Abstract
(1,3;1,4)-β-d-Glucans are widely distributed in the cell walls of grasses (family Poaceae) and closely related families, as well as some other vascular plants. Additionally, they have been found in other organisms, including fungi, lichens, brown algae, charophycean green algae, and the bacterium Sinorhizobium meliloti. Only three members of the Cellulose Synthase-Like (CSL) genes in the families CSLF, CSLH, and CSLJ are implicated in (1,3;1,4)-β-d-glucan biosynthesis in grasses. Little is known about the enzymes responsible for synthesizing (1,3;1,4)-β-d-glucans outside the grasses. In the present study, we report the presence of (1,3;1,4)-β-d-glucans in the exopolysaccharides of the Gram-positive bacterium Romboutsia ilealis CRIBT. We also report that RiGT2 is the candidate gene of R. ilealis that encodes (1,3;1,4)-β-d-glucan synthase. RiGT2 has conserved glycosyltransferase family 2 (GT2) motifs, including D, D, D, QXXRW, and a C-terminal PilZ domain that resembles the C-terminal domain of bacteria cellulose synthase, BcsA. Using a direct gain-of-function approach, we insert RiGT2 into Saccharomyces cerevisiae, and (1,3;1,4)-β-d-glucans are produced with structures similar to those of the (1,3;1,4)-β-d-glucans of the lichen Cetraria islandica. Phylogenetic analysis reveals that putative (1,3;1,4)-β-d-glucan synthase candidate genes in several other bacterial species support the finding of (1,3;1,4)-β-d-glucans in these species.
Similar content being viewed by others
Introduction
(1,3;1,4)-β-d-Glucans are unbranched homoglucans composed of both 3- and 4-linked β-glucopyranosyl (Glcp) residues. The occurrence of the (1,3;1,4)-β-d-glucans has been the subject of considerable interest1,2. They are best known in the cell walls of cereals and other grasses (family Poaceae). However, they have also been found in some other vascular plants, including those in families closely related to the Poaceae, the lycophyte Selaginella moellendorffii, and in horsetails (Equisetum spp.), as well as other monilophytes (ferns)3. Additionally, they have been occasionally reported in several other taxa, including charophycean green algae4, brown algae5, fungi6, and the fungal symbiont of lichens, such as Cetraria islandica (where the (1,3;1,4)-β-d-glucan is known as lichenin or lichenan)7, as well as the Gram-negative bacterium Sinorhizobium meliloti8. In the backbone of (1,3;1,4)-β-d-glucans, two and three consecutive (1 → 4) bonds [β-Glcp-(1 → 4)-β-Glcp-(1 → 4)-β-Glcp (a cellotriosyl unit) and β-Glcp-(1 → 4)-β-Glcp-(1 → 4)-β-Glcp-(1 → 4)-β-Glcp (a cellotetraosyl unit), respectively] are usually found separated by a single (1 → 3) bond; consecutive (1 → 3) bonds have never been found. The proportions of the cellotriosyl and cellotetraosyl units can be determined by treating the (1,3;1,4)-β-d-glucans with the enzyme lichenase (EC 3.2.1.73), which hydrolyzes them by breaking all the (1 → 4) bonds that immediately follow (1 → 3) bonds on the reducing end side, producing the triose β-Glcp-(1 → 4)-β-Glcp-(1 → 3)-β-Glc (abbreviated to G4G3G, DP3) and the tetraose β-Glcp-(1 → 4)-β-Glcp-(1 → 4)-β-Glcp-(1 → 3)-β-Glc (G4G4G3G, DP4) oligosaccharides that are easily separated and quantified. The molar ratio of these two oligosaccharides (DP3/DP4 ratio) varies widely with taxa, indicating variation in the molecular assembly of (1,3;1,4)-β-d-glucans1. (1,3;1,4)-β-d-glucans, with very high or very low DP3/DP4 ratios, such as lichenin that has a very high ratio of 20.2–24.6:1, are more regular and hence can more readily align over extended regions, resulting in them being insoluble7,9. (1,3;1,4)-β-d-glucans from cereals and other grasses have intermediate DP3/DP4 ratios of 2–4.5:1, suggesting these (1,3;1,4)-β-d-glucans are irregular and soluble in the matrix phase of primary cell walls of the Poaceae1,7,9.
Different genes and enzymes are involved in the biosynthesis of (1,3;1,4)-β-d-glucans, depending on the organisms involved. In the Poaceae, the (1,3;1,4)-β-d-glucans are synthesized by Carbohydrate Active enZyme (CAZy) family 2 glycosyltransferases (GT2) encoded by cellulose-synthase-like (CSL) gene families. To date, only the CSLF, CSLH, and CSLJ gene families have been characterized and are restricted to the Poaceae and other monocotyledons10,11,12. In the lycophyte Selaginella moellendorffii, (1,3;1,4)-β-d-glucan may be produced by a synthase encoded by an ortholog of the AGlcS gene that in the moss Physcomitrella patens encodes a GT2 enzyme that produces a mixed-linkage (1,3;1,4)-arabinoglucans13. Interestingly, these S. moellendorffii and P. patens GT2 enzymes resemble a fungal (1,3;1,4)-β-d-glucan synthase: the Three-Four Transferase 1 (Tft1) of Aspergillus fumigatus6. In brown algae, which have only cellotriosyl units separated by single (1 → 3) bonds in their (1,3;1,4)-β-d-glucans, their putative (1,3;1,4)-β-d-glucan synthase genes are genetically more closely related to the Bacterial cellulose synthase (Bcs) gene5. The enzymes responsible for the synthesis of bacterial exopolysaccharides (EPSs) are often encoded by the genes in the bacterial GT2 family, for example, the Bacterial cellulose synthase A (BcsA) and Curdlan synthase (CrdS)14. The structural composition of EPSs has a strong correlation with bacterial pathogenicity, including bacteria–host interaction, resistance to host immune response, and the formation of biofilms15. In the bacterium S. meliloti, the (1,3;1,4)-β-d-glucans in the EPS are unusual in having only cellobiosyl units [β-Glcp-(1 → 4)-β-Glcp] separated by single (1 → 3) bonds and are implicated in bacterial aggregation and biofilm formation8, as well as in host plant root attachment. Of the putative bacterial genes currently identified in (1,3;1,4)-β-d-glucan biosynthesis direct gain-of-function has not been demonstrated. The Bcs operon encodes for a protein, which contains D, D, D, QXXRW, and PilZ domains, responsible for the biosynthesis of bacterial cellulose16,17. It has therefore been suggested that the two genes operon bgsBA encoding protein (with “bgs” standing for “β-glucan synthesis”), which also contains D, D, D, QXXRW motifs, could be involved in the (1,3;1,4)-β-d-glucan biosynthesis in S. meliloti8. Even though BgsA does not have a PilZ domain, the expressed BgsA C-terminal peptide can still bind to the bacterial secondary messenger cyclic diguanylate (c-di-GMP)8, which is consistent with the finding of Cereibacter sphaeroides cellulose synthase (BcsA), in which activity is facilitated through c-di-GMP binding18,19,20,21,22.
Very few bacterial GTs have had their function biochemically characterized and, of these, most were from Gram-negative bacteria. The small intestine microbes are typically Gram-positive bacteria23, and their EPSs have a major influence on the microbiota population24, as well as helping commensals to prevent pathogen colonization25. Probiotic microorganisms contribute to the modulation of intestinal commensals and benefits in health. To this end, a commensal bacterium in a novel phylotype, named CRIB, was found to be increased by probiotic intake, which resulted in a decrease in the severity of pancreatitis and associated sepsis in an experimental rat model for acute pancreatitis studies26. Romboutsia ilealis CRIBT was then described as the type strain of the genus of Romboutsia. The genus Romboutsia is in the family Peptostreptococcaceae which has several members that are common intestinal microbes, including Clostridioides difficile27. Most Romboutsia species have an intestinal origin, but their roles in the gastrointestinal tract of humans and mammals are less well characterized.
Here we report the presence of (1,3;1,4)-β-d-glucans in the EPS of the Gram-positive bacterium R. ilealis CRIBT and investigate the role of the RiGT2 gene in (1,3;1,4)-β-d-glucan biosynthesis using a stably transfected Saccharomyces cerevisiae yeast model (Fig. 1). Our findings demonstrate the involvement of a prokaryotic GT2 in the synthesis of (1,3;1,4)-β-d-glucans. The heterologously synthesized (1,3;1,4)-β-d-glucans exhibit a high proportion of cellotriosyl units, similar to that found in lichenin7,28.
Results
(1,3;1,4)-β-d-glucan is a component of R. ilealis exopolysaccharides
We prepared R. ilealis CRIBT EPSs, which were chemically converted to partially methylated alditol acetates (PMAAs) mixture for glycosidic linkage analyses (Supplementary Fig. 1). All EPS preparations had 4-linked and 3-linked glucopyranosyl residues in relatively high abundance. The 4-linked Glcp: 3-linked Glcp ratio was approximately 4:1, which is higher than that of grass (1,3;1,4)-β-d-glucans (2.2:1–2.6:1)29.
EPS preparations were also treated with lichenase, and the enzyme-released oligosaccharides were separated by high-performance anion-exchange chromatography with pulsed amperometric detection (HPAEC-PAD). Retention times of the analytes were compared with those of a series of commercially available (1,3;1,4)-β-d-glucan oligosaccharides of known structures obtained by treating barley (1,3;1,4)-β-d-glucans with lichenase (Fig. 2). Although several unexpected oligosaccharides were detected in both the control (buffer only) and the lichenase-treated samples, two oligosaccharides released from R. ilealis CRIBT EPSs eluted at the same retention times as the commercially available DP3 (G4G3GR) and DP4 (G4G4G3GR), derived from barley (1,3;1,4)-β-d-glucan. The ratio of DP3 to DP4 was close to 2:1. This result confirmed that (1,3;1,4)-β-d-glucans are secreted by R. ilealis CRIBT in its EPSs.
R. ilealis CRIBT possesses a gene possibly encoding (1,3;1,4)-β-D-glucan synthase
After confirming (1,3;1,4)-β-d-glucans do exist in R. ilealis EPSs, we searched for a possible bacterial (1,3;1,4)-β-d-glucan synthase in the Carbohydrate Active Enzymes database (CAZY, http://www.cazy.org/)30 and National Center for Biotechnology Information protein database (NCBI) (Supplementary Table 1)16. Out of 5 putative GT2 proteins in the R. ilealis CRIBT database, only one GT2 protein (CED93608.1) contains the polysaccharide synthase signature sequence [D,D,D,Q(Q/R)XRW] signature. We, therefore, named the gene RiGT2 (R. ilealis glycosyltransferase family 2).
Expression of the RiGT2 in Saccharomyces cerevisiae resulted in the secretion of (1,3;1,4)-β-d-glucans
To characterize the function of the RiGT2 gene, the gene was cloned into the yeast-enhanced GFP-fusion vector pDDGFP-2 (modified from p424 GAL1, ATCC® 87329™) (Supplementary Fig. 2)31,32. The RiGT2 gene was placed under the control of the GAL1 promoter, with a Tobacco Etch Virus protease (TEVp)-eGFP-8His sequence added at the C-terminus. Membrane-bound RiGT2 was expressed in the quadruple knockout yeast strain (LoGSA), which exhibits low endogenous glucosyltransferase background activities due to deficient FKS1, GSY1, and GSY2 genes33, and RiGT2 gene homologs are absent in the yeast genome.
Following RiGT2 expression, the LoGSA cells were disrupted with glass beads. The membrane fractions were collected from the total cell lysate, and the membrane proteins were separated using reducing SDS–PAGE. This resolved a ~100 kDa protein band detected by in-gel fluorescence (Supplementary Fig. 3a), which corresponds roughly to the predicted molecular weight of RiGT2 protein with C-terminal GFP (102.7 kDa) (Supplementary Fig. 4). Total RiGT2 protein expression was determined by whole-cell fluorescence calibrated against a GFP standard31, and most colonies screened produced ~1 mg of RiGT2-GFP per liter culture (Supplementary Table 2). The identity of the RiGT2 protein was confirmed by LC–MS coupled with in-gel trypsin digestion (Supplementary Figs. 3b and 5).
We hypothesized that RiGT2 is involved in (1,3;1,4)-β-d-glucan biosynthesis, and hence, LoGSA expressing the RiGT2 gene is expected to produce (1,3;1,4)-β-d-glucans. The success of this approach relies heavily on the correct folding of the recombinant RiGT2, and the accessibility of the UDP-glucose and ancillary proteins, if any, that are needed for (1,3;1,4)-β-d-glucans synthesis in LoGSA. We then determined if synthesized (1,3;1,4)-β-d-glucans were present in the RiGT2 transgenic LoGSA EPS preparation (alcohol-insoluble residue, AIR) using oligosaccharide profiling after lichenase treatment (Fig. 1). We detected (1,3;1,4)-β-d-glucan with a DP3/DP4 molar ratio of 15:1 and no (1,3;1,4)-β-d-glucan oligosaccharides were detected in the EPS preparation (AIR) of control LoGSA carrying empty vector (EV) (Fig. 3, Supplementary Fig. 6). Compared to the oligosaccharide profile of the EPSs isolated from R. ilealis CRIBT, the (1,3;1,4)-β-d-glucan synthesized by RiGT2 LoGSA had a much higher DP3:DP4 ratio (15:1) than the (1,3;1,4)-β-d-glucan in R. ilealis EPS (2:1). MS analysis of the purified oligosaccharides from RiGT2 LoGSA gave ions with m/z [M+Na]+ 527, corresponded to DP3 (Hex3) (Fig. 4a, b). The result was further confirmed by linkage analysis of the DP3 oligosaccharides (Fig. 4c, Supplementary Fig. 7). The oligosaccharide alditol resulting from the NaBD4 reduction of DP3 was found to have a 1:1:1 ratio of 3-linked glucitol (reduced reducing end): 4-linked Glcp: t-Glcp (non-reducing end) (Supplementary Table 3), providing strong evidence for a DP3 structure of β-Glcp-(1 → 4)-β-Glcp-(1 → 3)-β-Glc (G4G3G). Moreover, a specific monoclonal antibody against (1,3;1,4)-β-d-glucan (BS400-3), was used with indirect immunofluorescence microscopy to locate (1,3;1,4)-β-d-glucan (Fig. 5). The confocal microscopy images showed that the monoclonal antibody targeting (1,3;1,4)-β-d-glucan labeled only the induced RiGT2 LoGSA cells (Fig. 5c), but was absent in non-induced LoGSA cells (Fig. 5i). The same conclusion was reached using the same antibody with indirect immunogold microscopy using a colloidal-gold-labeled secondary antibody in conjunction with transmission electron microscopy (TEM). Intense colloidal-gold labeling was found only over the cytoplasm of induced RiGT2 LoGSA cells (Supplementary Fig. 8).
To further verify the specific activity of RiGT2, the microsomal fractions (MFs) from RiGT2 LoGSA were isolated, and glucan synthase activity was measured by the radiometric in vitro assay34, using UDP-[14C]Glc as a radiotracer. The reactions were stopped by using an ethanol solution (66% aqueous ethanol) and the alcohol-insoluble polymers collected by filtration and the incorporated [14C]Glc quantified by scintillation counting. Figure 6a shows that alcohol-insoluble glucan polymers were synthesized in RiGT2-containing MFs in the presence of UDP-[14C]Glc and the [14C]-labeled polymers were degradable by lichenase, suggesting (1,3;1,4)-β-d-glucans were synthesized in RiGT2-containing MFs in vitro. Our gain-of-function yeast model successfully expressed the RiGT2 gene and was shown to encode the biosynthetic enzyme essential for (1,3;1,4)-β-d-glucan biosynthesis.
Structure analysis of theoretical RiGT2 models
To investigate the possible function of RiGT2 based on structural homology with other characterized GT2 enzymes, we generated two theoretical RiGT2 models using both non-template-based AlphaFold235 modeling and template-based modeling with SWISS-MODEL36. The theoretical model generated by AlphaFold2, RiGT2A, showed a high degree of reliability, as indicated by an overall high per-residue confidence metric (pLDDT; predicted local distance difference test) of 92.1, and a pTM score (predicted template modeling score) of 0.895. RiGT2 contains 637 amino acids that are predicted to have high fold similarity to C. sphaeroides cellulose synthase catalytic subunit CsBcsA (PDB 4HG6) (Fig. 7a, b). Two regions in the theoretical RiGT2 model displayed low or very low confidence scores (pLDDT < 70), namely residues 434–450 and the C-terminal region 609–637. Due to the poor prediction quality of the C-terminal region, residues 612–637 were deleted from the model. In CsBcsA, this region forms a long extension originating from the PilZ domain that does not interact directly with the polysaccharide-binding channel. Calculated on the structure-adjusted sequence alignment of the CsBcsA and RiGT2 sequences, the pair-wise sequence identity is 25.7%.
Although AlphaFold2 is a highly reliable tool for generating accurate structural models, we anticipated that the model generated by AlphaFold2 would be biased towards bacterial cellulose synthase CsBcsA (PDB 4HG6) due to the larger amount of experimental data available for cellulose synthases. Therefore, we generated a second model, RiGT2S, using SWISS-MODEL based on a single template, plant (1,3;1,4)-β-d-glucan synthase HvCslF6 (PDB 8DQK37), to provide complementary information to the RiGT2A model. The two models have inherited features from their respective parental structures. In the case of the AlphaFold2 model, this is not from a single template structure but the result of AlphaFold2 training on all available homologous templates available in the Protein Data Bank, which currently is dominated by bacterial cellulose synthase homologs. The RiGT2 model generated by SWISS-MODEL using HvCslF6 as a template has severe template bias, which is expected when only one template is used. Therefore, any similarities or differences observed for the two theoretical models (Supplementary Fig. 9) will in most parts only reflect the similarities and differences between the experimental structures of CsBcsA and HvCslF6.
The RiGT2A model displays most of the features of CsBcsA with an extramembrane catalytic glycosyltransferase (GT) domain, a transmembrane (TM) domain, and a C-terminal PilZ domain (Fig. 7a, b and Supplementary Fig. 10). The GT domain features seven β-strands and six α-helices and contains the [D,D,D,Q(Q/R)XRW] signature found in GT2 polysaccharide synthases38 (Fig. 7c and Supplementary Fig. 10). The first aspartate residue of the motif is Asp139, which has been assigned a role in coordinating the uridine ring in the UDP-Glc donor. The next aspartate of the motif is Asp188 in the DXD motif (186DAD188). Asp188 is expected to stabilize the diphosphate group of the donor via the metal cofactor (Mg2+), whereas Asp186 is predicted to stack with the ribose C4’–C5’ moiety of the donor’s UDP moiety. The last aspartate Asp280 is positioned in the conserved TED motif and is predicted to act as a catalytic base. The [Q(Q/R)XRW] motif is defined by 316QRDRW320 in RiGT2. Based on the CsBcsA structure, Trp320 in RiGT2 could stack with a polysaccharide chain as Trp383 stacks with Glc-17 at the acceptor site in CsBcsA (PDB 4HG619; PDB 4P00 and 4P0220; PDB 5EJ1 and 5EJZ18).
The TM domain in RiGT2A is predicted to contain seven TM helices (TMH2–8 in CsBcsA) that form a polysaccharide-binding channel. The model also includes three interface helices (IFH1–3). RiGT2 lacks the N-terminal TMH that coils with the periplasmic CsBcsB subunit in the CsBcsA–CsBcsB complex19. In the C. sphaeroides genome (GenBank CP000143.2), the genes coding for the catalytic subunit BcsA (gene RSP_0333) and the BcsB subunit (gene RSP_0332) are located in an operon, whereas in the R. ilealis genome (GenBank LN555523.1), the gene coding for RiGT2 (gene CRIB_856) does not cluster with a corresponding bcsB gene. Thus, based on the amino-acid sequence and the gene organization, RiGT2 is not likely to depend on a BcsB subunit, at least not of the same type as C. sphaeroides and related bacterial sugar polymerases where BcsB is required for polysaccharide synthesis39,40. In this respect, RiGT2 is more like the homotrimeric plant CesA741 and CesA842 which do not require a BcsB subunit, and whose catalytic domains also lack the N-terminal TMH. However, the trimeric plant enzymes instead have an extra TMH at the C-terminus that participates in oligomerization into homotrimers.
Bacterial sugar polymerases that produce biofilm polysaccharides contain a C-terminal PilZ domain that serves as an effector subunit that binds the secondary messenger cyclic-di-GMP for intracellular signaling, and in C. sphaeroides cellulose synthase, binding of cyclic-di-GMP stimulates the synthesis of bacterial cellulose20. In RiGT2, residues 522–594 are predicted to form a PilZ domain shaped as a six-stranded antiparallel β-barrel similar to that in CsBcsA (residues 584–676). The PilZ region displays very low overall sequence similarity to CsBcsA PilZ (Supplementary Fig. 10), but the PilZ [RXXXR] signature is present in RiGT2 (Supplementary Fig. 11), 517RTSER521, as well as the [D/NXSXXG] motif43, 554NLSEKG550.
A conserved gating loop in the GT domain of CsBcsA (residues 499–512) controls access to the active site20 (Fig. 7a, b and Supplementary Fig. 10). The gating loop in CsBcsA adopts different conformations (open or resting state) in response to the presence or absence of cyclic-di-GMP, where the resting state conformation blocks access to the active site, secured by a salt link between Arg580 in the PilZ domain and Glu371 in the GT domain. In the open state, Arg580 swings 180° to interact with cyclic-di-GMP, thereby breaking the salt link to Glu371. The corresponding gating loop in RiGT2 comprises residues 436–448. In both CsBcsA and RiGT2 the N-terminal region of the gating loop includes the signature [FXVTXK] (439FNVTLK444 in RiGT2 and 503FAVTAK508 in CsBcsA) while the rest of the residues of the gating loop are non-conserved. In the AlphaFold2 model of RiGT2, the gating loop region is one of two regions with a low confidence score. The residues forming the regulatory salt link in CsBcsA (Arg580 and Glu371) are not fully conserved in RiGT2 and correspond to Arg517 and Thr308, respectively.
The channel in CsBcsA spans ~50 Å and accommodates 11 β-1,4-linked D-glucosyl units (glucosyl 8–18 in PDB 4HG6) at the acceptor side. A superposition of the CsBcsA structure with the RiGT2A and RiGT2S models shows that RiGT2 would be able to accommodate 11 glucosyl units (glucosyl 8–18 in PDB 4HG6). The key side chains that constitute the individual glucosyl-binding sites were analyzed in CsBcsA, RiGT2, and HvCslF6 (Supplementary Table 4). While some side chains exist in all three proteins, there is no single glucosyl-binding site that is fully conserved. Since RiGT2 and HvCslF6 are both predicted to synthesize (1,3;1,4)-β-d-glucans, we expected the residues of the binding sites in RiGT2 to be more similar to HvCslF6 than to CsBcsA, but RiGT2 has more side chains in common with CsBcsA (Supplementary Table 4).
Since a polysaccharide chain, cellulose or (1,3;1,4)-β-d-glucan, needs to be translocated through the channel during synthesis, the side-chain interactions must not be too specific, and different types of side chains could fulfill the purpose. The more important structural feature of a channel that accommodates a nascent (1,3;1,4)-β-d-glucan, as opposed to cellulose, is the ability to adjust to the unique structural features introduced by the β-1,3 linkages. Due to the expected bias from cellulose synthase templates, such features are more difficult to analyze in the RiGT2A model. Indeed, the shape and size of the channel in the RiGT2A model are suspiciously similar to CsBcsA and appear too narrow to be able to accommodate a glucan with β-1,3 glycosidic bonds at every third linkage. Since the polysaccharide-binding channel is formed mainly by the TMHs, it is likely that the principal means by which to adapt the channel to (1,3;1,4)-β-d-glucan is by repositioning one or several TMHs. Comparison of the TMHs in CsBcsA and HvCslF6 indeed shows that the TMHs are somewhat displaced in HvCslF6 to expand the channel (also inherited by the RiGT2S model), which could be sufficient to accommodate a (1,3;1,4)-β-d-glucan chain (Supplementary Fig. 12). The slightly more spacious channel in HvCslF6, at least for some subsite, can be also seen by comparing the representations of the molecular surfaces for CsBcsA, HvCslF6 and the theoretical RiGT2 models (Supplementary Fig. 13).
By analyzing the sequences and models, experimental and predicted, for bacterial CsBcsA, RiGT2, and plant HvCslF6, our conclusion is that RiGT2 shares a higher degree of similarity with CsBcsA, but that the precise orientations of the TMHs may be more similar to the TM domain in HvCslF6. Thus, while the theoretical models cannot offer proof of RiGT2 synthesizing (1,3;1,4)-β-d-glucan, we find nothing that argues against this function at the structural level. However, the predicted structural features of RiGT2 would also be compatible with the synthesis of cellulose.
Evaluation of the PilZ domain in RiGT2 (1,3;1,4)-β-d-glucan biosynthesis
The similarities between CsBcsA and RiGT2 at the sequence level strongly suggest RiGT2 is also regulated by a PilZ domain. To the best of our knowledge, c-di-GMP is absent from the yeast secondary messenger system, but the RiGT2 LoGSA cells could still synthesize (1,3;1,4)-β-d-glucan. This apparent discrepancy required further investigation and the effect of different c-di-GMP concentrations on RiGT2 activity was examined by incubating RiGT2 LoGSA microsomal preparations with c-di-GMP (30, 60, or 90 μM) and UDP-[14C]Glc, and monitoring the (1,3;1,4)-β-d-glucan synthase activity with a radiometric assay. Surprisingly, there was no increase in synthase activity with an increase in c-di-GMP concentration (Fig. 6b), showing that (1,3;1,4)-β-d-glucan synthesis is not dependent on the presence of c-di-GMP. To further verify the direct binding of c-di-GMP to the PilZ domain of RiGT2, we expressed and purified the recombinant PilZ domain of RiGT2 with C-terminal GFP fusion protein (RiPilZ-GFP) and performed in vitro binding studies using Isothermal Titration Calorimetry (ITC) (Supplementary Fig. 14a) and Surface Plasmon Resonance (SPR) (Supplementary Fig. 14b), but no direct binding was detected. This lack of binding could be due to the requirement of interactions provided by the absent GT domain or interference from possible glycosylation in the yeast system. Alternatively, it could indicate that RiGT2 mediates the synthesis of (1,3;1,4)-β-d-glucan through a different mechanism that does not strictly require the binding of c-di-GMP.
To further verify that the PilZ domain is necessary for RiGT2 activity, we generated a RiGT2 with the PilZ domain deleted (RiGT2-PilZD) and transformed the gene into LoGSA. Interestingly, after digesting with lichenase the AIR prepared from transgenic yeast expressing the C-terminal truncated RiGT2, we found no (1,3;1,4)-β-d-glucan was synthesized (Supplementary Fig. 15a). This result was further supported by using the radiometric in vitro assay (Supplementary Fig. 15c), and we concluded that no (1,3;1,4)-β-d-glucan synthase activity was detected when the PilZ domain was cropped from RiGT2. Despite no interaction between the RiGT2 PilZ domain and the secondary messenger c-di-GMP, the presence of the C-terminal PilZ domain remains crucial for RiGT2 (1,3;1,4)-β-d-glucan synthase activity, possibly by stabilizing the protein structure or regulating enzymatic activity through conformational changes.
Phylogenetic analysis of (1,3;1,4)-β-d-glucan synthases
The phylogenetic tree was generated for (1,3;1,4)-β-d-glucan synthases and cellulose synthases of cereals and other grasses, and for Gram-positive and Gram-negative bacteria (Supplementary Fig. 16a). Each clade is very distant from each another. In a previous study, it was suggested that putative BgsA (1,3;1,4)-β-d-glucan synthases from Gram-negative bacteria are more likely to have evolved from curdlan (a 1,3-β-d-glucan) synthase (CrdS)8, whereas the grass (1,3;1,4)-β-d-glucan synthases CslF6, CslH, and CslJ are phylogenetically closer to cellulose synthase CesA44. However, there have been only a few phylogenetic studies on the GT2s of Gram-positive bacteria. Scott et al. (2020) reported a series of putative Clostridial cellulose synthase (Ccs) operons among the class Clostridia45. Information from this study could allow us to identify possible conserved regions that may be critical for bacterial (1,3;1,4)-β-d-glucan synthases, being evolved to mediate the synthesis of the two linkages. To this end, homologous sequences were obtained using a BLAST server with Clostridioides difficile CcsA (NCBI accession: WP_021391996.1) and RiGT2 (CED93608.1 or WP_180703307.1) sequences as individual queries. The identity of all selected sequences was greater than 30%, with over 70% query coverage. A neighbor-joining likelihood phylogenetic tree was then generated for a non-redundant subset of 188 sequences from RiGT2 and CcsA homologous (Supplementary Fig. 16b). The sequences are clustered into three main clades: clade 1 includes RiGT2 (1,3;1,4)-β-d-glucan synthase; clade 2 remains unknown, but with three proteins annotated as cellulose synthases (POB11864.1, RSU08770.1, GAF20063.1); and clade 3 contains three putative CcsAs45 (Supplementary Fig. 16b). Clade 1 sequence, which contains RiGT2, were therefore selected as (1,3;1,4)-β-d-glucan synthase candidates and renamed Cgs (Clostridia β-glucan synthase) because most of the sequences are in species in the class Clostridia (Fig. 8).
To identify conserved motifs that may play critical roles in bacterial (1,3;1,4)-β-d-glucan synthase, we aligned sequences from five Cgs and four CcsA proteins. The protein alignment reveals three possible conserved sequences in Cgs, marked in the red boxes in Supplementary Fig. 17. We can only speculate that these three marked regions could be bacterial (1,3;1,4)-β-d-glucan conserved motifs because this is the only bacterial (1,3;1,4)-β-d-glucan synthase that has been biochemically characterized in a Gram-positive bacterium. A conserved sequence referred to as “switch motif” is present in interface helix 3 (IFH3) of the plant (1,3;1,4)-β-d-glucan synthase HvCslF637. It was confirmed that Tyr787 in the switch motif of HvCslF6 was critical for introducing β-1,3 glycosidic bonds within (1,3;1,4)-β-d-glucan37. However, this motif is not present in RiGT2 or CsBcsA, where Tyr787 in HvCslF6 corresponds to Leu424 and Leu487, respectively. Therefore, further research is needed to determine whether IFH3 in Cgs could also regulate the ratio of 1,3 and 1,4 linkages in bacterial (1,3;1,4)-β-d-glucans.
(1,3;1,4)-β-d-glucans and (1,3;1,4)-β-d-glucan synthases are present in other Gram-positive bacteria species
The Gram-positive bacteria Clostridium bornimense, Clostridium ventriculi, Robinsoniella peoriensis and Clostridium tyrobutyricum were chosen to investigate if they contain (1,3;1,4)-β-d-glucan and (1,3;1,4)-β-d-glucan synthases based on the following: (1) they have RiGT2 homologs with high sequence similarities; (2) these species are commercially available, and (3) out of six species that are available, the four selected could be cultured in our anaerobic system, whereas others (Clostridium nigeriense and Niameybacter massiliensis) either grew very slowly or not at all and we were unable to harvest enough EPSs for the subsequent analyses. All these bacteria have identified RiGT2 homologs with high sequence homology and are positioned in the phylogeny within in same order Eubacteriales and class Clostridia. EPSs were isolated from the four selected species, and treated with lichenase, and the resulting oligosaccharides were analyzed by HPAEC-PAD. (1,3;1,4)-β-d-glucan DP3 and (1,3;1,4)-β-d-glucan DP4 were found in the hydrolysates from C. ventriculi, R. peoriensis, and C. tyrobutyricum in various ratios, whereas the hydrolysate from C. bornimense contained only (1,3;1,4)-β-d-glucan DP3 (Figs. 9 and S18). This result indicates that (1,3;1,4)-β-d-glucans are more widespread as components of EPSs than was previously thought, and the DP3:DP4 ratio may differ between species.
Discussion
EPSs of Gram-positive bacteria are crucial for their biofilm formation, cell adhesion, and pathogenicity. Many polysaccharides have been identified as EPSs, such as teichoic acid (TA) and poly-β-1,6-linked N-acetyl glucosamine (PNAG/PIA), and serve different functions46. However, no previous report has shown that (1,3;1,4)-β-d-glucans are part of the EPSs of Gram-positive bacteria. Here, we report (1,3;1,4)-β-d-glucans as a novel component of the EPS of R. ilealis CRIBT. When analyzed in secreted EPS, the (1,3;1,4)-β-d-glucans contained primarily DP3 and DP4 subunits, and their structures resemble those of grass (1,3;1,4)-β-d-glucans1,7,47,48. Rather surprising however, (1,3;1,4)-β-d-glucans produced by RiGT2 transgenic yeast consisted mainly of the DP3 (G4G3G) subunit. Our study suggests that the (1,3;1,4)-β-d-glucan produced by RiGT2 is structurally more closely related to lichenin (DP3 dominant) than to the (1,3;1,4)-β-d-glucans of Gram-negative bacteria S. meliloti (DP2 only). The discrepancy of DP3:DP4 ratio could be that the recombinant RiGT2 in the LoGSA cells are not in their native lipid environment, or the C-terminal fusion GFP could also impact the conformation and function of RiGT2. Lee and Hollingsworth (1997) reported that another Gram-positive bacterium, C. ventriculi, contains β-glucans with unique linkages49. We have further confirmed by phylogenetic analysis and polysaccharide analysis that C. ventriculi produce (1,3;1,4)-β-d-glucans and contains potential (1,3;1,4)-β-d-glucan synthases. Indeed, phylogenetic analysis showed the widespread occurrence of RiGT2 homologs in Gram-positive bacteria, especially in the Clostridiaceae family. However, we cannot rule out that other non-homologous enzymes are also involved in (1,3;1,4)-β-d-glucan biosynthesis. This has been shown in the biosynthesis of poly-β-1,6-N-acetylglucosamine (PNAG). For example, bacterial PNAG in Staphylococcus epidermidis is synthesized by the icaADBC operon50 whereas the PNAG in Bacillus subtilis is synthesized by the non-homologous epsHIJK operon51,52.
Cyclic-di-GMP is a potent inducer for biofilm formation in bacteria51,52. It activates the cellulose (BcsA-B) synthase complex and the poly-β-1,6-linked N-acetyl glucosamine synthase (Pga) for biofilm production39. In the absence of c-di-GMP, the resting state of the gating loop (residues 499–517) is stabilized by a conserved salt link in CsBcsA (PilZ Arg580-GT Glu371)19. The c-di-GMP-bound state of CsBcsA has a c-di-GMP dimer bound in the PilZ domain with the gating loop in an open conformation that allows entry to the substrate-binding pocket (PDB 4P00 and 4P0220). The transition from the resting to open state of the gating loop is controlled by conformational changes that break the salt link, more specifically, the two arginine residues of the RXXXR motif in the PilZ domain (Arg580 and Arg584) rearrange to stack with two of the guanylate rings in c-di-GMP monomer A, and at the same time form hydrogen bonds to the N7 and O6 atoms of the two guanylate rings of in c-di-GMP monomer B. The corresponding arginine residues in RiGT2 (Arg517 and Arg521) are conserved and are expected to be able to interact with a c-di-GMP dimer in a similar way in an open state (Supplementary Fig. 11).
Based on the theoretical RiGT2A model, Thr308 occupies the position of Glu371 in CsBcsA. Thr308 cannot form a salt link with Arg517, but a nearby glutamate residue, Glu520, could possibly provide this opportunity in a resting state of RiGT2. Interestingly, heterologously expressed RiGT2 can still synthesize (1,3;1,4)-β-d-glucan in the absence of c-di-GMP in the yeast system. The activity of RiGT2 was independent of c-di-GMP concentration based on a radiometry assay. Additionally, no binding of c-di-GMP to the purified C-terminal PilZ domain of RiGT2 was detected in our SPR and ITC binding studies. These results indicate that RiGT2-mediated (1,3;1,4)-β-d-glucan synthesis may be independent of c-di-GMP activation. However, since the RiGT2 PilZ domain has the key residues for c-di-GMP binding, we can not rule out the possibility of in vivo transcriptional regulation by c-di-GMP as, for example, the psl genes encoding Psl polysaccharide synthesizing machinery in P. aeruginosa are positively regulated by c-di-GMP at the transcriptional level53,54.
The functions of (1,3;1,4)-β-d-glucans can vary from species to species. Due to the distribution of β-1,3 linkages, “kinks” in the structure make (1,3;1,4)-β-d-glucans unable to lay precisely on each other and are thus more flexible and soluble compared to other linear polysaccharides55. (1,3;1,4)-β-d-glucan found in the cell walls of cereals and other grasses plays an essential role in cell expansion and in the mobilization of endosperm cell walls56. In fungi and charophycean green and brown algae, (1,3;1,4)-β-d-glucan has a significant architectural role in supporting cell wall structure5,57. It has also been proposed that the bacterial (1,3;1,4)-β-d-glucan secreted by S. meliloti is associated with the bacterium’s attachment to plant roots, which is consistent with bacterial biofilm formation mediating cell-to-cell and cell-to-surface interactions8. This, therefore, indicates that bacterial (1,3;1,4)-β-d-glucan may contribute to bacterial autoaggregation and biofilm production.
Some Gram-positive bacteria have been identified as pathogens with increased antimicrobial resistance because of their biofilm formation. The matrix of many known biofilms is composed predominately of cellulose, with PNAG and alginate as less abundant components58. The discovery of (1,3;1,4)-β-d-glucans in bacterial EPS introduces an important intervention target. More importantly, knowledge of the genes and proteins involved in EPS biosynthesis is fundamental to the study of biofilm formation and control58. The presence of (1,3;1,4)-β-d-glucans in the EPS of groups of bacterial species has been overlooked in the past, and knowledge is limited about their involvement in biofilm formation.
In summary, we have identified a (1,3;1,4)-β-d-glucan synthase in the Gram-positive bacterium R. ilealis CRIBT. Using the yeast platform (Fig. 1), we have provided unequivocal direct gain-of-function evidence for the participation of the prokaryotic RiGT2 gene in (1,3;1,4)-β-d-glucan biosynthesis. For a long time, it was believed that (1,3;1,4)-β-d-glucans did not exist in bacteria and were present only in plants and fungi (including lichens). It was not until (1,3;1,4)-β-d-glucans with repetitive DP2 units were found in bacterial species8 that it was realized that (1,3;1,4)-β-d-glucans are likely to participate in biofilm formation due to their gel-like physiochemical properties. Hence, identifying crucial biofilm biosynthetic enzymes can provide enzymatic inhibitory targets for the control of biofilm during infections. In addition, manipulation of the (1,3;1,4)-β-d-glucan synthase could be tuned to produce glycopolymers of desired quality, adding to the burgeoning field of biomaterial engineering towards more targeted physical and biological properties than their natural counterparts59,60.
To date, yeast has been one of the most successful heterologous overexpression systems in producing membrane proteins for high-resolution structural studies31. Having successfully expressed functional RiGT2, the overarching goal of our laboratory is to resolve the 3D structure of this polysaccharide synthase in order to understand its (1,3;1,4)-β-d-glucan biosynthesis at atomic resolution.
Methods
Bioinformatic analysis
All the protein sequences were searched by BLASTP in the National Center for Biotechnology Information database (NCBI) with Clostridioides difficile CcsA (WP_021391996.1) and RiGT2 (CED93608.1) sequences as individual queries for 500 target sequences. CD-HIT was used to decrease the redundancy61. The protein alignment and phylogenetic analysis were conducted using the EMBL-EBI web tool62 and MEGA version X63,64. The trees were constructed using the neighbor-joining likelihood method based on the Jones–Taylor–Thornton (JTT) model with the parameters as follows: a test of phylogeny, bootstrap method with 1000 replications; substitution type, amino acid; rates among sites, uniform rates. The trees were edited using iTOL online tool65. The alignment in Supplementary Fig. 10 was visualized using ESPript3.075 (https://espript.ibcp.fr/ESPript/ESPript/)66.
Anaerobic cultivation of R. ilealis CRIBT
The bacterium R. ilealis CRIBT (DSM 25109) was purchased from DSMZ GmbH (Leibniz Institute DSMZ-German Collection of Microorganisms and Cell Cultures GmbH, Germany). The inoculation procedures followed the instructions from DSMZ with some modifications. First, 50 mL PYG medium (10 g/L peptone, 5 g/L yeast extract, 5 g/L glucose, 0.5 g/L l-cysteine hydrochloride, 0.2 g/L Na2CO3, 1.5 g/L KH2PO4, 0.1 g/L MgSO4, 0.25 mg/L hemin, 0.5 mL/L resazurin, pH 7) was transferred to 125 mL Wheaton® serum bottles (DWK Life Sciences, Millville, NJ, USA) and covered with blue butyl rubber stoppers and crimp caps to prevent gas leakage. After autoclaving (20 min at 121 °C), the atmosphere in the bottles was exchanged by 20 cycles of alternating vacuum and purging (45 s per each cycle) with ultrapure N2 gas (N5.0) through a sterile syringe filter (0.2 µm polyethersulfone (PES)) to ensure an anaerobic environment. Next, the vial of freeze-dried bacteria from DSMZ was opened in a vinyl anaerobic chamber (Coy Laboratory Products, TG Instruments, Helsingborg, Sweden) containing 5% H2, 10% CO2, and 85% N2. The cell pellet was resuspended in medium (10 mL) and 0.1 mL was used to aseptically inoculate the prepared serum bottle that was then taken out of the anaerobic chamber, purged with 80% N2 and 20% CO2 (5 cycles) to provide CO2 for growth, and incubated at 37 °C for 3–5 days on an orbital shaker (100 rpm).
Stable transformation of RiGT2 and RiGT2-PilD in the LoGSA yeast strain
The PCR primers (Supplementary Table 5) were used to amplify fragments containing full-length open-reading frames of RiGT2 from synthetic cDNA prepared in pMA plasmid (GeneArt, ThermoFisher Scientific, USA). The cDNA of RiGT2 or the PilZ domain mutant (RiGT2-PilD) were cloned in the pDDGFP-2 vector by homologous recombination, transformed into the quadruple knockout yeast Saccharomyces cerevisiae LoGSA33 and expressed following the method described in Newstead et al.32 and Drew et al.31. Briefly, the gene was cloned into a pDDGFP-2 vector under the control of a galactose inducible promotor (GAL1) to produce a C-terminal GFP fusion protein. After yeast transformation, positive expression clones were screened by whole-cell and in-gel fluorescence. The gene sequence of the RiGT2 and the mutant were confirmed by sequencing (Eurofin Genomic) after DNA extraction (GeneJET Plasmid Miniprep Kit, Thermo Fisher Scientific).
Preparation of alcohol-insoluble residues (AIRs) of R. ilealis CRIBT and LoGSA yeast cells
The cultivation conditions of R. ilealis CRIBT are described in the instructions from DSMZ, and conditions for the cultivation of LoGSA cells are described in the next section. After centrifuging (3000×g, 10 min), the cells were collected and resuspended in 75% (v/v) aqueous ethanol solution, followed by sonication (Branson, Brookfield, USA) (40% amplitude, 30 s, 3 min total) for R. ilealis CRIBT and using a French Press (1000 psi; two rounds) for the yeast cells. The broken cells were washed with 100% ethanol and 100% acetone and collected by centrifuging (3000×g, 10 min) at room temperature (RT). After the washed AIRs had been completely air-dried, they were resuspended in deionized water, transferred to dialysis tubes (Cat. No. D0530-100FT, Sigma), and dialyzed with continuously running Milli-Q water for 16 h and freeze-dried.
Whole-cell and in-gel fluorescence
Selected LoGSA colonies were inoculated into 5 mL of -URA medium (2 g/L of yeast synthetic drop-out medium without uracil, and 6.7 g/L yeast nitrogen base without amino acids) containing 2% glucose, and were incubated overnight on an orbital shaker at 30 °C at 280 rpm. The overnight cultures were adjusted to OD600 0.12 in a total of 10 mL of -URA medium containing 0.1% glucose in 50 mL centrifuge tubes (Corning™ Mini Bioreactor Centrifuge Tube). The cultures were incubated at 30 °C at 280 rpm and induced with 20% galactose to make a final concentration of 2% galactose when the OD600 reached 0.6. The induced cultures were collected after 22 h and centrifuged (3000×g, 5 min) at 4 °C. The pelleted cells were resuspended in 200 μL of YSB buffer (50 mM Tris–HCl pH 7.6, 5 mM EDTA, 10% glycerol) and transferred to a black Nunc 96-well optical bottom plate. The whole-cell fluorescence emission was measured at 512 nm by excitation at 488 nm in a microplate spectrofluorometer (Clariostar, BMG Labtech). The clones with successful expressions were collected and transferred to Eppendorf tubes. The cells were broken by glass beads in a Tissue lyzer (Retsch GmbH, Haan, Germany) at 30 Hz for 7 min at 4 °C and pelleted by centrifugation (3000×g, 5 min). The supernatants were collected and centrifuged (17,000×g, 1 h), and the crude membranes were pelleted and mixed with sample loading buffer (Invitrogen). After incubation for 10 min at 37 °C, the mixtures were transferred into SDS–PAGE (Bio-Rad), and electrophoresis was carried out at 150 mV. The gel was analyzed with a CCD camera system (Las 1000 gel system, Fujifilm). All the steps were conducted at 4 °C unless indicated otherwise.
Neutral sugar linkage analysis of the oligosaccharides and polysaccharides
Linkage analyses of the polysaccharides and the oligosaccharides were conducted as previously reported67 with slight modifications. Briefly, the purified oligosaccharides (~1 mg) from bacteria and genetically modified yeast were reduced by NaBD4 (5 mg) in 0.5 mL of deionized water with magnetic stirring overnight, followed by the quenching of excess reductant with acetic acid, evaporating the resulting solution to dryness under nitrogen, and the repeated evaporation to dryness in 0.5 mL of 10% (v/v) acetic acid in methanol. The resulting oligosaccharide alditols were converted into their partially methylated alditol acetate (PMAA) derivatives by permethylation with methyl iodide in dimethyl sulfoxide and sodium hydroxide68, hydrolysis with 4 M trifluoroacetic acid (TFA) at 100 °C for 2 h, reduction with NaBD469, and peracetylation in a heated mixture of acetic anhydride/TFA (5:1, v/v)70. R. ilealis CRIBT AIRs were treated with α-amylase from porcine pancreas (Sigma-Aldrich®, USA) followed by extensive dialysis against deionized water and freeze-drying. The preparation of PMAAs from the α-amylase-treated polysaccharides was then conducted without treatment with NaBD4 before the methylation step. The PMAAs from the oligosaccharides and polysaccharides were analyzed on an Agilent 7890A-5977B GC-MS system (Agilent technologies, USA) with a Supelco SP2380 capillary GC column (100 m × 0.25 mm × 0.20 µm) with a constant column outlet helium flow rate of 1.2 mL/min and with an initial oven temperature of 140 °C, which was increased at 1.25 °C/min to 190 °C, and then at 8 °C/min to 250 °C (held for 30 min). Data were collected and analyzed by Agilent OpenLab CDS (version 2.6) software. The separated PMAAs were identified by their relative retention times and by comparing their MS fragmentation patterns with those of reference derivatives and by referring to the literature71. Two separate experiments were conducted for each sample.
Treatment with lichenase and (1,3;1,4)-β-d-glucan oligosaccharide purification
AIRs (2 mg) were resuspended in 200 μL of 20 mM phosphate buffer (pH 6.5) and treated with 50 μL of (1,3;1,4)-β-d-glucanase (lichenase) from Bacillus subtilis (Megazyme International Ireland Ltd)72. After 24 h of incubation at 50 °C, the mixtures were centrifuged (3000×g, 5 min), the supernatants loaded onto Bond Elut Carbon cartridges (Agilent Technologies, USA), and washed with H2O. The (1,3;1,4)-β-d-glucan oligosaccharides were eluted with 50% CH3CN and freeze-dried for subsequent analysis.
MALDI-TOF-MS
Samples were prepared by mixing 10 μL of the lichenase digest with 10 μL of 2,5-dihydroxybenzoic acid (DHB) [10 mg/L in 0.1% (v/v) TFA and 50% (v/v) CH3CN] and 6 μL of NaCl solution (10 mM). The mixtures were spotted onto a steel plate and air-dried. MALDI-TOF MS was performed with an Applied Biosystems 4800 MALDI instrument in the linear positive-ion mode. The crystalized analytes were ionized with an N2-laser and accelerated into a time-of-flight analyzer. The number of shots collected for the resulting spectra was determined by the response from the analytes.
High-performance anion exchange chromatography with pulsed amperometric detection (HPAEC-PAD)
The oligosaccharides were also analyzed by high-performance anion exchange chromatography with pulsed amperometric detection (HPAEC-PAD) using an ICS-3000 system (Dionex, Sunnyvale, CA, USA) and a CarboPac-PA1 analytical column (4 × 250 mm, Thermo Scientific, USA) fitted with a CarboPac PA1 guard column (4 × 50 mm, Thermo Scientific, USA). Before samples were injected, the column was washed with 200 mM NaOH for 10 min at a flow rate of 1 mL min−1 and equilibrated for 5 min with 100 mM NaOH and 5% (v/v) 1 M NaOAc. The samples were eluted with a constant 100 mM NaOH solution and a gradient of 1 M NaOAc from 5% to 25% over 25 min. Pure DP3 (G4G3G, M3) and DP4 (G4G4G3G, M4) (1,3;1,4)-β-d-glucan oligosaccharide standards, and pure oligosaccharide standards derived from laminarin (laminaribiose L2, laminaritriose L3, and laminaritetraose L4) and cellulose (cellotriose C3, cellotetraose C4, cellopentaose C5, and cellohexaose C6) were obtained from Megazyme.
Cultivation of selected anaerobic bacteria from class Clostridia
The medium used to culture all species in this study was prepared from the recipe on the DSMZ GmbH website (https://www.dsmz.de). Before opening the vials of bacteria (DSM 286 Clostridium ventriculi (syn. Sarcina ventriculi), DSM 2637 Clostridium tyrobutyricum, DSM 25664 Clostridium bornimense, DSM 102218 Clostridium nigeriense, DSM 100592 Niameybacter massiliensis, and DSM 106044 Robinsoniella peoriensis), medium was injected into 50 mL serum bottles, which were covered tightly with rubber to prevent gas leakage. After autoclaving, all the media were purged with 100% N2 to ensure an anaerobic environment. When the media were all prepared, the vials containing the bacteria were opened and inoculated in the anaerobic chamber. The inoculated media were purged with 100% N2 again and cultured at 37 °C for 3–5 days on an orbital shaker at 100 rpm. All the inoculation procedures followed the instructions from DSMZ GmbH.
Preparation of microsomal fractions (MFs) and RiGT2-GFP membrane protein purification
The selected LoGSA colonies were inoculated into 5 mL of -URA medium containing 2% glucose and were incubated overnight on an orbital shaker at 30 °C at 280 rpm. The overnight cultures were adjusted to OD600 0.12 in 1 L of -URA medium containing 0.1% (w/v) glucose in a 5 L flask. The cultures were incubated at 30 °C at 280 rpm, and induced with 20% (w/v) galactose to make a final concentration of 2% (w/v) galactose when the OD600 reached 0.6. The induced cultures were collected after 22 h and centrifuged (3000×g, 5 min, 4 °C). The cells were resuspended in a buffer (pH 7.4) containing 50 mM HEPES, 150 mM NaCl, 10% glycerol, and 1 tablet of cOmplete™ Protease Inhibitor Cocktail (Roche, Basel, Switzerland). The cell suspensions were then passed twice through a French Press (1000 psi) (SLM Aminco), and the cell debris was removed by centrifugation (3000×g, 5 min, 4 °C). The microsomal fractions (MFs) were collected by centrifugation (150,000×g, 1 h) in a Beckman Ti70 rotor, before being resuspended in 100 mM MOPS buffer (pH 7) containing 10% glycerol. The protein concentrations of the MFs were adjusted to 1.5 mg/ml for the radiometric in vitro assays. For the protein purification, the cultivation protocol was similar, but with some steps modified for the increase in scale: the overnight cultures were adjusted to an OD600 of 0.12 in 8 L of -URA medium containing 0.3% (w/v) glucose and incubated for 20 h before induction with 20% (w/v) galactose, with a final concentration of 2% (w/v) galactose. For the RiGT2-GFP protein solubilization, the pelleted MFs were solubilized in a buffer containing 50 mM HEPES (pH 7.4), 150 mM NaCl, 10 mM MgCl2 and 10% (v/v) glycerol, 20 mM imidazole and 1% (w/v) n-dodecyl-β-d-maltoside (DDM)/0.01% (w/v) cholesteryl hemisuccinate (CHS) for 1 h at 4 °C with gentle stirring. The insoluble material was separated by centrifugation (150,000×g, 1 h) in a Beckman Ti70 rotor. The soluble membrane fraction was then incubated with Ni-NTA His Bind Resin (Millipore, USA) at 4 °C for 1 h with gentle stirring. The resin was packed in a gravity flow chromatography column, washed with a wash buffer containing 50 mM HEPES (pH 7.4), 150 mM NaCl, 10 mM MgCl2 and 10% glycerol, 50 mM imidazole and 0.02% (w/v) DDM/0.002% (w/v) CHS in a 5-column volume, and eluted by an elution buffer containing 50 mM HEPES (pH 7.4), 150 mM NaCl, 10 mM MgCl2 and 10% (v/v) glycerol, 500 mM imidazole and 0.02% DDM/0.002%CHS. The eluted proteins were then separated by SDS–PAGE, and the identifications of the proteins were confirmed by in-gel tryptic digestion followed LC–MS.
Radiometric in vitro assay
A total of 100 μL MFs (1.5 mg/mL) from transgenic or non-transgenic LoGSA cells was added to a 100 μL solution containing 100 mM MOPS (pH 7), 18 mM MgCl2, 4 mM UDP-glucose, 0.2 μCi/ml UDP-[14C]-glucose. The c-di-GMP was added at the indicated concentrations, the reactions were incubated at RT for 1 h, and 95% (v/v) ethanol (400 μL) was added to terminate the reaction. The mixtures were then passed through Whatman® glass microfiber filters (binder-free). The remaining insoluble products on the filter were washed with ddH2O and 66% (v/v) ethanol, and then dried. The filter was submerged in Ultima Gold F scintillation reagent (PerkinElmer, Waltham, MA, USA) before being scintillation counted. All measurements were performed in triplicate. Error bars represent standard deviation (SD).
Surface plasmon resonance (SPR)
The RiPilZ-GFP was expressed in LoGSA yeast as indicated above and the harvested cells resuspended in PBS buffer (pH 7.4) with 1 tablet of cOmplete™ Protease Inhibitor Cocktail and disrupted by twice passing through a French Press (1000 psi). The cell debris was removed by centrifugation (3000×g, 5 min, 4 °C) and the supernatant was incubated with Ni-NTA His-Bind Resin (Millipore, USA) for 1 h at 4 °C with gentle stirring. The resin was packed in a gravity flow chromatography column, washed with PBS (pH 7.4) containing 10% glycerol and 50 mM imidazole in a 5-column volume, and eluted with PBS (pH 7.4) containing10% (w/v) glycerol and 500 mM imidazole. The eluted protein was then concentrated, and the buffer was switched to PBS (pH 7.4) to remove imidazole using Amicon Ultra-0.5 mL centrifugal filters (Millipore, USA). The ability of c-di-GMP to bind to RiPilZ-GFP fusion protein was determined using a Biacore T200 surface plasmon resonance instrument. RiPilZ-GFP was applied with sodium acetate buffer (pH 5.0) and immobilized to a level of about 12,000 response units on a CM5 sensor chip with an amine coupling kit (Cytiva, Sweden). Then, a two-fold serially diluted analyte, c-di-GMP, starting at a concentration of 40 µM was injected into the flow channels in PBS buffer. Each condition was implemented at a flow rate of 10 µL/min for 60 s at 25 °C. The signals obtained were subtracted from the reference channel that had not been coated with RiPilZ-GFP and plotted in a resonance unit against a time sensorgram.
Isothermal titration calorimetry (ITC)
The RiPilZ-GFP was expressed and purified as indicated above. Measurements were carried out at 25 °C in a MicroCal ITC200 system (MicroCal, Northampton, MA), with 40 μM RiPilZ-GFP in the cell and 400 μM c-di-GMP in the syringe. Samples (16) 2.5 μL in volume were injected at 1 min intervals with stirring at 1000 r.p.m. The data were fitted with Microcal PEAQ-ITC analysis software supplied by the manufacturer.
Indirect immunofluorescence microscopy using an inverted confocal microscope
Yeast cells from 10 mL cultures were centrifuged (4000×g, 10 min, 4 °C) and resuspended in PBS buffer (pH 7.4) resulting in a final volume of 300 µL. To prepare spheroplast cells, 30 µL of 10 mg/mL of Lyticase (from Arthrobacter luteus, Sigma-Aldrich®) in PBS and of β-mercaptoethanol (6 µL) were added and incubated on an orbital shaker at 30 °C at 200 rpm for 2 h. The spheroplast cells were collected by centrifugation (4000×g, 5 min, RT) and washed twice with PBS. The cells were then fixed with 4% (w/v) paraformaldehyde (PFA) with shaking at 200 rpm for 75 min at 30 °C. The fixed cells were washed with PBS and incubated in 0.1% (w/v) bovine serum albumin (BSA) in PBS for 15 min in RT and harvested by centrifugation (4000×g, 3 min, 4 °C). The cells were then resuspended in a solution containing the primary antibody, a mouse monoclonal IgG antibody against (1,3;1,4)-β-d-glucan (Biosupplies Australia Pty Ltd) (BS400-3) diluted 1:100 with 0.1% (w/v) BSA in PBS and incubated at 4 °C overnight. The cells were then washed with PBS (4 times) by centrifuging (4000×g, 3 min, 4 °C) and resuspended in the secondary antibody, goat anti-mouse IgG (H + L) labeled with the fluorophore Alexa FluorTM 555, (SuperclonalTM Recombinant Secondary Antibody, Alexa FluorTM 555) (ThermoFisher), diluted 1:100 with 0.1% (w/v) BSA in PBS and incubated for 1 h at RT in the dark. The cells were washed after the incubation with the primary antibody, and then pipetted onto microscope slides and mixed with 1 µg/mL DAPI (a fluorescent DNA stain) in SlowFadeTM antifade mountant (Thermo Fisher). The slides were incubated in the dark for at least 5 min and then examined using an inverted confocal microscope (Leica TCS SP5). Leica microsystem LAS AF software was used to view as well as export images of cells, and ImageJ and Fiji were used to analyze and generate images. The lasers used were as follows: diode 405 and 461 nm (blue); Ar 488, 518 nm (green); and DPSS (diode-pumped solid-state) 561 and 565 nm (red).
Indirect immunogold microscopy using transmission electron microscopy (TEM)
Embedding and sectioning
The yeast cells were briefly fixed in 8% (w/v) formaldehyde in 0.1 M phosphate buffer (pH 7.4), centrifuged (3000×g, 1 min), and replaced with a mixture of 5% (v/v) glutaraldehyde and 3% (w/v) paraformaldehyde in the same buffer for 1 h at RT, followed by washing in this buffer. To permeabilize the yeast cell walls, the samples were treated with 1% (w/v) sodium metaperiodate for 10 min. This was followed by extensive washing in phosphate buffer, dehydration in a graded series of concentrations of aqueous ethanol (30–80%), and infiltration with LR White Resin (Ted Pella, CA, USA). The resin was heat polymerized (55 °C, 48 h) the following day and ultrathin sections (~60 nm thick) were cut on an ultramicrotome (Leica EM UC7 Wetzlar, Germany) and placed on formvar-coated mesh grids.
Immunogold labeling
Grids were incubated with 50 mM glycine for 15 min followed by Aurion Blocking Solution (Aurion, Wageningen, Netherlands) for 30 min. After further incubation with BSA-C ™ (0.15%) (Aurion) for 15 min, the grids were placed on a drop of the mouse monoclonal antibody against (1,3;1,4)-β-d-glucan (5 µg/mL; BS400-3, Biosupplies Australia Pty Ltd) in BSA-C™ (0.15%) or only BSA-C™ for 1 h. The grids were washed in BSA-C ™ and then incubated (1 h) with a goat anti-mouse secondary antibody conjugated to 12 nm colloidal gold particles (1:100, Jackson ImmunoResearch, PA, USA). After extensive washing in BSA-C ™ and Milli-Q water, the grids were stained with 5% (w/v) uranyl acetate and Reynold’s lead citrate. The grids were examined using a Tecnai G2 Spirit BioTWIN transmission electron microscope (Tecnai/Thermofisher) operated at 80 kV and images were obtained with an Orius SC200 CCD camera (Gatan Inc/Blue-Scientific).
Theoretical modeling of the RiTG2 structure
Template selection
There are several experimental structures available for GT2 enzymes, including structures of bacterial and plant cellulose synthases, as well as a structure of the (1,3;1,4)-β-d-glucan synthase CslF6 from barley. Before deciding on the best strategy for model prediction, the RiGT2 sequence was compared with sequences for experimental structures using a template search in SWISS-MODEL36. SWISS-MODEL returned Cereibacter sphaeroides BcsA as the closest match with a global model quality estimate (GMQE) score of 0.65 and sequence identity of 25% (PDB 4P0020). The GMQE is a quality score that estimates the accuracy of a 3D model by combining properties from the target-template alignment and the template structure and ranges between 0 and 1 with a higher value indicating higher reliability. Cotton cellulose synthase had a GMQE score of 0.39 and sequence identity of 20% (PDB 7D5K41), while CslF6 had an even lower GMQE score of 0.36 and sequence identity of 21%. Based on these results, CsBcsA is expected to be the best structural template for the overall structure of RiGT2, while CslF6 could provide more accurate local structural information for binding of (1,3;1,4)-β-d-glucan.
Model building
The artificial intelligence software AlphaFold235 is superior for comparative modeling when sufficient training data exist (sequences and templates), which is the case for cellulose synthase, but tends to produce models that are biased towards the training sets. For scarce template data, such as for (1,3;1,4)-β-d-glucan synthase, traditional homology modeling using SWISS-MODEL could be useful to avoid such bias. We, therefore, decided to generate theoretical RiGT2 models using both non-template-based AlphaFold2 modeling and template-based modeling with SWISS-MODEL. Modeling using AlphaFold2 was performed as implemented in the Google Colaboratory resource (ColabFold73; AlphaFold2.ipynb). Briefly, the input included the full-length sequence for RiGT2 (UniProt A0A1V1I049), multiple sequence alignment (MSA) mode MMseqs2 (Many-against-Many sequence searching), unpaired_paired as pair mode (pair sequences from same species and unpaired MSA), and no template information. When using SWISS-MODEL, the structure of HvCslF6 was selected as a template (UniProt B1P2T4; PDB 8DQK37).
Alignment of the sequences for CsBcsA and RiTG2
The sequence of RiGT2 (UniProt A0A1V1I049) was aligned with that of the closest homolog with known experimental 3D structure, i.e., the catalytic subunit of C. sphaeroides cellulose synthase (CsBcsA; UniProt Q3J125). The structure of the (1,3;1,4)-β-d-glucan synthase CslF6 from barley (HvCslF6; UniProt B1P2T4; PDB 8DQK37) has been determined experimentally but has overall lower sequence similarity to RiGT2 and CsBcsA and was therefore not included in the alignment. An initial pre-alignment was generated using Clustal Omega74 (https://www.ebi.ac.uk/Tools/msa/clustalo/). The theoretical model of RiGT2 was then superposed with the experimental crystal structure of CsBcsA (PDB 4HG619) using SSM Superpose in Coot75,76 to produce an optimal three-dimensional match of the two coordinate sets. The pre-alignment was then manually optimized to match with the structural superposition. The final alignment was visualized with ESPript3.066 (https://espript.ibcp.fr/ESPript/ESPript/).
Reporting summary
Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.
Data availability
The data that support the findings of this study are available within the article and its Supplementary Information. Source data are provided with this paper.
References
Harris, P. J. & Fincher, G. B. Distribution, fine structure and function of (1,3;1,4)-β-glucans in the grasses and other taxa. In Chemistry, Biochemistry, and Biology of (1-3)-β-glucans and related Polysaccharides (eds Bacic, A. Fincher, G. B., & Stone, B. A.) 621–654 (Academic Press, 2009).
Trethewey, J. A. K., Campbell, L. M. & Harris, P. J. (1→3),(1→4)-β-d-Glucans in the cell walls of the Poales (sensu lato): an immunogold labeling study using a monoclonal antibody. Am. J. Bot. 92, 1660–1674 (2005).
Leroux, O. et al. Antibody-based screening of cell wall matrix glycans in ferns reveals taxon, tissue and cell-type specific distribution patterns. BMC Plant Biol. 15, 56 (2015).
Sørensen, I. et al. The charophycean green algae provide insights into the early origins of plant cell walls. Plant J. 68, 201–211 (2011).
Salmeán, A. A. et al. Insoluble (1→3),(1→4)-β-d-glucan is a component of cell walls in brown algae (Phaeophyceae) and is masked by alginates in tissues. Sci. Rep. 7, 2880 (2017).
Samar, D., Kieler, J. B. & Klutts, J. S. Identification and deletion of Tft1, a predicted glycosyltransferase necessary for cell wall β−1,3;1,4-glucan synthesis in Aspergillus fumigatus. PLoS ONE 10, e0117336 (2015).
Wood, P. J., Weisz, J. & Blackwell, B. A. Molecular characterization of cereal β-d-glucans. Structural analysis of oat β-d-glucan and rapid structural evaluation of β-d-glucans from different sources by high-performance liquid chromatography of oligosaccharides released by lichenase. Cereal Chem. 68, 31–39 (1991).
Pérez-Mendoza, D. et al. Novel mixed-linkage β-glucan activated by c-di-GMP in Sinorhizobium meliloti. Proc. Natl Acad. Sci. USA 112, E757–E765 (2015).
Cui, W. & Wood, P. J. Relationships between structural features, molecular weight and rheological properties of cereal β-d-glucans. in Hydrocolloids—Part 1 (ed Nishinari, K.) 159–168 (Elsevier, 2000).
Burton, R. A. et al. Cellulose synthase-like CslF genes mediate the synthesis of cell wall (1,3;1,4)-β-d-glucans. Science 311, 1940–1942 (2006).
Doblin, M. S. et al. A barley cellulose synthase-like CSLH gene mediates (1,3;1,4)-β-d-glucan synthesis in transgenic Arabidopsis. Proc. Natl Acad. Sci. USA 106, 5996–6001 (2009).
Little, A. et al. Revised phylogeny of the cellulose synthase gene superfamily: insights into cell wall evolution. Plant Physiol. 177, 1124–1141 (2018).
Roberts, A. W. et al. Functional characterization of a glycosyltransferase from the moss Physcomitrella patens involved in the biosynthesis of a novel cell wall arabinoglucan. Plant Cell 30, 1293–1308 (2018).
Schmid, J., Sieber, V. & Rehm, B. Bacterial exopolysaccharides: biosynthesis pathways and engineering strategies. Front. Microbiol. 6, 496 (2015).
Vu, B., Chen, M., Crawford, R. J. & Ivanova, E. P. Bacterial extracellular polysaccharides involved in biofilm formation. Molecules 14, 2535–2554 (2009).
Saxena, I. M. & Brown, R. M. Identification of cellulose synthase(s) in higher plants: sequence analysis of processive β-glycosyltransferases with the common motif’D, D, D35Q(R,Q)XRW‘. Cellulose 4, 33–49 (1997).
Saxena, I. M., Brown, R. M. Jr & Dandekar, T. Structure–function characterization of cellulose synthase: relationship to other glycosyltransferases. Phytochemistry 57, 1135–1148 (2001).
Morgan, J. L. W. et al. Observing cellulose biosynthesis and membrane translocation in crystallo. Nature 531, 329–334 (2016).
Morgan, J. L. W., Strumillo, J. & Zimmer, J. Crystallographic snapshot of cellulose synthesis and membrane translocation. Nature 493, 181–186 (2013).
Morgan, J. L. W., McNamara, J. T. & Zimmer, J. Mechanism of activation of bacterial cellulose synthase by cyclic di-GMP. Nat. Struct. Mol. Biol. 21, 489–496 (2014).
Slabaugh, E., Davis, J. K., Haigler, C. H., Yingling, Y. G. & Zimmer, J. Cellulose synthases: new insights from crystallography and modeling. Trends Plant Sci. 19, 99–106 (2014).
Bi, Y., Hubbard, C., Purushotham, P. & Zimmer, J. Insights into the structure and function of membrane-integrated processive glycosyltransferases. Curr. Opin. Struct. Biol. 34, 78–86 (2015).
Kastl, A. J. Jr, Terry, N. A., Wu, G. D. & Albenberg, L. G. The structure and function of the human small intestinal microbiota: current understanding and future directions. Cell. Mol. Gastroenterol. Hepatol. 9, 33–45 (2020).
Oerlemans, M. M. P., Akkerman, R., Ferrari, M., Walvoort, M. T. C. & de Vos, P. Benefits of bacteria-derived exopolysaccharides on gastrointestinal microbiota, immunity and health. J. Funct. Foods 76, 104289 (2021).
Jones, S. E., Paynich, M. L., Kearns, D. B. & Knight, K. L. Protection from intestinal inflammation by bacterial exopolysaccharides. J. Immunol. 192, 4813–4820 (2014).
Gerritsen, J. et al. Correlation between protection against sepsis by probiotic therapy and stimulation of a novel bacterial phylotype. Appl. Environ. Microbiol. 77, 7749–7756 (2011).
Gerritsen, J. et al. Characterization of Romboutsia ilealis gen. nov., sp. nov., isolated from the gastro-intestinal tract of a rat, and proposal for the reclassification of five closely related members of the genus Clostridium into the genera Romboutsia gen. nov., Intestini. Int. J. Syst. Evol. Microbiol. 64, 1600–1616 (2014).
Wood, P. J., Weisz, J. & Blackwell, B. A. Structural studies of (1→3),(1→4)-β-d-glucans by 13C-nuclear magnetic resonance spectroscopy and by rapid analysis of cellulose-like regions using high-performance anion-exchange chromatography of oligosaccharides released by lichenase. Cereal Chem. 71, 301–307 (1994).
Collins, H. M. et al. Variability in fine structures of noncellulosic cell wall polysaccharides from cereal grains: potential importance in human health and nutrition. Cereal Chem. 87, 272–282 (2010).
Drula, E. et al. The carbohydrate-active enzyme database: functions and literature. Nucleic Acids Res. 50, D571–D577 (2022).
Drew, D. et al. GFP-based optimization scheme for the overexpression and purification of eukaryotic membrane proteins in Saccharomyces cerevisiae. Nat. Protoc. 3, 784–798 (2008).
Newstead, S., Kim, H., von Heijne, G., Iwata, S. & Drew, D. High-throughput fluorescent-based optimization of eukaryotic membrane protein overexpression and purification in Saccharomyces cerevisiae. Proc. Natl Acad. Sci. USA 104, 13936–13941 (2007).
Pang, Z. et al. Analysis of a cellulose synthase catalytic subunit from the oomycete pathogen of crops Phytophthora capsici. Cellulose 27, 8551–8565 (2020).
Brown, C., Leijon, F. & Bulone, V. Radiometric and spectrophotometric in vitro assays of glycosyltransferases involved in plant cell wall carbohydrate biosynthesis. Nat. Protoc. 7, 1634–1650 (2012).
Jumper, J. et al. Highly accurate protein structure prediction with AlphaFold. Nature 596, 583–589 (2021).
Waterhouse, A. et al. SWISS-MODEL: homology modelling of protein structures and complexes. Nucleic Acids Res. 46, W296–W303 (2018).
Purushotham, P. et al. Mechanism of mixed-linkage glucan biosynthesis by barley cellulose synthase-like CslF6 (1,3;1,4)-β-glucan synthase. Sci. Adv. 8, eadd1596 (2022).
Stephens, Z., Wilson, L. F. L. & Zimmer, J. Diverse mechanisms of polysaccharide biosynthesis, assembly and secretion across kingdoms. Curr. Opin. Struct. Biol. 79, 102564 (2023).
Omadjela, O. et al. BcsA and BcsB form the catalytically active core of bacterial cellulose synthase sufficient for in vitro cellulose synthesis. Proc. Natl Acad. Sci. USA 110, 17856–17861 (2013).
Acheson, J. F., Ho, R., Goularte, N. F., Cegelski, L. & Zimmer, J. Molecular organization of the E. coli cellulose synthase macrocomplex. Nat. Struct. Mol. Biol. 28, 310–318 (2021).
Zhang, X. et al. Structural insights into homotrimeric assembly of cellulose synthase CesA7 from Gossypium hirsutum. Plant Biotechnol. J. 19, 1579–1587 (2021).
Verma, P., Kwansa, A. L., Ho, R., Yingling, Y. G. & Zimmer, J. Insights into substrate coordination and glycosyl transfer of poplar cellulose synthase-8. Preprint at bioRxiv https://doi.org/10.1101/2023.02.07.527505 (2023).
Amikam, D. & Galperin, M. Y. PilZ domain is part of the bacterial c-di-GMP binding protein. Bioinformatics 22, 3–6 (2006).
Schwerdt, J. G. et al. Evolutionary dynamics of the cellulose synthase gene superfamily in grasses. Plant Physiol. 168, 968–983 (2015).
Scott, W., Lowrance, B., Anderson, A. C. & Weadge, J. T. Identification of the Clostridial cellulose synthase and characterization of the cognate glycosyl hydrolase, CcsZ. PLoS ONE 15, e0242686 (2020).
Ruhal, R. & Kataria, R. Biofilm patterns in Gram-positive and Gram-negative bacteria. Microbiol. Res. 251, 126829 (2021).
Lazaridou, A., Biliaderis, C. G., Micha-Screttas, M. & Steele, B. R. A comparative study on structure-function relations of mixed-linkage (1→3), (1→4) linear β-d-glucans. Food Hydrocoll. 18, 837–855 (2004).
Cui, S. W. & Wang, Q. Cell wall polysaccharides in cereals: chemical structures and functional properties. Struct. Chem. 20, 291–297 (2009).
Lee, J. & Hollingsworth, R. I. Oligosaccharide beta-glucans with unusual linkages from Sarcina ventriculi. Carbohydr. Res. 304, 133–141 (1997).
Heilmann, C. et al. Molecular basis of intercellular adhesion in the biofilm‐forming Staphylococcus epidermidis. Mol. Microbiol. 20, 1083–1091 (1996).
Tamayo, R., Pratt, J. T. & Camilli, A. Roles of cyclic diguanylate in the regulation of bacterial pathogenesis. Annu. Rev. Microbiol. 61, 131–148 (2007).
Römling, U., Galperin, M. Y. & Gomelsky, M. Cyclic di-GMP: the first 25 years of a universal bacterial second messenger. Microbiol. Mol. Biol. Rev. 77, 1–52 (2013).
Baraquet, C., Murakami, K., Parsek, M. R. & Harwood, C. S. The FleQ protein from Pseudomonas aeruginosa functions as both a repressor and an activator to control gene expression from the pel operon promoter in response to c-di-GMP. Nucleic Acids Res. 40, 7207–7218 (2012).
Hickman, J. W. & Harwood, C. S. Identification of FleQ from Pseudomonas aeruginosa as a c-di-GMP-responsive transcription factor. Mol. Microbiol. 69, 376–389 (2008).
Kiemle, S. N. et al. Role of (1,3)(1,4)-β-glucan in cell walls: Interaction with cellulose. Biomacromolecules 15, 1727–1736 (2014).
Wilson, S. M. et al. Temporal and spatial appearance of wall polysaccharides during cellularization of barley (Hordeum vulgare) endosperm. Planta 224, 655–667 (2006).
Fontaine, T. et al. Molecular organization of the alkali-insoluble fraction of Aspergillus fumigatus cell wall. J. Biol. Chem. 275, 27594–27607 (2000).
Whitney, J. C. & Howell, P. L. Synthase-dependent exopolysaccharide secretion in Gram-negative bacteria. Trends Microbiol. 21, 63–72 (2013).
Rendueles, O., Kaplan, J. B. & Ghigo, J.-M. Antibiofilm polysaccharides. Environ. Microbiol. 15, 334–346 (2013).
Rinaudo, M. Main properties and current applications of some polysaccharides as biomaterials. Polym. Int. 57, 397–430 (2008).
Fu, L., Niu, B., Zhu, Z., Wu, S. & Li, W. CD-HIT: accelerated for clustering the next-generation sequencing data. Bioinformatics 28, 3150–3152 (2012).
Madeira, F. et al. Search and sequence analysis tools services from EMBL-EBI in 2022. Nucleic Acids Res. 50, W276–W279 (2022).
Stecher, G., Tamura, K. & Kumar, S. Molecular evolutionary genetics analysis (MEGA) for macOS. Mol. Biol. Evol. 37, 1237–1239 (2020).
Kumar, S., Stecher, G., Li, M., Knyaz, C. & Tamura, K. MEGA X: molecular evolutionary genetics analysis across computing platforms. Mol. Biol. Evol. 35, 1547–1549 (2018).
Letunic, I. & Bork, P. Interactive Tree Of Life (iTOL) v5: an online tool for phylogenetic tree display and annotation. Nucleic Acids Res. 49, W293–W296 (2021).
Robert, X. & Gouet, P. Deciphering key features in protein structures with the new ENDscript server. Nucleic Acids Res. 42, W320–W324 (2014).
Jones, D. R. et al. Analysis of active site architecture and reaction product linkage chemistry reveals a conserved cleavage substrate for an endo-alpha-mannanase within diverse yeast mannans. J. Mol. Biol. 432, 1083–1097 (2020).
Ciucanu, I. & Kerek, F. A simple and rapid method for the permethylation of carbohydrates. Carbohydr. Res. 131, 209–217 (1984).
Pettolino, F. A., Walsh, C., Fincher, G. B. & Bacic, A. Determining the polysaccharide composition of plant cell walls. Nat. Protoc. 7, 1590–1607 (2012).
Voiges, K., Adden, R., Rinken, M. & Mischnick, P. Critical re-investigation of the alditol acetate method for analysis of substituent distribution in methyl cellulose. Cellulose 19, 993–1004 (2012).
Carpita, N. C. & Shea, E. M. Linkage structure of carbohydrates by gas chromatography-mass spectrometry (GC-MS) of partially methylated alditol acetates. In Analysis of Carbohydrates by GLC and MS (eds Biermann, C. J. & McGinnis, G. D.) 157–216 (CRC Press, 1989).
Hsieh, Y. S. Y. & Harris, P. J. Xyloglucans of monocotyledons have diverse structures. Mol. Plant 2, 943–965 (2009).
Mirdita, M. et al. ColabFold: making protein folding accessible to all. Nat. Methods 19, 679–682 (2022).
Sievers, F. et al. Fast, scalable generation of high-quality protein multiple sequence alignments using Clustal Omega. Mol. Syst. Biol. 7, 539 (2011).
Emsley, P., Lohkamp, B., Scott, W. G. & Cowtan, K. Features and development of Coot. Acta Crystallogr. Sect. D 66, 486–501 (2010).
Pettersen, E. F. et al. UCSF Chimera—a visualization system for exploratory research and analysis. J. Comput. Chem. 25, 1605–1612 (2004).
Acknowledgements
This work was supported by the Knut and Alice Wallenberg Foundation, the KTH Royal Institute of Technology, and the National Science and Technology Council, Taiwan (MOST, 110-2636-M-038-001). We would like to thank Dr. Ann C.Y. Wong and Prof. Philip J Harris for critical reading and further assistance with the manuscript; Dr. David Drew for the pDDGFP-2 plasmid; Dr. Stefan Klinter for LoGSA yeast strain; Dr. Johan Larsbrink for expert advice and helpful discussion; Dr. Sheng-Wei Lin for assistance with the SPR; Chun Austin Changou, Huei-Min Chen, and the Taipei Medical University Core Facility Center for assistance with the confocal and electron microscopy; Dr. Monica Hodik and the BioVis platform of Uppsala University for assistance with the electron microscopy; Dr. Dayanand Kalyani for advice with protein purification; Dr. Lauren McKee for advice on radiometric in vitro assay, Dr. Vaibhav Srivastava for assistance with the identification of tryptic peptides.
Funding
Open access funding provided by Royal Institute of Technology.
Author information
Authors and Affiliations
Contributions
S.-C.C., Y.S.Y.H. conceived and designed the project. S.-C.C., M.-R.K., R.K.S., S.M.D.-M., X.X., J.Y., V.F., C.D. conducted the experiments. S.-C.C., R.K.S., X.X., F.V., D.W.A., V.F., C.D., and Y.S.Y.H. analyzed the data. S.C., R.K.S., X.X., V.F., C.D., and Y.S.Y.H. wrote the manuscript. All authors read and approved the final manuscript.
Corresponding author
Ethics declarations
Competing interests
The authors declare no competing interests.
Peer review
Peer review information
Nature Communications thanks the anonymous reviewer(s) for their contribution to the peer review of this work. A peer review file is available.
Additional information
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Supplementary information
Source data
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
About this article
Cite this article
Chang, SC., Kao, MR., Saldivar, R.K. et al. The Gram-positive bacterium Romboutsia ilealis harbors a polysaccharide synthase that can produce (1,3;1,4)-β-d-glucans. Nat Commun 14, 4526 (2023). https://doi.org/10.1038/s41467-023-40214-z
Received:
Accepted:
Published:
DOI: https://doi.org/10.1038/s41467-023-40214-z
- Springer Nature Limited