Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2014 May 1.
Published in final edited form as: Mol Microbiol. 2013 Apr 17;88(4):772–783. doi: 10.1111/mmi.12220

Phylum-wide general protein O-glycosylation system of the Bacteroidetes

Michael J Coyne 1, C Mark Fletcher 2, Maria Chatzidaki-Livanis 1, Gerald Posch 3, Christina Schaffer 3, Laurie E Comstock 1
PMCID: PMC3656502  NIHMSID: NIHMS463979  PMID: 23551589

Summary

The human gut symbiont Bacteroides fragilis has a general protein O-glycosylation system in which numerous extracytoplasmic proteins are glycosylated at a three amino acid motif. In B. fragilis, protein glycosylation is a fundamental and essential property as mutants with protein glycosylation defects have impaired growth and are unable to competitively colonize the mammalian intestine. In this study, we analyzed the phenotype of B. fragilis mutants with defective protein glycosylation and found that the glycan added to proteins is comprised of a core glycan and an outer glycan. The genetic region encoding proteins for the synthesis of the outer glycan is conserved within a Bacteroides species but divergent between species. Unlike the outer glycan, an antiserum raised to the core glycan reacted with all Bacteroidetes species tested, from all four classes of the phylum. We found that these diverse Bacteroidetes species synthesize numerous glycoproteins and glycosylate proteins at the same three amino acid motif. The wide-spread conservation of this protein glycosylation system within the phylum suggests that this system of post-translational protein modification evolved early, before the divergence of the four classes of Bacteroidetes, and has been maintained due to its physiologic importance to the diverse species of this phylum.

Keywords: Bacteroidetes, Bacteroides, protein glycosylation, glycan

Introduction

Although recognized for decades (Sleytr & Thorne, 1976), bacterial protein glycosylation has only recently been detected in a large number of diverse species. Most bacteria known to glycosylate proteins have specialized glycosylation systems whereby a few abundant polymeric surface proteins — such as flagellins, pilins, or S-layer proteins — are glycosylated (reviewed (Nothaft & Szymanski, 2010). Of the thousands of bacterial species recognized to date, relatively few have been demonstrated to have general protein glycosylation systems. Those systems that are documented are mainly confined to the genus/family in which the system was initially described. One of the most wide-ranging bacterial general protein glycosylation systems occurs in the Actinomyces. Several genera within the order Actinomycetales including Mycobacterium, Streptomyces and Corynebacterium have a common protein O-mannosylation system that resembles protein O-mannosylation in fungal and higher eukaryotes (Wehmeier et al., 2009, Dobos et al., 1996, Espitia & Mancilla, 1989, Cowlishaw & Smith, 2002). Campylobacter spp. were the first bacteria demonstrated to have a general protein glycosylation system in which target proteins are N-glycosylated (Szymanski et al., 1999). General protein O-glycosylation systems have been described in Neisseria spp. (Vik et al., 2009, Ku et al., 2009), Bacteroides spp. (Fletcher et al., 2009), Acinetobacter baumannii (Iwashkiw et al., 2012), Francisella tularensis (Egge-Jacobsen et al., 2011) and the Actinomycetes (reviewed (Espitia et al., 2010)).

The Bacteroides fragilis general O-glycosylation system shares a few characteristics with the general glycosylation systems of Campylobacter and Neisseria species in that for a protein to be glycosylated, it must be transported out of the cytoplasm. Also, the glycoproteins localize to various extracytoplasmic locations including the inner membrane, periplasm, outer membrane and outer surface. Despite these few broad commonalities, the systems are very distinct. A notable feature of the B. fragilis O-glycosylation system is that the glycan is added to serine and threonine residues within the three amino acid motif D(S,T)(A,L,V,I,M,T), in contrast to O-glycosylation in eukaryotes and to specific O-glycosylation systems in bacteria where there is no extended motif surrounding the glycosylated amino acid.

A recent mass spectrometry analysis demonstrated that the B. fragilis glycan comprises nine sugars with a molecular mass of 1550.6 Da (Posch et al., 2013). We previously demonstrated that this glycan contains fucose, and that the mutant strain Δgmd-fclΔfkp, with deletions of genes involved in both the de novo and salvage pathways for synthesis of GDP-fucose, is defective in protein glycosylation (Coyne et al., 2005, Fletcher et al., 2009). In addition, we identified a genetic region also involved in protein glycosylation termed LFG (locus of fragilis glycosylation) that contains multiple glycosyltransferases (GTs) and an oligosaccharide flippase. Deletion of this nine gene region resulted in a glycosylation phenotype indistinguishable from the Δgmd-fclΔfkp mutant in that glycoproteins have a reduced apparent molecular weight and are no longer modified with a fucose-containing glycan.

Protein glycosylation in B. fragilis is a fundamental and physiologically important property of the organism (Fletcher et al., 2009). Both ΔLFG and Δgmd-fclΔfkp are defective in their ability to competitively colonization the gnotobiotic mouse intestine, and each has impaired in vitro growth. The importance of protein glycosylation in B. fragilis is further supported by the fact that genes of the LFG locus are transcriptionally coupled to metG, thereby linking protein synthesis with protein glycosylation. In addition, there are likely hundreds of proteins glycosylated in B. fragilis, possibly accounting for more than half of the extracytoplasmic proteins (Fletcher et al., 2011). Several confirmed glycoproteins are involved in essential cellular processes including protein folding (Fletcher et al., 2009) and cell division (Fletcher et al., 2011).

The phylum Bacteroidetes contains four diverse classes of bacteria: the Bacteroidia, the Flavobacteriia, the Sphingobacteriia, and the Cytophagia. The Bacteroidia contain anaerobic species that associate with mammalian, vertebrate, and invertebrate hosts, largely as commensals, symbionts or pathogens, colonizing specific regions of the alimentary tract. Flavobacteriia include aerobic species occupying marine and soil habitats as well as endosymbiotic species. Sphingobacteriia species are found in marine and soil environments, and some Cytophagia species are adapted to survive in extreme salinity (Anton et al., 2002) and temperature (Sako et al., 1996).

We previously showed that five human intestinal Bacteroides species have general O-glycosylation systems similar to that of B. fragilis in that numerous proteins are glycosylated at the same three amino acid motif (Fletcher et al., 2009). Glycoproteins of these intestinal Bacteroides species are also modified with a fucose-containing glycan. In this study, we genetically and phenotypically characterize this O-glycosylation system and demonstrate that this general system of protein O-glycosylation is conserved in species of all classes of Bacteroidetes.

Results

ΔLFG and Δgmd-fcl Δfkp are glycosylated with a small core glycan

Previous analyses demonstrated that the glycoproteins of wild-type B. fragilis are larger than those in the Δgmd-fcl Δfkp and ΔLFG mutants (Coyne et al., 2005, Fletcher et al., 2009). In addition, the proteins in these mutants are not modified with a fucose-containing glycan (Coyne et al., 2005, Fletcher et al., 2009). To determine if these proteins are glycosylated at all, we processed SDS-PAGE gels containing cell lysates of wild type and the Δgmd-fclΔfkp and ΔLFG mutants with the Pro-Q Emerald Glycoprotein stain (Molecular Probes), which specifically stains glycosylated proteins. This analysis demonstrated that despite the altered sizes, proteins from these two mutants are glycosylated (Fig 1A). We purified three confirmed glycoproteins by means of a C-terminal His-tag from wild type or Δgmd-fclΔfkp and confirmed that despite the smaller sizes, each of these purified proteins is glycosylated in Δgmd-fclΔfkp (Fig 1B). The published structure of the B. fraglis glycan revealed a nine sugar molecule with a mass of 1550.6 Da. (Posch et al., 2013) (Fig. 1C). To understand how glycosylation is altered in these mutants, we performed further analyses using glycoprotein BF2494, a soluble periplasmic protein that is well expressed (Fletcher et al., 2011). We purified this protein from ΔLFG and Δgmd-fclΔfkp and used LC-MS/MS to determine the mass of tryptic glycopeptide DTILPQVAYYATLAADR, which contains a single glycosylation motif (underlined). This glycopeptide is 322.128 and 322.125 Da greater than the size of the protein component, from the ΔLFG and Δgmd-fclΔfkp backgrounds, respectively (Fig S1–2, Table S1). When compared to the structure of the wild type glycan (Fig. 1C), these data demonstrate that the glycan of these mutants is comprised of only the first two monosaccharides, a protein-linked hexose and branched O-methylated deoxyhexose. Based on our combined genetic and phenotypic data, this two monosaccharide glycan represents a “core glycan” that is built independently from the “outer glycan”. Unlike the outer glycan, the core glycan does not require genes of the LFG region for its synthesis. In addition, the methylated deoxyhexose sugar of the core glycan is not fucose and may be rhamnose, another common deoxyhexose of bacteria.

Figure 1.

Figure 1

Open in a new tab

Analysis of glycosylation in B. fragilis mutants. A. SDS-PAGE analysis of cell lysates of wild type B. fragilis and the Δgmd-fcl Δfkp and ΔLFG mutants processed using the Pro-Q Emerald Glycoprotein staining solution demonstrating that proteins in these mutants are still glycosylated but the glycoproteins are of different size relative to wild type. B. Same analysis this time showing that three purified His-tagged proteins isolated from wild type or Δgmd-fcl Δfkp are reduced in size relative to wild type. C. Preliminary structure of the WT glycan (Posch et al., in press) compared to the glycan of the Δgmd-fcl Δfkp and ΔLFG mutants revealed by LC-MS/MS analysis. The mutants synthesize a glycan with a mass consistent with only the first two monosaccharides of the wild type glycan.

LFG-like (outer glycan encoding) regions of Bacteroides species and phenotypes

The data described above demonstrate that the LFG region of B. fragilis is necessary for the synthesis of the outer glycan. We previously showed that the genomes of five other Bacteroides species have regions similar to LFG (LFG-like regions) where metG is followed by a flippase gene (wzx) and then by several other genes involved in glycan synthesis, which differ between species (Fletcher et al., 2009). Using publically available genomic sequences, we performed a comprehensive analysis of the genetic heterogeneity of LFG-like regions of 73 different Bacteroides strains encompassing 24 different species. We found a common genetic organization not only at the start of these loci where each region begins with metG (green), followed by wzx (blue), and also at the end of the region with orthologs of BF4305 and BF4306 (purple and red, respectively) encoding putative glycosyltransferases (Fig. 2).

Figure 2.

Figure 2

Open in a new tab

Genetic organization of the LFG-like regions of 24 Bacteroides species with genomic sequence. All regions begin with orthologs of metG (green) and wzx (blue) and terminate with orthologs of BF4305 (purple) and BF4306 (red). The yellow intervening genes are divergent between species.

As the LFG region of B. fragilis encodes products that synthesize an outer glycan, and the LFG-like regions differ between Bacteroides species, we expected that different Bacteroides species would synthesize distinct outer glycans. To test this, we created an antiserum specific to the B. fragilis outer glycan. We immunized rabbits with the acapsular mutant ΔungD1 ΔungD2ΔPSH (Coyne et al., 2008) to avoid generating an immune response to the capsular polysaccharides. Nearly all the antibodies in this antiserum are directed to the outer glycan, as this antiserum does not react with cell lysates of Δgmd-fclΔfkp or ΔLFG (Fig. 3A). This unexpected finding suggests that in rabbits, the outer glycan is the immunodominant antigen of the acapsular mutant. Using this outer glycan antiserum, we probed a blot containing cell lysates of five different B. fragilis strains and five other Bacteroides species. We found that all B. fragilis species analyzed reacted with this antiserum but the other Bacteroides species did not (Fig. 3B). These data demonstrate that these B. fragilis strains synthesize an immunologically similar outer glycan that differs from the outer glycans of other Bacteroides species.

Figure 3.

Figure 3

Open in a new tab

(A,B) Western immunoblot analyses of reactivity of antiserum raised to the B. fragilis acapsular mutant that is specific to the outer glycan. A. Cell lysates of wild type, Δgmd-fcl Δfkp, ΔLFG, and the acapsular mutant. B. Reactivity of cell lysates of six B. fragilis strains and the type strains of five other Bacteroides species. C. Western immunoblot analysis of B. ovatus, and B. thetaiotaomicron wild type and wzx insertional mutants probed with antiserum to unglycosylated BF2494-His purified from E. coli. D. Western immunoblot analysis of cell lysates of B. thetaiotaomicron strains and other Bacteroides species with an antiserum specific to molecule(s) absent in B. thetaaiotaomicron Ωwzx (outer glycan-specific).

To demonstrate that the LFG-like regions of other Bacteroides species are also involved in the synthesis of a glycan that modifies proteins, we made polar insertional mutations in wzx of the LFG-like regions of the type strains of B. ovatus and B. thetaiotaomicron. We used an antiserum to the protein component of BF2494 to probe western blots containing cell lysates of wild type and these wzx insertional mutants. Fig 3C shows that the B. ovatus and B. thetaiotaomicron orthologs of glycoprotein BF2494 are smaller in the wzx insertional mutants than they are in their wild type backgrounds (Fig. 3C).

Analysis of multiple strains of numerous Bacteroides species shows that with few exceptions, the LFG-like regions are conserved within a species (Fig. S4). These data are consistent with immunoblot analyses showing that the B. fragilis outer glycan antiserum recognizes all B. fragilis strains tested. To demonstrate that B. thetaiotaomicron strains synthesize an outer glycan that is similar within this species, we made an antiserum to the outer glycan of type strain VPI-5482 by immunizing rabbits with formalin-fixed wild type bacteria and then adsorbing the resulting antiserum with the B. thetaiotaomicron wzx mutant, removing all antibodies except those to the glycan absent in the mutant. Analysis of the reactivity of this antiserum demonstrates that all B. thetaiotaomicron strains analyzed synthesize an immunologically similar outer glycan, different from those of other Bacteroides species (Fig. 3D).

Glycosyltransferases BF4305 and BF4306

Due to the conservation of BF4305 and BF4306 in the LFG-like regions of all Bacteroides species, we predicted that they may function in the addition of the first and second monosaccharides of the outer glycan. To determine the order of monosaccharide addition by the LFG GTs, individual GT encoding genes were deleted (Fig. 4A) and the glycosylation phenotypes of these mutants were analyzed by western immunoblot. We used an antiserum to the protein component of BF0522, a glycoprotein with a single glycosylation motif (Fletcher et al., 2011), to determine the order in which the monosaccharides are added (Fig. 4B). The size of BF0522 from ΔBF4306 is indistinguishable from that produced by ΔLFG. LC-MS/MS analysis of the BF2494-His glycopeptide from ΔBF4306 revealed a mass addition of 322.131 Da (Fig S3, Table S1). This mass is nearly identical to the mass addition seen when this protein was recovered from both ΔLFG and Δgmd-fcl Δfkp, which are modified with only the core glycan. Therefore, BF4306 is necessary for the initial step of outer glycan synthesis. The next smallest form of BF0522 is from ΔBF4305, suggesting this GT adds the next monosaccharide of the outer glycan. As BF0522 from ΔBF4301 and ΔBF4300 are similarly sized, these GTs may each add a different monosaccharide to the same sugar, one as a side chain. Indeed, structural analysis of the wild type glycan did not fully rule out the possibility of branching in the outer glycan (Posch et al., 2013). BF4302 is annotated as hypothetical and its deletion does not affect glycoprotein synthesis (Fig 4B,C).

Figure 4.

Figure 4

Open in a new tab

LFG region glycosyltransferases. A. Genetic map of the B. fragilis LFG region showing deletion mutants created. B. Analysis of the size of glycoprotein BF0522 from deletion mutants with an antiserum to the protein component of BF0522-His purified from E. coli. C. Reactivity of mutants with and without the gene added back in trans with the α-outer glycan serum.

The conservation of BF4305 and BF4306 in LFG regions of Bacteroides species suggests that monosaccharides at the reducing end of the outer glycan are common between these intestinal species, yet there is no cross-reactivity of outer glycan sera between Bacteroides species. Using the panel of deletion mutants, we determined that the B. fragilis outer glycan antiserum does not react with truncated glycans, demonstrating specificity either for the full length glycan, or for sugars/linkages at the non-reducing end (Fig 4C), and explaining why this antiserum does not react with the outer glycans of other Bacteroides species. When BF4305 was placed in trans to the ΔBF4305 mutant, full reactivity with this antiserum was not restored. As BF4305 and BF4306 overlap, it was necessary to include the beginning of BF4306 in the construct. The first 12 amino acids of BF4306 will be synthesized from this construct, and may interfere with native BF4306, possibly contributing to the incomplete restoration of immunoreactivity.

Diverse species within the Bacteroidetes phylum O-glycosylate numerous proteins

Elizabethkinia meningoseptica and Pedobacter heparinus, from the Flavobacteriia and Sphinogbacteriia classes, respectively, have each been shown to synthesize a limited number of glycoproteins (Plummer et al., 1995, Huang et al., 1995). Structural analysis of the E. meningoseptica glycan revealed a heptasaccharide of which the first two sugars at the reducing end are mannose followed by a branched O-methylated rhamnose (Reinhold et al., 1995). These two sugars are consistent with the B. fragilis core glycan structure. In addition, structural analysis of the P. heparinus glycan similarly revealed an O-linked mannose with a branching rhamnose at the reducing end (Huang et al., 1995). Also, the second sugar of the B. fragilis core glycan, (Me) dHex, is likely methylated rhamnose as we have shown that the other common deoxy-sugar, fucose, is not present in the core glycan (Fletcher et al., 2009). Based on the similarities of the core glycans from two different Bacteroidetes classes to each other and to the core glycan of B. fragilis, we sought to determine if there were immunological commonalities in the core glycans from different classes of this diverse phylum. We generated an antiserum to the B. fragilis core glycan by immunizing rabbits with His-tagged glycoprotein BF2494 purified from the ΔLFG mutant (synthesizing only the core glycan) and adsorbed the resulting antiserum over a column containing BF2494-His purified from E. coli. This adsorption removed all the antibodies to the protein and His-tag rendering the antisera unable to react with the unglycosylated recombinant BF2494-His purified from E. coli (Fig 5A). This core glycan antiserum recognizes proteins modified with both the wild type glycan and the core glycan (Δgmd-fcl Δfkp, ΔLFG, ΔBF4306) Fig 5B. As expected, the molecules that are recognized by this antiserum are a different size when comparing wild type to the core-only mutants (Fig. 5B). In addition, this antiserum recognizes proteins of altered molecular sizes in the B. ovatus and B. thetaiotaomicron wzx mutants compared to the sizes of the molecules recognized in their wild type counterparts (Fig. 5C), showing that these Bacteroides species also glycosylate proteins with an immunologically similar core glycan. These data demonstrate that this antiserum is specific to the core glycan and recognizes not only the core glycan but also the wild type glycan of these Bacteroides species, which contains core glycan at the reducing end.

Figure 5.

Figure 5

Open in a new tab

Bacterioides species synthesize a common core glycan. A. Antiserum raised to BF2494-His from ΔLFG before (left panel) and after (right panel) adsorption with BF2494-His purified from E. coli. Lanes contain purified BF2494-His from E. coli or B. fragilis. The arrows show the purified protein from each background. The recombinant protein is larger from E. coli due to the addition of several amino acids between the protein and His-tag created by cloning in pET16b (B,C) Western immunoblot analysis using α-core glycan with cell lysates of B. B. fragilis wild type and mutants, C. Bacteroides wild type and wzx insertional mutants,

To extend this analysis to species within the phylum Bacteroidetes, we analyzed the reactivity of cell lysates of diverse species contained within all four classes of this phylum. The species we analyzed are arranged phylogenetically in Fig. 6A and include oral pathogens, marine and soil organisms, and the halophilic species Salinibacter ruber. Immunoblot analysis demonstrated that numerous bands from each Bacteroidetes species (colored) react with the core glycan antiserum (Fig 6B). In contrast, there is almost no reactivity with cell lysates of E. coli, Pseudomonas aeruginosa and Clostridium innocuum, with only a very faint low molecular weight band discernible in E. coli and P. aeruginosa (Fig 6B). These data demonstrate that the synthesis of glycoproteins is a general feature of the Bacteroidetes phylum where numerous proteins are glycosylated and that the glycans contains a core glycan similar to the B. fragilis core glycan. To determine if the core antiserum was reactive with O-glycosylated proteins of non-Bacteroidetes species, we tested this antiserum against three other species of bacteria with general protein glycosylation systems, C. jejuni, M. tuberculosis, and N. gonorrhea (Fig 6C). Several reactive bands were detected only in M. tuberculosis, an organism in which several proteins are O-mannosylated. The N-linked glycans (C. jejuni) and glycans O-linked with an N-acetylated sugar at the reducing end (N. gonorrhea) do not react with this antiserum.

Figure 6.

Figure 6

Open in a new tab

Common protein glycosylation system in Bacteroidetes. A. Phylogeny of the Bacteroidetes species used in this study. (B,C) Western immunoblot analysis using α-core glycan with cell lysates of B. diverse Bacteroidetes and non-Bacteroidetes species, green-Bacteroidiia, purple- Flavobacteriia, red- Sphingobacteriia, blue-Cytophagia species. C. 1- B. fragilis, 2- C. jejuni, 3- M. tuberculosis, 4- N. gonorrhoeae. D. Western immunoblot of whole cell lysates of E. coli and Bacteroidetes species synthesizing BF2494-His or the mutant BF2494-His with the triple T>A glycosylation site mutations, probed with α-His mAb.

As the data suggest that glycoproteins of diverse Bacteriodetes species are modified with a common core glycan, likely containing mannose and rhamnose or methylated rhamnose, it would be expected that the genomes of these diverse species would each contain genes that encode the four enzymes necessary for dTDP-rhamnose biosynthesis. These genes, rmlA, B, C, and D, encode products that convert glucose-1-P to dTDP-rhamnose, the nucleotide activated form necessary for the addition of rhamnose into glycans. Of the 18 Bacteroidetes strains analyzed in this study (Fig 6B), there are complete or draft genomes for 16 of them. We analyzed each of these 16 genomes for the four genes involved in dTDP-rhamnose biosynthesis and found that all genomes contain all four genes, in many cases multiple copies of each, and in most genomes, these genes are not clustered (Table S2).

Conservation of the glycosylation motif

We predicted that the three amino acid glycosylation motif D(S/T)(A/L/V/I/M/T) described in Bacteroides sp. would be a conserved glycosylation site in the phylum. To determine if phylogenetically distinct Bacteroidetes species also use this same motif, we transformed Sphingobacterium multivorum and Flavobacterium johnsoniae via electroporation with a plasmid containing B. fragilis BF2494-His, or BF2494-His with the threonines at the three glycosylation motifs of BF2494 changed to alanine (3T>A). This modified protein is not glycosylated in B. fragilis (Fletcher et al., 2009). In E. coli, both the BF2494-His and the BF2494-His 3T>A mutant protein show the same mobility in the Western blots, indicating that they have the identical apparent molecular mass. This is expected since E. coli does not have this glycosylation machinery (Fig 6D). In F. johnsoniae and S. multivorum, the wild type protein shows a laddering phenotype, similar to the glycosylation phenotype in B. fragilis, demonstrating that the protein is glycosylated in these divergent Bacteroidetes species. In contrast, expression of the triple glycosylation site mutant gene in F. johnsoniae and S. multivorum resulted in only a single, lower molecular weight protein compared to their wild type sizes, consistent with its unglycosylated size in B. fragilis. Therefore, these diverse Bacteroidetes species not only glycosylate proteins with a similar core glycan, but also use the same glycosylation motif.

We previously analyzed 20 different B. fragilis secreted proteins with glycosylation motifs and found that all were glycosylated. To date, we have not identified a secreted protein of B. fragilis that contains a glycosylation motif that is not glycosylated. Therefore, based on the number of predicted secreted proteins with glycosylation motifs in B. fragilis, there are 1021 candidate glycoproteins in this organism (Fletcher et al., 2011). The α-core glycan western immunoblots reveal that species of Sphingobacteriia, Flavobacteriia, and Cytophagia also glycosylate numerous proteins. Similar bioinformatics analyses were performed on all 266 non-endosymbiotic species of Bacteroidetes for which there is genome sequence and the numbers of putative glycoproteins are listed in Table S3. Similar to B. fragilis, the number of candidate glycoproteins in these species is high, with many strains predicted to glycosylate more than 1000 proteins.

Discussion

In this study, we describe a general protein O-glycosylation system that is conserved in phylogenetically diverse members of the phylum Bacteroidetes. All 18 of the Bacteroidetes species we analyzed synthesize multiple glycoproteins that are recognized with the antiserum to the O-linked core glycan. As this antiserum also reacts strongly with M. tuberculosis, in which proteins are modified with an O-linked mannose but do not contain the second monosaccharide of the core glycan, the glycans of these diverse Bacteroidetes species contain at least a common glycan epitope, likely an O-linked mannose or other hexose. Indeed, the glycans of two other Bacteroidetes species, E. meningosepticum (Reinhold et al., 1995) and P. heparinus (Huang et al., 1995), both have an O-linked mannose at the reducing end linked to a branched rhamnose or O-methylated rhamnose, supporting the conclusion that the core glycan may well be common to this phylum. In both of these species, the glycan chain extends from the O-linked mannose residue, as occurs in B. fragilis (Posch et al., 2013). A model of protein O-glycosyation in B. fragilis based on our current understanding is provided in Figure 7.

Figure 7.

Figure 7

Open in a new tab

Proposed model for glycan synthesis and addition to proteins in B. fragilis. The data suggests one of two models. In the model shown, the core and outer glycan are built separately on Und-P lipid carriers and flipped to the periplasmic side of the inner membrane by individual by Wzx flippases. The core glycan would be linked to the protein at the glycosylation motif by an as yet unidentified oligosaccharyltransferase (Otase). The outer glycan would then be linked to the core glycan by an O-antigen ligase-like protein. Alternatively, the outer glycan may be added to the core glycan in the cytoplasm and then a single flippase would transfer the entire glycan to the periplasm for addition to proteins. In this alternative model, a ligase to link core and outer glycan would not be necessary.

We show that S. multivorum and F. johnsoniae glycosylate proteins at the same motif as B. fragilis. Another Flavobacteriia species, E. meningosepticum, was previously shown to glycosylate three proteins at a motif containing DS or DT (Plummer et al., 1995). Our examination of the amino acid sequences that follow the glycosylated S or T of these three E. meningosepticum proteins reveal four different amino acids, all of which are tolerated at the third position of the B. fragilis glycosylation motif. Therefore, the glycosylation motif D(S/T)(A/L/V/I/M/T) is a phylum-wide protein O-glycosylation motif used by species of all classes of Bacteroidetes. Elucidation of this phylum-wide glycosylation motif allows us to calculate the number of candidate glycoproteins in sequenced Bacteroidetes species (Table S3). In many species, candidate glycoproteins include more than half of the extracytoplasmic proteins, consistent with the numerous glycoproteins revealed by immunoblot analysis.

The data strongly suggest that this general O-glycosylation system was present in an ancestral Bacteroidetes species prior to the divergence of the different classes. It is intriguing that such diverse Bacteroidetes species, including both anaerobes and aerobes, and organisms that live in varied environments including the mammalian alimentary tract, marine and soil habitats, and under conditions of extreme salinity, would all maintain this protein glycosylation system through their distinct evolutionary pathways. The phylum-wide conservation of this glycosylation system suggests that it is physiologically important to these bacteria, a feature not described to be imparted by glycosylation systems of other bacteria. Our initial analyses of both B. fragilis ΔLFG and B. fragilis Δgmd-fclΔfkp revealed that they had aberrant in vitro growth and could not competitively colonize the mammalian intestine (Coyne et al., 2005, Fletcher et al., 2009). These growth deficiencies were not due to complete lack of protein glycosylation, but rather the loss of the outer glycan. Due to the large number of candidate glycoproteins in B. fragilis, 43 of which are encoded by candidate essential genes (Fletcher et al., 2011), it is not surprising that these mutants would have defective growth and suggests that the glycosylation of proteins, with at least the core glycan, is likely essential to these organisms. Therefore, genes involved in core glycan synthesis should be essential. Among the Bacteroidetes species, essential gene analysis has only been performed for B. thetaiotaomicron (Goodman et al., 2009). Identifiable genes that should be required for core glycan synthesis are the four dTDP-rhamnose biosynthesis genes, and if present in single copy, should be essential genes. Although the B. thetaiotaomicron genome has multiple copies of each of these four genes, rmlA (BT2017) and rmlB (BT2016) were found to be essential genes (Goodman et al., 2009). These data further support that glycosylation of proteins with core glycan is an essential property of these bacteria.

Experimental Procedures

Bacterial strains and growth conditions

All Bacterodies, Parabacteroides, Porphyromonas, Prevotella, and Clostridium strains were grown in supplemented basal medium (Pantosti et al., 1991). Flavobacterium johnsoniae ATCC 17061, Elizabethkingia meningoseptica ATCC 13253 and Myroides odoratus DSM 2801 were grown in nutrient broth or BHIS (Brain Heart Infusion medium supplemented for Bacteroides growth) at 30°C and 37°C, respectively, Sphingobacterium multivorum ATCC 35656 and Sphingobacterium spiritivorum ATCC 33861 were grown in tryptic soy broth at 30°C, Salinibacter ruber BAA-605 was grown in ATCC medium 2402, and Pseudomonas aeruginosa PAO1 and E. coli strains were grown in L-broth. Campylobacter jejuni ATCC 33291 was obtained as a heat-killed pellet, Mycobacterium tuberculosis H37Rv was obtained as a methanol inactivated pellet, Neisseria gonorrhoeae ATCC 43069 was acquired as a heat killed pellet. Bacteroidales type strains are as follows: B. fragilis NCTC 9343, B. thetaiotaomicron VPI-5482, B. caccae ATCC 43185, B. ovatus ATCC 8483, B. uniformis ATCC 8492, B. vulgatus ATCC 8482, Parabacteroides distasonis ATCC 8503, Parabacteriodes merdae ATCC 43184, Porphyromonas gingivalis W83, Prevotella intermedia 17, Prevotella melaninogenica ATCC 25845, Tannerella forsythensis ATCC 43037. B. thetaiotaomicron strains CL01T03C02, CL03T03C07, and CL04T03C22, and Clostridium aff. innocuum CL03T06C07, were recovered from human feces as described (Zitomersky et al., 2011).

Insertional inactivation of the LFG-like loci of Bacteroides species

Plasmids were created to insertionally mutate wzx in B. fragilis NTCC 9343, B. ovatus ATCC 8483, and B. thetaiotaomicron VPI-5482. A 1,051-bp internal portion of the 1,455-bp wzx of B. ovatus (BACOVA_02186), and a 1,047-bp internal portion of the 1,449-bp wzx (BT2934) of B. thetaiotaomicron were amplified using the primers indicated in Table S4. Each PCR product was digested with BamHI and cloned into the Bacteroides suicide vector pNJR6 (Stevens et al., 1990). These plasmids were introduced into the appropriate Bacteroides species by conjugal transfer and cointegrates were selected by acquisition of erythromycin resistance.

Deletion of individual genes of the LFG locus of B. fragilis

DNA segments flanking genes to be deleted were amplified using primers indicated in Table S4. PCR products were digested with BamHI and either EcoRI or NcoI utilizing restriction sites engineered into the 5’ end of the primers, and ligated with BamHI-digested pNJR6. The ligations were used to transform E. coli DH5α and transformants selected on kanamycin plates. Correct clones were introduced into B. fragilis NCTC 9343 by conjugal transfer from E. coli. Selection for cointegrate strains and recovery of deletants was performed as previously described (Coyne et al., 2011).

Cloning of LFG region genes for complementation studies

BF4300, BF4301, BF4305, and BF4306 were PCR amplified using the primers indicated in Table S4. The PCR products were digested with BamHI and cloned into the BamHI site of expression vector pMCL140 (Chatzidaki-Livanis et al., 2010). The resulting plasmids were mobilized into the appropriate B. fragilis LFG region mutant via conjugal transfer.

Antisera production

Antibodies to the B. fragilis acapsular mutant (ΔungD1ΔungD2ΔPSH) (Coyne et al., 2008) and to B. thetaiotaomicron ATCC VPI-5482 were generated by immunizing rabbits with formalin-fixed whole bacteria using the EXPRESS-LINE protocol of Lampire Biological Laboratories (Pipersville, PA). To obtain antibodies specific to the molecule(s) absent in the B. thetaiotaomicron wzx insertional mutant, 100 µl of the wild-type whole organism antiserum was diluted in 10 ml phosphate buffered saline and mixed for one hour with two successive B. thetaiotaomicron wzx mutant bacterial pellets corresponding to 100 ml of an overnight culture. Antibodies to the core glycan were generated by immunizing rabbits using the EXPRESS-LINE protocol with BF2494-His purified from B. fragilis ΔLFG. Antibodies to the protein component and His-tag of BF2494-His were adsorbed from this antiserum by affinity adsorption with BF2494-His purified from E. coli immobilized on Ni-NTA agarose nickel-chelating resin (Life Technologies, Grand Island, NY). This adsorption was performed twice and the flow-through antiserum was confirmed to be devoid of antibodies to BF2494-His from E. coli. Generation of the antisera to the BF2494-His and BF0522-His proteins purified from E. coli was previously described (Fletcher et al., 2009). THE His Tag Antibody (GenScript USA Inc., Piscataway, NJ) mouse mAb was used to detect His-tagged fusion proteins.

Creation of an E. coli-Flavobacterium shuttle vector expressing wild type and mutant BF2494-His

Bacteroides vectors were shown to be unable to replicate in F. johnsoniae (Agarwal et al., 1997). Therefore, pCMF8 and pCMF62 were modified to carry the functions required for plasmid maintenance in F. johnsoniae. A 2.8-kb fragment was PCR amplified from the E. coli-F. johnsoniae shuttle vector pCP23 (Agarwal et al., 1997) using the primers indicated in Table S4, digested with SacI (the forward primer is situated upstream of a SacI site on pCP23 that defines one end of the 2.8-kb fragment), and cloned into a unique SacI site in pCMF8 and pCMF62, creating plasmids pMJC97 and pMJC98, respectively.

Electroporation of S. multivorum and F. johnsoniae

S. multivorum and and F. johnsoniae were grown aerobically in tryptic soy broth or supplemented BHI broth (BHIS), respectively, at 30°C until an OD550 of ~0.8. The cells were harvested by centrifugation at 2,000 × g for 10 min at 4°C, washed twice in an equal volume (100 ml) of ice-cold EB buffer (10% glycerol, 1 mM MgCl2), resuspended in 2.5 ml of ice-cold EB buffer, and held on ice. Pre-chilled plasmid DNA of either pCMF8 or pCMF62 (in the case of S. multivorum) or pMJC97 or pMJC98 (in the case of F. johnsoniae) was added to 100 µl of the cell suspension and transferred to pre-chilled E-shot 0.2-cm cuvettes (Life Technologies) and electroporated using a MicroPulser apparatus (Bio-Rad Laboratories) delivering 2.5 kV per pulse. Immediately after electroporation, 1 ml of media was added to the electroporation cuvette. The suspension was incubated aerobically at 30°C for one hour (S. multivorum) or two hours (F. johnsoniae), and plated on BHIS plates containing 10 µg/ml (S. multivorum) or 100 µg/ml (F. johnsoniae) of erythromycin.

Purification of His-tagged proteins for LC- MS/MS analyses

Cultures (3 l) of B. fragilis Δgmd-fcl Δfkp, ΔLFG, and ΔBF4306 mutants harboring pCMF141 encoding BF2494-His were grown to stationary phase, harvested, and washed twice with 500 ml of ice-cold MilliQ water to remove media components. The washed pellets were resuspended in 40 ml of guanidinium lysis buffer (pH 7.8) and the cells lysed by sonication using a Microson ultrasonic cell disruptor (model XL2000; Misonix). Fifteen-second pulses of 11–13 watts (RMS) were delivered six times to each sample, allowing the samples to cool on ice for 30 seconds between each burst. The His-tagged protein was purified under denaturing conditions using 0.5 ml of Ni-NTA agarose resin and the ProBond purification system, following manufacturer’s protocol (Life Technologies). Eluted samples (5 ml) were dialyzed against 10 liters (5 × 2 liters) of MilliQ water at 4°C. Samples were electrophoresed using 12% Bis-Tris gels (Life Technologies), and the gel was stained with SimplyBlue Safe Stain (Life Technologies) and bands were excised from the gels and processed as described below.

Glycan size analysis by LC-MS/MS

Gel slices containing purified BF2494-His isolated from B. fragilis NCTC 9343 Δgmd-fclΔfkp, ΔLFG, and Δ4306 mutants were reduced with 1 mM DTT for 30 min at 60°C and then alkylated with 5mM iodoacetamide for 15 min in the dark at room temperature, and subjected to a modified in-gel trypsin digestion procedure (Shevchenko et al., 1996). Gel pieces were washed and dehydrated with acetonitrile for 10 min. followed by removal of acetonitrile. Pieces were then completely dried in a speed-vac. Gel pieces were rehydrated with 50 mM ammonium bicarbonate solution containing 12.5 ng/µl modified sequencing-grade trypsin (Promega, Madison, WI) and incubated at 37°C overnight. Peptides were extracted by removing the ammonium bicarbonate solution, followed by one wash with a solution containing 50% acetonitrile and 1% formic acid, and dried. Samples were reconstituted in 5 – 10 µl of HPLC solvent A (2.5% acetonitrile, 0.1% formic acid). A nano-scale reverse-phase HPLC capillary column was created by packing 5 µm C18 spherical silica beads into a fused silica capillary (100 µm inner diameter × ~15–20 cm length) with a flame-drawn tip (Peng et al., 2001). After equilibrating the column, each sample was loaded via a Famos auto sampler (LC Packings, San Francisco CA) onto the column. A gradient was formed and peptides were eluted with increasing concentrations of solvent B (97.5% acetonitrile, 0.1% formic acid). Eluted peptides were subjected to electrospray ionization and then entered into an LTQ-Orbitrap mass spectrometer (Thermo Fisher, San Jose, CA). Eluting peptides were detected, isolated, and fragmented to produce a tandem mass spectrum of specific fragment ions for each peptide.

Locating LFG-like regions in Bacteroides species

Genomic DNA sequences from both completed and draft sequencing projects were obtained from Genbank. The NCBI programs mentioned below were contained in release 2.2.26 of the Windows version of the BLAST executables and were downloaded from the NCBI ftp site. If the Bacteroides species from which a genomic sequence was derived was not identified by the depositor, a BLASTable database was prepared using formatdb (Altschul et al., 1990, Altschul et al., 1997) and all high scoring segment pairs (HSPs) greater than or equal to 400 bp returned by bastn using the DNA sequence of one of the B. fragilis 9343 16S rRNA genes (BFr04; NCBI GeneID 3289048) as a query were retained. For each such organism, these DNA sequences were used as queries via the SOAP interface to the Seqmatch program of the Ribosomal Database Project (release 10.31, (Cole et al., 2009)) and the best matches returned for each segment were compared for species identification.

A pipeline of custom Perl scripts was utilized to locate and analyze the LFG-like region of each Bacteroides species. A BLASTable database was prepared for each organism using formatdb and protein translation information provided by the depositor. The closest orthologs to the B. fragilis 9343 BF4297 (MetG), BF4298 (Wzx), BF4305 and BF4306 LFG region proteins were located by blastp. If the orthologs detected by this method were clustered on the same contig, the genomic coordinates of the DNA sequence spanning the start of metG (encoding the BF4297 ortholog) and the end of the terminal glycosyltransferase family 2 gene (encoding the BF4306 ortholog) were calculated and used to retrieve the segment from NCBI in a GenBank format file. The features embodied in these GenBank files were used to generate the multi-panel open reading frame maps.

Detection of putative dTDP-rhamnose biosynthesis pathway proteins

The position-specific scoring matrices constituting NCBI's Conserved Domain Database (CDD, v. 3.09) (Marchler-Bauer et al., 2013) were retrieved, and the 4,873 COG (Tatusov et al., 2003) models were extracted and used to create an RPS-Blast searchable database using the formatrpsdb utility (v. 2.2.26) with the recommended settings (-o T -f 9.82 -S 100.0).

Complete and draft bacterial genomes of the strains under investigation were downloaded from NCBI. Each protein of each genome was compared to the local COG database via reverse position specific BLAST (Altschul et al., 1990) using the rpsblast utility (v. 2.2.26). Perl scripts were utilized to detect and enumerate proteins containing a match to one of the COG models of interest (COG1209, COG1088, COG1898, and COG1091 for RmlA, RmlB, RmlC, and RmlD, respectively) at an e-value less than or equal to 1 × 10–60. The data thus generated was used to compile Table S2.

Analysis of Bacteroidetes proteomes for enumeration of candidate glycoproteins

The proteome of each genome classified as a member of the Bacteroidetes Chlorobi group by NCBI was downloaded from GenBank (excluding endosymbionts of Blattabacterium spp. and from the category Candidatus) and analyzed using LipoP (Linux version 1.0a, (Juncker et al., 2003)) to predict proteins likely to be extracytoplasmic due the presence of a signal peptidase I or II (SpI or SpII) cleavage site. The proteomes were further analyzed using TMHMM version 2.0c for Linux (Krogh et al., 2001, Sonnhammer et al., 1998) to detect transmembrane proteins, and via a Perl regular expression for sequences containing the glycosylation motif D(S/T)(A/I/L/V/M/T). The data thus generated was used to compile Table S3.

Supplementary Material

Supp Fig S1-S4 & Supp Table S1-S4

Acknowledgements

We thank R. Tomaino for LC-MS/MS analyses, A. Tanner, A. Onderdonk, K. Leung and R. Husson for bacterial strains, and M.J. McBride for plasmid pCP23. This work was supported by Austrian Science Fund project P20605-B12 and NIH/NIAID grants R01AI067711 and R01AI081843.

Footnotes

The authors have no conflict of interest to declare.

References

  1. Agarwal S, Hunnicutt DW, McBride MJ. Cloning and characterization of the Flavobacterium johnsoniae (Cytophaga johnsonae) gliding motility gene, gldA. Proc Natl Acad Sci U S A. 1997;94:12139–12144. doi: 10.1073/pnas.94.22.12139. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Altschul SF, Gish W, Miller W, Myers EW, Lipman DJ. Basic local alignment search tool. J Mol Biol. 1990;215:403–410. doi: 10.1016/S0022-2836(05)80360-2. [DOI] [PubMed] [Google Scholar]
  3. Altschul SF, Madden TL, Schaffer AA, Zhang J, Zhang Z, Miller W, Lipman DJ. Gapped BLAST and PSI-BLAST: a new generation of protein database search programs. Nucleic Acids Res. 1997;25:3389–3402. doi: 10.1093/nar/25.17.3389. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Anton J, Oren A, Benlloch S, Rodriguez-Valera F, Amann R, Rossello-Mora R. Salinibacter ruber gen. nov., sp. nov., a novel, extremely halophilic member of the Bacteria from saltern crystallizer ponds. Int J Syst Evol Microbiol. 2002;52:485–491. doi: 10.1099/00207713-52-2-485. [DOI] [PubMed] [Google Scholar]
  5. Chatzidaki-Livanis M, Weinacht KG, Comstock LE. Trans locus inhibitors limit concomitant polysaccharide synthesis in the human gut symbiont Bacteroides fragilis. Proc Natl Acad Sci U S A. 2010;107:11976–11980. doi: 10.1073/pnas.1005039107. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Cole JR, Wang Q, Cardenas E, Fish J, Chai B, Farris RJ, Kulam-Syed-Mohideen AS, McGarrell DM, Marsh T, Garrity GM, Tiedje JM. The Ribosomal Database Project: improved alignments and new tools for rRNA analysis. Nucleic Acids Res. 2009;37:D141–D145. doi: 10.1093/nar/gkn879. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Cowlishaw DA, Smith MC. A gene encoding a homologue of dolichol phosphate-beta-D-mannose synthase is required for infection of Streptomyces coelicolor A3 (2) by phage (phi) C31. J Bacteriol. 2002;184:6081–6083. doi: 10.1128/JB.184.21.6081-6083.2002. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Coyne MJ, Chatzidaki-Livanis M, Paoletti LC, Comstock LE. Role of glycan synthesis in colonization of the mammalian gut by the bacterial symbiont Bacteroides fragilis. Proc Natl Acad Sci U S A. 2008;105:13099–13104. doi: 10.1073/pnas.0804220105. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Coyne MJ, Fletcher CM, Reinap B, Comstock LE. UDP-glucuronic acid decarboxylases of Bacteroides fragilis and their prevalence in bacteria. J Bacteriol. 2011;193:5252–5259. doi: 10.1128/JB.05337-11. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Coyne MJ, Reinap B, Lee MM, Comstock LE. Human symbionts use a host-like pathway for surface fucosylation. Science. 2005;307:1778–1781. doi: 10.1126/science.1106469. [DOI] [PubMed] [Google Scholar]
  11. Dobos KM, Khoo KH, Swiderek KM, Brennan PJ, Belisle JT. Definition of the full extent of glycosylation of the 45-kilodalton glycoprotein of Mycobacterium tuberculosis. J Bacteriol. 1996;178:2498–2506. doi: 10.1128/jb.178.9.2498-2506.1996. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Egge-Jacobsen W, Salomonsson EN, Aas FE, Forslund AL, Winther-Larsen HC, Maier J, Macellaro A, Kuoppa K, Oyston PC, Titball RW, Thomas RM, Forsberg A, Prior JL, Koomey M. O-linked glycosylation of the PilA pilin protein of Francisella tularensis: identification of the endogenous protein-targeting oligosaccharyltransferase and characterization of the native oligosaccharide. J Bacteriol. 2011;193:5487–5497. doi: 10.1128/JB.00383-11. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Espitia C, Mancilla R. Identification, isolation and partial characterization of Mycobacterium tuberculosis glycoprotein antigens. Clin Exp Immunol. 1989;77:378–383. [PMC free article] [PubMed] [Google Scholar]
  14. Espitia C, Servín-González L, Mancilla R. New insights into protein O-mannosylation in actinomycetes. Mol Biosyst. 2010;6:775–781. doi: 10.1039/b916394h. [DOI] [PubMed] [Google Scholar]
  15. Fletcher CM, Coyne MJ, Comstock LE. Theoretical and experimental characterization of the scope of protein O-glycosylation in Bacteroides fragilis. J Biol Chem. 2011;286:3219–3226. doi: 10.1074/jbc.M110.194506. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Fletcher CM, Coyne MJ, Villa OF, Chatzidaki-Livanis M, Comstock LE. A general O-glycosylation system important to the physiology of a major human intestinal symbiont. Cell. 2009;137:321–331. doi: 10.1016/j.cell.2009.02.041. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Goodman AL, McNulty NP, Zhao Y, Leip D, Mitra RD, Lozupone CA, Knight R, Gordon JI. Identifying genetic determinants needed to establish a human gut symbiont in its habitat. Cell Host Microbe. 2009;6:279–289. doi: 10.1016/j.chom.2009.08.003. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Huang L, Van Halbeek H, Eggimann B, Zimmermann J. Structural characterization of the novel O-linked carbohydrate structure of Flavobacterium heparinum heparinase I. Glycobiology. 1995;5:712. [Google Scholar]
  19. Iwashkiw JA, Seper A, Weber BS, Scott NE, Vinogradov E, Stratilo C, Reiz B, Cordwell SJ, Whittal R, Schild S, Feldman MF. Identification of a general O-linked protein glycosylation system in Acinetobacter baumannii and its role in virulence and biofilm formation. PLoS pathogens. 2012;8:e1002758. doi: 10.1371/journal.ppat.1002758. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Juncker AS, Willenbrock H, Von Heijne G, Brunak S, Nielsen H, Krogh A. Prediction of lipoprotein signal peptides in Gram-negative bacteria. Protein Sci. 2003;12:1652–1662. doi: 10.1110/ps.0303703. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Krogh A, Larsson B, von Heijne G, Sonnhammer EL. Predicting transmembrane protein topology with a hidden Markov model: application to complete genomes. J Mol Biol. 2001;305:567–580. doi: 10.1006/jmbi.2000.4315. [DOI] [PubMed] [Google Scholar]
  22. Ku SC, Schulz BL, Power PM, Jennings MP. The pilin O-glycosylation pathway of pathogenic Neisseria is a general system that glycosylates AniA, an outer membrane nitrite reductase. Biochem Biophys Res Commun. 2009;378:84–89. doi: 10.1016/j.bbrc.2008.11.025. [DOI] [PubMed] [Google Scholar]
  23. Marchler-Bauer A, Zheng C, Chitsaz F, Derbyshire MK, Geer LY, Geer RC, Gonzales NR, Gwadz M, Hurwitz DI, Lanczycki CJ, Lu F, Lu S, Marchler GH, Song JS, Thanki N, Yamashita RA, Zhang D, Bryant SH. CDD: conserved domains and protein three-dimensional structure. Nucleic Acids Res. 2013;41:D348–D352. doi: 10.1093/nar/gks1243. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Nothaft H, Szymanski CM. Protein glycosylation in bacteria: sweeter than ever. Nat Rev Microbiol. 2010;8:765–778. doi: 10.1038/nrmicro2383. [DOI] [PubMed] [Google Scholar]
  25. Pantosti A, Tzianabos AO, Onderdonk AB, Kasper DL. Immunochemical characterization of two surface polysaccharides of Bacteroides fragilis. Infect Immun. 1991;59:2075–2082. doi: 10.1128/iai.59.6.2075-2082.1991. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Peng SX, Branch TM, King SL. Fully automated 96-well liquid-liquid extraction for analysis of biological samples by liquid chromatography with tandem mass spectrometry. Anal Chem. 2001;73:708–714. doi: 10.1021/ac001036c. [DOI] [PubMed] [Google Scholar]
  27. Plummer TH, Jr, Tarentino AL, Hauer CR. Novel, specific O-glycosylation of secreted Flavobacterium meningosepticum proteins. Asp-Ser and Asp-Thr-Thr consensus sites. J Biol Chem. 1995;270:13192–13196. doi: 10.1074/jbc.270.22.13192. [DOI] [PubMed] [Google Scholar]
  28. Posch G, Pabs tM, Neumann L, Coyne MJ, Altmann F, Messner P, Comstock LE, Schäffer C. Cross-glycosylation of proteins in Bacteroidales species. Glycobiology. 2013;23:568–577. doi: 10.1093/glycob/cws172. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Reinhold BB, Hauer CR, Plummer TH, Reinhold VN. Detailed structural analysis of a novel, specific O-linked glycan from the prokaryote Flavobacterium meningosepticum. J Biol Chem. 1995;270:13197–13203. doi: 10.1074/jbc.270.22.13197. [DOI] [PubMed] [Google Scholar]
  30. Sako Y, Takai K, Ishida Y, Uchida A, Katayama Y. Rhodothermus obamensis sp. nov., a modern lineage of extremely thermophilic marine bacteria. Int J Syst Bacteriol. 1996;46:1099–1104. doi: 10.1099/00207713-46-4-1099. [DOI] [PubMed] [Google Scholar]
  31. Shevchenko A, Wilm M, Vorm O, Jensen ON, Podtelejnikov AV, Neubauer G, Mortensen P, Mann M. A strategy for identifying gel-separated proteins in sequence databases by MS alone. Biochemical Society transactions. 1996;24:893–896. doi: 10.1042/bst0240893. [DOI] [PubMed] [Google Scholar]
  32. Sleytr UB, Thorne KJ. Chemical characterization of the regularly arranged surface layers of Clostridium thermosaccharolyticum and Clostridium thermohydrosulfuricum. J Bacteriol. 1976;126:377–383. doi: 10.1128/jb.126.1.377-383.1976. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Sonnhammer EL, von Heijne G, Krogh A. A hidden Markov model for predicting transmembrane helices in protein sequences. Proc Int Conf Intell Syst Mol Biol. 1998;6:175–182. [PubMed] [Google Scholar]
  34. Stevens AM, Shoemaker NB, Salyers AA. The region of a Bacteroides conjugal chromosomal tetracycline resistance element which is responsible for production of plasmidlike forms from unlinked chromosomal DNA might also be involved in transfer of the element. J Bacteriol. 1990;172:4271–4279. doi: 10.1128/jb.172.8.4271-4279.1990. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Szymanski CM, Yao R, Ewing CP, Trust TJ, Guerry P. Evidence for a system of general protein glycosylation in Campylobacter jejuni. Mol Microbiol. 1999;32:1022–1030. doi: 10.1046/j.1365-2958.1999.01415.x. [DOI] [PubMed] [Google Scholar]
  36. Tatusov RL, Fedorova ND, Jackson JD, Jacobs AR, Kiryutin B, Koonin EV, Krylov DM, Mazumder R, Mekhedov SL, Nikolskaya AN, Rao BS, Smirnov S, Sverdlov AV, Vasudevan S, Wolf YI, Yin JJ, Natale DA. The COG database: an updated version includes eukaryotes. BMC Bioinformatics. 2003;4:41. doi: 10.1186/1471-2105-4-41. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Vik Å, Aas FE, Anonsen JH, Bilsborough S, Schneider A, Egge-Jacobsen W, Koomey M. Broad spectrum O-linked protein glycosylation in the human pathogen Neisseria gonorrhoeae. Proc Natl Acad Sci U S A. 2009;106:4447–4452. doi: 10.1073/pnas.0809504106. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Wehmeier S, Varghese AS, Gurcha SS, Tissot B, Panico M, Hitchen P, Morris HR, Besra GS, Dell A, Smith MC. Glycosylation of the phosphate binding protein, PstS, in Streptomyces coelicolor by a pathway that resembles protein O-mannosylation in eukaryotes. Mol Microbiol. 2009;71:421–433. doi: 10.1111/j.1365-2958.2008.06536.x. [DOI] [PubMed] [Google Scholar]
  39. Zitomersky NL, Coyne MJ, Comstock LE. Longitudinal analysis of the prevalence, maintenance, and IgA response to species of the order Bacteroidales in the human gut. Infect Immun. 2011;79:2012–2020. doi: 10.1128/IAI.01348-10. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supp Fig S1-S4 & Supp Table S1-S4
NIHMS463979-supplement-Supp_Fig_S1-S4___Supp_Table_S1-S4.pdf (763.8KB, pdf)

RESOURCES