Bifidobacterium bifidum Lacto-N-Biosidase, a Critical Enzyme for the Degradation of Human Milk Oligosaccharides with a Type 1 Structure

Abstract

Breast-fed infants often have intestinal microbiota dominated by bifidobacteria in contrast to formula-fed infants. We found that several bifidobacterial strains produce a lacto-N-biosidase that liberates lacto-N-biose I (Galβ1,3GlcNAc; type 1 chain) from lacto-N-tetraose (Galβ1,3GlcNAcβ1,3Galβ1,4Glc), which is a major component of human milk oligosaccharides, and subsequently isolated the gene from Bifidobacterium bifidum JCM1254. The gene, designated lnbB, was predicted to encode a protein of 1,112 amino acid residues containing a signal peptide and a membrane anchor at the N and C termini, respectively, and to possess the domain of glycoside hydrolase family 20, carbohydrate binding module 32, and bacterial immunoglobulin-like domain 2, in that order, from the N terminus. The recombinant enzyme showed substrate preference for the unmodified β-linked lacto-N-biose I structure. Lacto-N-biosidase activity was found in several bifidobacterial strains, but not in the other enteric bacteria, such as clostridia, bacteroides, and lactobacilli, under the tested conditions. These results, together with our recent finding of a novel metabolic pathway specific for lacto-N-biose I in bifidobacterial cells, suggest that some of the bifidobacterial strains are highly adapted for utilizing human milk oligosaccharides with a type 1 chain.

Bifidobacteria are gram-positive anaerobic bacteria that naturally colonize the human intestinal tract. Generally, the guts of breast-fed infants contain microflora that are dominated by bifidobacteria, in contrast with the contents of the guts of formula-fed infants (5, 14, 16, 46, 58). This increased population of bifidobacteria is believed to be important for infant health (11, 21, 36, 57). The selective growth of bifidobacteria observed in breast-fed newborns has been attributed to oligosaccharides contained in human milk (human milk oligosaccharides [HMOs]), but its molecular basis remains unclear (12, 13, 26, 54).

HMOs are present at 10 to 20 g/liter in human milk (29, 51) and are characterized by their complex structures (15, 27). More than 130 types of HMOs have been isolated so far, and among them, lacto-N-tetraose (Galβ1,3GlcNAcβ1,3Galβ1,4Glc), lacto- N-fucopentaose I (Fucα1,2Galβ1,3GlcNAcβ1,3Galβ1,4Glc), lacto-N-difucohexaose I (Fucα1,2Galβ1,3[Fucα1,4]GlcNAcβ1,3Galβ1,4Glc) and 2′-fucosyllactose (Fucα1,2Galβ1,4Glc) are known to be abundant, especially in colostrums (1, 7, 51). Recently, LoCascio et al. and Ward et al. (32, 54) demonstrated the ability of bifidobacteria to assimilate HMOs as the sole carbon source by conducting growth experiments and mass spectrometric analyses. However, the structural changes to HMOs that occur after fermentation have not been elucidated, and thus, the precise pathway by which HMOs are degraded by the organisms is still unclear.

We have recently found that bifidobacteria possess a unique metabolic pathway specific for lacto-N-biose I (LNB) (type 1 chain) and galacto-N-biose (Galβ1,3GalNAc) (GNB) (Fig. 1) (25). LNB is a building unit of the aforementioned three type 1 HMOs (lacto-N-tetraose, lacto-N-fucopentaose I, and lacto-N-difucohexaose I), and GNB is a core 1 disaccharide of O-glycans of mucin glycoproteins that are present in the human intestines and milk (31, 47). The GNB/LNB pathway involves the following five proteins/enzymes required for the uptake and degradation of the disaccharides: GNB/LNB transporter (50, 53), lacto-N-biose phosphorylase (LnpA) (25, 42), N-acetylhexosamine 1-kinase (NahK) (41), UDP-glucose-hexose 1-phosphate uridylyltransferase (GalT), and UDP-galactose epimerase (GalE). It is likely that LNB and GNB are first imported by GNB/LNB transporter into the cytoplasm and then phosphorolytically cleaved by LnpA into α-galactosylphosphate (Gal 1-P) and the respective N-acetylhexosamines. Finally, Gal 1-P and N-acetylhexosamines should enter glycolysis and the aminosugar metabolic cycle, respectively, via the Leloir-like pathway, consisting of NahK, GalT, and GalE (41). Considering the presence of this GNB/LNB pathway in bifidobacterial cells, the organisms should produce extracellular enzymes to liberate LNB and GNB from natural substrates. In fact, we previously identified an extracellular endo-α-N-acetylgalactosaminidase (EngBF) in Bifidobacterium longum JCM1217 that releases GNB from core 1-type O-glycans in mucin glycoproteins and found that this enzyme is widely distributed in bifidobacteria (10, 22, 24). Thus, examining the occurrence of an enzyme releasing LNB in bifidobacteria is important, not only for a better understanding of the GNB/LNB pathway, but also for elucidating the degradation pathway of HMOs with a type 1 structure. Here, we report the identification, molecular cloning, and characterization of lacto-N-biosidase (LnbB) from Bifidobacterium bifidum JCM1254. The presence of LnbB in several bifidobacterial strains provides insights into how they take nutritional advantage over other microorganisms to effectively colonize the guts of breast-fed newborns.

FIG. 1.

The GNB/LNB pathway and related enzymes. LNB (Galβ1,3GlcNAc) and GNB (Galβ1,3GalNAc) liberated from HMOs with a type 1 chain and mucin glycoproteins by LnbB and EngBF, respectively, might be captured by GL-BP, imported into the cell, and then phosphorolytically cleaved into Gal 1-P and the respective N-acetylhexosamines (GlcNAc and GalNAc). The former can undergo glycolysis via the Leloir-like pathway (GalE and GalT), and the latter can enter the aminosugar metabolic cycle after phosphorylation by NahK. Prior to the actions of LnbB and EngBF, α-1,2-fucosyl residues were removed by AfcA.

MATERIALS AND METHODS

Bacterial strains, media, and chemicals.

The bacterial strains, obtained from the Japan Collection of Microorganisms (JCM), RIKEN Bioresource Center, Japan, were grown anaerobically in GAM medium (Nissui, Japan) or a basal lactose medium consisting of 0.5% yeast extract, 1.0% peptone, 0.5% sodium acetate, 0.2% diammonium citrate, 0.02% magnesium sulfate, 0.2% dipotassium hydrogen phosphate, and 2% lactose with Anaeropack (Mitsubishi Chemical, Japan) at 37°C. The basal lactose medium was supplemented with 4% reduced reagent (2% l-cysteine hydrochloride, 34% l-ascorbic acid, and 11% sodium carbonate) prior to the inoculation of bacteria. LNB (Galβ1,3GlcNAc) and lacto-N-tetraose (Galβ1,3GlcNAcβ1,3Galβ1,4Glc) were purchased from Funakoshi, Japan; p-nitrophenyl (pNP) substrates were gifts from T. Usui of Shizuoka University or were purchased from Sigma; and pyridylamino (PA) sugars were from Takara, Japan, or prepared by the method of Hase et al. (18).

Isolation and sequencing of the LnbB gene (lnbB).

The LnbB gene (lnbB) was isolated from B. bifidum JCM1254 as follows. First, the internal region corresponding to nucleotides 778 to 1128 (the numbering starts at A of the initiation codon) was amplified by PCR using the genomic DNA of B. bifidum JCM1254 as a template and a pair of primers (5′-AGYCCNGGNCAYATG-3′ and 5′-NCCRTCRTTCCADATNCG-3′) that were designed based on the amino acid sequence of LnbB from Streptomyces sp. strain 142 (GenBank accession number U40488). Then, in order to isolate the upstream region, inverse PCR was carried out using BamHI-digested, circularized genomic DNA as a template and a primer pair (5′-GCTTGCCGACAACTCAGGCCGGAAGGATC-3′ and 5′-TACTCCGGGTAGTTCTCCAGCCAGACGTT-3′). The residual 3′ region was obtained by a standard colony hybridization method and cassette PCR using an LA PCR in vitro cloning kit (Takara, Japan) in which the oligonucleotide 5′-GCCACCGGCAACGAGCAGAACATC-3′ was used as a forward primer and the reverse primer was supplied by the manufacturer. These separately isolated DNA segments were sequenced and assembled in silico. To obtain a DNA fragment containing the entire lnbB gene, high-fidelity PCR involving KOD polymerase (Toyobo, Japan) was performed using the genomic DNA as a template and a primer pair (5′-CTCTCCCCGCTGATGTAGGT-3′ and 5′-AAGCTGACCGGCGTACTCTC-3′). The amplified fragment was inserted into the SmaI site of a low-copy-number plasmid, pMW118 (Nippon Gene, Japan), and sequenced.

Expression and purification of the recombinant LnbB.

A DNA fragment corresponding to amino acid residues 35 to 1064 of LnbB was amplified by high-fidelity PCR using the genomic DNA of B. bifidum JCM1254 as a template and a primer pair (5′-GGAATTCCATATGGCCGACGATAGTGCAGCCGGGTAC-3′ and 5′-CCGCTCGAGCTCCGTACCCGGTTTGGTCGG-3′; underlining indicates the designed NdeI and XhoI sites, respectively). The amplified fragment was digested with NdeI and XhoI and inserted into the corresponding sites of pET-23b (Novagen) to generate a C-terminally hexahistidine-tagged protein. After sequence confirmation, the resulting plasmid, pET23b-lnbB, was introduced into Escherichia coli Rosetta(DE3) pLacI (Novagen), and the transformants were incubated in LB medium containing 100 μg/ml ampicillin and 20 μg/ml chloramphenicol at 37°C until the optical density at 600 nm reached 0.5. Isopropyl 1-thio-β-d-galactopyranoside was added at a final concentration of 0.5 mM to induce protein expression. Following additional incubation for 3 h, the cells were harvested by centrifugation and suspended in Bugbuster protein extraction reagent (Novagen). The protein was purified by Ni²⁺-charged HiTrap chelating column chromatography, followed by Superdex 200 10/300 GL column chromatography (GE Healthcare). The protein concentration was determined using a Bio-Rad protein assay kit with bovine serum albumin as a standard. The purity of the protein was estimated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis, followed by Coomassie brilliant blue R250 staining.

LnbB assay.

The standard reaction mixture contained 50 mM citrate-phosphate buffer (pH 4.5), 0.25 mM pNP-Galβ1,3GlcNAc, and the enzyme in a total volume of 40 μl. After incubation for an appropriate time at 25°C, the reaction was stopped by adding 60 μl of 1 M sodium carbonate, and the amount of pNP released was determined by measuring absorbance at 405 nm. One unit of activity was defined as the amount of enzyme releasing 1 μmol of pNP per min. When PA oligosaccharides were used, reactions were carried out in 50 mM citrate-phosphate buffer (pH 6.0), and the products were analyzed by high-performance liquid chromatography (HPLC) using a Hitachi D-2000 Elite chromatograph system equipped with a TSKgel Amide-80 column (4.6 by 250 mm; Tosoh, Japan). Elution was carried out at a flow rate of 1.0 ml/min with a linear gradient between solvent A (90% acetonitrile containing 0.2 M triethylamine acetate, pH 7.3) and solvent B (50% acetonitrile containing 0.2 M triethylamine acetate, pH 7.3) in which the concentration of solvent B was increased from 40% to 70% in 25 min. PA sugars were detected by fluorescence with excitation and emission wavelengths of 310 and 380 nm, respectively.

The LnbB activities of various bacterial strains were examined using lacto-N-tetraose as the substrate. Aliquots (40 μl) of the bacterial cultures grown overnight on the basal lactose medium were added to a reaction mixture (40 μl) consisting of 100 mM potassium phosphate buffer (pH 7.0) and 4 mM substrate and incubated for 3 h to overnight. After the reaction was stopped by boiling, the mixture was deionized by Amberlite MB-3 and analyzed by HPLC with a Sugar-D column (4.6 by 250 mm; Nacalai Tesque, Japan) under a constant flow (1.0 ml/min) of 75% acetonitrile at 40°C. Elution was monitored by measuring the absorbance of the N-acetyl group at 214 nm.

The thermostability of the enzyme was examined by incubating the enzyme at different temperatures for 30 min prior to the assay. For determination of the optimal pH, citrate-phosphate buffer (pH 2.5 to 6.0), sodium phosphate buffer (pH 6.0 to 8.0), and Tris-HCl buffer (pH 8.0 to 9.0) were used at a final concentration of 50 mM. The reaction was initiated by adding the enzyme and continued for a short time (<2 min). The K_m and k_cat values were determined by a double-reciprocal plot of the data.

Transglycosylation and condensation reactions by LnbB.

The transglycosylation by LnbB was performed as described previously (2) using lacto-N-tetraose (1.3 mM) as a donor. The reaction was carried out in 20 mM citrate-phosphate buffer (pH 6.0) in the presence of various 1-alkanols (20%) and the enzyme (2.9 mU). The reaction products were analyzed by thin-layer chromatography (TLC) using a silica gel 60 aluminum sheet (Merck). The plate was developed in a solvent system of chloroform-methanol-water (7/6.2/2), and the carbohydrates were visualized by heating the plate after dipping it in diphenylamine-aniline-phosphoric acid reagent (3).

Synthesis of lacto-N-tetraose by transglycosylation was carried out in 20 mM citrate-phosphate buffer (pH 4.5) containing 5 mM pNP-LNB (donor), 0.5 M lactose (acceptor), and 9.6 mU of LnbB. The condensation reaction was performed in 20 mM citrate-phosphate buffer (pH 6.0) containing 1 M lactose, 0.1 M LNB, and 0.39 U of LnbB. After the reaction was stopped by boiling, the reaction mixture was analyzed by HPLC with a TSKgel Amide-80 column under a constant flow (1.0 ml/min) of 75% acetonitrile at 40°C, and the elution was monitored by measuring the absorbance of the N-acetyl group at 214 nm.

Electrospray ionization-mass spectroscopy (ESI-MS) analysis.

Mass spectra were obtained on an API-100 LC/MS system (Perkin-Elmer Sciex Instruments, Toronto, Canada). The samples were dissolved in 0.1% formic acid-acetonitrile (1/1) and injected at 3 μl/min with a microsyringe pump (Pump 22; Harvard Apparatus, MA). Scanning was done, in the positive mode, from m/z 300 to 1,000 for 1 min (six cycles).

Nucleotide sequence accession number.

The DNA sequence of the B. bifidum lnbB gene has been deposited in GenBank under accession number EU281545.

RESULTS

Occurrence of LnbB in several bifidobacterial strains.

First, we examined several animal- and food-derived bacteria for the ability to liberate LNB (Galβ1,3GlcNAc) from lacto-N- tetraose (Galβ1,3GlcNAcβ1,3Galβ1,4Glc) by directly adding the cultures to the reaction mixtures and analyzing the products by HPLC. Among the bifidobacterial species known to be infant intestinal colonizers, three strains of B. bifidum (JCM1254, JCM1255, and JCM7004) and three of B. longum (JCM1217, JCM1222, and JCM7054) showed activity, but strains of Bifidobacterium breve and Bifidobacterium catenulatum did not (Table 1). The LNB-releasing activity was not detected in the other bifidobacteria and the strains belonging to other genera, such as Bacteroides, Clostridium, and Lactobacillus. Although the results did not directly indicate the complete loss of activity in the negative strains listed in Table 1, because we used bacterial cultures grown under specific conditions, a species-limited occurrence of the activity was suggested by the results. Since B. bifidum JCM1254 exhibited the highest level of activity, we chose it for further studies.

TABLE 1.

Lacto-N-biosidase activities of various bacteria^a

Species	Strain	LNB-releasing activityb
Bifidobacterium adolescentis	JCM1275	−
Bifidobacterium adolescentis	JCM7046	−
Bifidobacterium angulatum	JCM7096	−
Bifidobacterium animalis	JCM10602	−
Bifidobacterium bifidum	JCM1254	+
Bifidobacterium bifidum	JCM1255	+
Bifidobacterium bifidum	JCM7004	+
Bifidobacterium breve	JCM1192	−
Bifidobacterium catenulatum	JCM1194	−
Bifidobacterium dentium	JCM1195	−
Bifidobacterium longum	JCM1210	−
Bifidobacterium longum	JCM1217	+
Bifidobacterium longum	JCM1222	+
Bifidobacterium longum	JCM7054	+
Bifidobacterium pseudocatenulatum	JCM1200	−
Bifidobacterium pseudolongum	JCM1205	−
Bifidobacterium scardovii	JCM12489	−
Bacteroides ovatus	JCM5824	−
Bacteroides thetaiotaomicron	JCM5827	−
Clostridium celatum	JCM1394	−
Clostridium hylemonae	JCM10539	−
Clostridium perfringens	JCM1290	−
Clostridium perfringens	JCM3816	−
Clostridium perfringens	JCM3817	−
Clostridium perfringens	JCM3818	−
Clostridium perfringens	JCM3819	−
Clostridium scindens	JCM6567	−
Enterococcus pseudoavium	JCM8732	−
Eubacterium limosum	JCM6421	−
Lactobacillus casei	JCM1134	−
Lactobacillus gasseri	JCM1130	−
Lactobacillus johnsonii	JCM8794	−
Lactobacillus paracasei	JCM1181	−
Lactobacillus plantarum	JCM1149	−
Lactobacillus reuteri	JCM1112	−
Propionibacterium acnes	JCM6425	−
Ruminococcus productus	JCM1471	−

^aOvernight culture of each bacterial strain grown in the basal lactose medium was examined for the ability to liberate LNB from lacto-N-tetraose. The reaction conditions are described in Materials and Methods.

^b+, detected; −, not detected.

Time-dependent degradation of lacto-N-tetraose into LNB and lactose was observed when we incubated the substrate with the washed cells of B. bifidum JCM1254 and subsequently analyzed the reaction products by TLC (Fig. 2A). Some of the liberated lactose was further cleaved into glucose and galactose. No degradation was observed when boiled cells were used. To verify the liberation of LNB, we performed ESI-MS, in which the molecular ion peak of m/z 406.0 (calculated for a sodium adduct of LNB, 406.1) was detected for the extract from the corresponding spot on the TLC plate (Fig. 2B). The results confirmed the presence of LnbB activity in B. bifidum JCM1254.

FIG. 2.

Liberation of lacto-N-biose I from lacto-N-tetraose upon incubation with B. bifidum JCM1254 cells. (A) TLC analysis of reaction products. Cells grown in GAM medium (1 ml) were harvested by centrifugation and incubated in the reaction mixture (1 ml) containing 0.9 mM lacto-N-tetraose and 10 mM potassium phosphate buffer (pH 7.0) at 37°C. Samples (5 μl) were withdrawn at the indicated times. Lacto-N-tetraose incubated with boiled cells was used as a control. Carbohydrates were visualized by heating the plate after dipping it in orcinol-H₂SO₄ reagent (20). The arrows indicate the positions of standard sugars. (B) ESI-MS analysis of the reaction product. The spot corresponding to LNB was extracted from the TLC plate, lyophilized, and then subjected to ESI-MS analysis. A molecular ion peak appeared at m/z 406.0, which is consistent with the calculated mass for a sodium adduct of LNB (406.1).

Molecular cloning and expression of LnbB from B. bifidum JCM1254.

To isolate the gene encoding LnbB from B. bifidum JCM1254, we designed various degenerated primers based on the amino acid sequence of LnbB from a soil actinomycete, Streptomyces sp. strain 142, which is the only enzyme identified so far (48). We used those primers in different combinations with the genomic DNA of B. bifidum JCM1254 as a template and obtained a 350-bp PCR product. We then analyzed the DNA sequence of this fragment and found that the amino acid sequence it encodes is quite similar to a part of Streptomyces LnbB. The flanking upstream and downstream regions were obtained by inverse PCR, colony hybridization, and cassette PCR and sequenced. Finally, the DNA fragment containing the entire gene was amplified by high-fidelity PCR. The detailed characterization of the gene product is given below.

The gene, designated lnbB, encoded a protein of 1,112 amino acid residues with a predicted molecular mass of 120 kDa. The deduced amino acid sequence contained a signal peptide and a membrane anchor at the N-terminal (amino acids [aa] 1 to 34) and C-terminal (aa 1082 to 1108) ends, as revealed by the SignalP 3.0 (http://www.cbs.dtu.dk/services/SignalP/) (4) and PSORT (http://psort.hgc.jp/form.html) (39) programs, respectively. These results, coupled with the results shown in Fig. 2, strongly suggested that LnbB is a membrane-tethered protein with a large extracellular region. A Pfam search (http://www.sanger.ac.uk/Software/Pfam/) (9) revealed the presence of the glycosyl hydrolase family 20 (GH20) domain (aa 179 to 496), the carbohydrate-binding module 32 (CBM32; aa 784 to 932), and the bacterial immunoglobulin (Ig)-like 2 domain (aa 960 to 1041) in that order from the N terminus (Fig. 3A). The BLAST search (http://www.ncbi.nlm.nih.gov/BLAST/BLAST.cgi) revealed that, among the fully characterized enzymes, the LnbB protein exhibited 38% amino acid identity to LnbB from Streptomyces sp. strain 142 and 20 to 25% identity to β-N-acetylhexosaminidases from Streptomyces plicatus (34, 55), Streptococcus gordonii FSS2 (17), and Homo sapiens (33) (Fig. 3B). β-N-Acetylhexosaminidases and LnbBs are members of GH20, and accordingly, the catalytically important residues identified in HexB (human) (33) and Hex (S. plicatus) (55) are well conserved in sequences of LnbB (Fig. 3B), but the residues involved in the recognition of O-3 and O-6 of N-acetylhexosamine in HexB and Hex are not conserved in LnbB, which might reflect the difference in substrate specificity between the two enzymes (N-acetylglucosamine versus galactosyl-β1,3-N-acetylglucosamine), although further structural and biochemical studies are required to draw a conclusion.

FIG. 3.

Primary structure of LnbB from B. bifidum JCM1254. (A) Schematic representation of the domain structure. The amino acid numbering starts at the probable initiation codon. The domain of GH20, CBM32, and bacterial Ig-like 2 domain are depicted as dark-gray, shaded, and light-gray boxes, respectively. The black bars at the N-terminal and C-terminal ends indicate a signal peptide and a membrane anchor, respectively. (B) Multiple alignment of GH20 domains of LnbB from B. bifidum JCM1254 and its homologues, created by TCOFFEE (44) and BoxShade 3.21. The numbering starts at the initiation codon of each protein. Identical residues and conserved substitutions are highlighted in black and dark gray, respectively. The general acid/base residues identified by the structural and biochemical analyses of N-acetylhexosaminidases are marked by asterisks, and the residues involved in the binding of O-3, O-4, and O-6 of N-acetylhexosamine and the stacking tryptophan at the −1 subsite are enclosed by boxes. The organisms and accession numbers are as follows; BB, LnbB from B. bifidum JCM1254; SS, LnbB from Streptomyces sp. strain 142 (U40488); SP, β-N-acetylhexosaminidase from S. plicatus (AF063001); HS, β-N-acetylhexosaminidase from H. sapiens (NM_000521); SG, β-N-acetylhexosaminidase from S. gordonii FSS2 (AY450645).

No homologues with significant similarity were found in the microbial genomes in the NCBI database, including B. longum strains (NCC2705 [49] and DJO10A) and B. adolescentis strains (ATCC 15703 and L2-32). However, of interest is the fact that a partial contig sequence with 87% identity was discovered in an infant gut metagenome (30), suggesting the presence of LnbB-positive bacteria in the environment, i.e., an infant gut.

We then expressed a truncated form of LnbB (aa 35 to 1064, missing the signal peptide and membrane anchor) in E. coli and purified it as the C-terminally His₆-tagged protein. The final protein preparation migrated as a single band with an apparent molecular mass of 110 kDa on sodium dodecyl sulfate-polyacrylamide gel electrophoresis (data not shown), which is in agreement with the calculated mass of 112 kDa.

Properties of the recombinant LnbB protein.

To verify the enzyme activity of recombinant LnbB, we first incubated lacto-N-tetraose with the purified protein and analyzed the reaction products by ESI-MS. In the control experiment without the enzyme, a molecular ion peak of m/z 730.8 appeared, which corresponds to a sodium adduct of lacto-N-tetraose (calculated, 730.2) (Fig. 4A). On the other hand, upon incubation with the enzyme, the peak of lacto-N-tetraose disappeared, and instead, two intense mass ion peaks appeared at m/z 365.5 and m/z 406.5, which could be sodium adducts of lactose (calculated mass, 365.1) and LNB (calculated mass, 406.1), respectively (Fig. 4B). This indicated that the isolated gene lnbB indeed encodes LnbB.

FIG. 4.

Hydrolysis of lacto-N-tetraose into LNB and lactose by the recombinant LnbB. Lacto-N-tetraose was incubated in the absence (A) and presence (B) of the purified enzyme, and the reaction mixtures were subjected to ESI-MS. The molecular ion peak at m/z 730.8 corresponds to a sodium adduct of lacto-N-tetraose (calculated, 730.2) (A), and the molecular ion peaks of 365.5 and 406.5 are sodium adducts of lactose (calculated, 365.1) and LNB (calculated, 406.1) (B), respectively.

Next, we analyzed the substrate preference of LnbB using various chromogenic and fluorescently labeled saccharides (Table 2). Among the pNP sugars tested, the enzyme was most active toward pNP-β-LNB and less active toward pNP-β-GNB (30% compared to pNP-β-LNB) but did not hydrolyze α-linked disaccharides. The LnbB protein did not act on β-linked pNP monosaccharides, including pNP-β-GlcNAc and pNP-β-GalNAc. With respect to PA oligosaccharides, the enzyme acted on lacto-N-tetraose, but not on the ganglioside GA1 structure with a β-linked GNB. This is in sharp contrast to the results obtained for the pNP disaccharides and may be due to inaccessibility of GA1 to the catalytic pocket because of a steric constraint formed by the linkage between C-1 of GalNAc and an axial hydroxyl group of d-galactose (Gal). The enzyme did not hydrolyze the fucosylated forms of lacto-N-tetraose (lacto-N-fucopentaose I and II) or lacto-N-neotetraose (type 2 chain). These results revealed that LnbB from B. bifidum has substrate preference for unmodified β-linked LNB.

TABLE 2.

Substrate preferences of the recombinant LnbB

Sugar	Substratea	Relative activity (%)
pNP sugars	Galβ1,3GlcNAcβ-pNP	100b
	Galβ1,3GalNAcβ-pNP	30
	GlcNAcβ-pNP	NDc
	GalNAcβ-pNP	ND
	Galβ1,3GlcNAcα-pNP	ND
	Galβ1,3GalNAcα-pNP	ND
PA sugars
LNT	Galβ1,3GlcNAcβ1,3Galβ1,4Glc-PA	100d
GA1	Galβ1,3GalNAcβ1,4Galβ1,4Glc-PA	ND
LNFP I	Fucα1,2Galβ1,3GlcNAcβ1,3Galβ1,4Glc-PA	ND
LNFP II	Galβ1,3(Fucα1,4)GlcNAcβ1,3Galβ1,4Glc-PA	ND
LNnT	Galβ1,4GlcNAcβ1,3Galβ1,4Glc-PA	ND

^aGal, d-galactose; Glc, d-glucose; Fuc, l-fucose; GalNAc, N-acetyl-d-galactosamine; GlcNAc, N-acetyl-d-glucosamine; LNT, lacto-N-tetraose; GA1, ganglioside GA1; LNFP I, lacto-N-fucopentaose I; LNFP II, lacto-N-fucopentaose II; LNnT, lacto-N-neotetraose.

^bThe value obtained with lacto-N-biosyl-β-pNP (250 μM) was taken as 100%.

^cND, not detectable.

^dThe value obtained with LNT-PA (1.0 μM) was taken as 100%.

The enzyme exhibited maximum activity at pH 4.5 for pNP-β-LNB and pH 6.0 for PA-lacto-N-tetraose, reflecting the difference in the leaving groups. The protein was stable between pH 4 and 9 and retained 88% activity after incubation at 40°C for 30 min. The K_m and k_cat values for pNP-β-LNB were determined to be 68 μM and 89 s⁻¹, respectively. The properties of LnbB from B. bifidum were quite similar to those of LnbB from Streptomyces sp. strain 142 (48), though the sequence identity was not so high between them, as mentioned above.

Transglycosylation and condensation activities of LnbB.

LnbB is a member of GH20, and therefore, its hydrolysis must proceed through a retaining mechanism (19). It is known that some retaining glycosidases show, in addition to hydrolytic activity, transglycosylation activity, in which the glycon moieties are transferred to appropriate acceptors with hydroxyl groups. The transglycosylation should serve as a powerful tool for creating neoglycoconjugates having biological activities (52, 56). We first examined the ability of LnbB to catalyze the transglycosylation by incubating lacto-N-tetraose (a donor substrate) with various 1-alkanols (acceptors). As expected, new spots appeared on the TLC plate when the acceptor molecules were added to the reaction mixtures, indicating the transglycosylation ability of LnbB (Fig. 5A). We also employed this activity for the synthesis of lacto-N-tetraose using pNP-β-LNB and lactose as a donor and an acceptor, respectively. When we incubated the reaction mixture in the presence of the enzyme and subsequently analyzed it by HPLC, two new peaks appeared at retention times of 12 and 46 min (Fig. 5B, a, bottom), in contrast to the control experiment without the enzyme (Fig. 5B, a, top). The peak at a retention time of 12 min was the liberated LNB, and that at 46 min could correspond to lacto-N-tetraose, since its retention time was consistent with that of the authentic sample of this compound (data not shown). We collected this peak fraction and subjected it to ESI-MS, detecting a mass ion peak of m/z 730.2, which corresponds to the sodium adduct of lacto-N-tetraose (calculated mass, 730.2) (data not shown). These results strongly suggest that LnbB is able to catalyze the synthesis of lacto-N-tetraose by transglycosylation.

FIG. 5.

Transglycosylation and condensation activities of the recombinant LnbB. (A) Transglycosylation to various 1-alkanols. Lacto-N-tetraose (donor) was incubated without (lane 2) and with (lanes 3 to 6) 1-alkanols in the presence of the enzyme. Lane 1, standard sugars (LNT, lacto-N-tetraose). The acceptors were as follows: lane 3, methanol; lane 4, ethanol; lane 5, 1-propanol; and lane 6, 1-butanol. (B) HPLC analyses of transglycosylation (a) and condensation (b) reactions of LnbB. pNP-β-LNB (donor) was incubated with lactose (acceptor) in the absence (top) and presence (bottom) of the enzyme (a), and LNB and lactose were incubated in the absence (top) and presence (bottom) of the enzyme (b). The elution was monitored by measuring the absorbance of the N-acetyl group at 214 nm.

Next, we examined the condensation activity of LnbB by incubating LNB and lactose in the presence of the purified enzyme. Interestingly, two new peaks appeared at retention times of 46 and 55 min (Fig. 5B, b, bottom), both of which gave molecular ion peaks of m/z 730.2 in ESI-MS analyses (data not shown), whereas no peak other than LNB appeared in the absence of enzyme (Fig. 5B, b, top). Judging from the HPLC profile, the earlier peak might correspond to lacto-N-tetraose, and the latter is an isomer probably having a β-1,6 linkage between the LNB residue and the lactose residue.

DISCUSSION

Physiological role of LnbB in bifidobacteria.

One of the widely reported functions of nondigestible HMOs is the selective promotion of the growth of bifidobacteria in the infant gut ecosystem (12, 13, 54). However, questions about what structure of HMOs exerts the bifidogenic effect, and how, have remained unanswered. In this study, by isolating the gene and examining the properties of LnbB from B. bifidum, we provide a possible explanation of how some of the bifidobacterial strains are highly adapted for utilizing HMOs with a type 1 chain.

As mentioned above, human milk is rich in lacto-N-tetraose, lacto-N-fucopentaose I, and lacto-N-difucohexaose I (1, 7, 51). The recombinant LnbB was highly specific to unmodified lacto-N-tetraose and did not hydrolyze when a substituent l-fucose was attached to lacto-N-tetraose. It is likely, therefore, that in the breast-fed infant's gut, lacto-N-fucopentaose I and lacto-N-difucohexaose I may be first digested by an extracellular α-l-fucosidase (AfcA) (Fig. 1) (23, 38) and then cleaved by extracellular LnbB into LNB and lactose. The liberated LNB can enter the GNB/LNB pathway in cells, and the remaining lactose can be either degraded by extracellular β-galactosidase (Fig. 2A) (35) or imported by a lactose transporter (28, 45), or perhaps both.

Clostridia, possible competitors of bifidobacteria in the gut ecosystem, have proteins homologous to LnpA from bifidobacteria, a key enzyme of the GNB/LNB pathway (6, 25). It has recently been elucidated that the homologue (CPF0553) from Clostridium perfringens ATCC 13124 showed a strong preference for GNB over LNB (a more than 50-fold higher k_cat/K_m value), while LnpA from B. longum JCM1217 shows identical k_cat/K_m values for the two disaccharides (40). We observed endo-α-N-acetylgalactosaminidase activities in some of the clostridial strains (JCM3816, JCM3818, and JCM3819) (data not shown) but did not observe LnbB activities under the tested conditions. Thus, if clostridia were to have a complete GNB/LNB pathway, it could be specific for the degradation of GNB liberated from sugar chains of mucin glycoproteins. It should be noted that GL-BP (the substrate-binding protein of the GLB/LNB transporter) homologues have not been found in any bacterial genomes, except for bifidobacteria (50, 53) and Propionibacterium acnes KPA171202. Taken together, the presence of LnbB and the GNB/LNB pathway in some bifidobacterial strains could give the organisms a nutritional advantage to increase their populations in the ecosystem of breast-fed newborns.

Structural features of LnbB.

LnbB activity was found mostly at the cell surface and slightly in the culture broth of B. bifidum JCM1254 (data not shown). Consistent with this observation, analysis of the primary structure of LnbB revealed the presence of a signal peptide and membrane-anchoring motif at its N and C termini, respectively. The catalytic domain of LnbB is likely to be located at amino acid residues 30 to 650, because this region contains the GH20 domain and has 38% identity with the full-length Streptomyces LnbB (640 amino acid residues, including a signal peptide).

CBM32 and bacterial Ig-like domain 2 are frequently found in the surface-located enzymes of bacteria. CBM32 is generally involved in binding to galactose-containing saccharides (8), suggesting that CBM32 of LnbB might interact with and affect affinity for its substrate type 1 HMOs. Though the role of the Ig-like domain in the activity of LnbB remains to be elucidated, it is likely that the domain at least acts to display the GH20 domain so that it protrudes from the cell surface, thereby enabling the cells to gain access to the substrates.

Synthesis of lacto-N-tetraose.

In spite of their biological significance, HMOs are rarely studied, primarily due to their enormous cost. The synthesis of lacto-N-tetraose by transglycosylation was first described by Murata et al., using partially purified LnbB from Aureobacterium sp. strain L-101 (37). In the present work, we found that LnbB from B. bifidum is also capable of synthesizing lacto-N-tetraose by both transglycosylation and condensation activities, though its efficacy is quite low at present. Because large-scale synthesis of LNB has been recently accomplished (43), the enzymatic synthesis of HMOs using LnbB is worth further exploration.

Concluding remarks.

Our studies propose that some bifidobacterial strains have a unique pathway for the degradation of HMOs with a type 1 chain (β-linked LNB). Among mammalian milk oligosaccharides, those of H. sapiens are known to be especially rich in the type 1 structure (1, 51). LnbB activity was found in the strains of B. longum and B. bifidum that naturally colonize infants' intestines but was not found in Bifidiobacterium animalis and Bifidiobacterium pseudolongum, which are frequently isolated from the guts of domestic animals, or in the other enteric bacteria. It is likely, therefore, that the ability to assimilate type 1 HMOs is limited to certain species of bifidobacteria, probably to certain strains of B. bifidum and B. longum, as mentioned by LoCascio et al. and Ward et al. (32, 54). Future studies must take into account whether the strains of bifidobacteria are isolated from breast-fed infants or bottle-fed infants, which may be correlated with the presence of LnbB, AfcA, and the GNB/LNB pathway.