A glycosyltransferase with a length-controlling activity as a mechanism to regulate the size of polysaccharides

Abstract

Cyclic β-1,2-glucans (CβG) are osmolyte homopolysaccharides with a cyclic β-1,2-backbone of 17–25 glucose residues present in the periplasmic space of several bacteria. Initiation, elongation, and cyclization, the three distinctive reactions required for building the cyclic structure, are catalyzed by the same protein, the CβG synthase. The initiation activity catalyzes the transference of the first glucose from UDP-glucose to a yet-unidentified amino acid residue in the same protein. Elongation proceeds by the successive addition of glucose residues from UDP-glucose to the nonreducing end of the protein-linked β-1,2-oligosaccharide intermediate. Finally, the protein-linked intermediate is cyclized, and the cyclic glucan is released from the protein. These reactions do not explain, however, the mechanism by which the number of glucose residues in the cyclic structure is controlled. We now report that control of the degree of polymerization (DP) is carried out by a β-1,2-glucan phosphorylase present at the CβG synthase C-terminal domain. This last activity catalyzes the phosphorolysis of the β-1,2-glucosidic bond at the nonreducing end of the linear protein-linked intermediate, releasing glucose 1-phosphate. The DP is thus regulated by this “length-controlling” phosphorylase activity. To our knowledge, this is the first description of a control of the DP of homopolysaccharides.

Osmoregulated periplasmic glucans are oligosaccharides present in the periplasm of certain Gram-negative bacteria. Common features of these oligosaccharides are the presence of glucose as the sole sugar constituent and the regulation of their synthesis by the osmolarity of the growth media. Osmoregulated periplasmic glucans may be cyclic, branched cyclic, or branched linear and, depending on the species, may be substituted with a variety of nonglycosidic residues (1, 2). Agrobacterium, Rhizobium, Sinorhizobium, and Brucella species synthesize osmoregulated periplasmic glucans of family II. Glucans of this family have 17–25 glucose residue cyclic β-1,2-backbones substituted with sn-1-phosphoglycerol, succinic acid, methylmalonic acid, or a combination of them (1–3).

Cyclic β-1,2-glucan synthase (Cgs), the enzyme responsible for the synthesis of cyclic β-1,2-glucans (CβG), is present in a restricted number of symbiotic or pathogenic bacteria, most of them belonging to the α-proteobacteria group, in which CβG are a symbiotic or virulence factor required for successful host interaction (4–9). Brucella abortus Cgs is a 320-kDa (2,867 amino acid residues) polytopic integral inner membrane protein with six transmembrane-spanning segments (TMSs) and with the N and C termini located on the cytoplasmic side of the membrane (10). Cgs, an enzyme using UDP-glucose as sugar donor and Mg²⁺ as cofactor, functions as an inverting processive β-1,2-glucosyltransferase that catalyzes the three enzymatic activities (initiation, elongation, and cyclization) required for synthesis of CβG. Synthesis is initiated by the transfer of the first glucose from UDP-glucose to a not yet identified amino acid residue in the same enzyme. Elongation is catalyzed by a UDP-Glc:β-1,2-oligosaccharide glucosyltransferase activity responsible for the addition of glucose residues to the nonreducing end of the linear β-1,2-oligosaccharide linked to the protein. Finally, cyclization, which is coupled to the release of the CβG from the protein, is probably a transglucosylation reaction during which the nonreducing end of the protein-linked oligosaccharide forms a β-1,2 linkage with the protein-linked reducing end of the polyglucose chain (6, 11).

Comparison of the Cgs sequence with those of other glycosyltransferases (GTs) led to the identification of the widely spaced D, DxD, E/D, (Q/R)xxRW motif, located between residues 475 and 818 (GT-84 domain), which is highly conserved in the active site of numerous GTs. Site-directed mutagenesis combined with in vitro and in vivo activity assays demonstrated that most of the amino acid residues of this motif are essential for Cgs activity, thus indicating that this region is implicated in the catalysis of initiation and/or elongation reactions required for the synthesis of the glucan β-1,2-glucose backbone. Overexpression of inactive mutants of the enzyme in wild-type backgrounds revealed that Cgs functions in the membrane as a monomer. Together, these results are compatible with a single addition model by which the identified D, DxD, E/D, (Q/R)xxRW motif forms a single center for substrate binding and glycosyl-transfer reaction (12).

Initiation, elongation, and cyclization reactions catalyzed by Cgs do not explain how the number of glucose residues of the CβG is controlled. Early studies revealed that the C-terminal region beyond residue 2008 is not required for synthesis of CβG. However, it was observed that C-terminal truncated forms of Cgs (at residues 2008 or 2179) produced CβG with a different TLC behavior (6).

Here we demonstrate that Cgs evolved a novel mechanism to control the number of glucose residues of the cyclic ring. This function is carried out by the C-terminal domain of the enzyme that, acting as a β-1,2-glucooligosaccharide phosphorylase simultaneous to chain elongation, catalyzes the phosphorolysis of the β-glycosidic bond at the nonreducing end of the linear β-1,2-glucooligosaccharide protein-linked intermediate releasing glucose 1-phosphate (glucose-1-P).

Results

The C-Terminal Region of Cgs Controls the Degree of Polymerization (DP) of the CβG.

To study whether the C-terminal region of Cgs is implicated in the control of the DP of the CβG, several truncated versions of B. abortus Cgs were constructed and expressed in an Agrobacterium tumefaciens Cgs-negative mutant (Fig. 1A). This heterologous expression strategy was used so that we could carry out our studies with the better-known Brucella enzyme but in a nonpathogenic context. Western blot analysis of membrane fractions revealed that all of the truncated proteins displayed the expected molecular weights and were inserted in the membrane [supporting information (SI) Fig. 7]. CβG synthesized by truncated mutants were analyzed by TLC, as described in Materials and Methods. As depicted in Fig. 1B, Cgs truncated at amino acid 1472 was inactive for the synthesis of CβG. The same result was obtained with proteins truncated upstream of this position (data not shown). On the other hand, Cgs truncated at position 1587 and all forms truncated downstream synthesized CβG but with a DP higher than the wild-type strain (Fig. 1B). Interestingly, even the protein truncated at position 2773 with only the 94 C-terminal residues deleted displayed the higher-DP phenotype (Fig. 1B). Furthermore, CβG produced by in-frame pentapeptide insertion mutants in the C-terminal region were analyzed (Fig. 1A). Four of six pentapeptide insertion mutants produced CβG with a DP similar to that observed with the C-terminal truncated proteins (Fig. 1C), indicating that some regions are less sensitive than others to the insertion of the pentapeptide.

Fig. 1.

The C-terminal region of Cgs controls the DP of the CβG. (A) Schematic representation of the wild type and representative truncated and pentapeptide insertion mutants of Cgs. Numbers above the bars indicate the position of residues of Cgs. Dark gray cylinders represent the TMSs. (B) TLC analysis of CβG produced by truncated mutants. (C) TLC analysis of CβG produced by in-frame pentapeptide insertion mutants. pBA24, plasmid expressing wild-type Cgs. * and **, migration of anionic and neutral CβG, respectively.

It was previously reported that ≈50% of B. abortus CβG showed an anionic behavior as they were substituted with succinic acid (2). Removal of substituents from the wild-type anionic CβG fraction recovered from the TLC plates by mild alkaline treatment yielded neutral glucans with the same DP of those of the nonsubstituted fraction, thus indicating that succinylation was not affected by ring size (2). Moreover, we have previously observed that a B. abortus cgm mutant unable to succinylate glucans produced only neutral CβG with the same size distribution as the wild-type strain (2).These evidences ruled out the possibility that mutants in the C-terminal region were affected in CβG modification.

Taken together, these results indicate that the Cgs region comprising residues 1–1587 represents the minimal region required for the synthesis of CβG (Fig. 1A) and reveal that the C-terminal region of Cgs is involved in the control of CβG DP.

Characterization of CβG.

CβG synthesized by C-terminal truncated and in-frame pentapeptide insertion mutants were further characterized by Bio-Gel P4 column chromatography, high-pH anion-exchange chromatography with pulse amperometric detection (HPAEC-PAD), and MALDI-TOF mass spectrometry analysis. As shown in SI Fig. 8 A and B, Bio-Gel P4 column chromatography confirmed results by TLC analysis as CβG synthesized by truncated or pentapeptide insertion mutants had a DP higher than the wild-type glucans.

To further characterize CβG produced by wild-type and mutant forms of Cgs, HPAEC-PAD and MALDI-TOF mass spectrometry analysis were carried out on neutral-CβG fractions. CβG recovered from the Bio-Gel P4 column were subjected to DEAE-Sephadex A-25 column chromatography, and neutral CβG were recovered (SI Fig. 8C). In HPAEC-PAD the retention time of a homologous series of carbohydrates increases with the DP. As shown in Fig. 2A, the chromatogram of neutral CβG produced by wild-type Cgs revealed the presence of six main peaks with retention times between 5 and 13 min. The elution profile of neutral CβG produced by the truncated mutant Cgs-1587stop and the insertion mutant Cgs-2319 revealed the same peaks, but with a different distribution, and the accumulation of molecules with higher retention times (between 14 and 28 min). These results are in agreement with those obtained by TLC analysis (SI Fig. 8C).

Fig. 2.

Characterization of CβG. HPAEC-PAD elution profiles (A) and MALDI-TOF mass spectrometry analysis (B) of neutral-CβG of A. tumefaciens A1045 strains harboring the indicated plasmid. pBA24, plasmid expressing wild-type Cgs. The numbers in B indicate the DP (in glucose units) corresponding to each m/z value.

Finally, neutral CβG were subjected to MALDI-TOF mass spectrometry analysis using a 2,5-dihydroxybenzoic acid matrix (Fig. 2B). The spectra of neutral CβG produced by the wild-type strain revealed the presence of six main signals at m/z 2,778.9, 2,939.7, 3,101.4, 3,263.1, 3,425.8, and 3,586.5. These ions as [M + Na]⁺ species had masses identical to those expected for CβG composed of 17–22 glucose residues, with the main species containing 17 and 19 glucose residues. Analysis of neutral CβG spectra produced by the truncated mutant Cgs-1587stop and the insertion mutant Cgs-2319 showed 12 sodium-cationized molecular ions, [M + Na]⁺. These molecular ion species had masses identical to those expected for CβG composed of 17–28 glucose residues, with the principal species containing 21 and 22 glucose residues. The m/z values of the signals corresponding to the glucans produced by the C-terminal mutants indicate that these are cyclic and not linear glucans. Peaks corresponding to the [M + K]⁺ ion with an m/z increased by 16 above the masses of the corresponding [M + Na]⁺ ions were also present.

As shown above, CβG of the wild-type strain are heterogeneous in size with DPs ranging from 17 to 22. Despite the length heterogeneity, the ring size distribution of CβG is not random, but with predomination of some species, suggesting the presence of a mechanism controlling the DP of the CβG. On the other hand, the size distribution of CβG produced by the C-terminal mutant forms of Cgs resembles a Gaussian distribution, indicating that such control mechanism is lost upon alteration of Cgs C-terminal region.

The Cgs C-Terminal Region Shows Similarity to GH-94 Glycoside Phosphorylase Family.

Bioinformatics analysis (see SI Text) revealed that Cgs C-terminal region (amino acids 1545–2867) displays significant sequence similarity to glycoside phosphorylases of family GH-94, such as cellobiose phosphorylases, chitobiose phosphorylases (ChBP), and cellodextrin phosphorylases (Fig. 3 A and B and SI Fig. 9). Those enzymes use inorganic phosphate as substrate and catalyze the phosphorolysis of β-glycosidic bonds generating glycosyl-phosphates with inversion of the anomeric configuration.

Fig. 3.

Sequence and site-directed mutagenesis analysis of B. abortus Cgs. (A) Modular organization of Cgs showing the GT-84 (GT family 84) and GH-94 (glycoside phosphorylase family 94) domains, and the conserved domains identified by National Center for Biotechnology Information-Conserved Domain Search (22). CBM-X, putative carbohydrate-binding domain; GH-94 AF, GH-94-associated family domain. Numbers above the bar indicate the position of residues of Cgs. TMSs I–VI are indicated. (B) Multiple sequence alignment of cyclic glucan synthases and the catalytic domain of GH-94 glycoside phosphorylases. The alignment was performed by using the ClustalW program (23) and edited with the Jalview 2.2 program (24). The length of proteins in amino acids is indicated on the left. Cgs, B. abortus cyclic glucan synthase (GenBank accession no. AF047823); ChvB, A. tumefaciens cyclic glucan synthase (GenBank accession no. NP_533395); NdvB, Sinorhizobium meliloti cyclic glucan synthase (GenBank accession no. P20471); ChBP, V. proteolyticus ChBP (GenBank accession no. BAC87867); CBP, Cellvibrio gilvus cellobiose phosphorylase (GenBank accession no. BAA28631); CDP, Clostridium stercorarium cellodextrin phosphorylase (GenBank accession no. AAC45511). (C) Schematic representation of the wild type and site-directed mutants of Cgs. The dark gray cylinders represent the TMSs. (D) TLC analysis of the CβG produced by site-directed mutants. pBA25, plasmid expressing wild-type Cgs. * and **, migration of anionic and neutral CβG, respectively. (E) Bio-Gel P4 chromatography of [¹⁴C]glucose-labeled glycopeptides. Glycopeptides were obtained and analyzed as described in Materials and Methods.

Based on this analysis we hypothesized that Cgs controls the DP of the CβG through a phosphorylase activity catalyzed by the C-terminal region.

Cgs C-Terminal Region Displays a β-1,2-Glucooligosaccharide Phosphorylase Activity.

As mentioned above, Cgs C-terminal region has significant sequence similarity with members of GH-94. Based on the multiple sequence alignment of the C-terminal region of Cgs and the catalytic domain of GH-94 glycoside phosphorylases, we pinpointed conserved residues Asp-2420 and Asp-2555 (Fig. 3B). These residues are analogous to Asp-350 and Asp-492 of Vibrio proteolyticus chitobiose phosphorylases (ChBP), respectively. Asp-350 contributes to fix the sugar moiety (GlcNAc) at the donor binding site (subsite −1), similar to the “fixer” residue of α-amylase (13), and is analogous to Asp-359 of Lactobacillus brevis maltose phosphorylase, whose essentiality for activity was confirmed by site-directed mutagenesis (14). Asp-492 of V. proteolyticus ChBP is assumed to be the catalytic residue (general acid) (15). D2420A and D2555A mutants were obtained and CβG was characterized by TLC (Fig. 3 C and D). Both replacements led to the production of CβG with a DP similar to that observed in the C-terminal truncated or in-frame pentapeptide insertion mutants. These results suggested that the control of the DP of CβG depends on the integrity of a phosphorylase activity present at the C-terminal region of the protein.

Glycopeptides can be obtained after extensive protease treatment of β-1,2-glucooligosaccharide protein intermediates (16). After this treatment most of the glycopeptides have only one amino acid attached to the reducing end of the β-1,2-glucooligosaccharide (17). Thus, the elution volume from a Bio-Gel P4 column is determined by the DP of the polyglucose chain originally linked to the protein. Glycopeptides from the wild-type Cgs and D2420A and D2555A mutants were obtained and subjected to Bio-Gel P4 chromatography as indicated in Materials and Methods. Mutants Cgs-D2420A and Cgs-D2555A yielded glycopeptides with a higher DP than the wild type (Fig. 3E), suggesting that the genuine substrate for the phosphorolysis reaction is the linear β-1,2-glucooligosaccharide protein-linked intermediate.

To investigate whether the C-terminal region of Cgs indeed displays a phosphorylase activity, we expressed this region from amino acid residue 1493 to amino acid residue 2867 (Cgs-CT for Cgs C-terminal region), as well as Cgs-CT displaying the site-directed replacements D2420A and D2555A (Cgs-CT-D2420A and Cgs-CT-D2555A), as N-terminal His-tagged fusion proteins (Fig. 4A). Recombinant soluble proteins were purified by using Ni-NTA affinity chromatography (SI Fig. 10). Phosphorylase activity was determined by measuring the formation of glucose-1-P by a specific coupled enzymatic assay, as described in Materials and Methods. Partially acid-hydrolyzed CβG were used as substrate for the reaction. As depicted in Fig. 4B, Cgs-CT displayed phosphorylase activity on this substrate (specific activity, 0.725 units/mg), but almost no activity was detected with the recombinant proteins Cgs-CT-D2420A and Cgs-CT-D2555A (specific activities, 0.025 and 0.095 units/mg, respectively). Furthermore, no activity was detected with Cgs-CT in the absence of inorganic phosphate or phosphoglucomutase in the reaction mixture. These results demonstrate that the C-terminal region of Cgs displays a phosphorylase activity on partially acid-hydrolyzed CβG.

Fig. 4.

Characterization of the recombinant C-terminal domain of Cgs. (A) Schematic representation of Cgs and the recombinant C-terminal proteins. Dark gray cylinders represent the TMSs. HT, histidine tag. (B) Phosphorylase activity as a function of protein concentration. Phosphorylase activity was determined by measuring the amount of glucose-1-P formed from phosphorolysis of the substrate, as described in Materials and Methods. For each reaction 400 μg of partial acid-hydrolyzed CβG was used as substrate. (C) Analysis of products released from [¹⁴C]glucose-labeled glycopeptides by the recombinant C-terminal domain (Cgs-CT) of Cgs. [¹⁴C]Glucose-labeled glycopeptides (3,400 cpm) were incubated with Cgs-CT (10 μg), and the reaction products were analyzed by DEAE-Sephadex chromatography. The column was eluted with 1 ml of water and with 1 ml each of 50, 100, 150, 200, 250, and 500 mM NaCl. Light gray columns, glycopeptides without Cgs-CT; dark gray columns, glycopeptides plus Cgs-CT; white columns, glycopeptides plus Cgs-CT in the absence of inorganic phosphate; black columns, [¹⁴C]glucose-1-P in the reaction mixture lacking glycopeptides.

To determine the substrate specificity, we assayed the activity of Cgs-CT with the following substrates: CβG (400 μg), α-d-sophorose (Glc-β-1,2-Glc, 30 mM), laminariobiose (Glc-β-1,3-Glc, 30 mM), laminarin ([Glc-β-1,3]_n, 200 μg), cellobiose (Glc-β-1,4-Glc, 30 mM), cello-oligosaccharides ([Glc-β-1,4]_n, 200 μg), malto-oligosaccharides ([Glc-α-1,4]_n, 200 μg), trehalose (Glc-α-1,1-Glc, 30 mM), isomaltotriose (Glc-α-1,6-Glc-α1,6-Glc, 30 mM), N,N′,N″-triacetylchitotriose (GlcNAc-β-1,4-GlcNAc-β-1,4-GlcNAc, 30 mM), and N,N′,N″,N‴-tetraacetylchitotetraose (GlcNAc-β-1,4-GlcNAc-β-1,4-GlcNAc-β-1,4-GlcNAc, 30 mM). No activity was detected with any of these substrates, indicating that Cgs may specifically catalyze the phosphorolysis of linear β-1,2-glucooligosaccharides.

To further characterize the product of Cgs-CT activity, [¹⁴C]glucose-labeled glycopeptides recovered from the Bio-Gel P4 columns (see Fig. 3E) were incubated with Cgs-CT and the radioactive products were analyzed by DEAE-Sephadex chromatography, as described in Materials and Methods (Fig. 4C). Incubation of glycopeptides with Cgs-CT and inorganic phosphate yielded a radioactive product that eluted from the column at the same NaCl concentration as glucose-1-P. HPAEC analysis revealed that the enzymatic radioactive product and authentic glucose-1-P eluted from the column at the same retention time (9 min). Under the same conditions the retention time of glucose 6-phosphate was 34 min. It may be concluded, therefore, that the C-terminal region of Cgs catalyzes the phosphorolysis of glycopeptides releasing glucose- 1-P from the nonreducing end.

Glucose-1-P Is Released During the Synthesis of CβG.

We propose that Cgs regulates the DP of the CβG by controlling the length of linear β-1,2-glucooligosaccharide protein-linked intermediates through the action of the C-terminal phosphorylase activity. To prove this hypothesis, we incubated total membranes of wild-type and mutant forms of Cgs with inorganic [³²P]phosphate in the presence or absence of the sugar donor substrate UDP-glucose. As shown in Fig. 5, wild-type Cgs released glucose-1-P only in the presence of UDP-glucose. No formation of glucose-1-P was observed when incubations were performed with the membrane fractions of truncated mutants Cgs-1587stop or Cgs-D2420A in the presence of UDP-glucose (Fig. 5). These results demonstrate that control of the cyclic glucan DP during the synthesis is mediated by the enzymatic activity present at the Cgs C-terminal region that catalyzes phosphorolysis of the linear β-1,2-glucooligosaccharide protein-linked intermediate.

Fig. 5.

Characterization of ³²P radioactive products generated in vitro. Total membranes fractions of A. tumefaciens A1045 strains harboring the indicated plasmid were incubated with inorganic [³²P]phosphate, and the radioactive products were analyzed by paper electrophoresis, as indicated in Materials and Methods. Radioactivity was detected by autoradiography. Glucose-1-P was differentiated from glucose 6-phosphate by mild acid hydrolysis (0.1 N HCl for 10 min at 100°C). Plus and minus signs indicate that 5 mM UDP-glucose was added or not to the reaction mixture, or that the sample was subjected or not to mild acid hydrolysis after the incubation. pBA24, plasmid expressing wild-type Cgs; Glc, glucose; Glc 1P, glucose-1-P; UDP-Glc, UDP-glucose; Pi, inorganic phosphate.

An additional radioactive product, migrating as a UTP standard, was only observed when the incubation was performed with a protein active for the synthesis of CβG in the presence of UDP-glucose (Fig. 5). The formation of this product depended on the synthesis of CβG regardless of the integrity of the Cgs C-terminal region. The pathway leading to the formation of the putative UTP deserves further studies.

Finally, total membranes of truncated Cgs-1587stop or Cgs-D2420A mutants were incubated with inorganic [³²P]phosphate in the presence of UDP-glucose and Cgs-CT. In both cases no formation of glucose-1-P was observed, indicating that the C-terminal domain must be part of the same polypeptide chain and/or anchored to the membrane to catalyze glucan phosphorolysis (data not shown).

Discussion

In this study we have elucidated the mechanism by which Cgs regulates the DP of CβG. The analyses of Cgs truncated form activity led to the identification of a stretch comprising residues 1–1587 as the minimal region required for catalysis of initiation, elongation, and cyclization reactions.

C-terminal truncated proteins beyond residue 1587 as well as some pentapeptide insertion mutants in the C-terminal region produced CβG with a higher DP, indicating that this region controls the cyclic glucan DP.

The Cgs C-terminal region shows significant sequence similarity with members of GH-94 glycoside phosphorylase family. Sequence analysis led to the identification of conserved amino acid Asp-2420 and Asp-2555 analogous to Asp-350 and Asp-492 of V. proteolyticus ChBP, respectively, which are essential for activity. Site-directed mutants Cgs-D2420A and Cgs-D2555A also produced CβG and linear glycopeptides with a higher DP than the wild-type enzyme. Recombinant C-terminal domain (Cgs-CT) displayed a phosphorylase activity that was specific for linear β-1,2-glucooligosaccharides whereas the recombinant proteins Cgs-CT-D2420A and Cgs-CT-D2555A almost completely lost phosphorylase activity. Furthermore, the recombinant C-terminal domain of Cgs catalyzed the phosphorolysis of glycopeptides releasing glucose-1-P from the nonreducing end.

Finally, incubation of total membrane fractions of the wild-type and mutant forms of Cgs with inorganic [³²P]phosphate in the presence or absence of UDP-glucose indicated that formation of glucose-1-P depended on the integrity of the C-terminal region and the presence of UDP-glucose, the sugar donor substrate for initiation and elongation reactions. These results demonstrate that Cgs regulates the DP of the CβG during the synthesis through a length-controlling activity present in the C-terminal region that catalyzes the phosphorolysis of the linear β-1,2-glucooligosaccharide protein-linked intermediate.

The mechanism shown in Fig. 6 for synthesis of CβG is proposed. Cgs itself acts as a protein intermediate and catalyzes the four enzymatic reactions required for the synthesis of CβG. The first glucose is transferred from UDP-glucose to a yet unidentified amino acid residue in the protein (initiation reaction). Successive glucoses are then transferred from UDP-glucose to the protein-bound glucose, thus elongating a linear polyglucose chain. The length of the β-1,2-oligosaccharide protein-linked intermediate results from a balance between the elongation reaction mediated by the GT domain (GT-84) and the phosphorolysis reaction exerted by the β-1,2-oligosaccharide phosphorylase domain (GH-94). Finally, the protein-linked β-1,2-oligosaccharide intermediate is cyclized, and CβG is released from the protein (cyclization reaction). From a bioenergetic perspective, it is interesting to note that the energy of the β-1,2-glucosidic bond is conserved during phosphorolysis in the form of glucose-1-P. Conversely, the energy of the glucosidic bond would be lost if the DP control were exerted by a β-glucosidase activity.

Fig. 6.

Proposed mechanism for the synthesis of CβG. Cgs itself acts as a protein intermediate and catalyzes the four enzymatic reactions required for the synthesis of CβG: initiation (A), elongation (B), phosphorolysis (C), and cyclization (D). At some point, probably after the glucan chain length reaches >17 glucose residues, the opposing activities of the glucosyltransferase and glucan phosphorylase control the DP. Cyclization puts an end to both reactions and releases the glucan from the protein. Light gray circle, glucose; dark gray circle, glucose at the nonreducing end of the polyglucose chain linked to the protein. Pi, inorganic phosphate. The GT-84 domain between TMSs II and III and the C-terminal GH-94 domain are indicated in light gray.

Cgs is the first dissected and well characterized example of a modular enzyme displaying at least two different types of activity on carbohydrates, that is, GT and glycoside phosphorylase. In addition, Cgs is the first described phosphorylase with activity toward β-1,2-glycosidic bonds.

The structural conservation predicted by Phyre (Protein Homology/analogY Recognition Engine program) analysis and the conservation of the catalytic and binding residues in the active site lead us to propose that Cgs phosphorolysis occurs trough a reaction mechanism similar to that proposed for ChBP of V. proteolyticus (15). Phosphorolysis may begin with a direct nucleophilic attack by phosphate to the glucosidic bond with the aid of Asp-2555 (general acid), which donates a proton to the glucosidic oxygen atom. The reaction may then proceed through an oxocarbenium cation-like transition state leading to the release of glucose-1-P with inversion of the anomeric configuration.

In V. proteolyticus ChBP, the N-terminal β-sandwich domain plays an important role in subunit interaction and substrate specificity (15). A distinctive feature of Cgs is that the region corresponding to the N-terminal region of GH-94 glycoside phosphorylases is duplicated with a carbohydrate-binding module (CBM-X) and a GH-94-associated family domain (GH-94 AF) in each region (Fig. 3A). Carbohydrate-binding modules (CBMs) have been classified into 48 different families based on amino acid sequence, binding specificity, and structure. Extensive data and classification can be found in the CAZy database (www.cazy.org). Based on structural and functional similarities, CBMs have been grouped into three types: “surface-binding” (type A), “glycan-chain-binding” (type B), and “small-sugar-binding” (type C). Type B CBMs have a β-sandwich structure, and the carbohydrate-binding sites comprise several subsites able to accommodate the individual sugar units of the polymeric ligand. The binding proficiency of this class of CBM is determined by the DP of the carbohydrate ligand; biochemical studies demonstrated increased affinities up to hexasaccharides and negligible interaction with oligosaccharides with DP of three or less (for review, see ref. 18). The CBM-Xs identified in Cgs display a predicted β-sandwich fold by Phyre sequence analysis. In addition, Cgs specifically catalyzes the phosphorolysis of linear β-1,2-glucooligosaccharides, but no activity was detected on a β-1,2-glucose disaccharide (sophorose). Therefore, the CBM-Xs could be considered type B (glycan-chain-binding) CBMs that may interact with the β-1,2-glucose backbone linked to the protein to target the catalytic module of phosphorylase to the substrate. Further work is required to determine the binding specificity of Cgs CBM and its relevance for the phosphorylase activity.

Pathogenic Brucella have developed strategies to persist for prolonged periods of time within host cells. Brucella controls the maturation of its vacuole to avoid innate immune host cell responses, such as lysosome fusion, and to reach its replicative niche (19). B. abortus CβG-deficient mutants are unable to prevent phagosome–lysosome fusion and are significantly defective in intracellular multiplication in professional and nonprofessional phagocytic cells (7, 9). CβG play a major role in circumventing host cell defense by interacting with cholesterol and reorganizing lipids rafts of host cell membranes (9). Thus, CβG are virulence factors that interact with lipid rafts and contribute to pathogen survival. Although the effect of ring size on this interaction is unknown, it may be suggested that cholesterol scavenging and lipid raft disruption activity might be affected by the ring size of the molecule.

Materials and Methods

Bacterial strains used in this study were grown as indicated in SI Text. Preparation of total membranes of A. tumefaciens strains was carried out as described previously (11). Protein concentration was determined by the method of Lowry et al. (20).

Linker scanning mutagenesis was performed by using the GPS-LS kit (New England Biolabs, Ipswich, MA) (21) (SI Text).

CβG were extracted by the ethanol method (70% ethanol, 1 h at 37°C) and analyzed by TLC as described previously (6). Isolation and purification of CβG were performed according to Briones et al. (11) with modifications described in SI Text.

HPAEC-PAD chromatographic analysis on a Carbopack PA-100 column and MALDI-TOF mass spectrometry analysis of CβG were performed as described in SI Text.

Site-directed mutagenesis was carried out by using the QuikChange Site-Directed Mutagenesis method (Stratagene, La Jolla, CA) (SI Text).

Cloning, expression, and purification of recombinant Cgs C-terminal domain was performed as indicated in SI Text.

Phosphorylase activity was determined by measuring the amount of glucose-1-P formed by phosphorolysis of the substrate by a coupled enzymatic assay using phosphoglucomutase and glucose- 6-phosphate dehydrogenase. Formation of NADPH was followed continuously with a spectrophotometer during 5 min at 25°C. One unit of enzyme activity was defined as the amount of enzyme that catalyzes the production of 1 μmol of glucose-1-P per min (SI Text).

Isolation and characterization of glycopeptides were done as described (11) (SI Text). [¹⁴C]Glucose-labeled glycopeptides were incubated with Cgs-CT for 1 h at 25°C, and the radioactive products were analyzed by chromatography on a DEAE-Sephadex A-25 column and by HPAEC on a Carbopack PA-10 column (SI Text).

Total membrane fractions of the different strains were incubated with NaH₂³²PO₄ (222,000 cpm; 1 μCi/μmol) for 1 h at 28°C, and the radioactive products in the supernatant were analyzed by paper electrophoresis (SI Text).

Abbreviations

CBM

carbohydrate-binding module

ChBP

chitobiose phosphorylase

CβG

cyclic β-1,2-glucans

Cgs

CβG synthase

DP

degree of polymerization

glucose-1-P

glucose 1-phosphate

GT

glycosyltransferase

HPAEC

high-pH anion-exchange chromatography

PAD

pulse amperometric detection

TMS

transmembrane-spanning segment.