Characterization of Three β-Galactoside Phosphorylases from Clostridium phytofermentans

Abstract

We characterized three d-galactosyl-β1→3-N-acetyl-d-hexosamine phosphorylase (EC 2.4.1.211) homologs from Clostridium phytofermentans (Cphy0577, Cphy1920, and Cphy3030 proteins). Cphy0577 and Cphy3030 proteins exhibited similar activity on galacto-N-biose (GNB; d-Gal-β1→3-d-GalNAc) and lacto-N-biose I (LNB; d-Gal-β1→3-d-GlcNAc), thus indicating that they are d-galactosyl-β1→3-N-acetyl-d-hexosamine phosphorylases, subclassified as GNB/LNB phosphorylase. In contrast, Cphy1920 protein phosphorolyzed neither GNB nor LNB. It showed the highest activity with l-rhamnose as the acceptor in the reverse reaction using α-d-galactose 1-phosphate as the donor. The reaction product was d-galactosyl-β1→4-l-rhamnose. The enzyme also showed activity on l-mannose, l-lyxose, d-glucose, 2-deoxy-d-glucose, and d-galactose in this order. When d-glucose derivatives were used as acceptors, reaction products were β-1,3-galactosides. Kinetic parameters of phosphorolytic activity on d-galactosyl-β1→4-l-rhamnose were k_cat = 45 s⁻¹ and K_m = 7.9 mm, thus indicating that these values are common among other phosphorylases. We propose d-galactosyl-β1→4-l-rhamnose phosphorylase as the name for Cphy1920 protein.

Phosphorylases are a group of enzymes involved in formation and cleavage of glycoside linkage together with glycoside hydrolases and glycosyl-nucleotide glycosyltransferases (synthases). Phosphorylases, which reversibly phosphorolyze oligosaccharides to produce monosaccharide 1-phosphate, are generally intracellular enzymes showing strict substrate specificity. Physiologically, such strict substrate specificity is considered to be closely related to the environment containing bacteria possessing them. For example, d-galactosyl-β1→3-N-acetyl-d-hexosamine phosphorylase (GalHexNAcP²; EC 2.4.1.211) from Bifidobacterium longum, an intestinal bacterium, forms part of the pathway metabolizing galacto-N-biose (GNB; d-Gal-β1→3-d-GalNAc) from mucin and lacto-N-biose I (LNB; d-Gal-β1→3-d-GlcNAc) from human milk oligosaccharides, both of which are present in the intestinal environment, with GNB- and LNB-releasing enzymes and GNB/LNB transporter (1–8). Another example is cellobiose phosphorylase from Cellvibrio gilvus, which is a cellulolytic bacterium. Cellobiose phosphorylase forms an important cellulose metabolic pathway with an extracellular cellulase system producing cellobiose (9, 10).

The reversible catalytic reaction of phosphorylases is one of the most remarkable features that make them suitable catalysts for practical syntheses of oligosaccharides. An oligosaccharide can be produced from inexpensive material by combining reactions of two phosphorylases, one for phosphorolyzing the material and the other for synthesizing the oligosaccharide, in one pot. Based on this idea, LNB is synthesized on a large (kg) scale using sucrose phosphorylase and GalHexNAcP (11). Practical synthesis methods of trehalose and cellobiose have also been developed (12, 13). However, only 14 kinds of substrate specificities have been reported among phosphorylases (13), thus restricting their use. Therefore, it would be useful to find a phosphorylase with novel activity.

GalHexNAcP phosphorolyzes GNB and LNB to produce α-d-galactose 1-phosphate (Gal 1-P) and the corresponding N-acetyl-d-hexosamine. To date, GalHexNAcP is the only phosphorylase known to act on β-galactoside. This enzyme was first found in the cell-free extract of Bifidobacterium bifidum (14) and then in B. longum (1, 15), Clostridium perfringens (16), Propionibacterium acnes (17), and Vibrio vulnificus (18). These studies revealed that GalHexNAcPs were classified into three subgroups based on substrate preference between GNB and LNB. These subgroups are as follows: 1) galacto-N-biose/lacto-N-biose I phosphorylase (GLNBP), showing similar activity on both GNB and LNB (B. longum and B. bifidum); 2) galacto-N-biose phosphorylase (GNBP), preferring GNB to LNB (C. perfringens and P. acnes); and 3) lacto-N-biose I phosphorylase (LNBP), preferring LNB to GNB (V. vulnificus) (18). The ternary structure of GLNBP from B. longum (GLNBP_Bl) has been revealed recently (19). Based on the similarity in ternary structures between GLNBP_Bl and β-galactosidase from Thermus thermophilus, which belongs to glycoside hydrolase family 42 (19, 20), GalHexNAcP homologs are classified as GH112 (glycoside hydrolase family 112), although phosphorylases are glycosyltransferases (21, 22).

Clostridium phytofermentans is an anaerobic cellulolytic bacterium. It is found in soil and grows optimally at 37 °C (23). Its whole genome sequence has been revealed (GenBank^TM accession number CP000885). The bacterium possesses three GalHexNAcP homologous genes (cphy0577, cphy1920, and cphy3030 genes; GenBank^TM accession numbers are ABX40964.1, ABX42289.1, and ABX43387.1, respectively). C. phytofermentans has the ability to utilize a wide range of plant polysaccharides (23), and substrate specificities of these three gene products (Cphy0577, Cphy1920, and Cphy3030 proteins) are considered to be responsible for this ability. Furthermore, the three proteins have not been clearly categorized as GLNBP, GNBP, or LNBP, based on the phylogenetic tree shown in Fig. 1.

FIGURE 1.

Phylogenetic tree of GalHexNAcP homologs in GH112. Multiple alignment was performed using ClustalW2 (available on the World Wide Web). A phylogenetic tree was constructed using Treeview version 1.6.6. The proteins characterized in this study are represented with boldface letters in boxes with a heavy outline. The other proteins are numbered serially in boxes. Characterized GLNBP, GNBP, and LNBP are represented with boldface black letters on a gray background, boldface white letters on a gray background, and boldface white letters on a black background, respectively. Organisms and GenBank^TM accession numbers of numbered proteins are as follows: 1, CPF0553 (C. perfringens ATCC13124, ABG83511.1) (16); 2, CPE0573 (C. perfringens str.13, BAB80279.1); 3, CPR0537 (C. perfringens SM101, ABG86710.1); 4, LnpA2 (B. bifidum JCM1254, BAE95374.1) (14, 15); 5, LnpA1 (B. bifidum JCM1254, BAD80752.1) (14, 15); 6, GLNBP_Bl (B. longum subsp. longum JCM 1217, BAD80751.1) (1, 16); 7, Blon_2174 (B. longum subsp. infantis ATCC 15697, ACJ53235.1); 8, BL1641 (B. longum NCC2705, AAN25428.1); 9, BLD_1765 (B. longum DJO10A, ACD99210.1); 10, GnpA (P. acnes JCM6473, AB468065) (17); 11, GnpA (P. acnes JCM6425, AB468066) (17); 12, PPA0083 (P. acnes KPA171202, AAT81843.1); 13, VV2_1091 (V. vulnificus CMCP6, AAO07997.1) (18); 14, VVA1614 (V. vulnificus YJ016, BAC97640.1); 15, Oter_1377 (Opitutus terrae PB90-1, ACB74662.1); 16, BCQ_1989 (B. cereus Q1, ACM12417.1); 17, BCAH187_A2105 (Bacillus cereus AH187, ACJ78918.1).

In this study, we characterized the three proteins. We reported that two of them were GalHexNAcPs and that the other was a β-galactoside phosphorylase showing unique substrate specificity.

EXPERIMENTAL PROCEDURES

Cloning, Expression, and Purification

C. phytofermentans was purchased from the American Type Culture Collection (Manassas, VA). C. phytofermentans was cultured anaerobically using Anaeropack (Mitsubishi Gas Chemical Co, Ltd., Tokyo, Japan) at 30 °C for 4 days on a modified GAM agar plate (Nissui Pharmaceutical Co., Tokyo, Japan). Genomic DNA of C. phytofermentans were obtained from colonies on the plate, using InstaGene^TM matrix (Bio-Rad). GalHexNAcP homologous genes were amplified by PCR using KOD plus DNA polymerase (Toyobo, Osaka, Japan) and genomic DNA of C. phytofermentans as the template. Primer pairs for cphy0577, cphy1920, and cphy3030 genes were forward primer 5′-ggtggtaaccatggaaaaagatacgatgtt-3′ and reverse primer 5′-acctaattgacactcgagctcatacc-3′, forward primer 5′-aggagaaatcatatggaacaacaaaaggag-3′ and reverse primer 5′-attatgtctcgagtaatggaagaattacgg-3′, and forward primer 5′-gataaataccatgggtgaaaaattaacagg-3′ and reverse primer 5′-gatgttaatctcgagatcctctaaaaatg-3′, respectively. These forward primers contained NcoI or NdeI sites (underlined), and the reverse primers contained XhoI sites (underlined). Amplified cphy0577 and cphy3030 genes were inserted into the NcoI and XhoI sites of pET28a(+) (Novagen, Madison, WI), and the amplified cphy1920 gene was inserted into the NdeI and XhoI sites of pET30a(+) (Novagen, Madison, WI) to add His₆ tags at the C termini. Escherichia coli BL21(DE3) was transformed by each constructed plasmid. Each transformant was grown in 150 ml of Luria-Bertani medium containing 30 μg/ml kanamycin at 37 °C to an absorbance of 0.8 at 660 nm in a 500-ml conical flask; the transformant was then treated with 0.1 mm isopropyl-β-d-thiogalactopyranoside. After induction, the transformant for the cphy0577 gene was cultured at 30 °C for 5 h, and the other two were cultured at 20 °C overnight with agitation at 160 rpm. Wet cells were collected by centrifugation at 5000 × g for 5 min, and then each sample was suspended in 20 mm sodium phosphate buffer (pH 7.0) containing 500 mm NaCl (buffer A). Each sample was disrupted by sonication and centrifuged at 15,000 × g for 15 min and was then loaded on an Ni²⁺-nitrilotriacetic acid-agarose (7 ml) (Qiagen, Hilden, Germany) column equilibrated with buffer A containing 20 mm imidazole. The column was washed with buffer A containing 20 mm imidazole until almost all unbound components were removed, and each recombinant protein was eluted with a linear gradient of 20–250 mm imidazole at a flow rate of 1.8 ml/min. The buffer of each sample was then changed to 20 mm MOPS buffer (pH 7.0) using an Amicon Ultra 10,000 molecular weight cut-off (Millipore, Billerica, MA). Finally, 22, 12, and 16 mg of recombinant Cphy0577, Cphy1920, and Cphy3030 proteins were obtained from 150 ml of culture medium, respectively. Single bands were detected at 80 kDa on SDS-PAGE, corresponding to the calculated molecular mass of recombinant proteins with His₆ tag (84864, 84232, and 84937 Da, respectively).

Measurement of Phosphorolytic and Synthetic Activity

Phosphorolytic activity was determined by quantifying the Gal 1-P produced from LNB, GNB, or d-galactosyl-β1→4-l-rhamnose (GalRha) (preparation method is described below) by the method of Nihira et al. (24). Synthetic activity was determined by measuring the increase in phosphate in the reaction mixture containing 10 mm Gal 1-P and 10 mm acceptor in 100 mm MOPS buffer (pH 7.0), using the method of Lowry and Lopez (25) as described below. The substrate solution (142.5 μl; 10 mm Gal 1-P and 10 mm acceptor in 100 mm MOPS buffer (pH 7.0)) was preincubated at 37 °C for 10 min and then mixed with 7.5 μl of an enzyme solution to start the reaction. An aliquot (12.5 μl) was added to 100 μl of 0.2 m sodium acetate (pH 4.0) to stop the enzymatic reaction. Then 12.5 μl of 1% ammonium molybdate containing 25 mm sulfuric acid and 12.5 μl of 1% ascorbic acid with 0.05% potassium bisulfate were mixed with the samples. The mixtures were incubated at 37 °C for 1 h, and the concentration of phosphate formed was quantified by measuring absorbance at 700 nm.

Temperature and pH Profile

Thermostability was defined as residual activity after incubation of enzyme (2.2 mg/ml) at various temperatures in 20 mm MOPS buffer (pH 7.0) for 30 min. Residual activities were determined by measuring synthetic activities using d-GlcNAc as an acceptor for Cphy0577 and Cphy3030 proteins and l-rhamnose (l-Rha) as an acceptor for Cphy1920 protein. When pH activity was determined, synthetic activity was measured in various 100 mm buffers (NaOAc, pH 4.5–5.5; MES, pH 5.5–6.5; MOPS, pH 6.5–7.5; Tris-HCl, pH 7.5–9.0). The acceptors used were the same as those in thermostability experiments. Samples were taken every 0.7 min for 4.9 min to determine phosphate concentration. The time course increase in phosphate concentration was fitted with linear regression to determine the reaction velocity.

Thin Layer Chromatography

An assay was performed using 10 μl of substrate solution. The reaction mixture (1 μl) was spotted onto a TLC plate (Kieselgel 60 F₂₅₄; Merck), and the sample was developed with a solution of acetonitrile/water (3:1, v/v). The TLC plates were briefly soaked in 5% sulfuric acid/methanol solution and heated in an oven until bands were sufficiently visualized.

Structural Analysis of the Products of Cphy1920 Protein

Each reaction mixture (1 ml) containing 50 mm Gal 1-P, 50 mm acceptor (l-Rha, l-mannose (l-Man), l-lyxose (l-Lyx), d-glucose (d-Glc), 2-deoxy-d-glucose (d-Glc2d), or d-Gal), and Cphy1920 protein (6.9 μg for l-Rha and 23 μg for other substrates) in 100 mm MOPS (pH 7.0) was incubated at 30 °C (overnight for l-Rha and 3 days for other substrates). The reaction mixtures were loaded onto a Toyopearl HW-40F column (2.6 cm ϕ × 90 cm; TOSOH, Tokyo, Japan) pre-equilibrated with distilled water, and the products were eluted by distilled water at a flow rate of 1.8 ml/min. Fractions containing a disaccharide were collected and desalted with Amberlite MB3 (Organo, Tokyo, Japan), followed by lyophilization. The amount of products obtained were 7.9, 9.1, 9.1, 10.1, 9.4, and 12.5 mg, respectively. The one-dimensional (¹H and ¹³C) and two-dimensional (double-quantum-filtered COSY, heteronuclear single-quantum coherence, and heteronuclear multiple-bond correlation) NMR spectra of the products were taken in D₂O, using a Bruker Avance500 spectrometer with 2-methyl-2-propanol as an internal standard. Proton signals were assigned based on double-quantum-filtered COSY spectra; ¹³C signals were assigned with heteronuclear single-quantum coherence spectra, based on assignment of proton signals. The position of linkage in each disaccharide was determined by detecting interring cross-peaks in each heteronuclear multiple-bond correlation spectrum.

Kinetic Analysis

In order to determine kinetic parameters, assays of phosphorolytic reactions were performed with various concentrations of LNB or GNB in the presence of 10 mm phosphate for Cphy0577 and Cphy3030 proteins and with various concentrations of GalRha and phosphate for Cphy1920 protein. Assays of synthetic reactions were performed with various concentrations of acceptors in the presence of 10 mm Gal 1-P. The kinetic parameters were calculated by curve fitting the experimental data with the theoretical equation, using Grafit version 4 (Erithacus Software, Middlesex, UK).

Anomeric Specificity of Cphy1920 Protein

The substrate solution was prepared by mixing 5 μl of 500 mm acceptors (l-Rha or d-Glc) and 25 μl of 100 mm Gal 1-P (adjusted pH at 7.2 with 1 n HCl) and keeping the mixture at room temperature (24 °C). The reaction was started by adding 20 μl of enzyme solution in 20 mm MOPS (pH 7.0; 0.48 mg/ml enzyme for l-Rha and 19 mg/ml enzyme for d-Glc), kept at room temperature, to the substrate solution (final concentration of acceptors and Gal 1-P was 50 mm). The anomeric forms of each substrate and product were quantified using high performance liquid chromatography (HPLC) under the following conditions. After 1- or 20-min incubation, 10 μl of the reaction mixture was injected onto a TSK amide-80 column (4.6-mm internal diameter × 25 cm; Tosoh, Tokyo, Japan) equilibrated with 75% acetonitrile. Samples were eluted with 75% acetonitrile at a flow rate of 1.5 ml/min. Saccharides were detected using a refractive index detector, RI model 504 (GL Science, Tokyo, Japan). Peaks of both anomers of the disaccharides were distinguished based on the existence ratio of the α-anomer over the β-anomer determined by NMR.

RESULTS

Sequence Analysis

Multiple alignment of GLNBP, GNBP, LNBP, and three GalHexNAcP homologs from C. phytofermentans is shown in Fig. 2. Because catalytic and substrate recognition residues of GLNBP_Bl have already been identified from its ternary structure (19), we compared the corresponding residues of these proteins. The catalytic proton donor of GLNBP_Bl (Asp³¹³) was conserved in Cphy0577, Cphy1920, and Cphy3030 proteins (Fig. 2). Recognition residues of galactose at subsite −1 (Asn¹⁶⁶, Asp³¹³, Tyr³⁶², and Phe³⁶⁴), phosphate-binding residues (Arg³², Arg²¹⁰, and Arg³⁵⁸), and residues that recognize the oxygen atom in the N-acetyl group of acceptors (Trp²³³) in GLNBP_Bl were also conserved in the other five proteins (Fig. 2). Val¹⁶² of GLNBP_Bl is a primary candidate for determining substrate preference between GNB and LNB. The residue is conserved in LNBP from V. vulnificus but is substituted by threonine in GNBP from C. perfringens (19) (Fig. 2). Because Val¹⁶² of GLNBP_Bl is conserved in Cphy0577 and Cphy3030 proteins, they were predicted to be GLNBP or LNBP. On the contrary, substrate specificity of Cphy1920 protein was unpredictable, because the corresponding residue of Cphy1920 protein was isoleucine, and Cphy1920 protein had low identity with the other proteins around the proton donor aspartate residue (Fig. 2).

FIGURE 2.

Multiple alignment of GalHexNAcP homologs. Multiple alignment was performed using ClustalW2 (available on the World Wide Web). N-terminal regions of the alignment containing a catalytic residue and substrate recognition residues are shown. Amino acids conserved with more than four sequences in the alignment are represented in white letters on a black background. GLNBP, GLNBP_Bl; GNBP, GNBP from C. perfringens; LNBP, LNBP from V. vulnificus; square, residues used for prediction of substrate specificity; star, proton donor of GLNBP_Bl; circle, phosphate-binding sites of GLNBP_Bl; inverse triangle, galactose-binding sites of GLNBP_Bl; triangle, N-acetyl group recognition site of GLNBP_Bl.

Cphy0577, Cphy1920, and Cphy3030 proteins are predicted not to possess N-terminal signal peptides as well as known GalHexNAcPs (16–18) in accordance with the SignalP3.0 server (available on the World Wide Web) (26, 27). This suggests that these three proteins are cytosolic.

Phosphorolytic Activity of Cphy0577, Cphy1920, and Cphy3030 Proteins

The phosphorolytic activity of recombinant Cphy0577, Cphy1920, and Cphy3030 proteins on LNB and GNB was examined. Specific activities of Cphy0577 protein on LNB and GNB at 37 °C were 19 and 20 units/mg, respectively. Cphy3030 protein showed a specific activity of 21 and 14 units/mg on LNB and GNB, respectively, at 30 °C (Cphy3030 protein is not stable at 37 °C). Cphy0577 and Cphy3030 proteins did not exhibit phosphorolytic activity either on lacto-N-tetraose (Galβ1→3GlcNAcβ1→3Galβ1→4Glc) or galacto-N-tetraose (Galβ1→3GalNAcβ1→4Galβ1→4Glc). These results indicate that both enzymes are GalHexNAcPs. However, Cphy1920 protein did not phosphorolyze either LNB or GNB, thus indicating that Cphy1920 protein is not GalHexNAcP.

Kinetic Parameters of Cphy0577 and Cphy3030 Proteins

Kinetic parameters for forward and reverse reactions of Cphy0577 and Cphy3030 proteins were examined. Cphy0577 protein exhibited similar kinetic parameters on LNB and GNB (Table 1). On the other hand, Cphy3030 protein showed similar k_cat values on LNB and GNB (Table 1). The value of K_m on LNB was ∼4 times less than that on GNB, and the k_cat/K_m value on LNB was ∼4 times more than that on GNB (Table 1). Kinetic parameters for the reverse reaction of both enzymes showed a similar tendency for the forward reaction (Table 1). Because both enzymes showed significant activities on both LNB and GNB, we identified them as GLNBPs.

TABLE 1.

Kinetic parameters of Cphy0577 and Cphy3030 proteins

Values were obtained by regressing data with the following equation using Grafit version 4.0.10: v = k_cat[E₀][S]/(K_m + [S]), where S represents LNB, GNB, d-GlcNAc, or d-GalNAc.

Protein	Reaction	Substrate	kcat	Km	kcat/Km
			s−1	mm	s−1mm−1
Cphy0577	Phosphorolysisa	LNB	40 ± 2	4.5 ± 0.4	8.8 ± 0.4
		GNB	50 ± 2	8.0 ± 0.6	6.3 ± 0.2
	Synthesisb	d-GlcNAc	49 ± 2	3.4 ± 0.3	14 ± 1
		d-GalNAc	48 ± 2	3.3 ± 0.2	14 ± 1
Cphy3030c	Phosphorolysisa	LNB	39 ± 1	2.8 ± 0.2	14 ± 1
		GNB	39 ± 2	10 ± 1	3.8 ± 0.2
	Synthesisb	d-GlcNAc	42 ± 3	1.9 ± 0.4	22 ± 4
		d-GalNAc	34 ± 2	3.3 ± 0.5	10 ± 2

^a Up to 10 mm LNB or GNB was used.

^b Up to 20 mm d-GlcNAc or d-GalNAc was used.

^c Assay was performed at 30 °C.

Comparison of Substrate Specificity among Cphy0577, Cphy1920, and Cphy3030 Proteins

We compared the specificities for the acceptor substrate of Cphy0577, Cphy1920, and Cphy3030 proteins in their reverse (synthetic) reactions. Although Cphy0577 and Cphy3030 proteins exhibited similar activity on d-GlcNAc and d-GalNAc in the presence of Gal 1-P, Cphy1920 proteins did not show activity on either of them (Table 2). Cphy1920 protein showed the highest activity on l-Rha and also exhibited activity on l-Man, l-Lyx, d-Glc, d-Glc2d, and d-Gal in this order (Table 2). However, Cphy0577 and Cphy3030 proteins did not show activity on either l-Rha, d-Glc, or d-Gal (Table 2). This result indicates that Cphy1920 protein is completely different from Cphy0577 and Cphy3030 proteins in substrate specificity.

TABLE 2.

Comparison of substrate specificity between Cphy0577, Cphy1920, and Cphy3030 proteins

Values in parenthesis represent specific activity (units/mg).

Substrate	Relative activitya (%)			Linkage of products by Cphy1920 protein
	Cphy0577	Cphy3030b	Cphy1920
	%	%	%
Acceptor
d-GlcNAc	100 (26)	100 (25)	NDc
d-GalNAc	100 (26)	72 (18)	ND
l-Rha	ND	ND	100 (43)	β1→4
l-Man	—d	—	37 (16)	β1→4
l-Lyx	—	—	8.8 (3.8)	β1→4
d-Glc	ND	ND	3.3 (1.4)	β1→3
d-Glc2d	—	—	1.8 (0.77)	β1→3
d-Gal	ND	ND	0.66 (0.28)	β1→3
d-Xyl	—	—	ND
d-Man	—	—	ND
l-Glc	—	—	ND
Donor
Gal 1-P	100 (26)e	100 (25)e	100 (43)f
Glc 1-P	NDe	NDe	NDf

^a Specific activity of Cphy0577 and Cphy3030 proteins on d-GlcNAc and specific activity of Cphy1920 protein on l-Rha in the presence of Gal 1-P were defined as 100% relative activity.

^b Assay was performed at 30 °C.

^c ND, less than 0.5% of relative activity.

^d —, not examined.

^ed-GlcNAc was used as an acceptor.

^fl-Rha was used as an acceptor.

We also investigated the specificities on the donor substrate of the three proteins. All of these proteins exhibited identical specificity (Table 2).

Basic Properties of Cphy0577, Cphy1920, and Cphy3030 Proteins

Cphy0577, Cphy1920, and Cphy3030 proteins were thermostable up to 45, 37, and 30 °C, respectively. This result indicates that thermostability differs by 15 °C between both GLNBPs. These three enzymes showed the highest activity at pH 6.0–7.0, thus indicating that they have optimum activity at neutral pH, as for other known GalHexNAcPs (16–18).

Identification of Products of the Synthetic Reaction of Cphy1920 Protein

After the synthetic reaction was performed in the presence of an acceptor (l-Rha, l-Man, l-Lyx, d-Glc, d-Glc2d, or d-Gal) and Gal 1-P, each reaction product was detected as a single spot on a TLC plate. The products were identified as GalRha, d-galactosyl-β1→4-l-mannose, d-galactosyl-β1→4-l-lyxose, d-galactosyl-β1→3-d-glucose (GalGlc), 2^I-deoxy-d-galactosyl-β1→3-d-glucose, and d-galactosyl-β1→3-d-galactose, respectively, based on the NMR spectra (Table S1). These results indicate that the disaccharides produced from the former three acceptors (l-Rha derivatives) differ from those produced from the latter three (d-Glc derivatives) in terms of the linkage position of the donor.

Kinetic Parameters of Synthetic Activity of Cphy1920 Protein

Kinetic parameters of Cphy1920 protein on six effective acceptors were determined. Only the K_m value on l-Rha was in the millimolar range and was 30–250 times smaller than those on the other acceptors (Table 3). On the contrary, the k_cat value on l-Rha was 0.4–3 times as large as those on the other acceptors (Table 3). This result indicates that k_cat/K_m values mainly depend on K_m values. The fact that the k_cat/K_m value on l-Rha was 15–800 times larger than those on the other acceptors indicates that l-Rha is the most effective acceptor for Cphy1920 protein (Table 3).

TABLE 3.

Kinetic parameters of synthetic activity of Cphy1920 protein

Values were obtained by regressing data with the following equation using Grafit Version 4.0.10: v = k_cat[E₀][S]/(K_m + [S]), where S is the acceptor.

Acceptor	kcat	Km	kcat/Km
	s−1	mm	s−1 mm−1
l-Rhaa	74 ± 2	2.4 ± 0.2	31 ± 2
l-Manb	170 ± 20	78 ± 16	2.2 ± 0.2
l-Lyxb	120 ± 13	220 ± 40	0.56 ± 0.03
d-Glcc	85 ± 14	460 ± 90	0.19 ± 0.01
d-Glc2dc	28 ± 3	290 ± 50	0.095 ± 0.004
d-Galc	24 ± 6	610 ± 160	0.039 ± 0.002

^a Up to 20 mm l-Rha was used.

^b Up to 100 mm acceptors were used.

^c Up to 150 mm acceptors were used.

Phosphorolytic Activity of Cphy1920 Protein

Double reciprocal plots of initial velocities against various initial concentrations of GalRha and phosphate gave a series of lines intersecting at a point (Fig. 3). This indicates that the phosphorolytic reaction of Cphy1920 protein follows a sequential bi-bi mechanism, as do inverting phosphorylases (16, 18, 28–34). Kinetic parameters of the enzyme determined by regressing data are shown in the legend of Fig. 3. The values of k_cat, K_mA, and K_mB were in the same range as other phosphorylases (16, 18, 28–34).

FIGURE 3.

Double-reciprocal plot of the phosphorolytic reaction of Cphy1920 protein with different concentrations of inorganic phosphate. Open circle, 0.5 mm P_i; closed circle, 1 mm P_i; open square, 2 mm P_i; closed square, 3 mm P_i; open triangle, 5 mm P_i. Kinetic parameters were as follows: k_cat = 45 ± 4 (s⁻¹), K_mA = 7.9 ± 0.8 (mm), K_mB = 1.9 ± 0.3 (mm), K_iA = 4.2 ± 0.7 (mm), and k_cat/K_mA = 5.7 ± 0.2 (s⁻¹ mm⁻¹), where A represents GalRha and B is P_i. Values were determined by regressing data with the following equation, using Grafit version 4.0.10: v = k_cat[E₀][A][B]/(K_iAK_mB + K_mA[B] + K_mB[A] + [A][B]).

Anomeric Specificity of Cphy1920 Protein

The anomeric specificity of Cphy1920 protein for acceptors was examined by detecting disaccharides produced from Gal 1-P and acceptors at the beginning of the synthetic reaction. When l-Rha was used as an acceptor, both anomers of the products were produced (Fig. 4A), indicating that Cphy1920 protein had no anomeric specificity on l-Rha. When d-Glc was used as an acceptor, the amount of α-d-Glc was reduced remarkably, and only the α-anomer of GalGlc was detected at the beginning of the reaction (Fig. 4B). This result indicates that Cphy1920 protein is specific for the α-anomer of d-Glc.

FIGURE 4.

HPLC profiles of the anomeric analyses in the synthetic reactions by GalRhaP with l-Rha (A) and d-Glc (B). A, upper lane, GalRha; middle lane, l-Rha; bottom lane, the reaction mixture after a 1-min incubation. B, upper lane, d-Glc and GalGlc; middle and bottom lanes, reaction mixture after 1- and 20-min incubation, respectively.

DISCUSSION

Classification of Cphy1920 Protein

Cphy1920 protein shows activity that has not been reported elsewhere, whereas Cphy0577 and Cphy3030 proteins are GLNBPs (Fig. 5). This is consistent with the prediction of substrate specificity of these three enzymes, based on the ternary structure of GLNBP_Bl. Furthermore, we compared acceptors of these enzymes in a synthetic reaction and found that Cphy1920 protein and GLNBPs did not exhibit activity on each other's acceptors (Table 2), thus suggesting that Cphy1920 protein is completely different from GalHexNAcPs in terms of substrate specificity. We propose d-galactosyl-β1→4-l-rhamnose phosphorylase (GalRhaP) as the name for Cphy1920 protein and suggest the need for a new EC number for GalRhaP (Fig. 5).

FIGURE 5.

Schematic representation of the reactions catalyzed by GalHexNAcP (A) and GalRhaP (B).

Even after comparing acceptors with all of the phosphorylases that have been reported, no phosphorylase acting on l-Rha has been identified until now, although phosphorylases acting on d-Glc (35), glucooligosaccharides (30), d-Fru (36), d-GlcNAc (32), d-GalNAc (16, 18), or d-glucose 6-phosphate (37) have been found. Therefore, this study is the first report of a phosphorylase acting on l-Rha.

Substrate Recognition of GalRhaP

GalRhaP synthesized β-1,4-galactosides from l-Rha derivatives and β-1,3-galactosides from d-Glc derivatives as acceptors in the presence of Gal 1-P. Activity on l-Rha derivatives was higher than that on d-Glc derivatives (Table 2), thus indicating that β-1,4-galactosides are the main substrates for GalRhaP.

Linkage position of the donor is one of the most notable differences between the substrates of GalRhaP and GalHexNAcPs. The difference can be explained by the ring conformations and orientations of substituting groups of l-Rha and d-Glc, because GalRhaP acts on β-1,3-galactoside as well as GalHexNAcPs when d-Glc derivatives are acceptors. Although l-Rha has a ¹C₄ chair conformation (38), the stable ring conformation of d-Glc is ⁴C₁. These ring conformations are conserved in the respective acceptor moieties of disaccharide products (Table S1). When hydroxyl groups of l-Rha and d-Glc at linkage positions of a donor are aligned, ring conformations of both acceptor moieties are stereochemically the same (Fig. 6A). We compared the orientation of substituting groups at corresponding positions of both l-Rha and d-Glc. The orientations of l-Rha 3-OH and C-6 methyl group are the same as those of d-Glc 2-OH and 4-OH, respectively (Fig. 6B). Moreover, both l-Rha 2-OH and the α-anomeric hydroxyl group of d-Glc are in an axial position (Fig. 6B). This orientation is very important, because GalRhaP did not exhibit activity on l-glucose (C-2 epimer of l-Man) (Table 2) and was specific for the α-anomer of d-Glc (Fig. 4B). In addition, GalRhaP was not specific for the anomer of l-Rha (Fig. 4A). This may be because the anomeric carbon of l-Rha corresponds to O-5 atom of d-Glc (Fig. 6B). Therefore, difference in the galactoside linkage position is rational in terms of ring conformations and orientations of substituting groups.

FIGURE 6.

Comparison of ring conformations (A) and substituents (B) between acceptors of GalRhaP. Pyranose rings are depicted to align with β-galactoside linkage positions that are shown in a box. Arabic numbers represent carbon numbers on hexoses. A, substituting groups except β-galactoside linkage positions are omitted. The dotted lines indicate ring conformations of hexoses. B, saccharides are shown as Haworth projections.

There are two other differences that should be noted between the substrates of GalRhaP and GalHexNAcPs in substituting groups. First, GalRhaP acts on d-Glc but not on d-GlcNAc, although GalHexNAcPs act on d-GlcNAc, thus indicating that GalRhaP cannot accept the N-acetyl group. Despite this fact, Trp²³³ of GLNBP_Bl, which recognizes the oxygen atom in the N-acetyl group of acceptors, is conserved in GalRhaP (Fig. 2). In GalRhaP, two residues (Ser-Asn) are inserted around the Trp residue (Fig. 2), corresponding to the loop region of GLNBP_Bl (19). Therefore, such an insertion might cause the corresponding Trp residue of GalRhaP to change its position and hinder acceptance of the N-acetyl group sterically. We also compared kinetic parameters of d-Glc and d-Glc2d to evaluate the contribution of d-Glc 2-OH to substrate recognition of GalRhaP. The value of k_cat/K_m on d-Glc2d was approximately half of that on d-Glc (Table 3), thus suggesting that d-Glc 2-OH is not important for d-Glc recognition. This result may be related to the fact that there is no hydrogen bond with the nitrogen atom in the N-acetyl group of acceptors in GLNBP_Bl (19). Second, the C-6 methyl group of l-Rha corresponds to the 4-OH of d-GlcNAc (Fig. 6B). Because GalRhaP showed a much smaller K_m value on l-Rha than that on l-Man and l-Lyx, where only the C-6 methyl group of l-Rha is substituted, the methyl group must be important for substrate recognition of GalRhaP (Table 3). Such a difference in substituting groups between l-Rha and d-GlcNAc may exist because the residue of GalRhaP corresponding to Val¹⁶² of GLNBP_Bl is not valine (GLNBP or LNBP type) or threonine (GNBP type) but isoleucine (Fig. 2).

Physiological Role of GalRhaP and GLNBPs from C. phytofermentans

Phosphorylases usually have strict substrate specificities that are closely related to the growth environment of bacteria possessing them. Because C. phytofermentans can utilize various kinds of plant polysaccharides for growth (23), it is speculated that GalRhaP and GLNBPs are involved in the metabolism of some plant polysaccharides.

GNBs from plant polysaccharides are little known, and LNB is known to exist in biosynthetic intermediates of N-linked glycoproteins during the processing of plant N-glycans in the Golgi apparatus, which is quite a minor component in plants (39). Moreover, there are no endo-α-N-acetylgalactosaminidase (7, 40, 41) and lacto-N-biosidase (5) (GNB- and LNB-releasing enzyme, respectively) homologs in the genome of C. phytofermentans. Physiological functions of GLNBPs are still uncertain.

GalRha is found in rhamnogalacturonan I (RG-I) contained in pectin as a structure with one galactose residue that binds l-Rha in the main chain of RG-I (42). No other pectinic structure containing GalRha was found in the Glycoscience.de data base (available on the World Wide Web). Because C. phytofermentans has some RG-I-metabolizing gene homologs (cphy1919 gene, unsaturated rhamnogalacturonyl hydrolase (GH105); cphy0343 gene, rhamnogalacturonan lyase (polysaccharide lyase family 11) (43)), GalRhaP is probably involved in the metabolism of RG-I.

GalRha structures are also found in some bacterial exopolysaccharides (44, 45) aside from plant polysaccharides. GalRhaP may be involved in salvage of exopolysaccharides produced by C. phytofermentans.

Conclusions

Phosphorylases are more suitable for synthesis of oligosaccharides than glycoside hydrolases in terms of reversibility of the reaction. Several large scale preparation methods of oligosaccharides taking advantage of this feature have been developed (11–13). However, the narrow variation of phosphorylases restricts their utility. In this study, we found a unique β-galactoside phosphorylase. Discovery of the enzyme will further expand utilization of phosphorylases.