest that several bifidobacteria possess a degradation system for mucin oligosaccharides. VII. REMARKS Since bifidobacteria are key commensals that promote a healthy intestinal tract, obtaining a better understanding of their specific metabolism is quite important for their effective and preferential colonization of the intestine. We have focused on the sugar catabolism of bifidobacteria, and have revealed, for the first time, the gene organizations and primary structures of 1,2-a-L-fucosidase and endo-a-N-acetylgalactosaminidase, both of which can act on and degrade naturally-occurring substrates including high-molecular-weight glycoproteins, indicating the physiological significance of these enzymes. In our laboratory, we have already cloned the gene encoding b-D-glucosidase, which has b-D-fucosidase activity, from B. breve clb (44). By using this recombinant enzyme, we synthesized a disaccharide that seemed to be specifically assimilated by bifidobacteria. The mechanism of catabolism of sugar in bifidobacteria seems to be very complicated. The accumulation of data on the genes of various glycosidases is necessary in order to elucidate the specific characteristics of bifidobacteria with respect to the assimilation of sugars. From this point of view, since 1,2-a-L-fucosidase and endo-a-N-acetylgalactosaminidase are unique enzymes that are not usually found in ordinary microorganisms, their utilization for specifically promoting the growth of bifidobacteria is being anticipated. Our results open the way to elucidating the roles of these specific enzymes in the intestinal ecosystem, and moreover, provide a basis upon which we hope new bifidus factors (prebiotics) will be developed in the near future.
VOL. 99, 2005
BIFIDOBACTERIAL GLYCOSIDASES ACTING ON SUGAR CHAINS
463
FIG. 4. Multiple alignment of the amino acid sequences of B. longum JCM1217 endo-a-GalNAc-ase with the homologous proteins. (A) Schematic diagram of endo-a-GalNAc-ase from B. longum JCM1217. FIVAR domain is presumably involved in binding of the protein to cell wall. (B) The gray boxes indicate comparatively conserved regions. (C) Multiple alignment of the amino acid sequences (shown in panel B as gray boxes). Identical residues and conserved substitutions are highlighted in black and dark gray, respectively. Asterisks indicate conserved acidic amino acids. The organisms and accession nos. (in brackets) are as follows: SC, S. coelicolor A3 (CAA20079); EF, E. faecalis V583 (AAO81568); BL1, B. longum JCM1217 (AY836679); BL2, B. longum NCC2705 (AAN24297); SP, S. pneumoniae R6 (AAK99132); and CP, C. perfringens 13 (BAB80399).
464
KATAYAMA ET AL.
ACKNOWLEDGMENTS This work was partly supported by a Grant-in-Aid for Scientific Research from Core Research for Evolutional Science and Technology (CREST) from the Japan Science and Technology Agency (JST), by a Grant-in-Aid for Scientific Research (B) from the Ministry of Education, Culture, Sports, Science and Technology, Japan, and by a Grant-in-Aid for Food Science Research from the Food Science Institute Foundation.
REFERENCES 1. Simon, G. L. and Gorbach, S. L.: The human intestinal microflora. Dig. Dis. Sci., 31, 147S–162S (1986). 2. Hooper, L. V. and Gordon, J. I.: Commensal host-bacterial relationships in the gut. Science, 292, 1115–1118 (2001). 3. Gibson, G. R. and Wang, X.: Regulatory effects of bifidobacteria on the growth of other colonic bacteria. J. Appl. Bacteriol., 77, 412–420 (1994). 4. Ouwehand, A., Isolauri, E., and Salminen, S.: The role of the intestinal microflora for the development of the immune system in early childhood. Eur. J. Nutr., 41, 32–37 (2002). 5. Femia, A. P., Luceri, C., Dolara, P., Giannini, A., Biggeri, A., Salvadori, M., Clune, Y., Collins, K. J., Paglierani, M., and Caderni, G.: Antitumorigenic activity of the prebiotic inulin enriched with oligofructose in combination with the probiotics Lactobacillus rhamnosus and Bifidobacterium lactis on azoxymethane-induced colon carcinogenesis in rats. Carcinogenesis, 23, 1953–1960 (2002). 6. Hoskins, L. C., Agustines, M., McKee, W. B., Boulding, E. T., Kriaris, M., and Niedermeyer, G.: Mucin degradation in human colon ecosystems. Isolation and properties of fecal strains that degrade ABH blood group antigens and oligosaccharides from mucin glycoproteins. J. Clin. Invest., 75, 944–953 (1985). 7. Larson, G., Falk, P., and Hoskins, L. C.: Degradation of human intestinal glycosphingolipids by extracellular glycosidases from mucin-degrading bacteria of the human fecal flora. J. Biol. Chem., 263, 10790–10798 (1988). 8. Schell, M. A., Karmirantzou, M., Snel, B. D., Vilanova, B., Berger, B., Pessi, G., Zwahlen, M. C., Desiere, F., Bork, P., Delley, M., Pridmore, R. D., and Arigoni, F.: The genome sequence of Bifidobacterium longum reflects its adaptation to the human gastrointestinal tract. Proc. Natl. Acad. Sci. USA, 99, 14422–14427 (2002). 9. Gum, J. R., Byrd, J. C., Hicks, J. W., Toribara, N. W., Lamport, D. T., and Kim, Y. S.: Molecular cloning of human intestinal mucin cDNAs. Sequence analysis and evidence for genetic polymorphism. J. Biol. Chem., 264, 6480– 6487 (1989). 10. Podolsky, D. K.: Oligosaccharide structures of human colonic mucin. J. Biol. Chem., 260, 8262–8271 (1985). 11. Slomiany, A., Zdebska, E., and Slomiany, B. L.: Structures of the neutral oligosaccharides isolated from A-active human gastric mucin. J. Biol. Chem., 259, 14743–14749 (1984). 12. Song, J.-F., Weng, M.-Q., Wu, S.-M., and Xia, Q.-C.: Analysis of neutral saccharides in human milk derivatized with 2-aminoacridone by capillary electrophoresis with laser-induced fluorescence detection. Anal. Biochem., 304, 126–129 (2002). 13. Salyers, A. A., West, S. E., Vercellotti, J. R., and Wilkins, T. D.: Fermentation of mucins and plant polysaccharides by anaerobic bacteria from the human colon. Appl. Environ. Microbiol., 34, 529–533 (1977). 14. Aminoff, D. and Furukawa, K.: Enzymes that destroy blood group specificity. I. Purification and properties of a-L-fucosidase from Clostridium perfringens. J. Biol. Chem., 245, 1659– 1669 (1970).
J. BIOSCI. BIOENG.,
15. Berg, J. O., Lindqvist, L., and Nord, C. E.: Purification of glycoside hydrolases from Bacteroides fragilis. Appl. Environ. Microbiol., 40, 40–47 (1980). 16. Marth, J. D.: O-Glycans, p. 101–113. In Varki, A., Cummings, R., Esko, J., Freeze, H., Hart, G., and Marth, J. (ed.), Essentials of glycobiology. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, New York (1999). 17. Kobata, A.: Use of endo- and exoglycosidases for structural studies of glycoconjugates. Anal. Biochem., 100, 1–14 (1979). 18. Yamamoto, K., Li, S.-C., and Li, Y.-T.: Microbial glycosidases, p. 497–511. In Ernst, B., Hart, G., and Sinay, P. (ed.), Carbohydrates in chemistry and biology. Wiley-VCH, New York (2000). 19. Bahl, O. P.: Glycosidases of Aspergillus niger. II. Purification and general properties of 1,2-a-L-fucosidase. J. Biol. Chem., 245, 299–304 (1970). 20. Tsuji, Y., Yamamoto, K., Tochikura, T., Seno, T., Ohkubo, Y., and Yamaguchi, H.: Purification and characterization of a-L-fucosidase from Bacillus circulans grown on porcine gastric mucin. J. Biochem., 107, 324–330 (1990). 21. Eneyskaya, E. V., Kulminskaya, A. A., Kalkkinen, N., Nifantiev, N. E., Arbatskii, N. P., Saenko, A. I., Chepurnaya, O. V., Arutyunyan, A. V., Shabalin, K. A., and Neustroev, K. N.: An a-L-fucosidase from Thermus sp. with unusually broad specificity. Glycoconj. J., 18, 827–834 (2001). 22. Ogata-Arakawa, M., Muramatsu, T., and Kobata, A.: a-LFucosidase from almond emulsin: characterization of the two enzymes with different specificities. Arch. Biochem. Biophys., 181, 353–358 (1977). 23. Khunsook, S., Alhadeff, J. A., and Bean, B. S.: Purification and characterization of human seminal plasma a-L-fucosidase. Mol. Hum. Reprod., 8, 221–227 (2002). 24. Fukushima, H., de Wet, J. R., and O’Brien, J. S.: Molecular cloning of a cDNA for human a-L-fucosidase. Proc. Natl. Acad. Sci. USA, 82, 1262–1265 (1985). 25. Fisher, K. J. and Aronson, N. N.: Isolation and sequence analysis of a cDNA encoding rat liver a-L-fucosidase. Biochem. J., 264, 695–701 (1989). 26. Occhiodoro, T. and Anson, D. S.: Isolation of the canine a-L-fucosidase cDNA and definition of the fucosidosis mutation in English Springer spaniels. Genome, 7, 271–274 (1996). 27. Muller-Taubenberger, A., Westphal, M., Noegel, A., and Gerisch, G.: A developmentally regulated gene product from Dictyostelium discoideum shows high homology to human a-L-fucosidase. FEBS Lett., 246, 185–192 (1989). 28. Nielsen, H., Engelbrecht, J., Brunak, S., and von Heijime, G.: Identification of prokaryotic and eukaryotic signal peptides and prediction of their cleavage sites. Protein Eng., 10, 1–6 (1997). 29. Nakai, K. and Horton, P.: PSORT: a program for detecting the sorting signals of proteins and predicting their subcellular localization. Trends Biochem. Sci., 24, 34–35 (1999). 30. Henrissat, B.: A classification of glycosyl hydrolases based on amino acid sequence similarities. Biochem. J., 280, 309– 316 (1991). 31. Shimizu, T., Ohtani, K., Hirakawa, H., Ohshima, K., Yamashita, A., Shiba, T., Ogasawara, N., Hattori, M., Kuhara, S., and Hayashi, H.: Complete genome sequence of Clostridium perfringens, an anaerobic flesh-eater. Proc. Natl. Acad. Sci. USA, 99, 996–1001 (2002). 32. Hoskins, J. A., Alborn, W., Arnold, J., Blaszczak, L., Burgett, S., DeHoff, B. S., Estrem, S., Fritz, L., Fu, D.-J., Fuller, W., and other 32 authors: Genome of the bacterium Streptococcus pneumoniae strain R6. J. Bacteriol., 183, 5709– 5717 (2001). 33. Xu, J., Bjursell, M. K., Himrod, J., Deng, S., Carmichael, L. K., Chiang, H. C., Hooper, L. V., and Gordon, J. I.: A genomic view of the human-Bacteroides thetaiotaomicron symbiosis. Science, 299, 2074–2076 (2003).
VOL. 99, 2005
BIFIDOBACTERIAL GLYCOSIDASES ACTING ON SUGAR CHAINS
34. Takami, H., Nakasone, K., Takaki, Y., Maeno, G., Sasaki, Y., Masui, N., Fuji, F., Hirama, C., Nakamura, Y., Ogasawara, N., Kuhara, S., and Horikoshi, K.: Complete genome sequence of the alkaliphilic bacterium Bacillus halodurans and genomic sequence comparison with Bacillus subtilis. Nucleic Acids Res., 28, 4317–4331 (2000). 35. da Silva, A. C. R., Ferro, J. A., Reinach, F. C., Farah, C. S., Furlan, L. R., Quaggio, R. B., Monteiro-Vitorello, C. B., Van Sluys, M. A., Almeida, N. F., Alves, L. M. C., and other 55 authors: Comparison of the genomes of two Xanthomonas pathogens with differing host specificities. Nature, 417, 459–463 (2002). 36. Huang, C. C. and Aminoff, D.: Enzymes that destroy blood group specificity. J. Biol. Chem., 247, 6737–6742 (1972). 37. Endo, Y. and Kobata, A.: Partial purification and characterization of an endo-a-N-acetylgalactosaminidase from the culture medium of Diplococcus pneumoniae. J. Biochem., 80, 1–8 (1976). 38. Bhavanandan, V. P., Umemoto, J., and Davidson, E. A.: Characterization of an endo-a-N-acetylgalactosaminidase from Diplococcus pneumoniae. Biochem. Biophys. Res. Commun., 70, 738–745 (1976). 39. Fan, J. Q., Kadowaki, S., Yamamoto, K., Kumagai, H., and Tochikura, T.: Purification and characterization of endoa-N-acetylgalactosaminidase from Alcaligenes sp. Agric. Biol. Chem., 52, 1715–1723 (1988).
465
40. Ashida, H., Yamamoto, K., Murata, T., Usui, T., and Kumagai, H.: Characterization of endo-a-N-acetylgalactosaminidase from Bacillus sp. and syntheses of neo-oligosaccharides using its transglycosylation activity. Arch. Biochem. Biophys., 373, 394–400 (2000). 41. Iwase, H., Ishii, I., Ishihara, K., Tanaka, Y., Omura, S., and Hotta, K.: Release of oligosaccharides possessing reduced-end N-acetylgalactosamine from mucus glycoprotein in Streptomyces sp. OH-11242 culture medium through action of endo-type glycosidase. Biochem. Biophys. Res. Commun., 151, 422–428 (1988). 42. Derensy-Dron, D., Krzewinski, F., Brassart, C., and Bouquelet, S.: b-1,3-Galactosyl-N-acetylhexosamine phosphorylase from Bifidobacterium bifidum DSM20082: characterization, partial purification and relation to mucin degradation. Biotechnol. Appl. Biochem., 29, 3–10 (1999). 43. Kitaoka, M., Tian, J., and Nishimoto, M.: A novel putative galactose operon involving lacto-N-biose phosphorylase found in Bifidobacterium longum. Appl. Environ. Microbiol. (2005). (in press) 44. Nunoura, N., Ohdan, K., Tanaka, K., Tamaki, H., Yano, T., Inui, M., Yukawa, H., Yamamoto, K., and Kumagai, H.: Cloning and nucleotide sequence of the b-D-glucosidase gene from Bifidobacterium breve clb, and expression of b-Dglucosidase activity in Escherichia coli. Biosci. Biotech. Biochem., 60, 2011–2018 (1996).