Search tips
Search criteria 


Logo of glycobLink to Publisher's site
Glycobiology. 2011 October; 21(10): 1266–1276.
Published online 2011 February 14. doi:  10.1093/glycob/cwr018
PMCID: PMC3167475

Structural characterization and MHCII-dependent immunological properties of the zwitterionic O-chain antigen of Morganella morganii


Morganella morganii is a commensal Gram-negative bacterium that has long been known to produce an antigen bearing phosphocholine groups. We determined the structure of this O-chain antigen and found that its repeating unit also contains a free amino group and a second phosphate: An external file that holds a picture, illustration, etc.
Object name is cwr01808.jpg This alternating charge character places the M. morganii O-chain polysaccharide into a small family of zwitterionic polysaccharides (ZPSs) known to induce T-cell-dependent immune responses via presentation by class II major histocompatibility complex (MHCII) molecules. In vitro binding assays demonstrate that this O-chain interacts with MHCII in a manner that competes with binding of the prototypical ZPS antigen PSA from Bacteroides fragilis, despite its lack of a helical structure. Cellular studies also showed that the M. morganii polysaccharide induces activation of CD4+ T-cells. Antibody binding experiments using acid hydrolyzed fragments representing the monomer and higher oligomers of the repeating unit showed that the phosphocholine group was the dominant element of the epitope with an overall affinity (KD) of about 5 × 10−5 M, a typical value for an IgM anti-carbohydrate antibody but much lower than the affinity for phosphocholine itself. These data show that the structure of the M. morganii polysaccharide contains a unique zwitterionic repeating unit which allows for immune recognition by T-cells, making it the first identified T-cell-dependent O-chain antigen.

Keywords: MHC interaction, Morganella morganii, phosphocholine antigen, polysaccharide structure, zwitterionic polysaccharide


Charged polysaccharide antigens from bacteria often have immunological properties that set them apart from their uncharged counterparts. For example, they can induce antibodies with higher-than-usual affinities for their oligosaccharide epitopes (Müller-Loennies et al. 2000) and these may occur in the form of helices (Brisson et al. 1997). More significantly, when positively and negatively charged groups are present, forming a zwitterionic polysaccharide (ZPS), the antigen may interact directly with the class II major histocompatibility complex (MHCII) proteins. This behavior has been extensively investigated with the prototypic polysaccharide, PSA from Bacteroides fragilis (Cobb et al. 2004; Cobb and Kasper 2008), and there is evidence for similar behavior by other ZPSs such as the teichoic acid from Streptococcus pneumoniae, termed C-substance or PnC (Tzianabos et al. 1993), and the S. pneumoniae type 1 capsular polysaccharide (Tzianabos et al. 1993; Velez et al. 2009). The structure of PnC includes phosphocholine groups, which are themselves zwitterionic, as well as an amino group and a second phosphate (Kulakowska et al. 1993).

The reactivity of several mouse myeloma proteins with PnC led to the finding that they were recognizing its phosphocholine moiety (Leon and Young 1971), and crystallography of the Fab fragment of one such myeloma protein, M603, with phosphocholine gave the first structure of an antibody with a bound hapten (Satow et al. 1986). The phosphocholine hapten is small when compared with the overall dimensions of the binding site, and hence additional interactions could occur between the M603 binding site and the antigen that bears the phosphocholine epitope. The identity of the original immunogen cannot be established with certainty, but Potter (1971) found that several organisms in the environment and flora of laboratory-raised Balb/c mice carried antigens that contained phosphocholine. These included Aspergillus species, the parasite Ascaris suum and a Gram-negative bacterium from the normal mouse flora, Morganella (Proteus) morganii. The M. morganii organism is an opportunistic pathogen in humans causing bladder infections and bacteraemia (for hospital surveys, see Kim et al. 2003; Falagas et al. 2006), and several O-antigen serotypes have been recognized by studies of organisms recovered from patients (Penner and Hennessy 1979). Immunization of mice with M. morganii cells led to hybridoma antibodies predominantly arising from the same VH and germline genes as M603 (Claflin et al. 1985), principally varying from it in the second complementarity-determining region of the H-chain (Claflin et al. 1987). However, M603 binds the M. morganii antigen less strongly (Claflin et al. 1985). The restriction to the M603 family was in contrast to immunization experiments with a rough strain of S. pneumoniae, R36A, rich in PnC, which gave hybridomas related to two other phosphocholine myeloma families T15 and M167/M511 as well as M603 (Claflin and Berry 1988); the three families use different germline genes.

The antibody properties found against the M. morganii phosphocholine-containing antigen prompted us to perform a structural analysis of the O-chain polysaccharide. We found not only phosphocholine, but also an amine and a second phosphate, extending its zwitterionic nature and raising the possibility for MHCII presentation and subsequent T-cell recognition as seen with other ZPS molecules. Here, we report its structure, demonstrate MHCII binding and T-cell activation and characterize the binding of repeating unit fragments by hybridoma antibodies. These findings reveal the first known example of an O-chain polysaccharide with the capacity to activate CD4+ T-cells via MHCII presentation, potentially identifying another commensal organism with the ability to promote immune system homeostasis (Mazmanian et al. 2005; Ochoa-Reparaz et al. 2010).


Determination of the polysaccharide structure

The M. morganii strain used in the above studies was serotyped by Dr. J. Penner and it belonged to the most common serotype, O:lab (Penner and Hennessy 1979). Exclusion experiments and tests for quelling reaction with the monoclonal antibody gave no evidence for capsular polysaccharide around organisms grown in liquid culture or on plates. Extraction with phenol in the standard manner for lipopolysaccharides (LPSs) gave only poor yields of antigens compared with extraction with sodium dodecyl sulfate (SDS)-citrate. When the latter extract was ultracentrifuged, both the supernatant and precipitate contained antigens. Fractionation of the supernatant on Sephadex G100 yielded pure polysaccharide antigen, as well as a lower molecular fraction that was predominantly enterobacterial common antigen. This was a linear form with a well-resolved nuclear magnetic resonance (NMR) spectrum (data not shown), which was fully assignable to the reported repeating unit structure of this polysaccharide, in contrast to the circular or lipid-attached forms previously reported (Dell et al. 1984).

NMR experiments on the LPS precipitate dissolved in deutero-SDS/ethylene diamine tetraacetic acid (EDTA) (Risberg et al. 1999) showed that it contained significantly more antigens, and mild acid hydrolysis of the LPS precipitate released antigenically active O-chain, which was separated by gel filtration from a core fraction and free Kdo. The total yield of antigens was approximately 3 mg/g of cells (wet weight). LPS was also isolated from a serotype O:1 reference strain, ATCC 499993, and NMR spectra obtained in deutero-SDS/EDTA were consistent with it having the same polysaccharide structure as the Potter strain (data not shown).

Complete acid hydrolysis of the O-chain fraction afforded d-galactose and d-galactosamine (at an approximate molar ratio of 1:2) together with low amounts of glycerol, as determined by gas–liquid chromatography (GLC) of the derived alditol acetate derivatives. The absolute stereochemistry of the glycoses was determined from the GLC retention times of the trimethylsilylated R-2-butyl glycoside derivatives. Colorimetric analysis for phosphate groups gave 20% (w/w) as PO4, consistent with the two groups found by NMR (see below).

The 13C NMR spectrum of the O-chain indicated the polymer to be composed of regular repeating units that contained three glycosyl residues (Figure 1). A strong signal corresponding to the choline methyl groups was evident at 55.1 ppm and substitution by O-acetyl groups was indicated by signals at 21.5 ppm (CH3) and 174.6 ppm (CAn external file that holds a picture, illustration, etc.
Object name is cwr018ileq1.jpgO). These latter signals were absent in the 13C NMR spectrum of the O-deacetylated O-chain (Figure 1B), which also showed that only one of the galactosamine residues was N-acetylated.

Fig. 1.
13C NMR spectrum (125 MHz) of the native (A) and O-deacetylated (B) M. morganii O-chain.

The 1H and 13C NMR spectra of the native and O-deacetylated polymers were fully assigned by 2D homo- and heteronuclear chemical shift correlation techniques (Figure 2A and B; Tables I and andII).II). From the chemical shifts and 1H–1H vicinal couplings, the subspectra corresponding to α- and β-linked d-GalpN residues and a β-d-Galp residue were identified. N-Acetylation of the polymer led to a substantial upfield shift of the α-d-GalpN H-1 resonance (ca. 0.4 ppm) which indicated this residue to be present as the free amino sugar in the native O–chain. Resonances for glycerol and choline spin systems were also identified in the 2D spectra. A comparison of the NMR data from the O-deacetylated and the native O-chains indicated that the 6-position of the β-d-GalpNAc was substituted by O-acetyl groups to the extent of 80%.

Table I.
1H NMR chemical shifts (ppm) and coupling constants (Hz) for O-deacylated M. morganii O-chaina
Table II.
13C NMR chemical shifts (ppm) for O-deacylated M. morganii O-chaina
Fig. 2.
2D NMR experiments. (A) Part of the heteronuclear 2D 1H–13C chemical shift correlation map of the O-deacetylated O-chain. (B) Contour plot of the 2D COSY of the O-deacetylated O-chain. Ring proton resonances (H-1 to H-4) of the phosphate-bearing ...

The occurrence of transglycosidic NOEs between H-1 of α-d-GalpN and H-3 of β-d-Galp, between H-1 of β-d-Galp and H-3 of β-d-GalpNAc and between H-1 of β-d-GalpNAc and the methylene protons of glycerol (H-3) established the arrangement of the residues within the O-chain repeating unit: An external file that holds a picture, illustration, etc.
Object name is cwr01809.jpg Methylation analysis confirmed the positions of the linkages.

The H-2 and H-4 resonances of the β-d-Galp residue showed additional coupling and were down-field shifted compared with the corresponding resonances of the methyl glycoside; this indicated phosphate substitution at the corresponding positions. The 31P NMR spectrum showed two resonances at −2.0 and 1.5 ppm. The high-field resonance was assigned to 31P of the phosphocholine since a correlation to the adjacent methylene protons was observed in the 1H–31P correlation experiment (Figure 2C). In addition, this resonance showed a strong correlation to H-2 of β-d-Galp which established the site of substitution at the C-2 position. The low-field 31P signal showed connectivities to C-4 of β–d-Galp and the methylene protons (H-1) of the glycerol moiety, indicating the O-chain to have the teichoic acid-like structure shown in Figure 3. This structure is also consistent with the ES-MS spectrum of the intact polysaccharide (Figure 4), which showed prominent ions from the acetylated and free repeating units, together with characteristic fragment ions for the components.

Fig. 3.
Structure of the M. morganii LPS O-chain (R = Ac, 80% or H, 20%) and comparison with the structures of two other ZPS species, S. pneumoniae PnC and B. fragilis PSA.
Fig. 4.
ES-MS spectrum of the intact polysaccharide. The annotated series of ions assigned to fragmentation of the repeating unit is paralleled by a second set with masses 18 Da lower, arising from initial loss of a water molecule. The major peaks to ...

Treatment of the N-acetylated O-antigen with 48% HF afforded the expected glycerol trisaccharide glycoside (1), together with significant amounts of the free trisaccharide. The acid lability of the glycosidic linkage of 3-linked β-d-GalpNAc has been previously observed (Jennings and Lugowski 1980). Heating of the O-antigen with 1% trifluoroacetic acid (TFA) at 80°C for 1 h led to cleavage at the same position, liberating the complete zwitterionic repeating unit.

Circular dichroism spectroscopy of the ZPS

NMR studies have shown that the ZPSs from B. fragilis (Wang et al. 2000) and S. pneumoniae (Choi et al. 2002), PSA and Sp1, respectively, are predominantly helical and that the immunologic properties of PSA depends on this helical content as measured by circular dichroism (CD; Kreisman et al. 2007). The far-UV CD spectra of the M. morganii ZPS and the monomer of its repeating unit were therefore determined. Molar ellipticities were calculated on the basis of the mass of the repeating unit, and these spectra showed only weak features in the 200–230 region (Figure 5). The ellipticities at 212 nm were −2.4 and −2.0 × 104 deg cm2 dmol−1, respectively, which are two orders of magnitude less than the corresponding value for the B. fragilis PSA, calculated from the data of Kreisman et al. (2007) to be −1.9 × 106 deg cm2 dmol−1 at 225 nm. These data suggest that the M. morganii ZPS is not significantly helical in nature, although a full NMR-based conformational analysis is needed to confirm these findings.

Fig. 5.
CD spectra of M. morganii ZPS (solid line) and its monomer fragment (dashed line). The spectra were calculated as molar ellipticities per monomer unit.

Interaction with a monoclonal antibody

In order to begin assessment of the immunologic properties of the M. morganii ZPS, the O-chain was hydrolyzed with mild acid to fragment the molecule into a mixture of the monomer and higher oligomers of its repeating unit, which were purified by gel filtration. Their binding affinities for an anti-M. morganii hybridoma, 180.7c9, were measured in an enzyme immunoassay inhibition assay, with antigen partially biotinylated on the amino groups of its d-galactosamine residues (Meikle et al. 1990). The oligomer fragments were better inhibitors than glycerophosphocholine (Figure 6 and Table III), but were weaker than phosphocholine.

Table III.
Inhibition of antigen binding by the hybridoma 180.7c9 in ELISA, by phosphocholine haptens and oligomers of the repeating unit of the ZPS
Fig. 6.
Inhibition enzyme immunoassays of the monoclonal antibody 180.7c9 and biotinylated M. morganii antigen. The antigen fragments and other haptens tested were, from the left, phosphocholine (♦), tetramer of the repeating unit (•), trimer ...

Interaction with MHCII

Purified recombinant human MHCII (HLA-DR1) was used in binding assays as described by Cobb and Kasper (2008) with 5 kDa M. morganii ZPSs biotinylated on the amino groups of its d-galactosamine residues (Meikle et al. 1990) to approximately 1 mol of biotin per mole of polysaccharide to determine whether the nature of ZPSs would enable MHCII binding. Binding between M. morganii ZPSs and MHCII was saturable with an apparent KD value of 3.66 ± 0.63 µM (Figure 7A), which is similar to the KD between HLA-DR1 and B. fragilis PSA (1.9 ± 0.4 µM; Cobb and Kasper 2008). Experiments using unlabeled PSA as a binding competitor with the biotinylated M. morganii ZPS showed that these two molecules appear to compete for the same binding site on MHCII (Figure 7B). In contrast, binding experiments with the M. morganii monomer or dimer forms as competitors failed to reduce the 5 kDa M. morganii ZPS binding, reflecting a length requirement for MHCII interactions beyond the basic repeating unit structure, as seen with PSA (Kreisman et al. 2007). However, both acetylated and control (native polysaccharide) effectively competed for binding with the labeled 5 kDa molecule (Figure 7C). Since the acetylated molecule has lost the positive charge on the primary amine, these data suggest that the zwitterionic nature of the phosphocholine group in itself is sufficient to retain the MHCII binding properties.

Fig. 7.
Analyses of M. morganii ZPS interactions with MHCII and downstream T-cell activation. (A) Binding curve of M. morganii ZPS with HLA-DR1, showing saturable interactions with a KD value of 3.66 ± 0.63 µM. The dotted ...

T-cell activation assays

Given the observed ability to bind with MHCII protein in vitro (Figure 7A–C), we sought to determine whether the M. morganii O-antigen could also mediate activation of CD4+ T-cells. Human peripheral blood mononuclear cells (PBMCs) from a hepatitis B peptide antigen (HepB)-immunized individual were used to isolate CD4+ T-cells using positive selection on magnetic beads as well as T-cell-depleted antigen-presenting cells (APCs) using CD3 negative selection. For activation, CD4+ T-cells and APCs were co-cultured at a 1:1 ratio with M. morganii ZPSs or B. fragilis PSA at 50 µg/mL. Controls included media alone (no antigen-negative control), superantigen [Staphylococcus enterotoxin B (SEB); positive control] or hepatitis B peptide (conventional antigen-positive control). On days 5, 6 and 7, culture supernatants were collected and analyzed for interferon-γ (IFNγ) by enzyme-linked immunosorbent assay (ELISA) as a measure of T-cell activation. The M. morganii ZPS activated the CD4+ T-cells to a similar extent as the B. fragilis PSA (3084 pg/mL INFγ vs. 5807 pg/mL IFNγ on day 7, respectively; Figure 7D). The stimulation index was calculated to illustrate fold IFNγ production over the negative media control (set to unity) and demonstrates a stimulation of 17-fold for the M. morganii ZPS antigen on day 7 (Figure 7E). To confirm that T-cell activation was due to presentation via MHCII, activation was compared with and without blocking antibody specific for HLA-DR (clone L243). We found that INFγ production was abrogated in the presence of anti-MHCII antibody (Figure 7F). Collectively, these data show for the first time that a bacterial O-antigen can activate CD4+ T-cells via MHCII presentation.


Since the discovery that ZPSs stimulate the T-cell-dependent immune system directly (Tzianabos et al. 1993), only a small number of them have had their interactions with the MHC determined by direct assay (Cobb and Kasper 2008; Velez et al. 2009). To date, all of the identified ZPS molecules are capsular polysaccharides and the most characterized example is PSA from B. fragilis, commensal Gram-negative organism. The M. morganii antigen is the first example of an O-chain polysaccharide that is zwitterionic in nature and shows similar immunological properties to those of PSA and other ZPS molecules. The structure we have determined was confirmed as the O1 form by examining the O-chain from a second strain, ATCC 49993, which has also been serotyped as O1. In contrast, it does not resemble in any way the M. morganii structure determined by Kilcoyne et al. (2002), whose serotype was not reported.

Investigation of the structure of the M. morganii O-chain not only confirmed the presence of phosphocholine, but also disclosed the presence of another pair of ions in the repeating unit, an amino group and a phosphate diester, adding to the zwitterionic character of the polysaccharide. The backbone phosphate group and the amino group on the α-d-GalpN are brought together by the conformation of the molecule, close enough for them to form a salt bridge, a most unusual feature in a polysaccharide.  The interaction of the polysaccharide with MHCII molecules and the resulting T-cell activation were very similar on a weight basis to those of the prototypic ZPS PSA. However, in the M. morganii case, the CD spectra were not intense enough to suggest a helical structure, a feature of PSA that has been linked to its ability to interact with MHCII molecules (Kreisman et al. 2007).

The M. morganii ZPS has the presence of phosphocholine, amino and phosphate groups in common with the S. pneumoniae antigen, PnC, which is also zwitterionic and activates T-cells (Tzianabos et al. 1993); however, the structures of these two molecules are quite distinct. For example, the antigenically important phosphocholine components are attached to the O2 of the d-Gal residue in the M. morganii polysaccharide and to the O6 of both d-GalNAc residues in PnC (Kulakowska et al. 1993). Thus, the phosphocholine is closer to the polysaccharide backbone in the M. morganii ZPS than in PnC, and alongside another substituent, the GalN moiety at O3. This more confined positioning may be relevant to the restriction of the immune response to this antigen to antibodies of the M603 family, unlike the PnC response. Phosphocholine occurs in several other pneumococcal polysaccharides, and our data with the N-acetylated ZPS suggest that this zwitterionic group might be sufficient for reactivity with MHCII molecules and the downstream T-cell activation; however, further studies with other carbohydrates will be needed to explore this possibility.

PnC is a teichoic acid of the ribitol phosphate type, and the M. morganii ZPS would be classified as a teichoic acid of the glycerol phosphate type, if it were not from a Gram-negative organism. The evidence at present indicates that the ZPS is an LPS O-chain, although O-chains with charged groups are uncommon. This evidence includes the (i) NMR spectrum of the ultracentrifuge precipitate, (ii) the “ladder” of bands found by silver-staining of deoxycholate polyacrylamide gel electrophoresis (PAGE) gels of it and by immunoblotting (data not shown), and (iii) the lack of evidence for a capsule. The overall charge on the polysaccharide is zero, since the two phosphate negative charges are counterbalanced by the positive charges on the d-GalN and choline constituents. Similarly, PnC is also neutrally charged, with an amino group on the trideoxy sugar offset by a phosphate group in the backbone, plus two phosphocholine groups.

The M. morganii ZPS is acid-labile, cleaving facilely at the β-d-GalNAc glycosidic bond to glycerol, rather than at the backbone phosphate as one might expect. This lability may explain two properties of the polysaccharide: the poor yield of the antigen obtained by phenol extraction, which may be due to the acidity of the extraction conditions, and the appearance of the antigen in both the supernatant and precipitate fractions from the ultracentrifugation. The facile cleavage enabled us to prepare oligosaccharides representing the repeating unit and oligomeric forms of it. Because the phosphate is not the site of cleavage, the charge arrangement within the repeating unit in the polymeric antigen is retained in the monomer. The monomeric fragment has the phosphocholine unit centrally placed, thus the sugar units would fill the center of the binding site around the phosphocholine pocket.

The inhibition experiments with antigen fragments show that the phosphocholine group is very much the immunodominant feature in antigen recognition by the hybridoma antibody 180.7c9. The antigen fragments had activities that were intermediate between those of the two simpler ligands, phosphocholine and glycerophosphocholine (Table III). The concentration for 50% inhibition approached 5 × 10−5 M, which approximates the dissociation constant in the type of enzyme immunoassay system used here. This affinity for antigens is within the range normally found for anti-carbohydrate IgMs, despite the high affinity for phosphocholine itself, and the antigen's charged groups do not lead to stronger binding in the manner of Kdo-containing antigens (Müller-Loennies et al. 2000). It appears that whatever new hydrogen bonds or other interactions are formed in the binding of the oligomers, increasing their affinities up to 7-fold compared with glycerophosphocholine, their effects are largely negated by other unfavourable interactions. The affinities of the oligomers are approximately proportional to their phosphocholine contents, suggesting that there are no extended interactions beyond one repeating unit. However, the restriction of the anti-M. morganii hybridomas to the M613 family shows that the structure that presents the phosphocholine moiety does have an important role, despite its limited effect on the binding constant, and there is an interdependence of this dominant epitopic feature and the presenting polysaccharide. It is interesting that the sequence differences in the CDR H2 regions of the M. morganii antibodies compared with M603 (Claflin et al. 1989) lead to a very different distribution of positive and negative sidechains around this loop, i.e., –53KNTHD- in 180.7c9 compared to –53NKGNK- in M603. This may allow more charge interactions with the phosphate/GalN region of the antigen compared with M603, which binds only weakly to the M. morganii antigen (Claflin et al. 1985).

M. morganii is a component of the normal human flora that can cause nosocomial infections, usually following surgery, much like Bacteroides. Remarkably, B. fragilis, through PSA and other polysaccharide components, has been postulated to play a role in the development of the mucosal immune system (Mazmanian et al. 2005). Recent studies have also shown that T-cells activated by ZPS molecules result in anti-inflammatory regulatory T-cells that can prevent the onset of inflammatory bowel disease (Mazmanian et al. 2008) and even central nervous system demyelinating disease (Ochoa-Reparaz et al. 2010) in animal models, suggesting that commensal bacteria that carry ZPS molecules can directly contribute to immune homeostasis and protect against inappropriate inflammation through T-cell activation. The occurrence of a T-cell-activating ZPS in the LPS of M. morganii, itself a commensal bacterium, not only suggests it could also contribute to the overall balance of the immune system, but also highlights the importance of including the O1 serotype M. morganii in the sequencing efforts to understand the human microbiome community.

Materials and methods

Growth of the organism

The M. morganii strain originally isolated by Potter (1971) and a monoclonal antibody to the phosphocholine antigen, 180.7C9.1, were kindly provided by Dr. L. Claflin. This strain is no longer available, but NMR experiments confirmed that the serotype O1 reference strain ATCC 49993 bears the same polysaccharide antigen. The organism was grown in 20 L batches in Bacto brain-heart infusion for 20 h at 37°C. The cells were harvested by centrifugation and extracted with 0.05% SDS, 0.1 M sodium citrate for 1 h at 56°C (Williams and Claflin 1980). The extract was centrifuged and the supernatant was extensively dialysed and then freeze-dried.

Purification of the polysaccharide

The extract was re-suspended in phosphate-buffered saline (PBS) and treated with nuclease and ribonuclease, followed by proteinase K. After dialysis and freeze-drying, it was re-suspended in PBS and ultracentrifuged at 100,000 × g for 16 h. The supernatant was fractionated by gel filtration on a 2.5 × 90 cm column of Sephadex G100 (GE Biosciences, Inc., Baie D'Urfé, Quebec, Canada) in 0.1 M ammonium acetate. Fractions were tested for the antigen by immunodiffusion in gel (Ouchterlony procedure) with the above monoclonal antibody. The LPS precipitate was re-suspended in 2% acetic acid and hydrolyzed at 100°C for 1.5 h. Insoluble material was removed by centrifugation and the supernatant was freeze-dried. The O-chain was separated from LPS core material and free Kdo by gel filtration on Sephadex G50, in pyridine-acetic acid buffer, 0.05 M, pH 4.7. The fractions were assayed for hexose/heptose (Dubois et al. 1956), hexosamine (Gatt and Berman 1965) and Kdo (Aminoff 1965).

Composition analyses

Hexose composition was determined by GLC separation of alditol acetate derivatives of the products of hydrolysis with 3 N HCl at 100°C for 3 h (Gunner et al. 1961). The absolute stereochemistry was determined by GLC of trimethylsilylated R-2-butyl derivatives (Gerwig et al. 1978). Phosphate content was measured colorimetrically (Chen et al. 1956). O-Acetyl groups were estimated from the NMR spectra and were removed by treatment with 5% NH3 at 37°C for 3 h, and free amino groups were acetylated by treatment with acetic anhydride at pH 10. Methylation analysis was done by the method of Hakomori (1964).

Oligosaccharide preparations

The core trisaccharide was obtained in dephosphorylated form by treatment with 48% aqueous HF. Hydrolysis with 1% trifluoroacetic acid led to production of the intact repeating unit, while milder treatments with 0.2% trifluoroacetic acid at 100°C for 1 h produced the di-, tri- and tetra-oligomers of this structure. The fragments were separated by gel filtration on a column (1 × 57 cm) of BioGel P4, 100–200 mesh (Bio-Rad, Mississauga, Ont., Canada) run in water. The sizes of the fragments and their structures were checked by mass spectrometry and proton NMR.

NMR spectroscopy

NMR spectra were obtained with a Bruker AMX 500 spectrometer using standard Bruker software. All measurements were made on solution at 320 K in 0.5 mL of D2O, subsequent to several lyophilizations with D2O. Proton spectra were obtained by using a spectral width of 10.6 KHz and a 90° pulse. Acetone was used as the internal standard, and chemical shifts were referenced to the methyl resonance (δH, 2.225 ppm; δC, 31.07 ppm). 31P NMR spectra were measured at 202 MHz using spectral width of 12.1 and 41.6 KHz and phosphoric acid (85%) was used as the external standard (δp, 0.0 ppm). Correlation spectroscopy (COSY) and nuclear overhauser effect spectroscopy (NOESY) experiments were performed as described earlier (Masoud et al. 1994); a mixing time of 400 ms was used for NOESY. Heteronuclear 2D 13C–1H and 31P–1H heteronuclear multiple quantum correlation (HMQC) correlation experiments were performed as previously described (Masoud et al. 1994). LPS samples were solubilized for NMR by adding perdeutero-EDTA (2 mM) and perdeutero-SDS (10 mg/mL) to the D2O solutions (Risberg et al. 1999). These deuterated compounds were obtained from Cambridge Isotope Laboratories, Inc. (Andover, MA).

Mass spectrometry

CE-MS was performed using a Prince CE system (Prince Technologies, Emmen, the Netherlands) coupled to a 4000 QTrap mass spectrometer (Applied Biosystems/Sciex, Concord, Ont., Canada). CE separation was obtained on a 90 cm length of bare fused silica capillary (365 μm OD × 50 μm ID) with CE-MS coupling using a liquid sheath-flow interface and isopropanol:methanol (2:1) delivered at a flow rate of 2 μL/min. An aqueous buffer consisting of 10 mM ammonium acetate, pH 7.0, was used for all experiments. For analysis of intact polysaccharide, in-source collision-induced dissociation was used to afford fragments representing repeating units of the O-chain in the intact polymer. This was accomplished by increasing the declustering potential in the skimmer region of the electrospray ionization source to 400 eV, thus applying extra internal energy to ionic species entering the mass spectrometer and generating fragments where signals corresponding to multiply-charged intact molecules would otherwise not be observed (Li et al. 2005).

MHCII binding assays

Purified recombinant human MHCII (HLA-DR1 allele) was used in binding assays as previously described for the B. fragilis antigen PSA (Cobb and Kasper 2008). M. morganii ZPS was reacted with aminohexyl-biotin-N-hydroxysuccinimide ester (Invitrogen Canada, Inc., Burlington, Ont., Canada) according to the manufacturer's directions, to attach biotin to the free amino groups of its d-galactosamine residues (Meikle et al. 1990). Its biotin content, determined with a FluoReporter Biotin Quantitation Kit (Invitrogen Canada, Inc.), was ~0.7 mol of biotin per mole of ZPSs, assuming that its molecular weight was 5 kDa. Binding of the labeled antigen to MHCII was compared with that of the B. fragilis antigen using detection by europium-conjugated streptavidin and time-resolved fluorescence (Cobb and Kasper 2008). Molar quantities and KD values were calculated based on an average molecular weight of 5 kDa.

T-cell activation assays

Human PBMCs were isolated on a Lymphoprep cushion as previously described. CD4+ T-cells were first isolated from PBMCs by positive selection using anti-CD4 magnetic beads according to the manufacturer's protocol (Miltenyi Biotec, Inc., Auburn, CA). The remaining CD4 cells were then depleted using anti-CD3 magnetic beads according to the manufacturer's protocol (Miltenyi Biotec, Inc.) to yield a T-cell-negative population of cells that were used as APCs. APCs were then treated with Mitomycin C to prevent proliferation. For activation, APCs and CD4+ T-cells were co-cultured at a 1:1 ratio (1 × 105 APCs and 1 × 105 T-cells) in 200 μL of RPMI1640 media supplemented with 10% fetal bovine serum. Antigens were added at the following concentrations: SEB (superantigen control) = 50 ng/mL, PSA (ZPS control) = 50 μg/mL, HepB (conventional protein antigen control; volunteer was previously hepatitis B virus vaccinated) = 5 μg/mL and M. morganii ZPS = 50 μg/mL. “Media” assays were performed without added antigens as a negative control. For antibody-blocking experiments, 25 µg of purified anti-HLA-DR monoclonal antibody (clone L243, purified from HB-55 hybridoma cells; ATCC, Manassas, VA) was added to each well at the same time as the antigens. On the indicated days, culture supernatants were collected and analyzed for IFNγ by ELISA as a measure of T-cell activation.

Enzyme immunoassays

Optimum amounts of monoclonal antibody 180.7c9 (provided by Dr. L. Claflin) for coating the plates, biotinylated antigens, prepared as above, and streptavidin-horseradish peroxidase conjugate were established in preliminary experiments, to obtain an approximate absorbance at 414 nm of 1.0 after 30 min development with the substrate 2,2-azino-bis(3-ethylbenz-thiazoline-6-sulfonic acid), di-ammonium salt (Sigma-Aldrich Canada Ltd., Oakville, Ont., Canada). Antibody solutions in PBS, 0.01 M sodium phosphate, 0.15 M NaCl, pH 7.2, were coated for 3 h, the plates were washed with 0.2% Tween 20 in PBS, blocked for 1 h with 0.1% skim milk in PBS-Tween and washed again with PBS-Tween. Biotinylated antigen in PBS-Tween milk and serial dilutions of the test inhibitors in the range 10−3 to 10−7 M were added simultaneously and the plates were incubated at room temperature overnight. After washing as before, the streptavidin conjugate and substrate were added. Colour development was measured with a Titertek Multiscan MC plate reader (Flow Laboratories, Mississauga, Ont., Canada).

CD spectroscopy

Spectra were obtained with a Jasco J-815 spectrometer (Jasco, Inc., Easton, MD), for 2.5 mg/mL solutions of the intact antigen and the monomer fragment, in PBS. The pathlength of the cell was 0.1 cm. The spectra were calculated as molar ellipticities per monomer unit.


This work was supported by USA National Institutes of Health grants OD004225 and GM082916 to B.A.C.

Conflict of interest statement

None declared.


We thank Dr. Latham Claflin (University of Michigan), Dr. John Penner (University of Toronto), Dr. Jianbing Zhang and Dr. Malcolm Perry for their contributions to this work, and Douglas Griffith, Rosemary Johnston, Perry Fleming, Sonia Leclerc and Anna Cunningham for excellent technical assistance.


APC, antigen-presenting cell; CD, circular dichroism; CE-MS, capillary electrophoresis - mass spectrometry; COSY, correlation spectroscopy; EDTA, ethylene diamine tetraacetic acid; ELISA, enzyme-linked immunosorbent assay; ES-MS, electrospray mass spectrometry; HepB, hepatitis B peptide antigen; HMQC, heteronuclear multiple quantum correlation; IFNγ, interferon-γ; LPS, lipopolysaccharide; MHCII, class II major histocompatibility complex; NOESY, nuclear Overhauser effect spectroscopy; NMR, nuclear magnetic resonance; PAGE, polyacrylamide gel electrophoresis; PBMCs, peripheral blood mononuclear cells; PBS, phosphate-buffered saline; PnC, Streptococcus pneumoniae C-substance; PSA, Bacteroides fragilis polysaccharide A; SDS, sodium dodecyl sulfate; SEB, staphylococcus enterotoxin B; Sp1, type I Streptococcus pneumoniae capsular polysaccharide; TFA, trifluoroacetic acid; ZPS, zwitterionic polysaccharide.


  • Aminoff D. Methods for the quantitative estimation of N-acetylneuraminic acid and their application to hydrolysates of sialomucoids. Biochem J. 1965;81:384–392. [PubMed]
  • Brisson J-R, Uhrinova S, Woods RJ, van der Zwan M, Jarrell HC, Paoletti LC, Kaspar DL, Jennings HJ. NMR and molecular dynamics studies of the conformational epitope of the type III group B Streptococcus capsular polysaccharide and derivatives. Biochemistry. 1997;18:3278–3292. doi:10.1021/bi961819l. [PubMed]
  • Chen PS, Toribara TY, Warner H. Microdetermination of phosphorus. Anal Chem. 1956;28:1756–1758. doi:10.1021/ac60119a033.
  • Choi YH, Roehrl MH, Kasper DL, Wang JY. A unique structural pattern shared by T-cell-activating and abscess-regulating zwitterionic polysaccharides. Biochemistry. 2002;41:15144–15151. doi:10.1021/bi020491v. [PubMed]
  • Claflin JL, Berry J. Genetics of the phosphocholine-specific antibody response to Streptococcus pneumoniae. J Immunol. 1988;141:4012–4019. [PubMed]
  • Claflin JL, Berry J, Flaherty D, Dunnick W. Somatic evolution of diversity among anti-phosphocholine antibodies induced with Proteus morganii. J Immunol. 1987;136:3060–3068. [PubMed]
  • Claflin JL, George J, Dell C, Berry J. Patterns of mutations and selection in antibodies to the phosphocholine-specific determinant in Proteus morganii. J Immunol. 1989;143:3054–3063. [PubMed]
  • Claflin JL, Hudak S, Maddelena A, Bender T. Antigen-specific anti-phosphocholine antibodies: Binding site studies. J Immunol. 1985;134:2536–2543. [PubMed]
  • Cobb BA, Kasper DL. Characteristics of carbohydrate antigen binding to the presentation protein HLA-DR. Glycobiology. 2008;18:707–718. doi:10.1093/glycob/cwn050. [PMC free article] [PubMed]
  • Cobb BA, Wang Q, Tzianabos AO, Kasper DL. Polysaccharide processing and presentation by the MHCII pathway. Cell. 2004;117:677–687. doi:10.1016/j.cell.2004.05.001. [PMC free article] [PubMed]
  • Dell A, Oates J, Lugowski C, Romanowska E, Kenne L, Lindberg B. The enterobacterial common antigen, a cyclic polysaccharide. Carbohyd Res. 1984;133:95–104. doi:10.1016/0008-6215(84)85186-1. [PubMed]
  • Dubois M, Gilles KA, Hamilton JK, Rebers PA, Smith F. Colorimetric method for determination of sugars and related substances. Anal Chem. 1956;28:350–356. doi:10.1021/ac60111a017.
  • Falagas ME, Kavvadia PK, Mantadakis E, Kofteridis DP, Bliziotis IA, Saloustros E, Maraki S, Samonis G. Morganella morganii infections in a general tertiary hospital. Infection. 2006;34:315–321. doi:10.1007/s15010-006-6682-3. [PubMed]
  • Gatt R, Berman ER. A rapid procedure for the estimation of aminosugars on a micro scale. Anal Biochem. 1965;15:167–171. doi:10.1016/0003-2697(66)90262-4. [PubMed]
  • Gerwig GJ, Kamerling JK, Vleigenthart JFG. Determination of the d and l configuration of neutral monosaccharides by high-resolution capillary GLC. Carbohyd Res. 1978;62:349–357. doi:10.1016/S0008-6215(00)80881-2.
  • Gunner SW, Jones JKN, Perry MB. The gas–liquid partition chromatography of carbohydrate derivatives. Part 1. The separation of glycitol and glycose acetates. Can J Chem. 1961;39:1892–1895. doi:10.1139/v61-254.
  • Hakomori S. Rapid permethylation of glycolipids and polysaccharides by methylsulfinyl cation in dimethylsulfoxide. J Biochem (Tokyo) 1964;55:205–208. [PubMed]
  • Jennings HJ, Lugowski C. Facile cleavage of some 2-acetamido-2-deoxy-β-d-gluco- and galactopyranosides using aqueous HF. Can J Chem. 1980;58:2610–2612. doi:10.1139/v80-417.
  • Kilcoyne M, Shashkov AS, Senchenkova SA, Knirel Y, Vinogradov EV, Radziejewska-Lebrecht J, Galimska-Stypa R, Savage AV. Structural investigation of the O-specific polysaccharides of Morganella morganii consisting of two higher sugars. Carbohyd Res. 2002;337:1697–1702. doi:10.1016/S0008-6215(02)00181-7. [PubMed]
  • Kim BN, Kim NJ, Kim NM, Kim YS, Woo JH, Ryu J. Bacteraemia due to tribe Proteeae: A review of 132 cases during a decade. Scand J Infect Dis. 2003;35:98–103. doi:10.1080/0036554021000027015. [PubMed]
  • Kreisman LSC, Friedman JH, Neaga A, Cobb BA. Structure and function relations with a T-cell-activating polysaccharide antigen using circular dichroism. Glycobiology. 2007;17:46–55. doi:10.1093/glycob/cwl056. [PMC free article] [PubMed]
  • Kulakowska M, Brisson J-R, Griffith DW, Young NM, Jennings HJ. High-resolution NMR spectroscopic analysis of the C-polysaccharide of Streptococcus pneumoniae. Can J Chem. 1993;71:644–648. doi:10.1139/v93-086.
  • Leon MA, Young NM. Specificity for phosphoryl­choline of six murine myeloma proteins reactive with Pneumococcus C polysaccharide and β lipoproteins. Biochemistry. 1971;10:1424–1429. doi:10.1021/bi00784a024. [PubMed]
  • Li J, Wang Z, Altman E. In-source fragmentation and analysis of polysaccharides by capillary electrophoresis/mass spectrometry. Rapid Comm Mass Spec. 2005;19:1305–1314. doi:10.1002/rcm.1927. [PubMed]
  • Masoud H, Perry MB, Brisson J-R, Uhrin D, Richards JC. Structural elucidation of the backbone oligosaccharide from the lipopolysaccharide of Moraxella catarrhalis serotype A. Can J Chem. 1994;72:1466–1477. doi:10.1139/v94-182.
  • Mazmanian SK, Liu CH, Tzianabos AO, Kasper DL. An immunomodulatory molecular of symbiotic bacteria directs maturation of the host immune system. Cell. 2005;122:107–118. doi:10.1016/j.cell.2005.05.007. [PubMed]
  • Mazmanian SK, Round JL, Kasper DL. A microbial symbiosis factor prevents intestinal inflammatory disease. Nature. 2008;453:620–625. doi:10.1038/nature07008. [PubMed]
  • Meikle PJ, Young NM, Bundle DR. O-antigen biotin conjugates: Preparation and use in direct competition enzyme immunoassays. J Immunol Meth. 1990;132:255–261. doi:10.1016/0022-1759(90)90037-V. [PubMed]
  • Müller-Loennies S, MacKenzie CR, Patenaude SI, Evans SV, Brade H, Brade L, Narang S. Characterization of high affinity monoclonal antibodies for chlamydial lipopolysaccharide. Glycobiology. 2000;10:121–130. doi:10.1093/glycob/10.2.121. [PubMed]
  • Ochoa-Reparaz J, Mielcarz DW, Wang Y, Begum-Haque S, Dasgupta S, Kasper DL, Kasper LH. A polysaccharide from the human commensal Bacteroides fragilis protects against CNS demyelinating disease. Mucosal Immunol. 2010;3:487–495. doi:10.1038/mi.2010.29. [PubMed]
  • Penner JL, Hennessy JN. O antigen grouping of Morganella morganii (Proteus morganii) by slide agglutination. J Clin Microbiol. 1979;10:8–13. [PMC free article] [PubMed]
  • Potter M. Antigen-binding myeloma proteins in mice. Ann NY Acad Sci. 1971;190:306–321. doi:10.1111/j.1749-6632.1971.tb13543.x. [PubMed]
  • Risberg A, Masoud H, Martin A, Richards JC, Moxon ER, Schweda EKH. Structural analysis of the lipopolysaccharide epitopes expressed by a capsule-deficient strain of Haemophilus influenzae Rd. Eur J Biochem. 1999;261:171–180. doi:10.1046/j.1432-1327.1999.00248.x. [PubMed]
  • Satow Y, Cohen GH, Padlan EA, Davies DR. Phosphocholine binding immunoglobulin Fab McPC603: An X-ray diffraction study at 2.7 Å J Mol Biol. 1986;190:593–603. doi:10.1016/0022-2836(86)90245-7. [PubMed]
  • Tzianabos AO, Onderdonk AB, Rosner B, Cisneros RL, Kasper DL. Structural features of polysaccharides that induce intra-abdominal abscesses. Science. 1993;262:416–419. doi:10.1126/science.8211161. [PubMed]
  • Velez CD, Lewis CJ, Kasper DL, Cobb BA. Type I Streptococcus pneumoniae carbohydrate utilizes a nitric oxide and MHCII-dependent pathway for antigen presentation. Immunology. 2009;127:73–82. doi:10.1111/j.1365-2567.2008.02924.x. [PubMed]
  • Wang Y, Kalka-Moll WM, Roehrl MH, Kasper DL. Structural basis of the abscess-modulating polysaccharide A2 from Bacteroides fragilis. Proc Natl Acad Sci USA. 2000;97:13478–13483. doi:10.1073/pnas.97.25.13478. [PubMed]
  • Williams KR, Claflin JL. Clonotypes of anti-phosphocholine antibodies induced with Proteus morganii (Potter) J Immunol. 1980;125:2429–2436. [PubMed]

Articles from Glycobiology are provided here courtesy of Oxford University Press