Glycobiology Advance Access originally published online on April 7, 2006
Glycobiology 2006 16(7):679-691; doi:10.1093/glycob/cwj113
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Identification of active site residues of the inverting glycosyltransferase Cgs required for the synthesis of cyclic ß-1,2-glucan, a Brucella abortus virulence factor
2 Instituto de Investigaciones Biotecnológicas-Instituto Tecnológico de Chascomús (IIB-INTECH), Consejo Nacional de Investigaciones Científicas y Técnicas, Universidad Nacional de General San Martín (CONICET-UNSAM), Buenos Aires, Argentina; and 3 Section of Microbial Pathogenesis, Yale University School of Medicine, Boyer Center for Molecular Medicine, New Haven, CT 065362
1 To whom correspondence should be addressed; e-mail: rugalde{at}iib.unsam.edu.ar
Received on February 7, 2006; revised on March 31, 2006; accepted on April 5, 2006
 |
Abstract |
|---|
 Top Abstract IntroductionResultsDiscussionMaterials and MethodsSupplementary DataConflict of interest statementAcknowledgmentsReferences |
|---|
Brucella abortus cyclic glucan synthase (Cgs) is a 320-kDa (2868-amino acid) polytopic integral inner membrane protein responsible for the synthesis of the virulence factor cyclic ß-1,2-glucan by a novel mechanism in which the enzyme itself acts as a protein intermediate. Cgs functions as an inverting processive ß-1,2-autoglucosyltransferase and has the three enzymatic activities required for the synthesis of the cyclic glucan: initiation, elongation, and cyclization. To gain further insight into the protein domains that are essential for the enzymatic activity, we have compared the Cgs sequence with other glycosyltransferases (GTs). This procedure allowed us to identify in the Cgs region (475–818) the widely spaced D, DxD, E/D, (Q/R)xxRW motif that is highly conserved in the active site of numerous GTs. By site-directed mutagenesis and in vitro and in vivo activity assays, we have demonstrated that most of the amino acid residues of this motif are essential for Cgs activity. These sequence and site-directed mutagenesis analyses also indicate that Cgs should be considered a bi-functional modular GT, with an N-terminal GT domain belonging to a new GT family related to GT-2 (GT-84) followed by a GH-94 glycoside hydrolase C-terminal domain. Furthermore, over-expression of inactive mutants results in wild-type (WT) production of cyclic glucan when bacteria co-express the mutant and the WT form, indicating that Cgs may function in the membrane as a monomeric enzyme. Together, these results are compatible with a single addition model by which Cgs acts in the membrane as a monomer and uses the identified motif to form a single center for substrate binding and glycosyl-transfer reaction.
Key words: Brucella abortus / cyclic ß-1,2-glucan / cyclic glucan synthase / glycosyltransferases / site-directed mutagenesis
|
Introduction |
|---|
|
TopAbstractIntroductionResultsDiscussionMaterials and MethodsSupplementary DataConflict of interest statementAcknowledgmentsReferences |
|---|
Brucella spp. are gram-negative, facultative intracellular bacteria that cause a chronic zoonotic disease worldwide known as brucellosis (Corbel, 1997
). Six species of Brucella with different host specificity and pathogenic characteristics have been described (Hoyer and McCullough, 1968; Verger et al., 1985). Brucella abortus, the etiological agent of bovine brucellosis, causes abortion and infertility in cattle and undulant fever in humans.
Brucella produces cyclic ß-1,2-glucan (Ugalde, 1999). This compound is a cyclic homopolysaccharide consisting of about 17–22 ß-1,2-linked glucose units. Isolation of the B. abortus cgs gene, which codes for the cyclic ß-1,2-glucan synthase (Cgs), and characterization of the synthesis of cyclic ß-1,2-glucan in this species have been reported (Iñón de Iannino et al., 1998). B. abortus cgs mutants showed reduced virulence in mice and defective intracellular multiplication, indicating that cyclic glucan is required for an effective host interaction (Briones et al., 2001). Isolation of the B. abortus cyclic glucan transporter gene (cgt) was described recently (Roset et al., 2004). Cgt mutants in B. abortus have also affected its interaction with the host. Thus, the transport of cyclic glucan to the periplasmic space and possible through the outer membrane is required in B. abortus for complete expression of virulence (Roset et al., 2004). Furthermore, cyclic ß-1,2-glucans synthesized by Brucella play a major role in circumventing host cell defense. Brucella cyclic glucan acts on host cell membranes at the level of lipid rafts controlling vacuole maturation, avoiding lysosome fusion, and allowing Brucella to reach the endoplasmic reticulum (Arellano-Reynoso et al., 2005). Thus, cyclic ß-1,2-glucan is a Brucella virulence factor required for intracellular survival; accordingly, Cgs may be a good target for developing new chemotherapy alternatives against this pathogen.
The enzymatic transfer of glycosyl residues is, quantitatively, the most significant reaction on earth. It is catalyzed by glycosyltransferases (GTs), the enzymes that transfer sugar moieties from activated donor molecules to specific acceptor molecules. Despite their catalytic diversity, GTs are subdivided into only two main classes: inverting GTs, whose product displays the opposite stereochemistry at the anomeric center to that of the activated donor, and retaining transferases, whose product retains the same anomeric configuration of the donor (Sinnott, 1990). GTs have been classified into families by amino acid sequence similarities (Campbell et al., 1997). This classification was then expanded into a hierarchical system of families, clans, and folds (Coutinho et al., 2003), and an updated classification of GTs is available from the Carbohydrate-Active enZymes database (URL: http://afmb.cnrs-mrs.fr/cazy/).
B. abortus Cgs is a 320-kDa (2868-amino acid) polytopic integral inner membrane protein with six transmembrane segments and the amino and carboxyl terminus located on the cytoplasmic side of the membrane (Ciocchini et al., 2004). This enzyme synthesizes cyclic ß-1,2-glucan by a novel mechanism in which the enzyme itself acts as a protein intermediate. Cgs uses uridine diphosphate (UDP)-glucose as a sugar donor, requires Mg2+ ion as cofactor, and has the three enzymatic activities required for synthesis of the cyclic polysaccharide (i.e., initiation, elongation, and cyclization). Initiation consists of the transfer of the first glucose from UDP-glucose to an unidentified amino acid residue of the enzyme. Elongation, catalyzed by a [UDP-Glc:ß-(1,2) oligosaccharide glucosyltransferase] activity, consists of the addition of glucose residues to the linear ß-1,2-oligosaccharide linked to the protein. Finally, Cgs catalyzes glucan cyclization, releasing the cyclic glucan from the protein, probably by a transglucosylation reaction during which the non-reducing end of the protein-linked oligosaccharide forms a ß-1,2-linkage with the protein-linked reducing end of the polyglucose chain (Briones et al., 1997; Iñón de Iannino et al., 1998). Consequently, Cgs acts as an inverting processive ß-1,2-autoglucosyltransferase.
Despite the characterization of Cgs activities and cloning of the Cgs genes, no information is available about the protein domains essential for enzyme activity such as active site(s). To gain further insight into the mechanism of action of Cgs, we have compared the Cgs sequence with other GTs, and we have identified the widely spaced D, DxD, D/E, (Q/R)xxRW motif that is highly conserved in all Cgs and in the active site of numerous GTs. By site-directed mutagenesis and in vitro and in vivo activity assays, we analyze the implication of this highly conserved motif as a possible active site of the enzyme.
|
Results |
|---|
|
TopAbstractIntroductionResultsDiscussionMaterials and MethodsSupplementary DataConflict of interest statementAcknowledgmentsReferences |
|---|
Sequence analysis of Cgs regions
Conventional BLAST sequence analysis reveals that Cgs displays high sequence similarity with the other Cgs such as Agrobacterium tumefaciens ChvB (53% of identity) and Sinorhizobium meliloti NdvB (54% of identity) (Zorreguieta and Ugalde, 1986
), but no significant homology with any other GT of the data base was detected. Furthermore, this analysis reveals that Cgs is a bimodular protein, with an N-terminal domain (from amino acids 1 to 1546) that is required for cyclic glucan synthesis and a glycoside hydrolase (GH) C-terminal domain (from amino acids 1547 to 2868) that is not required for glucan synthesis (Figure 1) (Iñón de Iannino et al., 1998 and unpublished data). It is remarkable that this region of the protein has a high level of similarity with cellobiose and cellodextrin phosphorylases and is highly conserved among all the Cgs sequenced so far. It remains to be established whether Cgs has cellobiose and/or cellodextrin phosphorylase enzymatic activity and whether this region is implicated in the in vivo regulation of cyclic glucan synthesis.
|
No recognizable domains or protein motifs of any Cgs enzyme have been identified before. We have previously determined the membrane topology of Cgs (Ciocchini et al., 2004) (Figure 1). In the proposed topological model, all the transmembrane-spanning segments (TMSs) are located in the half N-terminal region of the protein. The six TMSs determine three very small periplasmic loops and four large cytoplasmic regions. The cytoplasmic regions of Cgs were analyzed using position-specific iterated BLAST (PSI-BLAST) and three-dimensional position-specific scoring matrix (3D-PSSM) programs (Figure 1). Owing to its known sensitivity at a high divergence level, these programs allow to identify certain conserved motifs between proteins that share relatively little identity of their primary structure, as it occurs between Cgs and other GTs.
3D-PSSM analysis of Cgs region (1–418) reveals a predicted fold similar to that of sigma 70 subunit fragment of Escherichia coli RNA polymerase, peroxisomal targeting signal-1 receptor, and tetratricopeptide repeat (TPR) motif (Figure 1). All these protein folds are involved in protein–protein interaction. TPR is a protein–protein interaction module found in multiple copies in a number of functionally different proteins that facilitate specific interactions with a partner protein(s) (Blatch and Lassle, 1999). Further work is required to establish whether this region is implicated in the interaction of Cgs with other protein(s). An interesting candidate could be the Cgt, an integral inner membrane protein involved in the transport of cyclic ß-1,2-glucan into the periplasm (Roset et al., 2004).
PSI-BLAST analysis reveals that Cgs region (475–818) is distantly related to processive GTs of family GT-2 (see Identification of conserved amino acid residues in the Cgs region (475–818)). Furthermore, the analysis of this region by 3D-PSSM clearly reveals a predicted fold similar to that of SpsA (Biotext value, 0.37; certainty, 90%), also a GT-2 member whose crystal structure has been resolved (Charnock and Davies, 1999; Garinot-Schneider et al., 2000) (Figure 1). This protein is a putative GT involved in the synthesis of the spore coat in sporulating Bacillus subtilis, but its sugar donor and acceptor specificity remain undefined.
Finally, no significant hits were obtained with the Cgs regions (870–938) and (991–1546) (Figure 1). So, no function can be predicted for these regions.
Identification of conserved amino acid residues in the Cgs region (475–818)
As mentioned in Sequence analysis of Cgs regions, 3D-PSSM analysis of Cgs region (475–818) reveals that it has a predicted fold similar to that of SpsA, despite their low amino acid sequence similarity. The structure-based sequence alignment generated by 3D-PSSM with the secondary structure prediction is shown in Figure 2A. Pair-wise alignment of this Cgs region with the other GTs of the GT family 2, mainly processive GTs, can also be generated using PSI-BLAST (data not shown).
|
Cgs region between amino acid residues 475 and 665 is formed by alternating ß-strands and 
-helices and may correspond to the nucleotide sugar-binding domain (Figure 2A and C) (Charnock et al., 2001). Within this domain, we have identified Phe-505 and Asp-538, analogous to Tyr-11 and Asp-39 of SpsA, respectively. The crystal structure of SpsA revealed that these residues are implicated in the binding of the uracil base (Charnoc and Davies, 1999; Tarbouriech et al., 2001). Secondary structure prediction programs predict that, as the analogous residues of SpsA, Cgs Phe-505 and Asp-538 are located after ß1- and ß2-strands, respectively. We will refer to this region as conserved region I (CR I). We have also identified two conserved aspartic residues (Asp-635 and Asp-637), which form the so-called D635xD637 motif identified in many GT families (Breton and Imberty, 1999) (Figure 2A). In SpsA, the DxD motif is a TD98D99 sequence where Asp-98 binds the hydroxyl groups on the ribose moiety and Asp-99 binds the pyrophosphate via one molecule of Mn2+. Interestingly, these residues are located in the loop following the ß4-strand (Charnock and Davies, 1999; Tarbouriech et al., 2001), as predicted to be Asp-635 and Asp-637 of Cgs. We will refer to this region as CR II.
On the basis of pair-wise alignments generated by PSI-BLAST using the Cgs region (475–818) sequence as a query, we have constructed a multiple sequence alignment to analyze the C-terminal region that may correspond to the acceptor-binding domain (Figure 2B and C) (Charnock et al., 2001). From this alignment, we have identified Cgs Glu-743 that aligns with the predicted catalytic residue of other processive GTs, such as Saccharomyces cerevisiae Chs2 Asp-562 and mouse Has1 Asp-344, which were previously found to be essential for enzyme activity (Nagahashi et al., 1995; Yoshida et al., 2000). We have also identified that Cgs Glu-769 and Asp-770 form an EDY motif. This motif is not well conserved in other GTs as it is shown in the alignment of Figure 2B but is strictly conserved among all the Cgs sequenced so far (data not shown). A sequence analysis using PSI-BLAST carried out by Keenleyside and others (2001) revealed that this motif is conserved in the GT WbbE from Salmonella enterica as well as in other GTs of family GT-2. Site-directed mutagenesis experiments suggested that this motif contains the catalytic base residue. The region of Cgs containing the putative catalytic base residue was designated as CR III. In addition, we have identified a sequence in Cgs (RxxRW) (Figure 2B) very close to the motif QxxRW identified in domain B of some processive GTs (Saxena et al., 1995). The RxxRW sequence has also been identified in non-processive enzymes, such as the eukaryiotic glucosylceramide synthases from GT family 21 (GT-21) (Marks et al., 2001). However, it remains to be determined whether this motif has the same function in processive and non-processive enzymes. A number of kinetic and site-directed mutagenesis studies have confirmed that these residues are required for enzyme activity (Nagahashi et al., 1995; Yoshida et al., 2000). We will refer to this region as CR IV.
In brief, in the Cgs region (475–818), we have identified several amino acid residues that are highly conserved among all the Cgs sequenced so far and may be part of the D, DxD, D/E, (Q/R)xxRW motif (Figure 2C), which is a common motif identified in the active site of numerous GTs (Charnock et al., 2001). To test these hypotheses, we carried out site-directed mutagenesis of conserved residues of CRs I–IV.
In vitro and in vivo characterization of mutants
Each conserved amino acid residue within CRs I–IV, as well as selected neighboring residues (see in this section), was subjected to site-directed mutagenesis. Wild-type (WT) and the mutated proteins were subsequently expressed in B. abortus S19
cgs strain in which the 5' region of the cgs gene has been deleted and therefore produces no cyclic glucan.
The in vitro and in vivo activity was determined as described in Materials and Methods (Figures 3–6 and supplementary figure).
|
|
|
|
As it was mentioned above, Cgs is an integral inner membrane protein. To confirm the cellular localization of the mutated proteins, the total membrane fractions were prepared and analyzed by sodium dodecyl sulphate–polyacrilamide gel electrophoresis (SDS–PAGE). In all the cases, the protein was detected in the membrane fraction, indicating that the amino acid changes do not affect protein membrane integration (data not shown).
CR I substitutions.
Replacement of the conserved residue Phe-505 by alanine resulted in complete loss of in vitro activity (Figure 3A–C). This mutant, F505A, was incapable of restoring cyclic glucan production at WT levels in vivo (Figure 3D). Residues Ser-537, Trp-539, and Asp-541 surrounding the highly conserved Asp-538 were also substituted by alanine. The mutant S537A displayed no changes on Cgs activity. W539A and D541A displayed reduced activity in vitro and in vivo, but only the mutation of Asp-538, D538A, resulted in almost complete loss of activity. The conservative D538E mutant restored the activity in vivo but not in vitro (Figure 3). Amino acid residues Asp-571 and Asp-572, which are strictly conserved between Cgs and other Cgs but not between Cgs and GTs of family 2, could be substituted by alanine without affecting the in vitro and in vivo enzyme activity (Figure 3).
CR II substitutions.
Amino acid residues Asp-635 and Asp-637 that form a DxD motif were replaced by asparagine. The mutant D635N displayed almost complete loss of in vitro and in vivo activity. Similar results were obtained for the conservative change D635E. The mutant D637N also displayed a reduction in the activity but less important than that observed with D635N and D635E substitutions (Figure 4).
CR III substitutions.
The highly conserved Glu-743 was substituted by alanine or aspartic acid. These mutants, E743A and E743D, had no effect on Cgs activity (Figure 5). However, the mutant E769A displayed almost complete loss of in vitro and in vivo Cgs activity. The conservative change, E769D, restored partially the activity in vivo but not in vitro (Figure 5). On the contrary, when the neighboring residue D770 was changed by alanine (D770A), no significant reduction in the activity in vitro and in vivo was observed (Figure 5). Because the interpretation of the alignment of this region (Figure 2B) is ambiguous (domain B of GTs is highly variable reflecting the vast number of different acceptor species that may be accommodated within the active site), we decided to mutate the residues Asp-711, Asp-715, Asp-729, Asp-731, and Glu-734, all of which are strictly conserved among B. abortus Cgs and other Cgs. No significant reduction in enzyme activity was observed for D729A, D729E, D731A, and E734A mutants. In D711A and D715A mutants, the synthesis of cyclic glucan was significantly reduced (Figure 5C and D); however, the formation of the intermediate was less affected (Figure 5A and B), thus suggesting the cyclization reaction was being affected.
CR IV substitutions.
The residues of the conserved sequence motif RxxRW were subjected to site-directed mutagenesis analysis (Figure 6). R781A substitution showed reduced activity, whereas R784A substitution displayed almost complete loss of in vitro and in vivo activity. W785T substitution completely abolished Cgs activity in vitro and in vivo, whereas the Trp-785 replacement by alanine (W785A) conserved an enzymatic activity significantly higher than the W785T substitution.
Cgs is a monomeric enzyme
To discern whether Cgs functions as a monomeric or multimeric enzyme, we over-expressed non-functional D635E and W785T mutant forms of Cgs (in a plasmid of middle copy number) in the B. abortus WT background carrying a unique copy of cgs gene in the chromosome. As described above, D635E and W785T mutants failed to restore cyclic glucan production of S19cgs mutant (Figures 4D and 6D). The over-expression level of mutant forms with respect to WT Cgs was corroborated by Coomassie Blue-staining SDS–PAGE (Figure 7A).
|
The over-expression of inactive Cgs proteins does not exert any negative dominance, as judged by the WT level of cyclic glucan production of cells that co-express the mutant and the WT forms (Figure 7B), indicating that Cgs may function in the membrane as a monomeric enzyme.
|
Discussion |
|---|
|
TopAbstractIntroductionResultsDiscussionMaterials and MethodsSupplementary DataConflict of interest statementAcknowledgmentsReferences |
|---|
Sequence analysis of the B. abortus Cgs region between residues 475 and 818 led to the identification of conserved amino acid residues (F505, D538, D635xD637, E769, and R781xxR784W785) that may be part of the active site of the enzyme (Figure 2C). These residues are strictly conserved among all the Cgs sequenced so far, and we have demonstrated by site-directed mutagenesis that most of them are required for cyclic ß-1,2-glucan synthase activity. These results are the first step toward understanding the complex catalytic mechanisms of the enzyme responsible for the synthesis of cyclic ß-1,2-glucan.
Little is known about the catalytic mechanism of the three different enzymatic reactions (initiation, elongation, and cyclization) that participate during the synthesis of cyclic ß-1,2-glucan. Understanding the molecular mechanism that leads to the synthesis of this oligosaccharide required for subverting the natural immune cell response of the host may lead to the discovery and development of new inhibitors that may be used as novel chemotherapy against brucellosis.
The X-ray crystal structures of 12 inverting GTs, either free or bound to substrates, have been determined recently (reviewed by Breton et al., 2006). Despite their significant sequence diversity, GTs show great structural similarity. Two structural superfamilies, termed GT-A and GT-B, with distinguishable N-terminal and C-terminal domains (Unligil and Rini, 2000; Bourne and Henrissat, 2001) have been described for GTs. The enzymes of the GT-A fold have two dissimilar domains. The N-terminal domain, which recognizes the sugar-nucleotide donor, comprises several ß-strands that are each flanked by -helices as in a Rossmann-like fold, whereas the C-terminal domain, which contains the acceptor-binding site, consists largely of mixed ß-strands. By contrast, enzymes with the GT-B fold contain two similar Rossmann domains, with the N-terminal domain providing the acceptor-binding site and the C-terminal domain providing the donor-binding site. A third fold family has recently emerged with the elucidation of the crystal structure of a sialyltransferase (CstII) of Campylobacter jejuni (Chiu et al., 2004).
GT-2 SpsA (Charnock and Davies, 1999; Tarbouriech et al., 2001), GT-7 bovine ß-galactosyltransferase (Gastinel et al., 1999), GT-13 N-acetylglucosaminyl transferase I (Unligil et al., 2000), and GT-43 human ß-1,3-glucuronyltransferase I (Pedersen et al., 2000) proteins display similar 3D structures that correspond to the GT-A fold and near-identical catalytic machinery, although they have insignificant sequence similarity (Tarbouriech et al., 2001). According to the hierarchical family classification for GTs proposed by Coutinho and others (2003), these proteins were placed into Clan I that grouped inverting GTs with a GT-A fold. Structural alignments of all these structures reveal a strikingly similar catalytic environment, showing the conservation of the catalytic and binding residues in the active site and suggest a similar mechanism for all "inverting" GTs (Tarbouriech et al., 2001). Inverting transferases are assumed to use a single SN2-type displacement mechanism with nucleophilic attack by the acceptor species at the C-1 (anomeric) carbon of the donor sugar, leading to the formation of a glycosidic bond between the acceptor and sugar donor with the inversion of the configuration at C-1. Such a mechanism is generally believed to demand a general base (usually an aspartate or glutamate) to activate the sugar acceptor for nucleophilic attack by deprotonation because sugar hydroxyls are themselves quite poor nucleophiles. For most enzymes, the reaction also involves an additional carboxylate or carboxylates to coordinate a divalent metal ion (particularly Mn2+ or Mg2+) on the phosphate group(s) of the nucleotide and to facilitate the departure of the nucleotide-diphospho-leaving group.
The resolved X-ray crystal structures of the GTs mentioned above have shed light on the roles of the structurally invariant residues D, DxD, D/E present in CRs I–III, respectively (Figure 2C). Typically, four or five aspartate residues form a single center for glycosyl transfer. A single aspartate residue on strand ß-2 forms part of the recognition element for uracil or thymidine. The DxD motif situated on a loop at the C-terminal end of strand ß-4 coordinates the ribose molecule, the Mn2+, and potentially a hydroxyl group on the donor sugar. An additional carboxylate group, typically that of Asp but also Glu, acts as BrØnsted base, activating the acceptor for nucleophilic attack at C-1 by deprotonation. A number of kinetic and site-directed mutagenesis studies also confirm the role of these aspartate residues, or its equivalent, in catalysis by family GT-2 enzymes, such as S. meliloti glucosyltransferase ExoM (Garinot-Schneider et al., 2000) and WbbE from S. enterica (Keenleyside et al., 2001), among non-processive enzymes. These kinds of studies were also performed with some processive enzymes of this family, where an additional CR (Q/RxxRW) was described to be implicated in catalysis, such as S. cerevisiae chitin synthase 2 (Nagahashi et al., 1995), mouse hyaluronan synthase Has1 (Yoshida et al., 2000), and Acetobacter xylinium cellulose synthase (Saxena and Brown, 1997).
In CR I, we have identified two conserved residues, Phe-505 and Asp-538, that are essential for enzyme activity. These amino acid residues, as well as certain neighboring residues, were replaced by alanine. Only the mutants F505A and D538A display almost complete loss of in vitro and in vivo activity (Figure 3). Cgs Phe-505 and Asp-538 are analogous to Tyr-11 and Asp-39 of SpsA; hence, these residues may be implicated in UDP-glucose binding by aromatic stacking and coordinating the N-3 of the uracil base, respectively. The conservative D538E mutant restores the activity in vivo, which indicates that the presence of a charged amino acid residue, but not the side chain length, is critical for the activity. Furthermore, secondary structure prediction programs predict that Cgs Phe-505 and Asp-538 are located after ß1 and ß2-strands in domain A, respectively, similar to the localization of the analogous residues of SpsA.
The consensus sequence DxD, or its variants, has been identified in a wide variety of GTs (Breton and Imberty, 1999). Site-directed mutagenesis studies of several GTs have shown that the two conserved acidic residues of DxD motif are essential for GT activity. (Busch et al., 1998; Wiggins and Munro, 1998; Gulberti et al., 2003). Moreover, the recent resolution of the crystal structure of several GTs has clarified the role of the DxD motif in glycosyltransferases, indicating that it is directly involved in nucleotide sugar binding (Charnock and Davies, 1999; Gastinel et al., 1999; Pedersen et al., 2000; Unligil et al., 2000). In SpsA, this motif adopts the sequence xD98D99. SpsA Asp-98 binds the hydroxyl group O-2 on the ribose moiety, and SpsA Asp-99 interacts through Mn2+ with phosphate. In the equivalent region of Cgs (CR II), where DxD is DAD, substitution of Asp-635 abolishes enzymatic activity in vitro and in vivo, whereas substitution of Asp-637 results in reduced activity (Figure 4). These results suggest that Cgs Asp-635 is critical for Cgs activity and may be the residue that interacts with the ribose moiety of UDP-glucose. The mutant D635E, along with D635N, displays almost complete loss of in vitro and in vivo activity, indicating that both the size and the charge of the aspartate side chain are important for the interaction with UDP-glucose. Asp-637 may interact through Mg2+ (the divalent cation requires for Cgs activity) with phosphate and possibly stabilize the nucleotide diphosphate-leaving group during the nucleophilic attack. Interestingly, secondary structure prediction programs indicate that Cgs DxD motif is located in the loop following the ß4-strand, as it was described in SpsA.
The alignment of the Cgs region (475–818) with processive GTs shown in Figure 2B predicts that Cgs Glu-743, which aligns with the described or predicted catalytic residue of processive GTs, may be part of the catalytic site. However, the replacement of Glu-743 by alanine or aspartic acid has no effect on Cgs activity (Figure 5). To pursue the possibility that the ED(Y) motif forms part of the catalytic site, we performed site-directed replacement by alanine on Glu-769 and Asp-770. Enzymatic assays reveal that the mutant E769A displays almost complete loss of in vitro and in vivo Cgs activity, but D770A exhibits only a partial reduction of Cgs activity. The conservative change, E769D, restores only partially the activity in vivo but not in vitro (Figure 5). Together, these results support that Cgs Glu-769 is critical for Cgs glucosyltransferase activity and may act as the catalytic base during the transferase reaction. Catalytic residues are generally found on turns where they can move into position after the donor substrate and the acceptor have entered the catalytic pocket or cleft. Cgs Glu-769, according to the predicted secondary structure, is located in a loop (Figure 2B), which is in accordance with the hypothesis that Glu-769 may be the catalytic residue.
In the C-terminal of the Cgs region (475–818) (domain B), we have also identified the sequence motif RxxRW, and we have confirmed that the residues Arg-784 and Trp-785 are required for Cgs activity (Figure 6). It has been previously demonstrated that (Q/R)xxRW residues are required for GT activity (Nagahashi et al., 1995; Yoshida et al., 2000), and it has been suggested that this sequence motif may be important for the translocation of the elongating polysaccharide along the active site (Saxena and Brown, 1997). Only when the 3D structure of a GT with the full D, DxD, D/E, (Q/R)xxRW motif is resolved, the function of these residues will become clear.
Cgs has the three enzymatic activities required for the synthesis of cyclic glucan (i.e., initiation, elongation, and cyclization). Cgs region (475–818), where the D, DxD, E, RxxRW motif was identified, may be implicated in [UDP-Glc:ß-(1,2) oligosaccharide glucosyltransferase] activity being responsible for chain elongation during cyclic ß-1,2-glucan synthesis. However, we cannot rule out that this protein region may be also implicated in the initiation and cyclization reactions. In this regard, D711A and D715A mutants (particularly D715A) result in a reduced production of cyclic glucan in vitro and in vivo (Figure 5C and D), but the incorporation of glucose into Cgs is less affected (Figure 5A and B). These results indicate that in these mutants, glucose residues are incorporated into the protein, but the linear intermediate linked to the protein is not released efficiently through the cyclization reaction. Therefore, by the substitution of Asp-711 and Asp-715, two activities of this multifunctional protein were separated. Whether these residues are implicated in the cyclization reaction remains to be determined.
An important feature of Cgs is the ability to form a covalent intermediate (a reaction primer) during the synthesis of cyclic glucan. It can be considered, in terms of reaction mechanism, that Cgs may be related to the GT-8 enzyme glycogenin glucosyltransferase (EC 2.4.1.186
[EC]
), even though the glycogenin is a retaining enzyme. This enzyme is a self-glucosylating protein, which generates an oligosaccharide linked to the protein that serves as a primer for elongation by glycogen synthase. This oligosaccharide is composed of glucose residues in -1,4-glycosidic linkages, and the initial attachment of glucose to the protein is via a 1-O-tyrosine linkage (Smythe and Cohen, 1991). In spite of this similarity of the reaction mechanism, no sequence and structural similarities were detected between Cgs and glycogenin. Initiation of cyclic glucan synthesis consists of the transfer of the first glucose from UDP-glucose to an amino acid residue of the enzyme. It can be speculated that the initial linkage of glucose to the Cgs protein may be also through a 1-O-tyrosine linkage. Work is in progress to identify the amino acid residue and the type of linkage of the intermediate.
To relieve perceived problems both with torsional strain and with the need to synthesize polysaccharides with different alternating sugars, a two active-center model has been proposed for processive GT-2 members (Saxena et al., 1995; Saxena and Brown, 1997, 2000). These enzymes are responsible for the synthesis of polymers such as cellulose, chitin, and hyaluronic acid. The two catalytic-center model is based on the assumption that the residues of the D, DxD, D/E motif, spread over domains A and B, are sufficient to constitute two discrete glycosyl-transfer centers. Yet, the recent crystal structures of GTs revealed that all these residues are required to form a single viable catalytic center (Charnock and Davies, 1999; Gastinel et al., 1999; Pedersen et al., 2000; Unligil et al., 2000). An alternative two-center model could result from the dimerization of two polypeptide molecules acting in a concerted way. In Cgs, we have identified only one conserved D, DxD, D/E, (Q/R)xxRW motif located in the Cgs region (475–818) that may form a single catalytic center. Furthermore, to know whether Cgs self-association in the membrane is required for the synthesis of cyclic glucan, we over-expressed Cgs-inactive mutants in the WT background. The results of this suppression activity assay suggest that Cgs molecules do not self-assemble in the membrane, indicating that Cgs may function in the membrane as a monomer and not as an oligomer (Figure 7). It means that during the synthesis of cyclic glucan, the glucosyl transfer is intramolecular and not intermolecular. It has been reported that Pasteurella multocida hyaluronan synthase protein also functions as a monomer (Jing and DeAngelis, 2000), but it is not known for cellulose synthases and other processive GTs of GT family 2. Together, these results are compatible with a single addition model by which Cgs acts in the membrane as a monomer and uses the A and B domains (Figure 2C) to form a single center for substrate binding and the glycosyl-transfer reaction.
An additional feature of many GTs is their modularity (Coutinho et al., 2003). Previously, Cgs was not classified into any GT family. Instead, owing to the homology of the C-terminal domain with cellobiose and cellodextrin phosphorylases, Cgs was placed into the GH family 94. Cgs acts as an inverting GT, uses UDP-glucose as sugar donor, and requires Mg2+ ion as cofactor. Furthermore, the sequence- and site-directed mutagenesis analyses performed in this study indicate that Cgs possesses some of the motif characteristics of GT-2 family and may belong to this family, but the very low level of sequence similarity with GT-2 members prevents its classification into this family. Therefore, B. abortus Cgs and the other Cgs were classified into a new GT family related to GT-2 (GT-84). In summary, Cgs should be considered a bi-functional modular GT, with an N-terminal GT domain belonging to a new GT family related to GT-2 (GT-84) followed by a GH-94 glycoside hydrolase C-terminal domain (Figure 1).
Because at present no structural information is available for a processive-inverting GT and considering that most of them are membrane proteins, it will be interesting to determine the crystal structure of the globular Cgs domain (475–818) to understand in detail the mechanism of glycosyl transfer in this class of GTs.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionResultsDiscussionMaterials and MethodsSupplementary DataConflict of interest statementAcknowledgmentsReferences |
|---|
Bacterial strains, plasmids, and growth conditions
The bacterial strains and plasmids used in this work are summarized in Table I. E. coli was grown at 37°C in Luria–Bertani broth (Sangari et al., 1994
). Brucella strains were grown at 37°C in Brucella Broth (Difco Laboratories, Detroit, MI). If necessary, media were supplemented with appropriate antibiotics as follows: ampicillin, 100 µg/mL for E. coli; nalidixic acid, 5 µg/mL for B. abortus; and kanamycin, 50 µg/mL for E. coli and B. abortus.
|
All the protocols using live Brucellae strains were performed in a biosafety level-3 laboratory facility.
Sequence analysis
BLAST (Altschul et al., 1990) and PSI-BLAST (Altschul et al., 1997) searches were performed at the National Center for Biotechnology Information (NCBI) server. Protein fold recognition was performed using the 3D-PSSM program (Imperial College of Science, Technology & Medicine [http://www.sbg.bio.ic.ac.uk]) (Kelley et al., 2000). The secondary structure predictions were created by accessing the Jpred2 server (Cuff et al., 1998).
Construction of a cgs deletion mutant
To obtain a clean deletion of the cgs gene, plasmid pBC52, containing a 5.2-kb EcoRI fragment of the cgs gene, was digested with AvrII-BlpI, blunted with T4 DNA polymerase (New England Biolabs, Beverly, MA), and ligated to a dicistronic cassette, sacB-accI (Ugalde et al., 2003), digested with EcoRV. The resulting plasmid, named pBC52SG, has a fragment of the cgs gene interrupted by the cassette. This plasmid was electroporated into B. abortus S19; gentamicin-resistant, chloranfenicol-sensitive, and sucrose-sensitive colonies were selected; and double recombination events were confirmed by PCR. The resulting strain was designated B19SG. Plasmid pBC52 was digested with AvrII-BlpI, blunted, and religated; the resulting plasmid, which has a deletion of 2.2 kb, was named pBC30. This plasmid was digested with EcoRI, and the resulting 3.0-kb fragment was cloned into vector pK18mob (Schafer et al., 1994), yielding pK30. This plasmid was conjugated into strain B19SG, and sucrose-resistant and gentamicin- and ampicillin-sensitive colonies were selected. Double recombination events were confirmed by PCR using appropriate oligonucleotides. The resulting strain was designated S19cgs. The insertion of the gentamicin resistance cassette into the S19 strain was performed to generate an intermediate strain that was used to obtain a clean deletion of cgs. Afterward, the intermediate gentamicin-resistant strain B19SG was destroyed.
Plasmid construction and site-directed mutagenesis
Plasmid pBBR1MCS-2 (Kovach et al., 1995) was digested with SfiI, treated with T4 DNA polymerase (New England Biolabs), and religated, yielding pBBS. An 11.3-kb XhoI–BamHI fragment from pBA19 (Iñón de Iannino et al., 1998), containing the cgs gene and its own promoter, was ligated in pBBS digested with XhoI and BamHI; the resulting plasmid was named pBA22.
Site-directed mutagenesis of the cgs gene was carried out by PCR using the "megaprimer" method (Sarkar and Sommer, 1990) and a method based on the QuickChange kit from Stratagene (La Jolla, CA). The latter method, using pGTA as the template, was used to generate the desired change in amino acid residues Glu-769, Asp-770, Arg-781, Arg-784, and Trp-785. The resulting plasmid was digested with AatII, and the excised 1,2-kb AatII cgs fragment containing the desired mutation was ligated into pBA22 digested with the same enzyme.
All other mutants were generated by the "megaprimer" method, using pBA22 as the template and synthetic oligonucleotides with the appropriate nucleotide changes. A mutagenic primer was used in a PCR with a reverse primer. The first PCR product was purified with the Qiaex kit (Qiagen, GmBH, Hilden, Germany), and a second PCR was performed using this PCR product as megaprimer together with a forward primer. The second PCR product was purified and digested with SfiI and BlpI, and the excised 980-pb SfiI–BlpI fragment containing the desired mutation was ligated into pBA22 digested with the same enzymes.
Mutated DNA was sequenced with an automated model 373 DNA sequencer (Perkin–Elmer Applied Biosystems Division, Foster City, CA) according to manufacturer’s instructions to confirm PCR fidelity.
Plasmids containing the WT (pBA22) and the mutated cgs gene variants (pBA22 derivatives with the desired mutation) were introduced into B. abortus S19cgs strain by electroporation.
Preparation of permeabilized cells and total membranes
The preparation of permeabilized cells and total membranes of B. abortus strains was carried out as described previously (Briones et al., 1997). Protein concentration was determined by the method of Lowry and others (1951) using bovine serum albumin as a standard.
In vitro activity of Cgs
Cgs activity was determined as described previously (Briones et al., 1997). Briefly, permeabilized cells were incubated with UDP-[14C]glucose (500,000 cpm; 300 µCi/µmol) in 50 mM Tris–Cl buffer, pH 8.2, 5 mM MgCl2 at 28°C for 5, 10, and 20 min. Incorporation of [14C]glucose into Cgs protein (trichloroacetic acid [TCA]-insoluble fraction) and cyclic glucan (water-soluble fraction) was quantified by liquid scintillation. When comparing the activity of WT (S19cgs harboring pBA22) and mutant forms (S19cgs harboring pBA22 derivatives with the desired mutation) of Cgs, radioactivity recovered from S19cgs strain was taken as background and subtracted from reported values. Enzymatic activity was normalized by the relative level of protein expression and expressed as a percentage of that of the WT Cgs.
TCA-insoluble fractions were also subjected to SDS–PAGE (Laemmli, 1970). Proteins were stained with Coomassie Blue to determine the relative level of Cgs protein expression during the assay by using the 1D Image Analysis software (Kodak Digital Science, New Haven, CT, USA), and radioactivity was detected by fluorography, as described previously (Briones et al., 1997), to verify the specific incorporation of the radioactivity into Cgs.
To check the specific incorporation of [14C]glucose into the cyclic glucan, 10,000 cpm from water-soluble fractions was subjected to thin-layer chromatography (TLC) and radioactivity detected by autoradiography (Briones et al., 1997). Each of the reaction products was found to co-migrate with the authentic cyclic ß-1,2-glucan (data not shown).
In vivo activity of Cgs
Cells from cultures of B. abortus strains were grown for 48 h and harvested by centrifugation. Cyclic ß-1,2-glucans were extracted from cell pellets with ethanol (70% ethanol, 1 h at 37°C). Ethanolic extracts were centrifuged, and supernatants were dried in a speed-vac centrifuge. Extracted glucans were dissolved in 70% ethanol and submitted to TLC as described previously (Iñón de Iannino et al., 1998). TLC plates were developed by spraying with 5% sulfuric acid in ethanol and heating for 10 min at 125°C. Cyclic glucans were obtained from the same number of cells, and SDS–PAGE analysis confirmed no significant difference in the amounts of enzyme synthesized by the different strains.
Suppression activity assay
Plasmids (pBA22 derivatives) expressing the mutant forms D635E and W785T of Cgs were introduced into B. abortus S19 strain by electroporation. Transconjugants were selected in plates containing kanamycin (50 µg/mL), and cyclic glucan production was analyzed as described above. The expression level of WT and mutant forms of Cgs was analyzed by Coomassie Blue-staining SDS–PAGE.
|
Supplementary Data |
|---|
|
TopAbstractIntroductionResultsDiscussionMaterials and MethodsSupplementary DataConflict of interest statementAcknowledgmentsReferences |
|---|
Supplementary data are available at Glycobiology online (http://glycob.oxfordjournals.org/).
|
Conflict of interest statement |
|---|
|
TopAbstractIntroductionResultsDiscussionMaterials and MethodsSupplementary DataConflict of interest statementAcknowledgmentsReferences |
|---|
None declared.
|
Acknowledgments |
|---|
|
TopAbstractIntroductionResultsDiscussionMaterials and MethodsSupplementary DataConflict of interest statementAcknowledgmentsReferences |
|---|
We thank R.A. Geremia for scientific advice, Susana Raffo for the synthesis and purification of UDP-[14C]glucose, and Pablo Briones for technical assistance.
This work was supported by a grant from the Agencia Nacional de Promoción Científica y Técnica, SECyT, Buenos Aires, Argentina, PICT 2000 no. 01-09194, the Consejo Nacional de Investigaciones Científicas y Técnicas, CONICET, Buenos Aires, Argentina, PIP 2346, and the Universidad Nacional de General San Martín, Buenos Aires, Argentina, PIA S-05/21. A.E.C. and M.S.R. are fellows of Consejo Nacional de Investigaciones Científicas y Técnicas, CONICET, Buenos Aires, Argentina. N.I.D.I. and R.A.U. are members of the research carrier of CONICET.
|
Abbreviations |
|---|
3D-PSSM, three-dimensional position-specific scoring matrix; Cgs, cyclic glucan synthase; CR, conserved region; GTs, glycosyltransferases; PSI-BLAST, position-specific iterated BLAST; SDS–PAGE, sodium dodecyl sulphate–polyacrilamide gel electrophoresis; TCA, trichloroacetic acid; TLC, thin-layer chromatography; TMS, transmembrane-spanning segment
|
References |
|---|
|
TopAbstractIntroductionResultsDiscussionMaterials and MethodsSupplementary DataConflict of interest statementAcknowledgmentsReferences |
|---|
Altschul, S.F., Gish, W., Miller, W., Myers, E.W., and Lipman, D.J. (1990) Basic local alignment search tool. J. Mol. Biol., 215, 403–410.[CrossRef][ISI][Medline]
Altschul, S.F., Madden, T.L., Schaffer, A.A., Zhang, J., Zhang, Z., Miller, W., and Lipman, D.J. (1997) Gapped BLAST and PSI-BLAST: a new generation of protein database search programs. Nucleic Acids Res., 25, 3389–3402.
Arellano-Reynoso, B., Lapaque, N., Salcedo, S., Briones, G., Ciocchini, A.E., Ugalde, R., Moreno, E., Moriyon, I., and Gorvel, J.P. (2005) Cyclic beta-1,2-glucan is a Brucella virulence factor required for intracellular survival. Nat. Immunol., 6, 618–625.[CrossRef][ISI][Medline]
Blatch, G.L. and Lassle, M. (1999) The tetratricopeptide repeat: a structural motif mediating protein–protein interactions. Bioessays, 21, 932–939.[CrossRef][ISI][Medline]
Bourne, Y. and Henrissat, B. (2001) Glycoside hydrolases and glycosyltransferases: families and functional modules. Curr. Opin. Struct. Biol., 11, 593–600.[CrossRef][ISI][Medline]
Breton, C. and Imberty, A. (1999) Structure/function studies of glycosyltransferases. Curr. Opin. Struct. Biol., 9, 563–571.[CrossRef][ISI][Medline]
Breton, C., Snajdrova, L., Jeanneau, C., Koca, J., and Imberty, A. (2006) Structures and mechanisms of glycosyltransferases. Glycobiology, 16, 29R–37R.
Briones, G., Iñón de Iannino, N., Roset, M., Vigliocco, A., Paulo, P.S., and Ugalde, R.A. (2001) Brucella abortus cyclic beta-1,2-glucan mutants have reduced virulence in mice and are defective in intracellular replication in HeLa cells. Infect. Immun., 69, 4528–4535.
Briones, G., Iñón de Iannino, N., Steinberg, M., and Ugalde, R.A. (1997) Periplasmic cyclic 1,2-beta-glucan in Brucella spp. is not osmoregulated. Microbiology, 143 (Pt 4), 1115–1124.[Abstract]
Busch, C., Hofmann, F., Selzer, J., Munro, S., Jeckel, D., and Aktories, K. (1998) A common motif of eukaryotic glycosyltransferases is essential for the enzyme activity of large clostridial cytotoxins. J. Biol. Chem., 273, 19566–19572.
Campbell, J.A., Davies, G.J., Bulone, V., and Henrissat, B. (1997) A classification of nucleotide-diphospho-sugar glycosyltransferases based on amino acid sequence similarities. Biochem. J., 326 (Pt 3), 929–939.[ISI][Medline]
Charnock, S.J. and Davies, G.J. (1999) Structure of the nucleotide-diphospho-sugar transferase, SpsA from Bacillus subtilis, in native and nucleotide-complexed forms. Biochemistry, 38, 6380–6385.[CrossRef][Medline]
Charnock, S.J., Henrissat, B., and Davies, G.J. (2001) Three-dimensional structures of UDP-sugar glycosyltransferases illuminate the biosynthesis of plant polysaccharides. Plant Physiol., 125, 527–531.
Chiu, C.P., Watts, A.G., Lairson, L.L., Gilbert, M., Lim, D., Wakarchuk, W.W., Withers, S.G., and Strynadka, N.C. (2004) Structural analysis of the sialyltransferase CstII from Campylobacter jejuni in complex with a substrate analog. Nat. Struct. Mol. Biol., 11, 163–170.[CrossRef][ISI][Medline]
Ciocchini, A.E., Roset, M.S., Iñón de Iannino, N., and Ugalde, R.A. (2004) Membrane topology analysis of cyclic glucan synthase, a virulence determinant of Brucella abortus. J. Bacteriol., 186, 7205–7213.
Corbel, J.M. (1997) Brucellosis: an overview. Emerg. Infect. Dis., 2, 213–221.
Coutinho, P.M., Deleury, E., Davies, G.J., and Henrissat, B. (2003) An evolving hierarchical family classification for glycosyltransferases. J. Mol. Biol., 328, 307–317.[CrossRef][ISI][Medline]
Cuff, J.A., Clamp, M.E., Siddiqui, A.S., Finlay, M., and Barton, G.J. (1998) JPred: a consensus secondary structure prediction server. Bioinformatics, 14, 892–893.
Garinot-Schneider, C., Lellouch, A.C., and Geremia, R.A. (2000) Identification of essential amino acid residues in the Sinorhizobium meliloti glucosyltransferase ExoM. J. Biol. Chem., 275, 31407–31413.
Gastinel, L.N., Cambillau, C., and Bourne, Y. (1999) Crystal structures of the bovine beta4galactosyltransferase catalytic domain and its complex with uridine diphosphogalactose. EMBO J., 18, 3546–3557.[CrossRef][ISI][Medline]
Gulberti, S., Fournel-Gigleux, S., Mulliert, G., Aubry, A., Netter, P., Magdalou, J., and Ouzzine, M. (2003) The functional glycosyltransferase signature sequence of the human beta 1,3-glucuronosyltransferase is a XDD motif. J. Biol. Chem., 278, 32219–32226.
Hoyer, B.H. and McCullough, N.B. (1968) Homologies of deoxyribonucleic acids from Brucella ovis, canine abortion organisms, and other Brucella species. J. Bacteriol., 96, 1783–1790.
<